AIMarketsInIVD 2020 Kalorama

Artificial Intelligence Markets
in IVD
April 2020
© Kalorama Information
Artificial Intelligence Markets in IVD
Table of Contents
1: EXECUTIVE SUMMARY ............................................................................................................................. 1
THE ARTIFICIAL INTELLIGENCE BOOM...................................................................................................................... 1

THE FDA IS GETTING READY................................................................................................................................. 3
FDA Awards DNAnexus Contract to Power precisionFDA Collaborative Omics Environment in the Cloud 4
COVID-19 AND ARTIFICIAL INTELLIGENCE............................................................................................................... 5
Deep Learning to Discover Origins of COVID-19 ....................................................................................... 6
AI-Assisted App Will Direct Individuals Most At Risk for Having COVID-19 .............................................. 7
FDA Allows Remote Use of Philips IntelliSite Pathology Solution during COVID Crisis .............................. 8
JLK Inspection: RT-PCR COVID-19 Testing with AI-Enhanced Imaging Aid ................................................ 9
Persivia: Ruleset to Triage COVID-19 Testing ............................................................................................ 9
Surgisphere Rapid Test Uses AI to Diagnose COVID-19 .......................................................................... 10
Biofourmis' AI-Powered Remote Monitoring Platform to Fight COVID-19 ............................................. 10
METHODOLOGY ............................................................................................................................................... 12
2: ARTIFICIAL INTELLIGENCE IN DIAGNOSTICS MARKETS............................................................................ 13
AI IN CLINICAL PRACTICE .................................................................................................................................... 13

SELECTED AI-BASED INITIATIVES / COLLABORATIONS / TESTS .................................................................................... 16
Table 2-1: Selected AI-based Initiatives / Collaborations / Tests ................................................................... 17
IBM Watson Health ................................................................................................................................. 23
BLOOD TYPING AND SCREENING .......................................................................................................................... 25
CANCER .......................................................................................................................................................... 25
Freenome: Computational Biology and Machine Learning for Cancer Detection ................................... 25
CompCyst: Big Increase in Identification of Benign Cysts ....................................................................... 27
Grail: Machine Learning on Cancer Mutations ....................................................................................... 27
Lung Cancer Diagnostics – LungLife AI .................................................................................................... 27
CARDIAC DISEASE ............................................................................................................................................. 28
Eko’s AFib and Heart Murmur Detection Algorithms .............................................................................. 28
Machine Learning to Assess the Likelihood of Acute Myocardial Infarction ........................................... 29
Bay Echo Labs: FDA-cleared EF Calculation ............................................................................................ 30
GENETIC TESTING ............................................................................................................................................. 30
GRAIL ....................................................................................................................................................... 31
April 2020 © Kalorama Information i

Rady Children’s Institute for Genomic Medicine ..................................................................................... 31

Pathway Genomics.................................................................................................................................. 33
GLUCOSE MONITORING ..................................................................................................................................... 34
Medtronic’s Guardian Connect CGM System .......................................................................................... 34
University of Warwick Develops A.I. Technology for Tracking Blood Glucose ........................................ 35
HEMATOLOGY .................................................................................................................................................. 36
HISTOLOGY...................................................................................................................................................... 39
PathAI and Gilead Evaluating Machine Learning Approaches to Liver Histology Assessment ............... 41
INFORMATION TECHNOLOGY............................................................................................................................... 42
Google ..................................................................................................................................................... 43
Electronic Medical Records: Predictive Analytics and AI ......................................................................... 45
KIDNEY DISEASE ............................................................................................................................................... 45
KidneyIntelX: Improved Kidney Testing ................................................................................................... 45
LIQUID BIOPSY ................................................................................................................................................. 46
MICROBIOLOGY ................................................................................................................................................ 47
MICROBIOME .................................................................................................................................................. 50
Viome ...................................................................................................................................................... 51
MOLECULAR DIAGNOSTICS ................................................................................................................................. 51
Seegene: AI-Inspired Molecular Diagnostics ........................................................................................... 51
Predicine and Flagship Biosciences Collaborate ..................................................................................... 52
Molecular Colorectal Diagnosis .............................................................................................................. 52
NEXT-GENERATION SEQUENCING ........................................................................................................................ 52
Siemens ................................................................................................................................................... 53
Curetis Group Company Launches AI-powered Molecular Antibiotic Susceptibility Test ........................ 54
Oxford Nanopore .................................................................................................................................... 55
BGI Genomics .......................................................................................................................................... 56
Qiagen ..................................................................................................................................................... 56
PARASITOLOGY ................................................................................................................................................. 56
ARUP-Techcyte: Detection of Parasites and Ova .................................................................................... 56
PATHOLOGY .................................................................................................................................................... 56
LabCorp and Philips Collaborate ............................................................................................................. 58
POINT OF CARE ................................................................................................................................................ 62
Siemens Strengthens Technology Position in Immunoassay Point-of-Care Testing ................................ 62
Reliant Immune Diagnostics: In-Home Rapid Diagnostic Screening Using AI ......................................... 62
TRANSPLANT RISK TESTING ................................................................................................................................. 65
CareDX and Predigraft ............................................................................................................................ 65
URINALYSIS ..................................................................................................................................................... 66
Clever Culture: FDA Cleared Automation of Sample Interpretation ........................................................ 66
WHAT THE FUTURE HOLDS ................................................................................................................................. 66
ii April 2020 © Kalorama Information

3: MARKET ANALYSIS: ARTIFICIAL INTELLIGENCE IN DIAGNOSTICS ............................................................ 68
SIZE AND FORECAST ESTIMATE ............................................................................................................................ 68

Figure 3-1: AI in Diagnostics Estimated Market and Forecast, 2019-2024 ($M) ............................................ 69
MARKET BY APPLICATION ................................................................................................................................... 70
Figure 3-2: Artificial Intelligence Market in Diagnostics: Market Estimate by Application (Histology-based AI,
Microbiology, Hematology, Glucose Testing, Biomarker Diagnostic Aids) (%) .............................................. 70
COMPONENT MARKET ANALYSIS ......................................................................................................................... 71
Table 3-1: Artificial Intelligence Market in Diagnostics: Market by Component, 2019 (Hardware, Software,
Services) ($M, %)............................................................................................................................................ 71
April 2020 © Kalorama Information iii

1: Executive Summary
The Artificial Intelligence Boom
Artificial intelligence (AI) is becoming more and more common in our everyday lives. But
the role of AI in clinical diagnostics is the subject of intense controversy. The naysayers
cite that nothing can replace human experience in the interpretation of test data. On the
other hand proponents of AI point to the many benefits that AI in all its forms have brought
to the diagnostics arena including workplace efficiencies and improved diagnostics. At the
heart of AI is machine learning, a term coined in 1959 by IBM employee, Arthur Samuel,
who defined it as “the field of study that gives computers the ability to learn without being
explicitly programmed.” The objective of machine learning is to develop algorithms that
inform post-test actions either by the operator or by the instrument itself. Today, machine
learning includes supervised and unsupervised learning and has widely been applied to
areas, including clinical data management, automation of diagnoses, and biomarker
discovery.
The development of deep learning and neural networks has led to artificial intelligence
gaining learning capabilities and according to some the new AI tools are now better than
human eyes at recognizing patterns. This is paving the way for the emergence of a new
generation of medical diagnostic devices that are sometimes capable of surpassing the
detection skills of the best medical practitioners.
AI in healthcare has huge and wide reaching potential and the future of healthcare and the
future of machine learning and artificial intelligence in medicine are deeply interconnected.
Artificial intelligence in healthcare refers to the use of complex algorithms designed to

perform certain tasks in an automated fashion. The algorithms can review, interpret and
even suggest solutions to complex medical problems. So using this notion of AI, the
earliest iteration of AI information technology (IT) has been present in clinical lab
medicine for quite a while. Lab information systems and high-end chemistry and
hematology instruments employ reflex testing algorithms to verify and improve outlier test
results. Reflex testing and Intelligent Routing dynamic system software bring automated
patient-centric workflow to the laboratory. By understanding the tests requested, sample
April 2020 © Kalorama Information 1

volume available and real-time analyzer capacity and status, it continuously calculates the
most expeditious route for every patient sample repeats and STAT and routine scheduling.
Interactive glucose meters are the most commonplace application of AI-incorporated

devices. Test results and trends inform the user as to follow-on actions required. There is
general consensus that this application of AI has greatly improved the life of many people
with diabetes. AI-driven glucose management devices help diabetic users interpret and use
their test results to manage their diabetes.
Over the past several years, healthcare has witnessed a transformation with a shift from
paper-based records systems to electronic records, and incorporation of digital health
monitoring devices and advanced patient screening systems. These advances have resulted
in a data explosion, which can best be manipulated and analyzed using artificial intelligence
(AI) technology. Thus digitization of test data is an intrinsic part of information technology
that has given rise to lab information systems, electronic health records and digital
pathology initiatives that can be found in many healthcare facilities.
The past few years has seen extraordinary advances in artificial intelligence in clinical
medicine. More products have been cleared for clinical use, more new research use only
applications have come to market and many more are in development. In recent years
diagnostics companies in collaboration with AI companies have begun implementing
increasingly sophisticated machine learning techniques to improve the power of data
analysis for patient care. The goal is to use developed algorithms to standardize and aid
interpretation of test data by any medical professional irrespective of expertise. This way
AI technology can assist pathologists, laboratorians and clinicians in complex decision-
making.
New applications of AI are in development to predict outcomes such as diagnosis, death or

hospital readmission; improve upon standard risk assessment tools; elucidate factors that
contribute to disease progression; or to advance personalized medicine by predicting a
patient’s response to treatment. AI tools are used to review data, and to uncover patterns in
the data that can be used to improve analyses, uncover inefficiencies.
Digital pathology products and diabetes management devices were the first to come to
market with data interpretation applications. The last few years have seen the use of AI
interpretation apps extended to a broader range of products including microbiology, disease
genetics and cancer precision medicine. This report will review some of the AI-linked tests
2 April 2020 © Kalorama Information

and test services that have come to market and others that are in development and will
finish with a perspective for the future.
The FDA Is Getting Ready
With the proliferation of AI-driven medical tests and devices, the regulation of these
products for clinical testing is still a work in progress. It is generally accepted that an
increasing number of AI-powered healthcare tools will seek regulatory clearance in the
years to come. The FDA is preparing for such a future and weighing how to assess and
certify such medical devices. In April 2019, the U.S. FDA presented a proposed plan for
how it might regulate medical devices armed with artificial intelligence — also known as
software-as-a-medical-device (SaMD) — whose algorithms can change based on machine
learning and possibly affect people in ways for which they were not approved or cleared.
SaMD is a product without a device body in the sense that it is pure software that operates
on a platform.
Therefore, a new generation of diagnostics will exist as pure software code. How clinical
laboratorians, and regulators, will tackle this new data stream is up for discussion. A first
example of a market cleared SaMD device Apple’s electrocardiogram app on its Apple
Watch Series 4 (September 2018). The full statement on the FDA’s plan can be found @
www.fda.gov/news-events/press-announcements/statement-fda-commissioner-scott-
gottlieb-md-steps-toward-new-tailored-review-framework-artificial.
The FDA is also developing improved strategies for regulation and launching new projects
at the FDA's internal data science incubator. The FDA Information Exchange and Data
Transformation (INFORMED) incubator will explore opportunities for machine learning
and AI to improve existing clinical practices, such as open-access platforms and shared,
continually reconciled database exchange technologies such as blockchain, which can
enable the widespread and secure exchange of health data.
Among the ongoing projects is collaboration with Project Data Sphere

(www.projectdatasphere.org), a nonprofit open-access cancer data repository, aimed at
developing algorithms for classification of tumor dynamics using medical imaging data.
The FDA has also launched a fellowship program with Harvard University (Boston, MA)
on AI and machine learning, which is focused on designing, developing, and implementing

algorithms for regulatory science applications. One such example is a clinical decision
support software that uses AI algorithms to help alert neurovascular specialists of brain
deterioration.
FDA Awards DNAnexus Contract to Power precisionFDA Collaborative Omics Environment in

the Cloud
In December 2019, DNAnexus, Inc., a leader in enterprise omics technology, announced

that the FDA had awarded DNAnexus a five-year, $20 million contract to power the
precisionFDA Collaborative Omics Environment in the Cloud. PrecisionFDA has enabled
significant advances in regulatory science for next-generation sequencing (NGS)-based
drugs and devices, bringing together government, academic, and industry to crowd-source a
reference standard approach to the validation of bioinformatics pipelines assessing genomic
variation.
PrecisionFDA was launched in December 2015, and through collaborative engagement

with regulatory stakeholders in scientific challenges, the site established a gold standard for
germline variant comparison. Under this award, DNAnexus, with partner Deloitte, LLC,
will use its FedRAMP-authorized platform-as-a-service to enhance precisionFDA
capabilities for sponsor-reviewer interaction, add support for multi-omics, and provide a
library of analytical, statistical, and machine learning applications, accessible to reviewers
and powerful enough for bioinformaticians.
DNAnexus previously developed the precisionFDA Challenge Construction Toolkit that

FDA and partners use to design, deploy, and present cloud-based scientific challenges,
creating a community of more than 4,000 users across 1,200 government, commercial,
academic, patient advocacy, and healthcare organizations. The science-driven applications
cover a wide range of bioinformatics topics including: variant calling accuracy; food-borne
pathogen and biothreat detection; representation and reproducibility of NGS pipelines;
assessment of tumor mutation burden; and use of machine learning and artificial
intelligence models using multi-omics data to identify biomarkers and predict brain cancer
patient outcomes.

COVID-19 and Artificial Intelligence
At the time of writing of this report, most healthcare authorities in the world are focused on
COVD-19. It is a testament to AI’s adaptability that several companies are already
employing it to help meet this diagnostic challenge.
There are numerous ways that AI has contributed and could contribute to the detection of
COVID-19, based on what is being done now and what Kalorama has seen in AI
applications for other infectious diseases:
x Developing of Immunoassays – Serological test are in high demand as they could

add testing volume in an inexpensive way. They could also potentially test for
immunity in a patient, and allow patients to go back to work or pursue group
activities when social distancing barriers are removed, which remains an open
question as of this writing. But immunoassays require modeling of antibodies that
could develop after exposure to the virus, and that is complicated. AI can assist
finding models more likely to work, and perhaps speed up immunoassay
development. As reported in Nature magazine, An-Suei Yang, research fellow at the
Genomics Research Center, Academia Sinica, in Taipei, Taiwan, and colleagues
claim to be first to develop a mAb against the nucleocapsid (N) protein of SARS-
CoV-2,.The group developed the test in 19 days, by working with artificial-
intelligence models of antibody–antigen interactions.
x Developing of Molecular Tests - South Korean firm Seegene was one of the first to
develop a test. But it did not have a sample of COVID-19. Instead AI was put to
work as the company tried to discovery tests that would react properly to the virus.
Without the computer, it would have taken the team two to three months to develop
such a test. Instead Seegene developed a test in a matter of weeks. It was only the
third time the company had used its super computer – rather than its research and
development team working manually – to design a test. It had previously used the
system to make diagnosis kits for urethritis, the inflammation of the urethra.
x Triaging Testing – With testing at a premium, deciding who to test, and who to
save resources on is important. AI systems could help decide who needs immediate
resources and who can wait. Even when testing is no longer triage, those at risk for
disease could be predicted with AI-based searches of electronic records.

x Genomic Interpretation AI has been used to more efficiently search genomes to

determine the origins of SARS-CoV-2, which can be critical to developing tests and
possibly predicting future strains for test development.
x AI-Enhanced Imaging could be added to an IVD test to enhance its detection

accuracy. Imaging is an area where AI solutions have already been well
documented in effectiveness, given the strength of vector imaging and other
approaches.
Indeed, while there have been many failures in the China's outbreak occurred despite a
warning from an AI system designed to pick up pandemics. While China dealt with the
initial epidemic, a computer server halfway across the world started issuing worrying alerts
of a potential new outbreak. The server runs software by BlueDot, a company based in San
Francisco that uses AI to monitor infectious disease outbreaks for signs of early trouble.
The machine-learning system flagged articles in Chinese that reported 27 pneumonia cases
associated with a market that had seafood and live animals in Wuhan. In addition to the
alert, BlueDot correctly identified the cities that were highly connected to Wuhan using
things like global airline ticketing data to help anticipate where the infected might be
traveling. The international destinations that BlueDot anticipated would have the highest
volume of travelers from Wuhan were: Bangkok, Hong Kong, Tokyo, Taipei, Phuket,
Seoul, and Singapore. 11 of these cities at the top of their list were the first places to see
COVID-19 cases.
Deep Learning to Discover Origins of COVID-19
A new report published on March 22 in The Journal of Proteome Research suggests that
while bats are a likely natural reservoir, an intermediate host between the mammals and
humans is probable. Lead author Chengxin Zhang and colleagues from the department of
computational medicine and bioinformatics at the University of Michigan addressed the
origins of SARS-CoV-2 by examining two recent studies. The first study reported that the
SARS-CoV-2 spike protein had four unique insertions that shared "uncanny similarities"
with an HIV-1 protein. The controversial paper, which was published online on a preprint
server and has since been removed, was met with skepticism in the scientific community
but caused rumors and conspiracy theories to circulate among the general public. The
researchers reanalyzed the structural location and sequence homology of the four spike

protein insertions described in the withdrawn paper. Deep-learning modeling demonstrated

that all four of the insertions were located outside of the receptor binding domain of the
spike protein.
They found that the four insertions are not unique to the novel coronavirus and HIV-1.
They concluded that the similarities in the sequence-based alignments of the very short
fragments were statistically insignificant, according to basic local alignment search tool
(BLAST) E values. In fact, the similarities in the fragments were shared among many other
viruses including bat coronaviruses. They did conclude that the sequence alignment of the
fragments indicated a close evolutionary relation to the bat coronavirus. The researchers
also scrutinized a second study that analyzed the relative synonymous codon usage (RSCU)
of the virus in eight vertebrates and found that snakes have the smallest codon usage
difference. This conclusion is controversial among virologists due to lack of evidence that
coronaviruses can infect animals other than mammals and birds.
To address several concerns in the study, the researchers reimplemented the RSCU
comparison algorithm with 102,367 vertebrate species. They found that snakes did not have
the lowest RSCU distances with SARS-CoV-2, suggesting that the analysis conducted in
the aforementioned study was incomplete. Interestingly, two kinds of frogs had the
smallest RSCU distances among all vertebrates with sufficient sequences. The researchers
suggest that RSCU analysis failed to identify an intermediate host because it is not specific
enough to discriminate coronaviruses from different vertebrate hosts. This is in comparison
to intermediate hosts for SARS-CoV and MERS-CoV, which have almost no difference in
RSCU.
AI-Assisted App Will Direct Individuals Most At Risk for Having COVID-19
A smartphone app coupled with artificial intelligence and machine intelligence will
increase access to at-home COVID-19 risk assessments, which will help provide health
officials with real-time information to better target potential COVID-19 patients, according
to experts from the Medical College of Georgia. The report featured in the Infection
Control & Hospital Epidemiology said that the app will direct individuals most “at-risk” of
having COVID-19 to officials who will also be able to see emerging demographics of
vulnerable patients to enhance prevention and treatment initiatives. The app will ask
individuals questions such as where they live, their gender, age, and race, as well as about
any recent contact with someone who has or had coronavirus or any recent travel history. It

will also ask about common symptoms of COVID-19 such as fever, cough, shortness of
breath, fatigue, sputum production, headache, diarrhea, and pneumonia. For those
individuals unable to fill out their own survey, the app will feature a collection of similar
questions for a friend or relative to fill out on their behalf. Artificial intelligence will use
an algorithm to rapidly assess the individual’s information, send them a risk assessment,
and alert the nearest facility with testing ability that a health check is necessary. A mobile
check will be arranged if the patient is unable to travel. The collective information of
many individuals will aid rapid and accurate identification of geographic regions, including
cities, counties, towns, and villages, where the virus is circulating, and the relative risk in
that region so healthcare facilities and providers can prepare. Investigators also
highlighted that the app will help them learn more about how the virus is spreading, which
is vital to gaining control over the continuing outbreak.
FDA Allows Remote Use of Philips IntelliSite Pathology Solution during COVID Crisis
Philips announced that the FDA has granted a waiver for the use of readily available
consumer monitors at home with the Philips IntelliSite Pathology Solution. Pathologists,
who may not be able to work on-site due to the current COVID-19 emergency, can start
working from home to review pathological cases and so prevent any delay in critical patient
care.
Pathologists play a crucial role in the detection and diagnosis of a broad range of diseases,
including cancer, and consequently guide treatment decisions. The already scarce resources
of pathologists are under enormous pressure to maintain high-quality patient care. Recently,
both pathologists' societies and industry associations successfully advocated a waiver from
remote pathology restrictions during this public health pandemic.
The FDA temporarily exempts restrictions for the remote use of the Philips IntelliSite
Pathology Solution from March 28, 2020, onwards under specific conditions and device
specifications. Pathologists are allowed to access whole slide images of patient tissue
remotely through a secure VPN high-speed internet connection. And using a viewing web-
based browser and monitor meeting specified minimum requirements, they can make a
definitive diagnosis of histology cases from their home office.

JLK Inspection: RT-PCR COVID-19 Testing with AI-Enhanced Imaging Aid
South Korean Firm JLK Inspection Inc. has added another dimension to RT-PCR testing
for COVID-19 with AI and imaging enhanced diagnostic aids. The system starts with a
PCR test and then AI-enhanced X-ray analysis to continue to inform providers of disease
progress or remission.
The company's product is its AI Total COVID-19 Care based on its AIHuB. AIHuB
conducts analysis from MRI, CT, X-ray and mammography imagery with an AI-enabled
technology that encompasses a wide variety of diagnostic techniques to pinpoint illnesses
such as stroke, Alzheimer’s and cancer. This has been applied to COVID-19 in South
Korea, where JLK has connected their system to an RT-PCR test. The product triages
positive results as severe or mild to allow for prioritizing and providing the proper
treatment to the infected. The company’s AI Total COVID-19 Care is comprised of three
products, including AI Software that triages the pneumonia patients developed by nCov for
more than 5-7 days.
The AI-based lung disease analysis solution developed by the company is a device that
helps medical staff diagnose lung disease through chest X-ray image analysis. The
company says the advantage is that it improves portability through a method of coupling to
a portable X-ray device and its VIEWER X software.
Persivia: Ruleset to Triage COVID-19 Testing
In March 2020, Persivia, a provider of real-time bundled payment, value-based care,

population health, and quality management solutions, announced the launch of a new
COVID-19 surveillance module within its Soliton AI engine. Delivered as part of its
CareSpace platform, this module uses data from multiple sources to identify patients who
may have a COVID-19 novel Coronavirus infection. The ruleset may help healthcare
professionals more effectively target patients who need to be tested. The new functionality
is available to Persivia’s install base of 20 million covered lives across the U.S. The ruleset
is being offered at no cost to public health departments across the country to help them
manage the ongoing public health crisis.
The new Soliton module targets suspected COVID-19 cases by processing structured and
unstructured data that Persivia acquires and normalizes from disparate sources, including
real-time EHR data. The Soliton AI engine uses evidence-based algorithms to identify

specific concepts and keywords in structured and unstructured data to identify potential
cases of interest. Patients who are identified as having three or more relevant symptoms, as
identified by CDC guidelines, such as fever, shortness of breath, and cough, are flagged
with a level one alert. At this lowest level, the patient’s physician is prompted to educate
the patient about COVID-19 and monitor their status. If Persivia also identifies a travel or
exposure history the alert gets raised to a level two, where a CDC test is recommended.
Patients receive a level three alert when they are presumed positive for COVID-19 and
level four when they are confirmed with symptoms. Medical professionals are also able to
search for all patients flagged as possibly having COVID-19, and can filter them by alert
level and symptoms.
It took less than 10 days for Persivia to bring this ruleset to production on its flexible,
modular AI system. Persivia will be able to adapt and add to the ruleset as more is
discovered about the disease, its symptoms, and progression. Similar rulesets can be
developed for any health state, from other infectious diseases, to behavioral health and
heart failure.
Surgisphere Rapid Test Uses AI to Diagnose COVID-19
In March 2020, Surgisphere Corporation (Palatine, Illinois) announced the availability of

its new rapid diagnostic test for COVID-19 that uses artificial intelligence.
With advanced machine learning, Surgisphere has developed an intelligent tool that uses
three common lab tests to identify patients likely to have COVID-19 infection.
With its cloud-based healthcare data analytics platform (QuartzClinical), Surgisphere

identified patients diagnosed with COVID-19 in five countries. The company then created a
machine learning model using decision tree analysis and clinical and lab data on 430
patients to develop the decision support tool.
The test is available for public use.
Biofourmis' AI-Powered Remote Monitoring Platform to Fight COVID-19
In March 2020, Biofourmis, a provider of digital therapeutics that power personalized

predictive care, announced that its technology is being leveraged in a remote monitoring

and disease surveillance program in Hong Kong involving patients with diagnosed or
suspected novel Coronavirus disease (COVID-19). The program – administered by The
University of Hong Kong – also includes Hong Kong-based Harmony Medical Inc., which
is Biofourmis' joint venture partner for the China region.
Diagnosed and potentially COVID-19-infected patients are being monitored with

Biofourmis' Biovitals Sentinel platform, a turn-key solution that the company built in less
than two weeks specifically for this initiative by customizing its industry-leading, FDA-
cleared artificial intelligence (AI)-powered Biovitals Analytics platform. Biovitals
Sentinel's 24/7 remote monitoring technology and analytics are providing clinicians
involved in the COVID-19 program with clinical decision support for early identification of
any physiological changes that could indicate deterioration, and to enable earlier
interventions for better outcomes.
The goal of the program is to leverage Biovitals Sentinel to remotely monitor patients and
identify COVID-19-related physiological biomarkers that indicate deterioration in patients.
The hope is that the effort also will rapidly lead to a better epidemiological understanding
of COVID-19 to improve outcomes as more people become infected.
Patients with COVID-19 deterioration commonly exhibit symptoms such as fever, cough
and shortness of breath, all of which can be closely monitored through related physiological
parameters via Biofourmis' clinical-grade biosensor Everion, which is to be worn on the
arm by patients quarantined in their homes or clinical settings. Clinicians and researchers
are still learning how this strain of the Coronavirus affects the body, and the research will
help elucidate that.
Everion has several medical-grade sensors, including optical, temperature, electrodermal,

accelerometer and barometer. Everion can be comfortably worn for 24 hours a day for
several days and has much higher-than-average patient satisfaction and adherence rates. AI-
powered Biovitals Analytics will derive more than 20 physiological signals from the sensor
data, including temperature, heart rate, blood pulse wave, heart rate variability, respiration
rate, inter-beat-interval and others. These signals are then fed through advanced AI and
machine-learning techniques to flag key physiological changes that could indicate disease
progression.
The Biovitals Sentinel solution also includes a smartphone app to collect qualitative data
reported by patients as they respond to AI-generated "smart nuggets." Treating physicians,

through the Biovitals Sentinel clinician-facing web dashboard, will observe and be alerted
to any significant physiological changes and adverse events so they can intervene when
necessary.
At the time of this report’s release, Biofourmis is in conversations with government

agencies and other potential partners in various regions about applying its technology to
help even more patients and to learn more about COVID-19 so that it can be treated more
effectively.
Methodology
The information presented in this report is the result of data gathered from government
reports, company product literature and other corporate brochures and documents, as well
as information found in the scientific and trade press.

2: Artificial Intelligence in Diagnostics Markets
AI in Clinical Practice
Similar to the product commercialization process that is evident in the liquid biopsy market,
IVD companies’ efforts have been supported by the U.S. FDA’s efforts to give products
expedited review. This helps patients have more timely access to devices and technology
that provide more effective treatment or diagnosis for life-threatening diseases.
Digital pathology products and diabetes management devices were the first to come to
market with data interpretation applications. As shown in the table below, the last few
years have seen the use of AI interpretation apps extended to a broader range of products
including microbiology, disease genetics and cancer precision medicine. The following are
of particular interest.
In December 2018, Adaptive Biotechnologies (Seattle, WA) received FDA clearance for its
ClonoSEQ assay, a next generation sequencing (NGS)-based test for minimal residual
disease (MRD) in patients with acute lymphoblastic leukemia (ALL) or multiple myeloma.
ClonoSEQ, which uses NGS technology to assess disease burden, is the first and only assay
to be cleared by the FDA for MRD assessment in any lymphoid cancer and the first FDA-
cleared diagnostic assay powered by immunosequencing. ClonoSEQ is an in vitro
diagnostic that uses multiplex PCR and NGS to identify and quantify certain gene
sequences in DNA extracted from bone marrow from patients with ALL or multiple
myeloma. The ClonoSEQ assay measures the amount of MRD and is capable of detecting
MRD at levels below 1 in 1 million cells.
For microbiology, microbiology automated sample processing leader, Copan Diagnostics

(Murrieta, CA) markets its PhenoMATRIX culture plate reader with AI analysis. Their
Phenotypic Colony Recognition algorithm examines colonies present on the plate and
compares them against a library of thousands of colony images to match the phenotypic
characteristics to assign a predictive value. The physician confirms the presumptive ID
assigned by the software and then performs further work up. The system's User-Defined
Expert Rules Algorithm, which uses pre-programmed rules and analyzes patient
demographics to determine if the growth on a plate is significant and relevant. For example,

the software can be programmed to check to see if the patient is pregnant or a female of
child bearing age and alert for growth that is characteristic of Group B Strep.
Clever Culture (Switzerland) received FDA clearance to market its APAS platform for
culture-plate screening and interpretation.
In May 2019, AI-powered kidney disease and transplant management diagnostic developer
RenalytixAI (New York, NY) received an FDA breakthrough device designation for its
KidneyIntelX artificial-intelligence powered diagnostic for kidney disease. The company
claimed that the FDA clearance as the first ever for an AI-powered kidney disease
diagnostic. The KidneyIntelX is designed to diagnose and improve clinical management of
patients for patients with Type II diabetes and fast-progressing kidney disease. The
diagnostic uses machine learning algorithms to assess the combination of blood-based
biomarkers and electronic health record info to identify progressive kidney disease.
KidneyIntelX has been developed in collaboration with the Mount Sinai Health System
(New York, NY).
Qiagen’s (Venlo, Netherlands) bioinformatics teams are developing new software solutions
and adding proprietary cloud-based content to support the latest research and clinical trends
in molecular testing, especially the interpretation of large volumes of data from next-
generation sequencing. In addition, Qiagen is integrating these digital technologies with
instruments and molecular content to provide seamless Sample to Insight workflows with
the acquisition of OmicSoft. In January 2017, Qiagen acquired OmicSoft Corporation
(Germantown, MD) providing access to OmicSoft's multi-omics data management
infrastructure solution as well as expertly curated 'omics' data sets that complement
Qiagen's bioinformatics portfolio that offer insights into complex biological data.
OmicSoft is recognized for its suite of software solutions that enable scientists and
researchers to efficiently analyze and visualize their own data sets and compare them to
massive volumes of publicly available 'omics' data sets - such as The Cancer Genome Atlas
(TCGA) operated by the U.S. National Cancer Institute - and share results with colleagues.
In January 2020, Qiagen launched Qiagen QCI Precision Insights in Europe. QCI Precision
Insights is a clinical interpretation service provided by a team of molecular biologists and
oncologists who translate patient-specific molecular data into clinical insights and
therapeutic options. Qiagen said that the service enables molecular pathologists and
oncologists in Europe to match complex genomic test data to relevant, geographically

matched treatment options in minutes for evidence-based, individualized treatment

recommendations.
Royal Philips (Amsterdam, Netherlands) showcased its oncology informatics solutions at

the 2018 American Society of Clinical Oncology Annual Meeting (ASCO) in Chicago, IL.
The company’s seamless digital pathology, genomics and tumor surveillance technology
support tools, IntelliSpace Genomics and IntelliSpace Oncology, help support and scale
precision medicine programs to assist oncologists with treatment decisions. In February
2019, Philips announced that The Miami Cancer Institute signed a multiyear agreement to
use Royal Philips' IntelliSpace Precision Medicine Genomics software platform in its new
genomics laboratory. The partnership will support the oncology genomic testing workflow
at the institute, which is part of the Baptist Health South Florida system.
Siemens Healthineers is working on its Digital Ecosystem, a digital platform that combines
imaging, in-vitro diagnostics and medical documentation.
Nikon Instruments (Melville, NY) has developed the NIS.ai microscopy image analysis and
processing module that leverages deep learning and artificial intelligence to extract
unbiased data from vast amounts of microscopy datasets. The module is a suite of AI-
based processing tools that utilizes convolutional neural networks (CNNs) in order to learn
how to read images from small training datasets supplied by the user. The training results
can then are applied to process and analyze huge volumes of data, allowing researchers to
increase throughput and expand their application limits. The NIS.ai includes a suite of
applications for predictive imaging, image segmentation, and processing.
In November 2019, Sophia Genetics (Boston, MA) launched its AI platform service.
Hospitals send their data to Sophia’s AI platform for genomic variant detection, analysis,
and interpretation. Their deliverable is a report that is issued to the hospital serving as the
clinician’s diagnosis. The AI platform uses statistical analysis, pattern recognition, and
machine learning to analyze the large data sets. Sophia reports that its machine learning
capability is currently analyzing and learning with over 380,000 genomic profiles and
analyzing more than 15,000 cases a month.
GenomeSmart (Los Altos, CA) has developed GenomeBrain that uses artificial intelligence
to match a person's personal and medical history, ethnicity and age, with national
guidelines to recommended medically actionable genetic tests. The platform then generates
a report that clinicians and their patients can use to determine a testing plan. The company

is partnering with several companies to integrate GenomeBrain with their test platforms.
December 2019, the company announced that Aspira Labs, a subsidiary of Vermillion
(Austin, TX) will start a pilot project and offer GenomeBrain to a select group of its
customers to improve identification of patients who are at risk for hereditary cancer or
passing hereditary diseases to their children. In November 2019, GenomeSmart signed an
agreement with NorthBay Healthcare (California) to use GenomeBrain in its breast cancer
program.
In October 2019, Genomenon (Ann Arbor, MI) announced a partnership with Sophia
Genetics (Saint Sulpice, Switzerland) to incorporate the Mastermind Genomic Search
Engine into the Sophia artificial intelligence platform and Alamut Suite for genomic
analysis. Sophia has developed a machine-learning tool that helps healthcare professionals
detect and characterize genomic variants linked to cancers and hereditary disorders.
Sophia’s Alamut Suite is a genome browser that graphically displays complex genetic
information and assesses the pathogenic status of disease mutations.
In October 2019, Sysmex Corp. (Tokyo, Japan) and artificial intelligence company OPTiM
(Tokyo, Japan) established a joint venture aimed at the development and operation of
services and platforms for digital medicine. The companies are already collaborating on
initiatives to combine Sysmex’s global sales and service network with OPTiM’s artificial
intelligence and data management technologies. Now they plan to develop and test new
diagnostic methods that combine image information from gene tests with AI analysis, as
well as to use AI for image processing of various data obtained from existing Sysmex
analyzers.
Selected AI-based Initiatives / Collaborations / Tests
The following are recent developments in artificial intelligence in diagnostics. While each
does not necessarily represent a unique event (there are many vendors operating and
proposing products in AI at the present time); they are representative efforts.

Table 2-1: Selected AI-based Initiatives / Collaborations / Tests
Company Location Area Details Status
Adaptive clonoSEQ immunosequencing with AI, min residual

U.S. ca MRD FDA
Biotechnologies disease
With AI developer Techcyte, announced they have co-

ARUP
U.S. parasitology developed an AI-augmented tool for improved detection dev
Laboratories
of parasites and ova
With Cibiltech (Paris, France), partnering to

commercialize Predigraft, a data analysis tool that
CareDx U.S. transplant mkt
provides an early prediction of an individual's risk of
allograft rejection and transplant loss
Caris Life Sciences U.S. cancer MI GPS (Genomic Profiling Similarity), tumor types rsch
Clever Culture Switzerland micro APAS platform culture-plate screening & interpretation FDA
Copan Diagnostics U.S. micro PhenoMATRIX culture plate analysis AI mkt
With Translational Software AI drug response,

Diagnomics U.S. pers med dev
metabolism, and treatment efficacy
FDA clears Eko’s AFib and heart murmur detection

Eko U.S. cardiac FDA
algorithms
Euroclone
Italy liq biopsy Precipio's Ice Cold-PCR DNA, liquid biopsy, EU mkt
Diagnostica
Freenome plans to get its colorectal cancer screening

blood test through a validation study and expand its
blood-based, AI-driven testing platform into other tumor
Freenome U.S. cancer rsch
types, as well as other disease areas, with the help of
$160 million raised in a series B financing round in July
2019.
GenomeSmart U.S. prec med GenomeBrain link patient with appropriate genetic test RUO
In a collaboration with PathAI, Gilead is evaluating

Gilead U.S. pathology machine learning approaches to liver histology dev
assessment
Gnosis Data
Greece genome Just Add Data (JAD) Bio system AI for multi-omics RUO
Analysis
Google U.S. histo AI research and digital pathology rsch
Google's DeepMind research shows that AI could

Google U.S. kidney rsch
accurately predict acute kidney injury (AKI)

Began September 2018 a large-scale study, using

machine learning to set up a huge library of healthy
Grail U.S. cancer rsch
mutations and cancer mutations, using data from white
blood cell genomes and cfDNA
Ibex Medical Israel histo Second Read (SR) system, AI digital pathology mkt
IBM Watson
U.S. alz early Alzheimer's detection rsch
Health
AI added to IndivuType multi-omics cancer research

Indivumed Germany cancer RUO
platform
Inspirata U.S. histo Dynamyx digital pathology software mkt
Institut Curie France genome Intel AI tools for a genomics database dev
IntegraGen France sequencing sequencing data with Google Cloud mkt
E360 analytics technology genotypic-phenotypic

IQVIA U.S. genome rsch
database
CompCyst test incorporates the analysis of genetic

Johns Hopkins U.S. cancer rsch
mutations in pancreatic cyst fluid and AI algorithms
Johns Hopkins U.S. liq biopsy pan cancer liquid biopsy Dx with AI analysis rsch
machine learning with diagnostics, cancer, Alzheimer’s,

Maxwell Plus Australia genome rsch
etc.
Medtronic Ireland diabetes Sugar.IQ app with IBM Watson IQcast AI mkt
Nikon Instruments U.S. microscopy NIS.ai microscopy image analysis rsch
Paragon
U.S. sequencing Paragon CleanPlex NGS with Sophia Genetics AI platform mkt
Genomics
Phase Genomics U.S. microbiome computational platform for microbiome samples rsch
Philips Netherlands cancer Clinical Pathways, a clinical decision platform, oncology rsch
Philips Netherlands genome IntelliSpace Precision Medicine Genomics software rsch
LabCorp and Royal Philips collaborate in digital pathology

Philips Netherlands pathology through the implementation of the Philips IntelliSite mkt
Pathology Solution
Predicine GeneRADAR liquid biopsy & Flagship

Predicine U.S liq biopsy mkt
Biosciences, AI digital pathology

PredictImmune UK IBD PredicSure IBD CE
Qiagen Clinical Insight analysis and interpretation

Qiagen Netherlands sequencing RUO
solutions
Will pair Freenome's AI genomics platform with various

Qiagen Netherlands CDx Qiagen molecular analysis platforms and co-market the rsch
technologies to pharma companies
RenalytixAI U.S. kidney KidneyIntelX AI diagnostic for kidney disease FDA
Announced January 2018 an AI-based assay

Seegene Korea molecular development system that develops (MDx) assays in only mkt
4 days
Siemens
Germany sequencing Fast Track cycler & FastFinder software rsch
Healthineers
Siemens Acquires Minicare BV to strengthen its portfolio in AI

Germany POC mkt
Healthineers POC immunoassay testing
Sophia Genetics Switzerland cancer Solid Tumor Solution molecular with AI analysis CE
Sophia Genetics U.S. genome AI platform service srvc
UgenTec Belgium PCR Fastfinder real-time PCR interpretation system CE
Xifin U.S. prec med VisualStrata precision medicine informatics platform mkt
Source: Kalorama Information, company reports
In spite of a still uncertain regulatory pathway, the IVD industry and its partners have
embarked in full force in the advancement of AI driven tests and devices. In the UK, in
May 2018, Prime Minister Theresa May announced the country is aiming to cut cancer
deaths by 10% using artificial intelligence (AI) as the key driver of improved health
outcomes. The plan calls for the National Health Service (NHS, London), the AI industrial
sector, and health charities to use data and AI to transform the diagnosis of chronic
diseases, with the goal of seeing around 22,000 fewer people dying from cancer each year
by 2033.
The plans will see at least 50,000 people each year diagnosed at an early stage of prostate,
ovarian, lung, or bowel cancer through the use of emerging technologies that will cross-
reference people’s genetics, habits, and medical records with national data to spot those at

an early stage of cancer. A goal is to integrate all patient data into single unified software
framework that will enable faster handling of medical information and will lay the
foundation for more efficient interaction between different specialties and enable more
precise and personalized clinical decisions.
Hitachi Ltd (Tokyo, Japan) is taking traditional machine learned algorithms one step further
– AI enables machines to predict medical events. According to Hitachi, doctors and nurses
using the algorithm will be able to tell exactly why a certain patient is at high risk for
hospital admission, and what they can do about it.
In December 2017, Hitachi Ltd. and Partners Connected Health (Somerville, MA)
announced a collaboration to develop artificial intelligence technology, which can very
accurately predict the risk of hospital readmissions within 30 days among patients with
heart failure. The AI technology helps choose appropriate patients to participate in a
readmission prevention program after hospital discharge and provides an explanation as to
why the patients have been identified as high-risk ones. Hitachi and the Partners
Connected Health Innovation team plan to jointly conduct a prospective study to evaluate
the prediction program by clinicians, and study how to integrate this within clinical
workflows.
Partners Connected Health, at Partners HealthCare, is leveraging information technology –

mobile phones, tablets, wearables, sensors and remote health monitoring tools – to deliver
quality patient care outside of traditional medical settings. The Connected Health team
creates and deploys mobile technologies in a number of patient populations and care
settings and is conducting innovative research studies to test the effectiveness of mobile
health technologies in various clinical applications, including medication adherence, care
coordination, chronic disease management, prevention and wellness.
In April 2019, PredictImmune (Cambridge, UK) entered a commercial agreement with

Addenbrooke's Hospital, part of Cambridge University Hospitals National Health Service
Foundation Trust, to provide the company's PredictSure IBD prognostic assay to clinicians
and gastroenterologists in the UK and Ireland. PredictImmune's CE-IVD PredictSure
platform integrates a blood-based biomarker assay and a proprietary algorithm to identify
biomarkers linked to irritable bowel disease (IBD), including Crohn's disease and ulcerative
colitis. In March 2019, PredictImmune’s PredictSure IBD test was approved for the UK’s
NICE's Diagnostic Assessment Programme, which evaluates new diagnostic technologies
for the UK's National Health Service.

In April 2019, consumer wellness startup Onegevity Health (New York, NY) signed a
licensing agreement with Japan-based Kirin Holdings and Mitsui to launch Onegevity's
testing system in Japan and to develop unique nutritional products using the company's AI-
based platform. As part of the collaboration, the partners will start offering Onegevity's
GutBio assay, which combines metagenomic sequencing with AI-based personalized
recommendations to increase the user's gut health, in Japan. Onegevity Health's AI
platform integrates multiple health-related data sources and generates insights from
machine learning.
In October 2018, GE Healthcare (Chicago, IL) announced the company is collaborating

with hospital service providers in Europe and the U.S. to implement an artificial
intelligence (AI)-powered command center to effectively and efficiently synchronize all
elements of a patient's hospital experience. The AI-powered system will support caregivers
to enhance patient safety, orchestrate seamless care delivery and get patients back home
sooner.
According to GE, up to 20 staff based in the command center will monitor a “wall of
analytics” that constantly pulls in streams of real-time data from the multiple systems at the
hospital. Advanced algorithms will help staff to anticipate and resolve bottlenecks in care
delivery before they occur, recommending actions to enable faster, more responsive patient
care and better allocation of resources. The data will be displayed on multiple high
definition screens in the command center - as well as on tablets and mobile devices,
providing 24-7 support to busy medical teams across the hospital.
In the U.S., GE Healthcare has joined forces with CHI Franciscan Health (Tacoma, WA) to
implement a NASA-style "Mission Control" command center, making it the first hospital
system in the state of Washington, and the fifth globally, to utilize the technology. The
mission control center will use the power of AI and predictive analytics to optimize care
coordination, speed care delivery, and improve the patient experience, while maintaining
patient privacy. The system works by looking at each individual hospital as part of a larger
system, continually examining real-time data and using machine learning to recommend
actions that can predict and prevent risk, balance staff workload, and streamline the
discharge process so patients can get home sooner.
In January 2019, researchers from Case Western Reserve University (Cleveland, OH,)
reported they are using artificial intelligence and smart-imaging computer technology for
viewing routine medical images — MRIs, CAT scans, tissue biopsies — to determine the

best type of cancer treatment. In addition to detecting patterns beyond what the human eye
can see, AI and computers can also determine the likelihood that a patient is going to
respond to chemotherapy. The researchers suggest that a significant reduction can be
brought about in the number of surgeries and of biopsies in patients who do not require
them because they have benign nodules. Their research has been able to determine whether
breast cancer patients, and more recently lung cancer patients, need chemotherapy.
In December 2018, the OWKIN Loop Network Breakthrough Medical Moments Powered
by AI (New York, NY, https://owkin.com/) was launched. It includes over 30 international
hospitals and research institutions across the United States and Europe and has been
designed to help researchers train predictive models on real-world data at scale and transfer
the accrued knowledge to a collective intelligence. Members include Cleveland Clinic (OH,
USA), Mount Sinai (New York, NY, USA), and Groupe AP-HP, a cluster of 39 hospitals in
France, among others.
The Loop Network creates an ecosystem that shares the collective knowledge, benefitting
research organizations, partner hospitals, and pharmaceutical companies in advancing
research and development in oncology, cardiovascular, neurodegenerative, and
autoimmune diseases. Projects undertaken by the network include the training of a
predictive model identifying new quantitative biomarkers associated with prognosis in a
rare cancer, the prediction of brain age from magnetic resonance imaging (MRI), and
prediction of gene expression profiles from slide images as a marker of response to
immunotherapy.
The state of Kentucky has one of the highest rates of preterm birth in the U.S, according to
the Centers for Disease Control and Prevention. A big challenge has been identifying
mothers at risk early enough in their pregnancy in order to design and implement effective
intervention strategies. So, in May 2018, Passport Health, a state-based Medicaid
managed-care insurer started a pilot that uses artificial intelligence to identify mothers' at-
risk factors. The plan uses a maternity analytics platform developed by Kentucky-based
data management solutions company Lucina Health to stratify medical and socio-economic
risk factors to allow Passport to pinpoint how best to help those at greatest risk.
In July 2018, BC Platforms (Basel, Switzerland) and Tieto (Uusimaa, Finland) announced
their plan to create a platform for integrating clinical and genomic data with an eye toward
improving predictive and precision medicine. The partners indicated that they will target
healthcare providers in the Nordic countries of Europe. The companies will use Tieto's

Lifecare line of healthcare technology for lab information management systems and
electronic medical records offerings to ensure that genomics data together with clinical data
will be used to more accurately guide individual diagnoses and treatments, but also provide
support of preventive and predictive medical decision making.
IBM Watson Health
The best-known application of artificial intelligence in medicine is IBM’s Watson. First

introduced in 2012, IBM marketed its Watson supercomputer as having the potential to
“revolutionize” cancer diagnosis and treatment. However, according to an article in the
Wall Street Journal in August 2018, IBM has fallen far short of that goal during the past
seven years. The article states, “More than a dozen IBM partners and clients have halted or
shrunk Watson’s oncology-related projects. Watson cancer applications have had limited
impact on patients, according to dozens of interviews with medical centers, companies and
doctors who have used it, as well as documents reviewed by the Wall Street Journal.”
Anatomic pathologists who use tumor biopsies to diagnose cancer have always been
skeptical and have wondered if IBM’s Watson would actually help physicians do a better
job in the diagnosis, treatment, and monitoring of cancer patients. The findings of the Wall
Street Journal show that Watson has yet to make much of a positive impact when used in
support of cancer care.
Nevertheless, IBM and partners continue to explore applications for Watson as a tool in
diagnosis and treatment. According to IBM Watson Health “by coupling clinical data with
genomic data, there is an exceptional opportunity to make polygenic risk scoring more
robust and powerful, and ultimately transformative for patient care. Such transformation
could never happen without these kinds of partnerships."
Building on this, MIT-IBM Watson Lab announced in 2018, the company’s collaborations
are part of IBM Watson Health’s longstanding commitment to scientific research and “our
belief that working together with the world’s leading institutions is the fastest path to
develop, advance, and understand practical solutions that solve some of the world’s biggest
health challenges.”
In February 2019, IBM Watson Health announced an expansion to its partnership with the
Broad Institute (Cambridge, MA), providing artificial intelligence to help clinicians mine
genomic data to predict patient risk for cardiovascular disease. Under the new three-year

agreement, IBM Watson and the Broad will build algorithms to extract knowledge from
population-wide and hospital-based biobank data, genomic information, and electronic
health records to develop genetic risk scores. The ultimate goal is to create computational
models to analyze phenotypic and genotypic data to identify risk factors within individual
genomes so clinicians can intervene early and prevent potential heart attacks and atrial
fibrillation.
IBM Watson and the Broad already are a little more than two years into a five-year, $50
million research initiative that seeks to discover the basis for drug resistance in cancers.
The Broad is a partner in IBM Watson Genomics from Quest Diagnostics, a service that is
available to oncologists nationwide that focuses on oncology diagnostics and providing
options for evidence-based methods of treating cancer.
In February 2019, IBM Watson also announced that it would put a total of $50 million into
a 10-year research collaboration with the Harvard-affiliated Brigham and Women's
Hospital and with Vanderbilt University Medical Center to apply AI to public health. The
scientific collaborations will focus on critical health problems that are ideally suited for AI
solutions. Initial areas of study are expected to include the use of AI to improve the utility
of electronic health records (EHRs) and claims data to address significant public health
issues like patient safety, precision medicine and health equity. The research will also
explore physician and patient user experience and interactions with AI technologies.
In March 2019, IBM reported on its research whereby a blood-based AI could help with
early Alzheimer's detection. IBM suggests that the test can detect a person's risk for the
disease decades before symptoms start showing up. IBM researchers have harnessed the
powers of machine learning to figure out a way to detect the universally recognized
Alzheimer’s marker, amyloid beta with a simple blood test as an early substitute for
analysis of cerebrospinal fluid. IBM said: While testing actual spinal fluid will obviously
be more accurate, IBM believes its blood test can help predict Alzheimer's risk with an
accuracy of up to 77%. The method is still in the very early stages of testing, but in the
future, scientists could also use it to find appropriate subjects for programs developing a
cure. The research can be found in Nature, Scientific Reports, March 11, 2019.

Blood Typing and Screening
Automated blood typing systems employ similar biochemical mechanisms to classify blood
but replace the human component with artificial intelligence. Several devices are currently
available, which generally utilize microtiter plates to determine ABO and Rh types, solid
phase red cell adherence assays for antibody detection and/or modified solid phase red cell
adherence plates for IgG cross matches. Interpretation of the reactions is done by artificial
intelligence software.
Automated blood typing has several advantages over manual typing in terms of cost,
efficiency and errors. First, a shortage of health care personnel often requires transfusion
services to utilize cross-trained, less-experienced staff from other areas of the laboratory.
Use of automation may allow the blood bank to conduct tests without need for costly
specialized technical staff. Furthermore, since interpretation of the immunohematology
reaction is done by the instrument without subjective interpretation by the technologist,
human errors are reduced.
Cancer
Freenome: Computational Biology and Machine Learning for Cancer Detection
Freenome is a privately held startup company which is developing artificial intelligence

(AI)-based liquid biopsy tests for the early detection of cancer.
Using thousands of cancer-positive blood samples, the company’s AI platform is being

trained to recognize biomarker patterns that represent cancer type, stage, and effective
treatment options. Freenome’s AI assesses multiple analytes from blood, cfDNA, cfRNA,
and circulating proteins, to identify complex patterns associated with the body’s response to
specific tumor types, which may include immune response signals.
Freenome’s multiomics platform detects key biological signals from a routine blood draw.
The platform integrates assays for cell-free DNA, methylation, and proteins with advanced
computational biology and machine learning techniques to identify additive signatures that
improve the accuracy for early cancer detection given the molecular subtypes of cancer are
heterogeneous in nature. This strategy incorporates a multidimensional view of both tumor-

and immune-derived signatures that enables the early detection of cancer, instead of relying
only on tumor-derived markers, which may miss the early signs of cancer.
Freenome is developing a liquid biopsy test for the detection of colorectal cancer. The test
is based on the company’s AI platform, which is being trained to recognize biomarker
patterns that represent cancer type, stage, and effective treatment options.
In October 2018, the company showed initial clinical data from a large international,
multicenter study involving 1,253 samples. Freenome’s AI platform was able to detect
colorectal cancer with relatively high sensitivity (82%) and specificity (85%). The company
is conducting further studies for colorectal cancer, and is currently enrolling patients for its
AI-EMERGE multicenter study.
Freenome plans to get its colorectal cancer screening blood test through a validation study
and expand its blood-based testing platform driven by artificial intelligence (AI) into other
tumor types, as well as other disease areas, with the help of $160 million raised in a series
B financing round. Including the latest financing round -- led by RA Capital Management
and Polaris Partners – the company has raised a total of $238 million. Colorectal cancer is
the first target for the company's AI platform for detecting cell-free biomarkers in blood. In
a poster presentation at Digestive Disease Week, the company reported that its prototype
blood-based test using cell-free DNA (cfDNA) had high sensitivity and specificity for
detecting stage I/II cancers. The AI-EMERGE validation study is evaluating Freenome's
approach prospectively.
Results will be submitted for parallel review by the U.S. Food and Drug Administration
and the U.S. Centers for Medicare and Medicaid Services. A test that incorporates the
analysis of genetic mutations in pancreatic cyst fluid and artificial intelligence (AI)
algorithms is helpful for making the call on benign findings, with potential to spare patients
from unnecessary surgeries, researchers reported in Science Translational Medicine July
2019.
In May 2018, a partnership between Freenome and QIAGEN for the accelerated
development and commercialization of NGS-based companion diagnostic tests was
announced. The collaboration pairs Freenome’s AI technologies with QIAGEN’s Sample to
Insight solutions to address the need for molecular diagnostic tests to predict the therapeutic
benefit of immune-oncological therapies for individual patients. Under the partnership
agreement, the two companies will co-market the resulting solutions.

CompCyst: Big Increase in Identification of Benign Cysts
A test that incorporates the analysis of genetic mutations in pancreatic cyst fluid and
artificial intelligence (AI) algorithms is helpful for making the call on benign findings, with
potential to spare patients from unnecessary surgeries, researchers reported in Science
Translational Medicine, July 2019. The test – dubbed comprehensive cyst or CompCyst –
incorporates mutation status, loss of heterozygosity, aneuploidy, protein markers, and
clinical data. An artificial intelligence algorithm is used to help identify which patients
should have surgery and which ones can be followed safely.
In a retrospective, multicenter study of 862 patients with pancreatic cysts who had surgery,
the test added useful information beyond what is used to manage patients in standard care --
that is, clinical features, nonspecific fluid tests, and imaging studies -- for determining if
cysts were benign or malignant, reported Dr. Bert Vogelstein, co-director of the Ludwig
Center at Johns Hopkins Kimmel Cancer Center, and colleagues in Science Translational
Medicine. Standard methods for characterizing cysts identified 19% of patients with a
benign cyst, whereas adding CompCyst resulted in the correct identification of 60%. For
those with mucin-producing types of cysts and no dysplasia or invasive cancer, standard
clinical management identified 34% of benign cysts, compared with 49% for CompCyst.
Grail: Machine Learning on Cancer Mutations
GRAIL Inc. (Menlo Park, CA), a cancer genomics company, has begun what they call a
large-scale study, using machine learning to set up a huge library of healthy mutations and
cancer mutations, using data from white blood cell genomes and cfDNA. The data will be
used to train their new cancer screening algorithms. GRAIL is combining high-intensity
sequencing, leading-edge computer science, and large population-scale clinical studies to
develop a blood test for early-stage cancer detection. GRAIL aims to develop a blood test
to detect cancer early before symptoms appear offering higher survival rates compared to
late-stage diagnosis.
Lung Cancer Diagnostics – LungLife AI
LungLife AI, founded in 2008 and based in Thousand Oaks, California (formerly
Cynvenio), is a cancer diagnostics company focused on transforming cancer diagnosis and
management through the artificial intelligence (AI) enabled molecular analysis of cancer

biomarkers in blood. The company's testing methodology is designed to deliver actionable

information to clinicians via minimally invasive blood draw. Their tests span all stages of
lung cancer, from an aid in diagnosis for patients with suspicious lung nodules as well as
treatment stratification and monitoring in late-stage lung cancer.
LungLB test is LungLife AI’s liquid biopsy-based molecular cancer assay for early stage
diagnosis and confirmation for clinicians whether to refer onto biopsy where the low-dose
computed tomography (LDCT) scan reveals suspicious nodules.
LungLife AI’s LiquidBiopsy Platform performs isolation, enumeration, and molecular

analysis of rare cells and cell-free DNA (cfDNA) from whole-blood samples; with a
turnaround time of 48 hours from blood sample to genomic data. The system incorporates
the isolation and molecular characterization by immunofluorescence labeling of multiple
analytes, namely CTCs, ctDNA, and germline DNA, to provide a more comprehensive
evaluation of a liquid biopsy sample.
LungLife AI OncoPrep isolation kits perform detection and molecular characterization of

circulating tumor cells (CTCs) in the background of normal and other nucleated cells.
In December 2019, LungLife AI announced that the company had received certification
from the Clinical Laboratory Improvement Amendments of 1988 (CLIA). Under the
“deemed status” provision of the US Department of Health and Human Services (HHS)
Center for Medicare and Medicaid Services (CMS), the state of California has granted the
company’s initial CLIA clinical laboratory license.
LungLife AI is currently developing a series of companion diagnostics, for the prediction of

likely therapeutic responses to later stage treatments in patients with advanced cancer.
Cardiac Disease
Eko’s AFib and Heart Murmur Detection Algorithms
In January 2020, Eko, a digital health company applying artificial intelligence in the fight
against heart disease, announced that the U.S. Food and Drug Administration had cleared a
suite of algorithms that, when combined with Eko’s digital stethoscopes, will enable
healthcare providers in the U.S. to more accurately screen for heart conditions during

routine physical exams. If left undiagnosed, these heart conditions can lead to stroke and
heart failure.
Eko’s AI is able to identify heart murmurs, a leading symptom of valvular heart disease,
with 87% sensitivity and 87% specificity. In comparison, a recent study (European Heart
Journal, Volume 38, Issue suppl_1, August 2017, ehx493.P5437,
https://doi.org/10.1093/eurheartj/ehx493.P5437) revealed that using traditional
stethoscopes, primary care physicians had a sensitivity of 43% and a specificity of 69% for
detecting significant valvular heart disease, which affects over 5 million Americans.
The AI is able to detect AFib with 99% sensitivity and 97% specificity when analyzing the
1-lead ECG tracing from the Eko DUO stethoscope. The integration of ECG into the
stethoscope enables providers to quickly screen patients for the serious arrhythmia during a
standard physical exam.
The algorithm also reports heart rate and QRS duration and identifies tachycardia and
bradycardia, abnormally fast and slow heart rates, which can be indicative of heart disease
or other health conditions such as thyroid disease.
Eko’s AFib and Murmur screening algorithms are the first in a suite of cardiac screening
algorithms Eko plans to combine with its digital stethoscope devices to assist providers in
the detection of cardiovascular conditions. In December 2019, Eko announced the FDA had
granted the company breakthrough status for a novel ECG-based algorithm that, if FDA-
cleared, could provide an easily accessible screening test for heart failure.
Machine Learning to Assess the Likelihood of Acute Myocardial Infarction
Abbott is continuously utilizing new technologies, such as AI and machine learning, to

create innovative solutions in healthcare. In September 2019, Abbott announced new
research, published in the journal Circulation, which found its algorithm could help doctors
in hospital emergency rooms more accurately determine if someone is having a heart attack
or not, so that they can receive faster treatments or be safely discharged.
In this study (Than, MP et al. Circulation. 2019; published online Sept 10.
https://doi.org/10.1161/CIRCULATIONAHA.119.041980), researchers from the U.S.,
Germany, U.K., Switzerland, Australia and New Zealand looked at more than 11,000
patients to determine if Abbott's technology developed using AI could provide a faster,

more accurate determination that someone is having a heart attack or not. The study found
that the algorithm provided doctors a more comprehensive analysis of the probability that a
patient was having a heart attack, particularly for those who entered the hospital within the
first three hours after their symptoms started.
Bay Echo Labs: FDA-cleared EF Calculation
Bay Echo Labs develops software algorithms that learn to select best image quality clips for
Left Ventricular Ejection fraction calculations. The system received 510(k) clearance from
the U.S. Food and Drug Administration (FDA) for the fully automated clip selection and
calculation of left ventricular ejection fraction (EF). EF is the single most widely used
metric of cardiac function and used as the basis for many clinical decisions. The EchoMD
AutoEF algorithms eliminate the need to manually select views, choose the best clips and
manipulate them for quantification, an often time-consuming and highly variable process.
Genetic Testing
Although there are approximately 6,000 genes associated with disease, just a fraction are
considered medically actionable or have a therapeutic option. According to research by the
American College of Medical Genetics and Genomics (ACMG), some 59 genes can be
considered medically actionable genes. That is, clinically significant changes found in one
of the genes greatly increase a person’s likelihood of developing the associated health risk,
which range from high cholesterol to breast, ovarian or colon cancer. At the same time,
there are more than 76,000 different genetic tests available to consumers and health care
professionals.
Most disease processes involve a number of genes and cell pathways. New multi-gene
diagnostics tests by whole genome and next generation sequencing and which examine
thousands of specific gene sequences might one day hold the key to assessing disease risk,
diagnosing diseases, and guiding precision medicine treatment decisions. However,
artificial intelligence algorithms are needed to interpret the mass of test data produced by
these tests.

GRAIL
GRAIL Inc. (Menlo Park, CA), a cancer genomics company, has begun what they call a
large-scale study, using machine learning to set up a huge library of healthy mutations and
cancer mutations, using data from white blood cell genomes and cfDNA. The data will be
used to train their new cancer screening algorithms. GRAIL is combining high-intensity
sequencing, leading-edge computer science, and large population-scale clinical studies to
develop a blood test for early-stage cancer detection. GRAIL aims to develop a blood test
to detect cancer early before symptoms appear offering higher survival rates compared to
late-stage diagnosis.
Rady Children’s Institute for Genomic Medicine
In April 2019, researchers at Rady Children’s Institute for Genomic Medicine (RCIGM; the
research arm of Rady Children’s Hospital-San Diego) have utilized automated machine-
learning and clinical natural language processing (CNLP) to diagnose rare genetic diseases
in record time. This new method is speeding answers to physicians caring for infants in
intensive care and opening the door to increased use of genome sequencing as a first-line
diagnostic test for babies with cryptic conditions.
A new study documenting the process was published in the journal Science Translational
Medicine. The workflow and research were led by the RCIGM team in collaboration with
leading technology and data-science developers — Alexion, Clinithink, Diploid, Fabric
Genomics and Illumina.
The RCIGM team has pioneered a rapid Whole Genome Sequencing process to deliver
genetic test results to neonatal and pediatric intensive care (NICU/PICU) physicians to
guide medical intervention.
By reducing the need for labor-intensive manual analysis of genomic data, the supervised
automated pipeline provided significant time-savings. In February 2018, the same team
achieved the Guinness World Record™ for fastest diagnosis through whole genome
sequencing. Of the automated runs, the fastest times – averaging 19 hours – were achieved
using augmented (artificial) intelligence.

RCIGM has optimized and integrated several time-saving technologies into a rapid Whole
Genome Sequencing (rWGS) process to screen a child’s entire genetic makeup for
thousands of genetic anomalies from a blood sample.
Key components in the rWGS pipeline come from Illumina, a global leader in DNA
sequencing, including Nextera DNA Flex library preparation, whole genome sequencing
via the NovaSeq 6000 and the S1 flow cell format. Speed and accuracy are enhanced by
Illumina’s DRAGEN (Dynamic Read Analysis for GENomics) Bio-IT Platform.
Other pipeline elements include Clinithink’s clinical natural language processing platform
CliX ENRICH that quickly combs through a patient’s electronic medical record to
automatically extract comprehensive patient phenotype information.
Another core element of the machine learning system is MOON by Diploid. The platform
automates genome interpretation using AI to automatically filter and rank likely pathogenic
variants. Deep phenotype integration, based on natural language processing of the medical
literature, is one of the key features driving this automated interpretation. MOON takes five
minutes to suggest the causal mutation out of the 4.5 million variants in a whole genome.
In addition, Alexion’s rare disease and data science expertise enabled the translation of
clinical information into a computable format for guided variant interpretation.
As part of this study, the genetic sequencing data was fed into automated computational
platforms under the supervision of researchers. For comparison and verification, clinical
medical geneticists on the team used Fabric Genomics’ AI-based clinical decision support
software, OPAL (now called Fabric Enterprise) – to confirm the output of the automated
pipeline. Fabric software is part of RCIGM’s standard analysis and interpretation
workflow.
The study titled “Diagnosis of genetic diseases in seriously ill children by rapid whole-
genome sequencing and automated phenotyping and interpretation,” found that automated,
retrospective diagnoses concurred with expert manual interpretation (97 percent recall, 99
percent precision in 95 children with 97 genetic diseases).
Researchers concluded that genome sequencing with automated phenotyping and

interpretation – in a median 20:10 hours – may spur use in intensive care units, thereby
enabling timely and precise medical care.

An estimated four percent of newborns in North America are affected by genetic diseases,
which are the leading cause of death in infants. Rare genetic diseases also account for
approximately 15 percent of admissions to children’s hospitals.
The RCIGM workflow is engineered to speed and scale up genomic data interpretation to
reduce the time and cost of whole genome sequencing. The team’s goal is to make rWGS
accessible and available to any child who needs it.
Increased automation of the process removes a barrier to scaling up clinical use of WGS by
reducing the need for time-consuming manual analysis and interpretation of the data by
scarce certified clinical medical geneticists. There were fewer than 1,600 of these experts
nationwide in 2017, according to the American Board of Medical Genetics and Genomics.
In March 2020, Blue Shield of California became the first health plan in the United States
to cover rapid and ultra-rapid Whole Genome Sequencing to help critically ill babies and
children in intensive care with unexplained medical conditions receive precision care.
Rady Children’s Institute began performing genomic sequencing in July 2016. As of the
end of March 2019, the team had completed testing and interpretation of the genomes of
more than 750 children. One-third of those children have received a genetic diagnosis with
25 percent of those benefitting from an immediate change in clinical care based on their
diagnosis.
Pathway Genomics
Founded in 2008, Pathway Genomics offers digital healthcare and genetic testing. Based in
San Diego, their CLIA- and CAP-accredited clinical laboratory provides physicians and
patients in 40+ countries with actionable and accurate precision healthcare information to
improve health and wellness. The company’s program with IBM Watson is a smartphone
app that merges artificial intelligence and deep learning with personal genetic information.
The app provides users with personalized health and wellness information based on the
individual’s health history.
In January 2018, Pathway Genomics signed an agreement with United Crest Healthcare.
United Crest Healthcare is a leading mobile health platform developer and operator, with its
imedic Cloud Hospital used by thousands of doctors and patients in the US, Russia, China,

and other Asian countries. imedic is powered by advanced artificial intelligence and
knowledge-based technology for comprehensive analysis of users’ health data.
As part of United’s strategic development imedic Cloud Hospital eco-system, United is

partnering with Pathway to offer personalized DNA tests as part of the imedic healthcare
services to their patients in working with hospitals, doctors and patients in Asia.
The DNA tests are easily collected with buccal swab at home or at the nearest participating
clinics to be sent to Pathway’s clinical laboratory for processing. Pathway produces results
of the DNA test with a patient report in 2-3 weeks on the mobile APP of imedic or online at
Pathway.com.
In August 2017, Pathway Genomics forged a distribution agreement with Medikonia, a

solution-based genomics service provider in Hong Kong, to bring its wide range of clinical
genetic testing with advanced artificial intelligence analysis to Hong Kong.
In January 2016, Pathway Genomics and IBM unveiled the Pathway Genomics OME™
app, powered by Watson. The Pathway Genomics OME app merges cognitive computing
and deep learning with precision medicine and genetics to enable Pathway Genomics to
provide consumers with personalized wellness information.
Glucose Monitoring
Interactive glucose meters are the most commonplace application of AI-incorporated

devices. There is general consensus that this application of AI has greatly improved the life
of many people with diabetes. AI-driven glucose management devices help diabetic users
interpret and use their test results to manage their diabetes.
Smart glucose systems comprise 18% of the market and that is expected to increase as more
continuous glucose monitoring (CGM) systems incorporate AI features that consumers
adopt. Discussed here are two recent developments in AI-driven glucose management.
Medtronic’s Guardian Connect CGM System
In March 2018, the Food and Drug Administration approved Medtronic’s Guardian
Connect continuous glucose monitoring (CGM) system for use in people with diabetes
between the ages of 14 and 75 years old.

The first CGM to use artificial intelligence for this purpose, the Guardian Connect system
is intended to help people with diabetes who use multiple daily injections of insulin to
manage their diabetes by helping them prevent hyperglycemia and hypoglycemia,
according to a statement from Medtronic.
Guardian Connect utilizes a predictive algorithm that alerts patients of significant swings in
blood glucose levels up to 60 minutes prior to the event. When combined with the Guardian
Sensor 3, which is placed on the abdomen to monitor blood glucose levels, the Guardian
Connect system was accurate and was able to alert patients about 98.5% of hypoglycemic
events, according to the results of a clinical trial. This information can also be shared and
monitored with care takers and family member in real time or via text message.
The Guardian Connect system provides access to the Sugar.IQ smart diabetes assistant.
Utilizing artificial intelligence technology from IBM Watson Health, the Sugar.IQ assistant
continually analyzes how patient’s blood glucose levels respond to factors like food intake,
insulin dosages, and daily routines. The combination of continuous monitoring and real-
time analysis may help to identify patterns and lead to personalized insights that will help
patients with diabetes keep their blood glucose levels under control.
University of Warwick Develops A.I. Technology for Tracking Blood Glucose
A new technology for detecting low glucose levels via ECG using a non-invasive wearable
sensor, which with the latest artificial intelligence can detect hypoglycemic events from
raw ECG signals, has been made by researchers from the University of Warwick.
Dr Leandro Pecchia’s team at the University of Warwick in January 2020 published results
in a paper titled “Precision Medicine and Artificial Intelligence: A Pilot Study on Deep
Learning for Hypoglycemic Events Detection based on ECG” in the Nature Springer
journal Scientific Reports, proving that, using the latest findings of artificial intelligence
(i.e., deep learning), they can detect hypoglycemic events from raw ECG signals acquired
with off-the-shelf non-invasive wearable sensors.
Two pilot studies with healthy volunteers found the average sensitivity and specificity
approximately 82% for hypoglycemia detection, which is comparable with the current
CGM performance, although non-invasive.

Hematology
On a par with clinical chemistry as far as product maturity is hematology. Today and over
the past 50 years, the complete blood count (CBC) has made an enormous impact on patient
care. The information clinicians derive from the CBC is used in the evaluation of just
about every medical condition.
Similar to their histology peers, hematology lab managers and hematologists look to
technology that can reduce costs and shorten turnaround time while maintaining or
increasing quality. Digital imaging designed for hematological analyses is a technology by
which a camera, combined with appropriate optics and illumination, creates a numerical
representation of a physical object, in this case blood cells. Once in a digital format, the
image can be analyzed, displayed, printed, or manipulated in ways that might not otherwise
be possible.
Hematology analyzers participate in the automation and AI movement by adding new

parameters to a basic instrument—such as platelet counting, reticulocyte counting and
CD4/CD8 analysis. Another advantage is the expansion into body fluids other than blood.
The third avenue is in automation strategies that include integration onto a core lab
automation track, the addition of an automated slide maker and stainer, cell imaging
technology and bioinformatic software. These strategies have already been put in place by
all of the major hematology companies.
Thus hematology vendors have perfected high-end hematology analyzers to perform a

number of specialized tests that are now available on-demand and almost instantaneously.
Using the cell staining and counting capabilities of high-end hematology analyzers allows
labs to automate fluid analyses such as CSF. Abbott Diagnostics, Beckman
Coulter/Danaher, Siemens Healthineers, Sysmex, and Horiba offer specialized platelet, red
blood cell and white blood cell analysis and automatic slide makers/stainers. This feature
decreases the number of differential slides that have to be made to only those samples with
really abnormal cells
Leader in this area is Sysmex (Tokyo, Japan) that markets the Hematoflow automatic
hematology analyzer that includes a DxH 800 Coulter Cellular Analysis System and FC
500 Flow Cytometer from Beckman Coulter (Brea, CA and now part of Danaher). The
target is for leukemia testing but in effect the combined instrument could test for a whole
range of cell markers including: CD4/CD8/T & B lymphocytes, and CD34. The instrument

can also be used for chemotherapy monitoring for leukemia and lymphoma; HLA Typing,
chromosomal abnormalities and fetal cell analysis.
In the past few years hematology has joined the march to more automation and
management IT solutions, especially for the differential count. Facilitating digital
hematology is automated slide staining that optimizes quality and enables standardization
to help ensure accurate identification of white blood cells, red blood cell morphology and
platelet assessment. Standardized stained slides are immediately available for digitization
and viewing upon completion of staining,
The lead company in this area is CellaVision (Lund, Sweden). CellaVision has distribution
agreements with the major hematology companies including: Abbott Diagnostics, Beckman
Coulter/Danaher, Horiba, Siemens Healthineers and Sysmex.
The company’s major products are the CellaVision DM series of cell imaging and analysis
systems that were introduced in 2000. The latest instrument, the CellaVision DM1200 cell
image analysis system locates, identifies, and pre-classifies blood cells by means of
automatic microscopy and advanced image processing and thus presents a new way of
handling manual differentials traditionally performed on a microscope. The analyzer pre-
classifies the white blood cells, pre-characterizes parts of the red morphology and provides
functionality for platelet estimation.
For small to mid-sized labs CellaVision developed the CellaVision Image Capture System,
a cell morphology system for networked hospitals and associated laboratories. The System
is designed to utilize a laboratory's existing microscope. CellaVision software helps the
technologist to find, focus and capture cell images, and then transmits the digital cells to a
CellaVision DM96 or DM1200 in a main lab at another location where the review is
performed. The product gives laboratories within the same network the ability to share,
review and store their manual differential counts.
A second major image analysis system for hematology is offered by Medica Corp.
(Bedford, MA). Unlike Cellavision, Medica also sells offers blood testing analyzers
including: the EasyLyte, ISE Module, EasyBloodGas, EasyElectrolytes, EasyStat blood gas
analyzer, EasyRA chemistry analyzer and EasyCell assistant hematology imaging system.
Medica also markets the EasyCell assistant cell image analyzer for hematology
laboratories. The EasyCell solution automates and simplifies the cell differential process
by locating and preclassifying white blood cells. The EasyCell assistant uses optical

pattern recognition software to automatically locate white cells on a blood smear. It

digitally stores their images, pre-classifies them, and then presents them for review —
grouped by cell type — on an LCD display. The assistant also presents images of red cells
and platelets, enabling red cell morphology and platelet estimates.
Optional EasyCell Remote Review Software allows the networking of multiple EasyCell
assistant workstations, both onsite and in outside clinics and practices.
CellaVision’s success has attracted other entrants. West Medica (Austria) has developed the
Vision line of solutions for digital microscopy for hematology, cytology, cytogenetics,
microscopy and slide preparation. The company’s Vision Hema Capture can be installed
on the lab’s existing personal computer and microscope. This system allows the
technologist to capture the images of blood cells: leucocytes, erythrocytes, and platelets.
The technologist forwards the captured cell images to centralized laboratory for further
processing on Vision Hema hematology imaging analyzers.
Fully digital hematology received a boost when Roche Diagnostics acquired Constitution
Medical, Inc. (CMI, Westborough, MA). Constitution Medical has developed the
Bloodhound Integrated Hematology Analyzer, a totally digital concept. The instrument
addresses the entire core lab hematology station workflow – slide making; slide staining;
complete blood count; differentials; and streamlined, software-assisted manual review of
slided, stained and imaged cells. The Bloodhound System combines a digital image-based
cell locator, cell classifier and cell counter with its own slide maker and stainer. It produces
a CBC and 5-part differential and automatically analyzes reticulocytes. Following
processing, a comprehensive set of results are provided to medical technologists through
the Bloodhound Viewing Station, an integrated, interactive display that features cell
galleries that sort white blood cells, red blood cells and platelets. The system also isolates
unclassified cells of interest and presents them for classification.
Roche offers the cobas m 511 integrated hematology analyzer that features Bloodhound
technology. The cobas m 511 combines a digital morphology analyzer, cell counter and
classifier into one streamlined instrument preparing, staining and analyzing microscopy
blood slides.
Since the late 1990s hematology cell counting progressed from impedance-based
technology (the basis of the Coulter Principle) by borrowing from flow cytometry.
Incorporation of laser excitation and fluorescence detection into the hematology analyzer

has grown the instruments’ cell discrimination abilities to include blood cell components
that required labor-intensive manual methods.
Histology
Histology slide analysis is highly labor intensive and subjective in nature. At least this was
the scenario for this market segment until the commercialization of automation and digital
image analysis. Now histology is following the trend already in action in other test
segments and is investing in end-to-end automation of the entire tissue process. From slide
preparation through to analysis, all the way to result reporting, combined solutions enable
faster turn-around time, while providing standardized results.
The growing number of cancer patients is a major factor, but a number of technological
advancements also contribute to the growth. Sample preparation of histology samples has
always been fairly automated, but companies have invested to improve upon this with new
faster tissue fixatives and systems that improve nucleic acid extraction from tissue samples.
Of note is new instrumentation to automate RNA detection in fixed tissues. This is
important for advanced companion tests.
Automation of histology has become fairly routine in the developed world and will
gradually spread to the larger urban centers in developing countries. Automated sample
processing and slide strainers have been available for some 10 to 15 years; however, end-
to-end automation, digital optics and the artificial intelligence algorithms and high capacity
computer power necessary to really make these systems work efficiently have only recently
become available.
Innovation in immunohistochemistry (IHC) staining has evolved significantly over the last
two decades. The process of staining has shifted from labor-intensive, manual techniques
toward semi-automated instruments with off-line processes such as deparaffinization and
antigen retrieval, and now to a new generation of baking-through-staining automated
instrumentation systems that enable standardization, improved staining consistency, and
expedited turnaround times.
The Dako Omnis automation solution comprises hardware, software and reagents. The
Dako Omnis offers full automation to satisfy the needs of large-scale laboratories, hospitals
and universities. The Dako Omnis automates IHC and ISH processes, meaning entire

patient cases can be processed simultaneously with continuous loading, batch loading and
overnight run options. Also, processes such as slide processing and instrument maintenance
can be traced to individual operators. The system was developed in collaboration with The
Tecan Group. Dako also offers the DakoLink software for use in Dako’s Autostainer and
Artisan instruments.
Sakura Finetek (Japan) has been a leader in tissue sample preanalytical processing. Sakura
Finetek introduced the Tissue-Tek AutoTEC a120 Automated Embedder, a fully automated
tissue embedder that features a cassette barcode reader for LIS (laboratory information
system) connectivity and traceability of specimen blocks, and SMARTair technology that
optimizes block quality. The AutoTEC technology combined with the Paraform
Sectionable Cassette System comprises a fully automated instrument—offering high-
throughput of up to 120 cassettes per hour—that locks the orientation of specimens, thereby
increasing patient safety by eliminating the risk of orientation mistakes and tissue loss.
In June 2019, Sakura Finetek USA launched 12 new ready-to-use antibodies for the Tissue
Tek Genie Advanced Staining System.
Up until recently there is one part of the tissue analysis process that had yet to be
automated; that is tissue block cutting to prepare sections that are placed on a glass slide to
be stained and analyzed by microscopy. MEDITE Cancer Diagnostics, Inc. (Germany) has
developed the patented Histology "Lab-in-one Automation" System, which represents a
huge leap in automation, once available. The system uses robotic and laser technology
combined into a new closed and computer-controlled system including an automated
sectioning device.
MEDITE's patent serves again as the forerunner in transitioning the industry towards
complete automation. The company’s "Lab-in–One Automation" technology (once more
widely distributed) will revolutionize the automation of pathology laboratories by
combining many disparate systems into one complete automated histology laboratory.
Biopsies are automatically processed into microscopy slides ready for diagnosis and
digitalization. The "Lab-in-one Automation" fully automates 24 by 7 tissue processing,
embedding, sectioning, staining, cover-slipping and digital slide scanning. The company
expects to start marketing the "Lab-in-One Automation" systems in the next 2-3 years
contingent upon the availability of developmental resources.

Hørsholm, Denmark-based Visiopharm introduced its Oncotopix and Biotopix software

module portfolios, which offer analytical algorithms that streamline and improve tissue
pathology through automation of histology readings. The former is optimized explicitly for
cancer diagnostic applications and can also be used for cancer research and drug discovery.
Among the modules included in the suite is Visiopharm’s CE-marked Virtual Double
Staining (VDS) Ki67 module, initially designed for use in breast cancer applications.
BioGenex Laboratories, Inc. (Fremont, CA) announced in January 2017 the launch of the
universal, one-step FFPE sample preparation system, EZ-AR2 Elegance, for high-
throughput applications, handling de-waxing, hydration, and antigen retrieval, all in a
single step. EZ-AR2 is capable of processing 100 slides in 30 minutes; when used in
conjunction with the EZ-Retriever microwave system or the fully automated Xmatrx
system, it can produce excellent IHC staining results for over 400 antibodies of varying
immunoglobulin isotypes and species.
The introduction of information technology and smart phone applications has come to
histology. Market leader Roche Diagnostics/Ventana Medical further solidified its position
with the introduction of workflow tools that allow pathologists and lab managers to track
the progress of a slide from preparation to digitization along with other critical slide
information. This facilitates the exchange of information between pathologists and lab
managers, resulting in a more streamlined workflow.
PathAI and Gilead Evaluating Machine Learning Approaches to Liver Histology Assessment
November 2019: In a collaboration with PathAI, a leader in AI-powered research in

pathology, Gilead is evaluating machine learning approaches to liver histology assessment
for use in the diagnosis and staging of Non-alcoholic steatohepatitis (NASH) and
monitoring of treatment response in clinical trials. A study of images from liver biopsies
from patients screened for the Phase 3 STELLAR program compared the staging and
characterization of liver disease as assessed by experienced pathologists and by the PathAI
research platform. The pathologists scored biopsies using the NASH Clinical Research
Network (CRN) and Ishak fibrosis classifications, and the PathAI research platform, a
convolutional neural network, evaluated these biopsies following training on more than
68,000 annotations from 75 board-certified pathologists.
The results showed that the machine learning models and the consensus of readings from
the independent pathologists demonstrated high consistency for the key histologic features

of NASH. Importantly, for the staging of fibrosis, the predictions of the machine learning
model were highly correlated with those of the central pathologist for both the NASH CRN
(rs=0.83) and Ishak (rs=0.86) staging systems.
In a separate analysis, machine learning models were developed to recognize patterns

associated with each fibrosis stage, using slide-level pathologist Ishak fibrosis stages.
Images of liver biopsies from 674 patients with compensated cirrhosis (F4) enrolled in the
Phase 3 STELLAR-4 clinical trial demonstrated that machine learning models are
predictive of disease progression, illustrate the heterogeneity of fibrosis in NASH cirrhosis,
and correlate with noninvasive markers of fibrosis. These data highlight the potential of
machine learning models to characterize patients with cirrhosis beyond conventional
histological staging.
In April 2019, PathAI announced that it had raised $60 million in Series B funding. The
round was led by new investor General Atlantic, a leading global growth equity firm, with
strong participation from General Catalyst and other existing investors. The new capital is
to fuel PathAI’s continued expansion as the company seeks to advance the medical
discipline of pathology to advance its mission of offering faster, safer, and more powerful
solutions for the diagnosis and sub-typing of diseases like cancer.
PathAI’s plan for this new capital is to enhance offerings to existing partners, drive
continuous improvement of its flagship pathology research platform, meet market demands,
and fuel research and development into new tools and medical devices. Its growing
partnerships with global pharmaceutical companies seek to accelerate drug development in
life-saving therapeutics, and its partnerships with leading diagnostic laboratories aim to
support pathologists in bringing faster, more accurate, and more predictive diagnostics to
patients.
Information Technology
Laboratory IT and clinical scoring systems and algorithms have been used in medical
practice for some time; recently there has been an increase in the application of machine
learning (artificial intelligence) to improve these tools. While traditional algorithms require
all calculations to be pre-programmed, machine-learning algorithms deduce the optimal set
of calculations by searching for patterns in large collections of patient data.

There have been extraordinary advances in artificial intelligence, the end-game for
information technology. Similar to the product commercialization process that is evident in
the liquid biopsy market, IVD companies’ efforts have been supported by the US FDA’s
efforts to give products expedited review. This helps patients have more timely access to
devices and technology that provide more effective treatment or diagnosis for life-
threatening diseases. More products have been cleared for clinical use, more new research
use only applications have come to market and many more are in development – companies
are collaborating to improve the power of data analysis for patient care. Over the past
several years, healthcare has witnessed a transformation with a shift from paper-based
records systems to electronic records and incorporation of digital health monitoring devices
and advanced patient screening systems. These advances have resulted in a data explosion,
which can best be manipulated and analyzed using artificial intelligence (AI) technology.
Evidence to support AI's use in diagnostics is building: an August 2019 in JAMA Network
Open found that a novel artificial intelligence algorithm outperformed pathologists in
differentiating ductal carcinoma in situ (DCIS) from atypical hyperplasia, a job that is
"considered the most diagnostically difficult task in clinical practice," researchers wrote in
the paper published online. The researchers compared the performance of machine learning
with the interpretation of digitized images by 87 practicing pathologists. A machine-
learning algorithm that analyzed two features (tissue distribution and structure) had lower
sensitivity than the pathologists when it came to distinguishing between invasive and
noninvasive cancer.
Google
Mayo Clinic, Google Partner on AI-Based Diagnostics, Cloud Computing
Mayo Clinic and Google announced a 10-year strategic partnership in September 2019.
This partnership hopes to redefine how health care is delivered and accelerate the pace of
health care innovation through digital technologies.
Mayo Clinic selected Google Cloud to be the cornerstone of its digital transformation.
Mayo will use advanced cloud computing, data analytics, machine learning and artificial
intelligence (AI) to redefine health care delivery, bringing together global providers and
consumers to make health care better.

The extensive collaboration will combine the world-class clinical knowledge and expertise
from Mayo Clinic with the industry-leading digital technology and data security expertise
from Google to develop solutions for people who want medicine to deliver on the promise
of scientific and technological advances.
With the help of Google, Mayo Clinic will transform the way it advances virtual care with
AI-enabled digital diagnostics. Mayo also will leverage Google technology to boost its
ability to conduct medical research. Through this partnership, Mayo Clinic will be able to
develop and deploy new machine learning models designed to improve treatment precision
and clinical outcomes of diseases.
Google Cloud will secure and store Mayo Clinic's data, while working with Mayo Clinic to
apply AI and other cloud computing technologies to solve complex healthcare problems.
However, Mayo Clinic will continue to control access and use of its patient data by using
Google's cloud technologies. Mayo Clinic will specifically authorize the use of data in
projects to create new healthcare insights and solutions in conjunction with partners,
including Google. Mayo Clinic will manage access to all data using rigorous long-standing
institutional controls.
Google DeepMind: Breaking the Black Box of Clinical Practice
Among many other projects, Google's DeepMind applied AI technology to a

comprehensive de-identified electronic health record dataset collected from a network of
over a hundred VA sites. The research shows that the AI could accurately predict acute
kidney injury (AKI) in patients up to 48 hours earlier than it is currently diagnosed.
Importantly, the model correctly predicted 9 out of 10 patients whose condition deteriorated
so severely that they then required dialysis. How physicians are making decisions based
on is often a black box. Deep Mind's model provides the clinical information that was most
important in making its predictions of deteriorating kidney function and provides predicted
future results for several relevant blood tests. This information may help clinicians
understand the reasoning behind the AI-enabled alert and anticipate future patient
deterioration.

Electronic Medical Records: Predictive Analytics and AI
Predictive analytics and machine learning are two areas that can go hand-in-hand in
improving healthcare, including increasing the accuracy of diagnoses by using predictive
algorithms.
With the vast amounts of information that are contained in electronic medical records
(EMRs) and the intersecting of artificial intelligence to compile that information,
physicians, clinicians, nurses, patients and others will be able to make more informed
decisions about healthcare diagnosis, delivery and personal wellness and disease
management.
In February 2019, Cerner launched ChartAssist, a new AI-enabled workflow, to join a suite
of solutions designed to reduce physician burnout. Cerner’s suite of AI-enabled solutions
goes beyond traditional processing and focuses on user efficiency to identify gaps and
inconsistencies within the patient record. These advancements will help providers address
patient care and validate diagnoses, ultimately helping to reduce the physician workload,
while supporting the financial strength of the health system.
One of the more interesting applications of healthcare AI is the prevention of negative

consumer health outcomes via predictive behavioral modelling. Certain insurers, such as
Premera Blue Cross, are investing in AI analytics that can predict with a high degree of
certainty when a member is about to have a severe health event. Similarly, Cigna has
invested in Prognos, which applies AI to lab diagnostics. Analyzing a database of 14 billion
medical records, Prognos can predict when a customer is most likely to visit an emergency
room, need hip or knee replacement, or become at risk for depression. This potentially
allows the insurer to intervene with pre-emptive care.
Kidney Disease
KidneyIntelX: Improved Kidney Testing
In July 2019, RenalytixAI's KidneyIntelX test for predicting rapid kidney function decline
using blood-based biomarkers received a Current Procedural Terminology (CPT) code from
the American Medical Association (AMA) for use in submitting private and Medicare
insurance claims, effective in October 2019.

The company expects that the new CPT Proprietary Laboratory Analyses (PLA) code –
0105U – will help the U.S. rollout of its lead product enabled by artificial intelligence (AI).
A Medicare payment rate for the new code will be developed through the 2019 Clinical
Laboratory Fee Schedule process, and RenalytixAI expects that this will take effect at the
beginning of 2020.
The KidneyIntelX product tests for predictive biomarkers, including soluble tumor necrosis
factor receptor 1 (sTNFR1), soluble tumor necrosis factor receptor 2 (sTNFR2), and kidney
injury molecule 1 (KIM1), and it also takes aspects of patient history into account for its
predictive scoring. The goal is to identify rapid kidney function decline and address it early
enough to avoid kidney failure. The KidneyIntelX test received breakthrough device status
from the U.S. Food and Drug Administration in May.
RenalytixAI was founded in 2018 with an exclusive multiyear license with the Icahn
School of Medicine at Mount Sinai involving the commercialization of AI products to
detect, manage, and treat kidney disease in people with diabetes and other common
conditions. Among other things, the deal covers the use of deidentified data from more than
3 million patient health records from Mount Sinai, which is a shareholder in the company.
RenalytixAI is also developing the AI-enabled FractalDx blood test for use in predicting
early acute kidney rejection and personalizing the administration of immunosuppression
therapy in kidney transplant patients.
Liquid Biopsy
The testing of cell free and tumor DNA (liquid biopsy) is becoming an important tool for
early and metastatic cancer detection. The technology is also playing a part in pathogen
epidemiology.
Oncology faces two major challenges: detecting cancers when they are small and easier to
treat and offering cancer patients personalized targeted cancer care. The use of machine
learned algorithms are a perfect complement to liquid biopsy that aims to answer the
challenge by detecting cell free and tumor DNA in the sea of a patient’s peripheral blood.
A liquid biopsy test can detect traces of DNA, the meaning of, which are not always
obvious to the oncologist. Since all tumors are unique, liquid biopsy methods have yet to
accurately detect most cancers with absolute certainty.

Further, many cancers lack a simple set of known mutations that can identify the tumor and
so determine the best treatment option. Also, the development of the analysis of tumor
DNA and circulating tumor cells with sequencing techniques produces a large amount of
test data, some of which is not generally associated with the tumor site in question. This is
where machine learning comes into the picture.
Machine intelligence algorithms (AI) use neural networks to interpret data. These networks
can process large tracts of data and independently identify patterns of interest, similar to
how the human doctor’s brain would do, but faster and with more accuracy.
So far, machine-learning algorithms designed to work in conjunction with liquid biopsies

have performed well in clinical validation studies, but no AI algorithms been approved for
clinical use.
The potential for this technology mix is very strong. AI algorithms can improve in their
ability to perform as more data is put into the system. They rarely get confused. This gives
the approach of using AI to boost liquid biopsy an even clearer future in cancer diagnosis
and treatment.
Of particular interest is CareDx (Brisbane, CA) that is preparing to market at home and
internationally. CareDx markets several blood-based donor-derived cell-free DNA (dd-
cfDNA) test for the assessment of transplant organ health. In May 2019, the company and
Cibiltech (Paris, France) announced that they are partnering to commercialize Predigraft, a
data analysis tool that provides an early prediction of an individual's risk of allograft
rejection and transplant loss. Under the terms of the agreement, CareDx will hold the
exclusive rights to commercialize Predigraft in the U.S. and will also become a minority
equity owner of Cibiltech. The companies have also agreed to collaborate to advance the
development of artificial intelligence in transplant care. Predigraft was developed from
Cibiltech's proprietary software algorithm iBox, which was developed using outcomes data
from tens of thousands of transplant patients.
Microbiology
Quick and appropriate test results are of course important in all aspects of lab medicine.
However the imperative to respond to patient needs when an infection is present is most
critical for patient care. The collection of hospital-acquired infection rates, identification of

antibiotic-resistant organisms, and detection of inappropriate antibiotic use as well as the

selection of the most appropriate antibiotics are also dependent on information obtained
from the microbiology laboratory. The use of computers in medicine is relatively new but
many of the functions of the microbiology laboratory have already been improved by
computer methods. The integration of microbiology data to hospital information systems
and the application of artificial intelligence allow hospital personnel to continuously
monitor patients and identify situations that would not be feasible by manual methods.
Artificial intelligence in microbiology is now being used to produce faster results for labs.
Computer generated algorithms are helping to identify and count colonies and increase the
speed with which labs can accurately assess growth cultures.
A basic component of AI in microbiology is full laboratory automation. All automated

microbiology systems offer specimen processing, track, image acquisitions systems, smart
incubation, and workstations. The key difference in the systems revolves around the
strength and technology behind algorithms and the extent of use of artificial intelligence
(AI).
Automation aside, once a pathogen is isolated microscopes enhanced with artificial

intelligence could help clinical microbiologists diagnose pathogens and potentially deadly
blood infections faster. An automated AI-enhanced microscope system can identify images
of bacteria on microscope slides quickly and accurately. This approach illustrates the use
of machine learning in microbiology.
Company developed algorithms and AI can recognize and interpret colonies in a mixed
culture. They can then flag the sample and pass it on to the technologist for additional
investigation. This works by counting colonies accurately and then querying a
phenotypical database. User-defined results are then applied to the specimen based on its
source, location, or the individual patient’s demographic information. Proponents suggest
that this technology could form the basis of a future diagnostic platform that augments the
capabilities of the microbiology lab to improve patient care.
Microbial identification and antimicrobial susceptibility testing (ID/AST) is the end-point

of the majority of work a microbiology lab performs on a daily basis and is very time
consuming. Since the mid 1990s ID/AST vendors have responded to the need for quicker
test systems, and in fact for many sample types these systems can have a definitive test
result in 6 hours versus 1-2 days with non-automated systems. A differentiating factor

among these system is the reagent mix contained in the panels and bioinformatics that
provides more specific analysis if bacterial identification to the genus/species level.
The first and crucial step in automated microbiology is sample transport and preparation
steps so that the patient sample can be processed by automated systems. Microbiology
companies have invested in automation that speeds up the process, especially the pre-
analytical sample preparation steps.
The leader in this regard that has absolutely revolutionized ID/AST automation is Copan
Diagnostics (Murietta, CA). Copan is the leading supplier of microbiology sample
collection and transport systems in the world. The company’s products include patented
Flocked Swabs, ESwab, and UTM Viral Transport. Copan has increased its influence in
microbiology testing with the introduction of its Walk-Away Specimen Processor (WASP)
and WASPLab. WASPLab can be partnered with the company’s PhenoMATRIX and
Colibrí colony picking module. PhenoMATRIX provides artificial intelligence for rules-
based sorting that highlights culture plates of clinical interest. WASPLab also offers the
ability to view, workup, and report cultures remotely from anywhere with Internet access.
Copan’s WASPDT Walk-Away Specimen Processor automates all facets of specimen

processing in microbiology, including planting and streaking, Gram slide preparation, broth
inoculation and disk application to aid in rapid, direct identification.
The major microbiology companies have recognized the potential of full microbiology
automation and artificial intelligence. Some have partnered with Copan and others have
gone their own route. In all cases, the net result will be faster, more standardized ID/AST
testing.
bioMérieux (France) and Copan are working to further AI in microbiology. Their

partnership combines Copan’s strength in automation and pre-analytics with bioMérieux’s
leadership in microbiological diagnostics. The companies are dedicated to the development
of additional workflow optimization, imaging and software algorithms for new
microbiology solutions.
Siemens Healthineers (now Danaher/Beckman Coulter) and Copan have also entered
announced a partnership whereby laboratories can add Copan’s Walk-Away Specimen
Processor (WASP) to their Siemens MicroScan system. The company and bioMérieux
have developed information technology systems that connect ID/AST systems to the
laboratory information system.

Thermo Fisher Scientific markets the Trek Sensititre OptiRead Automated Fluorometric
Plate Reading System. The system provides fast plate reads and electronic transfer of test
results for processing, interpretation and report generation. The company’s Sensititre AIM
Automated Inoculation Delivery System is compatible with most 96 well microtitre plates
that work with dry or frozen antibiotic susceptibility formats. Thermo Fisher has also
partnered with Copan to automate all aspects of specimen processing. The Thermo
Scientific Automated Workflow Solution includes the COPAN WASPDT instrument.
Microscopy-based diagnostic parasitology has remained mostly unchanged for decades.

August 2019, ARUP Laboratories (Salt Lake City, UT), a national reference laboratory, and
Techcyte (Lindon, UT,), a developer of artificial intelligence (AI) based image analysis
solutions for the diagnostics industry, announced they have developed (what they report is)
the world’s first AI-augmented ova and parasite detection tool. Techcyte’s digital
diagnostics platform applies convolutional neural networks to pre-classify the fecal sample
images captured by a 3DHISTECH (Budapest, Hungary) Panoramic 250-Flash III scanner.
Pre-classifying the images using the Techcyte tool allows ARUP’s technologists to
efficiently read stained glass slides manually and improves the accuracy of parasite
detection. The companies have reported that the ova and parasite tool is the first among
several projects being co-developed by ARUP and Techcyte. They expect to produce high
quality algorithms that can be developed and applied to future unmet laboratory needs.
Adopters of full laboratory automation and microbiology AI say that the benefits are both
real and impressive. They include decreased turnaround times, reduced labor costs, and
increased antimicrobial stewardship. Although it’s stressed that these advances will never
replace human expertise or skill, more and more labs are recognizing the benefits of AI
software to both their labs and the patients that they serve.
Microbiome
The continuous search of the etiology of diseases as part of the body’s systemic response to
change has led to a consideration of the gut microbiome as part of disease processes. While
still in the early phase of development, a number of companies have developed assays for
the study of the microbiome and there is a proliferation of microbiome-related therapeutics.

Viome
Viome (New York, NY) is a wellness service that applies artificial intelligence to complex
biological data to provide personalized diet, nutrition and lifestyle recommendations for
healthy living. Viome tailors its service to evaluate and pinpoint potential issues in the
body related to the Microbiome, which tunes the immune system, and reduces
inflammation. Viome specifically targets the bacteria, viruses, yeast, fungus, and mold in
the gut that works symbiotically with the body to provide the nutrition needed to feel
energetic and stay healthy. The company detects all of the nutrients and toxins being
produced by the body’s microbiome.
Molecular Diagnostics
Seegene: AI-Inspired Molecular Diagnostics
Seegene is a leader in advanced multiplex endpoint and real-time PCR chemistries with
technology that is able to produce highly sensitive and specific panel results for infectious
disease, pharmacogenetics, and oncology testing. The company has partnered with IVD
heavyweights for the development of multiplex assay kits. Seegene seeks out prospective
clients among IVD companies developing molecular POC diagnostic platforms and
previously signed an agreement with Akonni Biosystems. Seegene has additionally
developed a multiplex microarray technology applicable to molecular POC testing.
Reagent kits available from Seegene are in the areas of qPCR and PCR respiratory panel
testing; qPCR and PCR flu typing; multiplex pneumonia testing; Middle East Respiratory
Syndrome Coronavirus (MERS-CoV) testing; qPCR and PCR GI panel testing; qPCR and
PCR STD or STI panel testing; HSV-1&2 testing; CT/NG testing; high-risk HPV screening
and HPV genotyping, among other areas.
Seegene assay panels primarily target viral and bacterial infections and include panels for
viral respiratory infections, STDs, herpes, diarrhea, HPV, viral meningitis, and sepsis. The
company’s Quantplex STI-7 Assay differentiates and quantifies seven sexually transmitted
pathogens.
In January 2018, Seegene announced that, utilizing a newly created artificial intelligence
(AI) based assay development system, the company succeeded in developing molecular

diagnostics (MDx) assays in only four days. These assays are high multiplex real-time PCR
reagents developed in their entirety by AI and are designed to simultaneously detect eight
different DNA targets for each of meningitis and sexually transmitted infections.
This AI-based automated assay development system, called the Seegene Digitalized
Development System (SG-DDS), incorporates Seegene’s proprietary know-how and
patented technologies accumulated over the last 15 years in MDx assay development.
Predicine and Flagship Biosciences Collaborate
In September 2018, Predicine (Hayward, CA) and Flagship Biosciences (Westminster, CO)
are excited to announce a strategic venture to combine Predicine’s GeneRADAR molecular
insights platform with Flagship’s artificial intelligence-enabled cTA digital pathology
platform. The partners aim to provide comprehensive biomarker profiling to empower
clinical trials in the immuno-oncology space.
Molecular Colorectal Diagnosis
A research team from University Hospital Zurich and the University of Oxford has
developed a method to predict the molecular classification of colorectal cancer from digital
pathology slides. Colorectal cancer is the third most common malignant tumor in men and
women with approximately 1.8 million new cases globally per year, including around 4,000
in Switzerland. Surgery, radiation and chemotherapy as well as precision therapeutics are
the established treatment options, but are associated with relevant side effects. Precise
information about the molecular subtype of the tumor using RNA sequencing can support
patient stratification for personalized therapy. Yet cancer classification through RNA
sequencing remains a resource-intensive, costly process: examining a single sample costs
over CHF 1,000. Further, up to 20 percent of samples cannot be conclusively classified due
to insufficient availability of material or ambiguous results.
Next-Generation Sequencing
Next-generation sequencing accompanied by data analysis algorithms can allow clinicians

and researchers to uncover the hidden aspects of antibiotic resistance, cancer pathways and
rare and chronic diseases. This could mean precision medicine is beginning to put one-

size-fits-all medicine in the archives of therapeutics. Their usage will be a large factor in
the future growth of diagnostics.
On average, more than 14 Genetic Testing Units (GTUs) per day are entering the
commercial market, a pace faster than in previous high-growth years. Most disease
processes involve a number of genes and cell pathways. New multi-gene diagnostics tests
by whole genome and next-generation sequencing that examine thousands of specific gene
sequences might one day hold the key to assessing disease risk, diagnosing diseases and
guiding precision medicine treatment decisions. However, artificial intelligence algorithms
are needed to interpret the mass of test data produced by these tests.
Siemens
In April 2018, Siemens launched a new molecular thermocycler, the Fast Track cycler
(FTC), and the complementary FastFinder software. According to Siemens, the compact
FTC allows laboratories of all sizes to implement molecular testing with simplicity and
speed, while the FastFinder software improves workflow using Artificial Intelligence (AI)-
powered automation.
The FastFinder software, developed in collaboration with UgenTec (Hasselt, Belgium) is a

fully automated real-time PCR interpretation platform that is compatible with all
multiplexing kits from Fast Track Diagnostics. The software allows users to fully automate
and standardize analysis of kits with the Applied Biosystems 7500 and LightCycler 480
PCR platforms. FastFinder uses its built-in statistical machine learning models to analyze
every amplification curve, then automatically combines these results and processes them
through an assay-specific decision tree, leading to overall and conclusive results.
May 2016, Siemens acquired NEO New Oncology (Cologne, Germany) that provided
Siemens an entry point into NGS-based genomic testing and expands its capabilities in
precision medicine a genetic make-up of tumors. The Neo next generation sequencing
technology selects and analyzes relevant parts of the tumor DNA at high resolution. The
resulting large volumes of data are processed by software developed by Neo New
Oncology using a quality-assured system.
NEO New Oncology is developing molecular profiling assays based on NGS (Next
Generation Sequencing), for tissue specimens and body liquids. This includes NEOliquid,
a liquid biopsy test for the analysis of genomic profiles of solid tumors from a simple blood

sample. NEO New Oncology’s high quality 3rd generation hybrid capture technology
allows for the analysis of circulating tumor DNA.
Siemens reported: “the acquisition of NEO New Oncology provides Siemens Healthineers
an entry point into NGS-based genomic testing and expands its capabilities in precision
medicine and companion diagnostics. Furthermore, Siemens Healthineers establishes a
business prospect in the field of molecular tests and services, with the plan to provide
testing and enablement services to physicians, hospitals and laboratories.”
Curetis Group Company Launches AI-powered Molecular Antibiotic Susceptibility Test
In October 2019, Curetis N.V., a developer of next-level molecular diagnostic solutions,

announced that its fully-owned subsidiary Ares Genetics had launched an early access
program for its novel, artificial intelligence (AI) powered, next-generation sequencing
(NGS) based molecular antibiotic susceptibility test (AST).
Information on antibiotic susceptibility of pathogens is of utmost importance for clinical

practice, epidemiology and public health purposes as well as for the development of
pharmaceutical products in the infectious disease sector. Ares Genetics therefore has
developed a molecular AST that will be marketed under the brand name ARESupa –
Universal Pathogenome Assay. It is an expanded version of the NGS-based ARESupa
initially launched in August 2019 for the identification of pathogens and resistance genes.
The new version of the ARESupa is now capable of also accurately predicting antibiotic
susceptibility via AI-powered interpretation of high-throughput DNA sequencing data.
ARESupa is based on whole-genome sequencing of bacterial strains isolated from clinical

specimens, combined with data analysis and interpretation powered by ARESdb, Ares
Genetics’ unique, proprietary reference database on genetic antimicrobial resistance
markers. ARESdb covers genomes of about 40,000 bacterial strains and associated
susceptibility data for more than 100 different antibiotics.
A blinded evaluation study with a leading global IVD player demonstrated a predictive
performance of the AI-powered ARESupa in line with FDA requirements for over 50
drug/pathogen combinations. Further blinded evaluation studies with leading academic
medical centers were anticipated for Q4/2019.

With the early access program gaining momentum, broader commercialization of the AI-
powered antibiotic susceptibility test is anticipated to begin in early 2020. The test will be
initially offered for non-diagnostic applications in epidemiology, infection control, and
outbreak analysis for customers in the public health sector and the pharmaceutical industry.
A laboratory-developed test (LDT) on native patient samples for human diagnostic use in
indications in which current culture-based diagnostic practice is inherently challenging is
planned. Furthermore, Ares Genetics entered into a multi-phase strategic partnership with
an undisclosed leading global in vitro diagnostics corporation to jointly develop diagnostic
solutions for infectious disease testing based on the ARESupa.
In September 2019, QIAGEN, in a strategic partnership and under a license from Ares
Genetics, launched ARESdb as part of its CLC Microbial Genomics Module for general
explorative antibiotic research by the life science community.
Oxford Nanopore
In December 2019, a large-scale population genomics project was announced by the

Department of Health in Abu Dhabi, UAE. The program aims to improve the
understanding of the genomics of individuals in the region as part of a personalized
medicine initiative. The project is also exploring ways to deliver participants their own
genomes as part of the drive towards better health for the region.
The sequencing project will be fulfilled by G42 in Abu Dhabi. Oxford Nanopore's
sequencing technology will be used in this population-scale project, providing information-
rich sequencing at ultra-high scale. With the recent launch of the PromethION 48 device,
Oxford Nanopore offers technology that can sequence long or ultra-long fragments of
DNA, at the scale required for population-scale projects. One PromethION device can
deliver more than 7Tb of sequence data in a single run of 48-flow cells.
Nanopore sequencing supports all aspects of genetic discovery, including genome

assembly, the analysis of larger genetic variants, the ability to reach previously inaccessible
areas of the genome, and direct methylation analysis. All of these, unlocked by the ability
to sequence native DNA/RNA, the ability to sequence the longest fragments of DNA/RNA,
and fully scalable electronics-based technology, are critical to understanding the breadth
and depth of genetic diversity.

BGI Genomics
In January 2018, BGI Genomics (formerly Beijing Genome Institute) and the Federation of
Shenzhen Commerce announced that they would develop a BGI Global Innovation Center
in San Jose, CA and Seattle, WA that would focus on healthcare, precision medicine,
artificial intelligence, and big data.
Qiagen
Venlo, Netherlands-based Qiagen’s bioinformatics teams are developing new software

solutions and adding proprietary cloud-based content to support the latest research and
clinical trends in molecular testing, especially the interpretation of large volumes of data
from next-generation sequencing.
Parasitology
ARUP-Techcyte: Detection of Parasites and Ova
U.S. reference laboratory ARUP Laboratories and artificial intelligence (AI) developer
Techcyte announced they have co-developed an AI-augmented tool for improved detection
of parasites and ova. The tool was designed to advance diagnostic capabilities at the
parasitology lab of ARUP, a nonprofit arm of the University of Utah, and will allow for
quicker turnaround times, lower costs, and improved patient care. The companies also said
the new technology will mean less eye fatigue and neuromuscular tension for lab workers.
Because the tool quickly screens out negative results, workers spend less time looking
through microscopes and more time analyzing results, according to ARUP and Techcyte.
Development of the AI tool for ova and parasite testing is part of a larger collaboration
between ARUP and Techcyte on digital technologies for laboratories.
Pathology
Digital pathology is the technique of analyzing high-resolution digitally scanned histology

images, which may then be analyzed using computational tools and algorithms. In the first
step, the whole histology glass slide is scanned with the help of a high-resolution image

scanner. This image information is then reviewed by a pathologist. The role of computer-
based imaging is very important and helps pathologists in making a decision. Therefore,
digital pathology is an intersection of pathology and computers that is capable of replacing
the conventional microscope-based diagnosis in the near future.
Digital pathology is enabled in part by virtual microscopy, a method of capturing

microscopic images and transmitting them over computer networks. This allows
independent viewing of images by large numbers of people in diverse locations.
A combined, end-to-end, automated solution for histology brings workflow efficiency that
is a necessity now that workloads have increased and staffing is either scarce or too
expensive. And digital pathology enables a next step in pathology: the ''big data'' era of
medical care that is referred to as computational pathology.
Histology slide analysis is highly labor intensive and subjective in nature. At least this was
the scenario until the commercialization of automation and digital image analysis. Now
histology is following the trend already in action in other testing areas and is investing in
end-to-end automation of the entire tissue process. From slide preparation through to
analysis; all the way to result reporting, combined solutions enable faster turn-around time
while providing standardized results.
The growing number of cancer patients is a major factor, but a number of technological
advancements also contribute to the growth. Sample preparation of histology samples has
always been fairly automated, but companies have invested to improve upon this with new
faster tissue fixatives and systems that improve nucleic acid extraction from tissue samples.
Of note is new instrumentation to automate RNA detection in fixed tissues. This is
important for advanced companion tests.
Automated sample processing and slide strainers have been available for some 10 to 15
years, however end to end automation, digital optics and the artificial intelligence
algorithms and high capacity computer power necessary to really make these systems work
efficiently have only recently become available.
Innovation in immunohistochemistry (IHC) staining has evolved significantly over the last
two decades. The process of staining has shifted from labor-intensive, manual techniques
toward semi-automated instruments with off-line processes such as deparaffinization and
antigen retrieval, and now to a new generation of baking-through-staining automated

instrumentation systems that enable standardization, improved staining consistency, and

expedited turnaround times.
LabCorp and Philips Collaborate
Through strategic investments in R&D, an acquisition, partnerships and technology

licenses, Philips is building a leading portfolio for the digital pathology.
Philips lntelliSite pathology solution is an automated digital pathology image creation,

management and viewing system comprised of an ultra-fast pathology slide scanner, an
image management system and dedicated software tools. This solution contains advanced
software tools to manage the scanning, storage, presentation, reviewing, and sharing of
images. Philips comments that by supporting the transition to digital workflows, the
company seeks to help pathology laboratories simplify access to histopathology
information and implement more efficient workflows.
The Philips IntelliSite Pathology Solution is CE Marked for in vitro diagnostic use. In
Canada, the Philips IntelliSite Pathology Solution is licensed by Health Canada for in vitro
diagnostic use. In the U.S., Philips reported that IntelliSite Pathology is the first and
currently only digital pathology solution to be marketed for primary diagnostic use. It aids
pathologists in the review and interpretation of digital images of surgical pathology slides
that are prepared from formalin-fixed paraffin embedded (FFPE) tissue.
In the United States, the Philips IntelliSite Pathology Solution is indicated for in vitro
diagnostic use for Manual Read of the Digital HER2 Application. The Philips IntelliSite
Pathology Solution is also registered for in vitro diagnostic use in Australia, Singapore and
Middle East.
In June 2018, LabCorp and Royal Philips (Amsterdam, Netherlands) announced a

collaboration in digital pathology. Through the implementation of the Philips IntelliSite
Pathology Solution, LabCorp will incorporate digitized workflows within its anatomic
pathology services, which are an integral component of its comprehensive clinical
laboratory and end-to-end drug development services.
Philips IntelliSite Pathology Solution is the first, and currently the only, digital pathology
solution marketed for primary diagnostic use in the U.S. It aids pathologists in the review
and interpretation of digital images of surgical pathology slides that are prepared from

formalin-fixed paraffin embedded (FFPE) tissue. LabCorp will initially implement the
system in four of its laboratories, and it will also work with its interested customers to
evaluate their potential transition to digital pathology.
Philips is working with organizations that have developed AI algorithms to incorporate

machine learning into the IntelliSite platform.
Digital pathology systems from Philips also include the new advanced computational
pathology solution, TissueMark, The TissueMark system aids in identifying insufficient
samples for molecular tests with accurate cellularity guidance. The AI-enabled application
allows region of interest detection and cellular profile estimation in whole slide images
(WSI) of lung histology, lung cytology, colon and breast FFPE, and H&E tissue samples in
60 seconds. Interoperable with the Philips IntelliSite Pathology Solution, TissueMark
provides a high throughput, intuitive molecular test workflow that helps labs reduce
variability in tumor estimation. TissueMark’s deep learning AI technology accumulates
information from several hundred images and applies that knowledge to accurately identify
tumor samples and estimate tumor percentages, This advanced platform is intended to
enhance the quality of macrodissection and ultimately reduce the costs caused by
insufficient tumor input.
Philips Wellcentive is a platform that uses machine learning and AI to identify and predict
clinical data anomalies, enhance workflows and identify complex medical situations.
Philips Respironics DreamMapper uses AI to enable people to take control over their own
health by improving sleep therapy.
Philips CareSage is a predictive analytics technology that employs AI to help prevent

hospital readmissions and ER visits in the elderly.
All electronic devices are prone to several risks, from system failures to power outage,
which could be a matter of life or death. The e-Alert IoT-driven solution by Philips aims to
solve this. Rather than waiting for a device to fail, the new system takes a proactive
approach through virtual monitoring of medical hardware, alerting hospital staff if there is a
problem.

Histology market leaders Ventana Medical Systems (now part of Roche Diagnostics), Dako
(now part of Agilent Technologies) and Sakura Finetek Europe have invested in advancing
the automation of histology.
Ventana’s first foray into automated histochemistry began with a number of automated
market cleared stainers: VENTANA System for Primary Diagnosis, digital H&E, HE 600
system, a fully automated H&E staining, BenchMark Special Stains platform and
VENTANA Gram Stain BenchMark Special Stains instrument. Now the Roche Digital
Pathology solution combines hardware, software and the VENTANA Companion
Algorithm image analysis software to form an end-to-end solution.
Digital imaging specialists Aperio and BioImagine (now part of Roche Diagnostics) have
commercialized IT tools that transfer digital images directly to a patient’s electronic
medical record and also allow pathologists to view slide images from remote locations.
The Dako Omnis automation solution comprises hardware, software and reagents. That
automates IHC and ISH processes, meaning entire patient cases can be processed
simultaneously with continuous loading, batch loading and overnight run options. Also
processes such as slide processing and instrument maintenance can be traced to individual
operators. The system was developed in collaboration with The Tecan Group. Dako also
offers the DakoLink software for use in Dako’s Autostainer and Artisan instruments.
More recently, Dako and Omnyx, LLC (a joint venture of GE Healthcare and the
University of Pittsburgh Medical Center) developed a clinical image analysis for digital
pathology in Europe. As part of the collaboration, Dako used its expertise in staining and
image analysis to develop algorithms that are incorporated into the Omnyx digital
pathology platform.
Sakura Finetek (Japan) has been a leader in tissue sample preanalytical processing. The
company markets the Tissue-Tek AutoTEC a120 Automated Embedder, a fully automated
tissue embedder that features a cassette barcode reader for LIS (laboratory information
system) connectivity and traceability of specimen blocks, and SMARTair technology that
optimizes block quality.
Sakura also offers the VisionTek M6 Digital Microscope, a robotic imaging system that
pathologists can use like a conventional microscope. VisionTek M6 incorporates a
motorized brightfield optical system, together with multiple cameras to convert optical
fields of view of specimen on glass slides, at various magnifications, into high-resolution

digital images that are displayed in real time on a high-resolution monitor for ergonomic
viewing. The system provides real-time slide review for anatomical pathology,
cytopathology and hematopathology.
Leica Microsystems (Wetzlar, Germany) markets the Digital Image Hub Enterprise, its
clinical workflow solution for digital pathology. It provides image management,
integration and communication capabilities, in conjunction with the client viewer, SlidePath
Gateway. Digital Image Hub Enterprise utilizes slide barcodes to identify and
automatically consolidate multiple slides within a case into a work list. The system can then
automatically associate the slides with appropriate case information by querying the
laboratory information management system, thus presenting both case information and
digital slides concurrently. Pathologists can also view their digital slides through multiple
modalities including Internet browsers and iPad, anytime and anywhere.
Leica Biosystems, in December 2017, entered into a collaboration with the UK Office of
Life Sciences to advance their mission through expanded use of Leica Biosystems Aperio
Digital Pathology Solutions. Leica announced that the agreement draws substantial
investment into the sector including the development of a digital pathology program
leveraging Artificial Intelligence (AI).
Applied Spectral Imaging (Carlsbad, CA) has developed GenASIs Pathology, an automated
IHC and CISH imaging and analysis solution. Biocare Medical (Concord, CA) provides
automated instrumentation for IHC and ISH staining. The companies reported that they
have tested and confirmed the combined solution’s effectiveness in performing end-to-end
IHC and CISH, imaging, analysis and reporting.
December 2019, Philips Healthcare (Netherlands) announced plans to begin offering Paige
AI's (New York, NY) prostate cancer detection technology alongside its own digital
pathology products. Paige.AI received FDA's breakthrough device designation and CE
Mark in March 2019 and a CE mark for its module in February. Paige was spun out from
technology developed at Memorial Sloan Kettering Cancer Center (New York, NY). Under
Paige.AI's 2018 licensing agreement with MSK, the company receives de-identified images
of digitized slides to help it develop artificial intelligence tools that can assist pathologists
in identifying cancer. Philips IntelliSite Pathology Solution is a whole-slide imaging
system with digital surgical pathology slide review features.

January 2020, researchers from the Karolinska Institutet (Stockholm, Sweden) reported that
an artificial intelligence system could be used to detect and grade prostate cancer in prostate
needle biopsy samples. The research was online Jan. 8, 2020 in The Lancet Oncology.
December 2019, researchers from University of Texas Southwestern (Dallas, TX)

announced they have developed ConvPath; a software tool that uses artificial intelligence to
recognize cancer cells from digital pathology images, allowing clinicians to predict patient
outcomes. The process of manually identifying all the cells in a pathology slide is
extremely labor intensive and error-prone. The researchers report that ConvPath,
overcomes these obstacles by using AI to classify cell types from lung cancer pathology
images. The ConvPath algorithm can “look” at cells and identify their types based on their
appearance in the pathology images using an AI algorithm that learns from human
pathologists. The algorithm effectively converts a pathology image into a “map” that
displays the spatial distributions and interactions of tumor cells, stromal cells and
lymphocytes in tumor tissue.
August 2019, researchers from the University of Washington (Seattle, WA) and University
of California (Los Angeles, CA) reported they have developed an artificial intelligence
system that could help pathologists read breast biopsies more accurately.
Point of Care
Siemens Strengthens Technology Position in Immunoassay Point-of-Care Testing
In July 2019, Siemens Healthineers enhanced its technology position with the acquisition of
Minicare BV, bolstering its portfolio in AI POC immunoassay testing. Minicare BV is a
privately held company with nearly 15 employees and is based in Eindhoven, Netherlands.
Minicare BV’s handheld technology offers Siemens Healthineers the capability to
miniaturize immunoassay testing for point-of-care applications. The deal closed on July 2,
2019. Financial details of the transaction were not disclosed.
Reliant Immune Diagnostics: In-Home Rapid Diagnostic Screening Using AI
Reliant Immune Diagnostics is a startup based in Austin, TX. The company’s mission is to
provide greater access to healthcare at lower costs through in-home rapid diagnostic

screening using AI and machine learning. The company is currently building a suite of
over-the-counter diagnostic tests for conditions like colon cancer, influenza, STIs, strep,
and urinary tract infection. The tests will be operated with a mobile app, which is also in
development.
Most notably, in 2018, ReliantID licensed the TRI Analyzer, with plans to integrate it into
its services. The device, developed at the University of Illinois at Urbana-Champaign, is a
3D-printed cradle that attaches to a smartphone, converting its camera into a visible-light
spectrometer. An optical fiber directs the light from the LED on the phone’s camera into a
microfluidic chamber, where it passes through a liquid sample, reflects off of a mirror, and
passes through a second optical fiber onto a CMOS pixel sensor. An app on the phone
analyzes the reading from the sensor for color changes. The TRI Analyzer appears to have
three proof-of-concept applications: it detects fetal fibronectin, as an indicator of preterm
birth, newborn phenylketonuria screening, and an assay for the soluble transferrin receptor,
an indicator of iron deficiency. Presently, ReliantID is working to raise investment funds to
develop a commercial product using the technology.
On its own, ReliantID currently has two patents, with three more in the application phase:
x Arbovirus indicative birth defect risk test (#9,857,372; granted January 2018): “A
system for providing immunoassay test results for multiple medical conditions,
comprising a testing device having thereon an alignment target and having a
plurality of immunoassay test strips, the plurality of immunoassay test strips each
including a sample pad capable of receiving a biologic sample, a conjugate pad
containing particles for conjugating with antibodies or antigens present in the
biologic sample, and a membrane strip having a test line and a control line, wherein
the test line and the control line are viewable, and a mobile device having a camera,
a viewing screen, and a software application stored thereon, wherein the software
application provides executable instructions to capture an image of the testing
device, process an image to determine pixel count and line intensity of the test line
of each of the plurality of immunoassay test strips, and present test results on the
viewing screen.”
x Pregnancy test to assess disease risk (#9,857,373; granted January 2018): “A system
for providing pregnancy testing in conjunction with disease risk testing, comprising
a testing device having an alignment target and having a plurality of immunoassay
test strips, the plurality of immunoassay test strips each including a sample pad, a

conjugate pad containing particles for conjugating with antibodies or antigens

present in the biologic sample, and a membrane strip having a test line and a control
line, wherein the test line and the control line are viewable, and a mobile device
having a camera, a viewing screen, and a software application stored thereon,
wherein the software application provides executable instructions to determine a
risk value for each of at least one disease risk tested using the biologic sample,
assess the risk value for each of the at least one disease risk tested, and present, if
immediate medical action is required, a medical action indicator on the viewing
screen.”
x System and method for variable function mobile application for providing medical
test results (application submitted November 2017): “A method for providing
variable function medical tests, comprising providing by a mobile device
application a plurality of selectable medical test functions, receiving information
from the mobile device application regarding test results from a test performed
using a testing device, wherein the testing device includes an alignment target
disposed on the testing device and a plurality of immunoassay test strips receiving
at the server an image of the testing device from the mobile device application,
determining by the server RGB values for a plurality of pixels of the image,
normalizing by the server the RGB values into a single value, comparing by the
server the single value to a control value stored on the server, and providing by the
server a risk indicator, wherein the risk indicator indicates a likelihood of a presence
of a medical condition.”
x System and method for variable function mobile application for providing medical
test results using visual indicia to determine medical test function type (application
submitted November 2017): “A method for image analysis of medical test results,
comprising receiving information from a mobile device application regarding a test
performed using a testing device, wherein the testing device includes a plurality of
immunoassay test strips and at least one test function indicator on a surface thereof,
wherein the mobile device application is configured to recognize the at least one test
function indicator to trigger performance of one or more of the plurality of medical
test functions, receiving at the server an image of the testing device from the mobile
device application, determining by the server RGB values for a plurality of pixels of
the image, normalizing by the server the RGB values into a single value, comparing
the single value to a control value, and providing by the server a risk indicator,

wherein the risk indicator indicates a likelihood of a presence of a medical

condition.”
x System and method for image analysis of medical test results (application submitted
November 2017): “A method for image analysis of medical test results, comprising
receiving, at a server, information from a mobile device regarding test results from a
test performed using a testing device, wherein the testing device includes an
alignment target disposed on the testing device and a plurality of immunoassay test
strips, receiving at the server an image of the testing device from the mobile device,
determining by the server RGB values for a plurality of pixels of the image,
normalizing by the server the RGB values into a single value, comparing by the
server the single value to a control value stored on the server, and providing by the
server a risk indicator, wherein the risk indicator indicates a likelihood of a presence
of a medical condition.”
Transplant Risk Testing
CareDX and Predigraft
CareDx markets several blood-based donor-derived cell-free DNA (dd-cfDNA) test for the
assessment of transplant organ health. In May 2019, the company and Cibiltech (Paris,
France) announced that they are partnering to commercialize Predigraft, a data analysis tool
that provides an early prediction of an individual's risk of allograft rejection and transplant
loss. Under the terms of the agreement, CareDx will hold the exclusive rights to
commercialize Predigraft in the U.S. and will also become a minority equity owner of
Cibiltech. The companies have also agreed to collaborate to advance the development of
artificial intelligence in transplant care. Predigraft was developed from Cibiltech's
proprietary software algorithm iBox, which was developed using outcomes data from tens
of thousands of transplant patients.

Urinalysis
Clever Culture: FDA Cleared Automation of Sample Interpretation
In May 2019, LBT Innovations Limited, a leader in medical technology automation using
artificial intelligence, announced that its 50% owned joint venture company, Clever Culture
Systems (CCS), had received 510(k) clearance from the United States FDA for its APAS
Independence instrument with associated urine analysis module (AM). APAS
Independence platform can process more than 200 culture plates at a time and
automatically screen, interpret, and sort them in response to the growing demands of
laboratories and the fast-growing automated pathology market. It noted that the APAS
Independence walkaway system screens culture plates that show significant bacterial
growth from those that do not. The clearance, achieved ahead of schedule, paves the way
for sales of the platform and associated urine analysis module, Clever Culture Systems
General Manager Peter Bradley said in a statement. With urine accounting for more than
50 percent of all types of specimens, a significant number of laboratories in the US have
expressed interest in the product, and the firm expects to roll it out to targeted customers
before the end of this year.
What the Future Holds
Artificial intelligence has been getting a lot of attention in recent years. As the technology
advances, laboratory scientists are making great strides in the capabilities and sophistication
of AI. While some people may worry about patient confidentiality and other scenarios,
much of the AI that is in use today can actually end up helping humanity. Artificial
intelligence has been heavily promoted as the next big thing in healthcare for nearly a
decade. AI was expected to revolutionize how doctors treat patients. That it would bring
about a new age clinical decision support tools.
Based on the earlier discussion of IBM’s Watson, we see that Watson’s lofty programs may
not have been successful. However, scientists have devised a number of AI-based programs
and products that have teased out complex medical and clinical information to address the
most pressing disease diagnostics.

In spite of a still uncertain regulatory pathway, the diagnostics laboratories and companies
have embarked in full force in the advancement of AI driven tests and devices. Based on
the examples given in this report, we see that a number have already proven successful and
that many more similar AI-based initiatives are yet to come.

3: Market Analysis: Artificial Intelligence in Diagnostics
3: Market Analysis: Artificial Intelligence in

Diagnostics
Size and Forecast Estimate
This section provides estimates for the revenues earned by the aggregate of companies in
the market for AI in diagnostics. By embracing digitalization, AI algorithms and machine
learning, pathologists can cut down on potential errors and leave more time for human
analysis. Innovations that have revolutionized pathology include whole slide imaging,
which is the scanning of a complete microscope slide and the generation of a high-
resolution file. The ability to read a whole slide image, visually annotated and enriched
with information from AI algorithms, will provide a faster and more accurate diagnosis.

Figure 3-1: AI in Diagnostics Estimated Market and Forecast, 2019-2024 ($M)

Market by Application
The AI in Diagnostics market can roughly be divided into several application categories.
Smart glucose systems comprise 18% of the market and that is expected to increase as more
continuous glucose monitoring (CGM) systems incorporate AI features that consumers
adopt. The remainder of revenue is estimated from systems that use AI-read genetic or
other biomarkers to aid in microbiology, hematology or liquid biopsy and other diagnosis.
Figure 3-2: Artificial Intelligence Market in Diagnostics: Market Estimate by Application (Histology-based
AI, Microbiology, Hematology, Glucose Testing, Biomarker Diagnostic Aids) (%)

Component Market Analysis
The market consists largely of software (66%), though hardware remains an important
component.
Table 3-1: Artificial Intelligence Market in Diagnostics: Market by Component, 2019 (Hardware, Software,
Services) ($M, %)
Component 2019 %Mkt
Hardware 30 23%
Software 86 66%
Services 14 11%
Total 130 100%

AIMarketsInIVD 2020 Kalorama

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

AIMarketsInIVD 2020 Kalorama

Uploaded by

Copyright:

Available Formats

Artificial Intelligence Markets

1: EXECUTIVE SUMMARY ............................................................................................................................. 1

THE ARTIFICIAL INTELLIGENCE BOOM...................................................................................................................... 1

2: ARTIFICIAL INTELLIGENCE IN DIAGNOSTICS MARKETS............................................................................ 13

AI IN CLINICAL PRACTICE .................................................................................................................................... 13

April 2020 © Kalorama Information i

Rady Children’s Institute for Genomic Medicine ..................................................................................... 31

ii April 2020 © Kalorama Information

3: MARKET ANALYSIS: ARTIFICIAL INTELLIGENCE IN DIAGNOSTICS ............................................................ 68

SIZE AND FORECAST ESTIMATE ............................................................................................................................ 68

April 2020 © Kalorama Information iii

The Artificial Intelligence Boom

Artificial intelligence in healthcare refers to the use of complex algorithms designed to

April 2020 © Kalorama Information 1

Interactive glucose meters are the most commonplace application of AI-incorporated

New applications of AI are in development to predict outcomes such as diagnosis, death or

2 April 2020 © Kalorama Information

The FDA Is Getting Ready

Among the ongoing projects is collaboration with Project Data Sphere

April 2020 © Kalorama Information 3

FDA Awards DNAnexus Contract to Power precisionFDA Collaborative Omics Environment in

In December 2019, DNAnexus, Inc., a leader in enterprise omics technology, announced

PrecisionFDA was launched in December 2015, and through collaborative engagement

DNAnexus previously developed the precisionFDA Challenge Construction Toolkit that

4 April 2020 © Kalorama Information

COVID-19 and Artificial Intelligence

x Developing of Immunoassays – Serological test are in high demand as they could

April 2020 © Kalorama Information 5

x Genomic Interpretation AI has been used to more efficiently search genomes to

x AI-Enhanced Imaging could be added to an IVD test to enhance its detection

Deep Learning to Discover Origins of COVID-19

6 April 2020 © Kalorama Information

protein insertions described in the withdrawn paper. Deep-learning modeling demonstrated

April 2020 © Kalorama Information 7

8 April 2020 © Kalorama Information

JLK Inspection: RT-PCR COVID-19 Testing with AI-Enhanced Imaging Aid

Persivia: Ruleset to Triage COVID-19 Testing

In March 2020, Persivia, a provider of real-time bundled payment, value-based care,

April 2020 © Kalorama Information 9

Surgisphere Rapid Test Uses AI to Diagnose COVID-19

In March 2020, Surgisphere Corporation (Palatine, Illinois) announced the availability of

With its cloud-based healthcare data analytics platform (QuartzClinical), Surgisphere

The test is available for public use.

Biofourmis' AI-Powered Remote Monitoring Platform to Fight COVID-19

In March 2020, Biofourmis, a provider of digital therapeutics that power personalized

10 April 2020 © Kalorama Information

Diagnosed and potentially COVID-19-infected patients are being monitored with

Everion has several medical-grade sensors, including optical, temperature, electrodermal,

April 2020 © Kalorama Information 11

At the time of this report’s release, Biofourmis is in conversations with government

12 April 2020 © Kalorama Information

2: Artificial Intelligence in Diagnostics Markets

2: Artificial Intelligence in Diagnostics Markets

For microbiology, microbiology automated sample processing leader, Copan Diagnostics

April 2020 © Kalorama Information 13

2: Artificial Intelligence in Diagnostics Markets

14 April 2020 © Kalorama Information

2: Artificial Intelligence in Diagnostics Markets

matched treatment options in minutes for evidence-based, individualized treatment

Royal Philips (Amsterdam, Netherlands) showcased its oncology informatics solutions at

April 2020 © Kalorama Information 15

2: Artificial Intelligence in Diagnostics Markets

Selected AI-based Initiatives / Collaborations / Tests