Professional Documents
Culture Documents
net/publication/368848373
CITATIONS
0
10 authors, including:
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Pralhad Mudalkar on 28 February 2023.
As the world's population ages and the prevalence of patients with numerous chronic illnesses rises, there is an increasing demand for high-
quality healthcare in all cultures. Clinical diagnoses and treatment recommendations are mostly made using computational approaches in
medical artificial intelligence (AI). Today, applications of artificial intelligence have been created to address some of the most urgent issues
facing health organizations today. There are many data-rich processes in the healthcare industry, and the availability of vast volumes of data,
along with advancements in computing power and artificial intelligence approaches, have opened up a world of possibilities.
Keywords: Artificial intelligence, Health care, Digital health,Business studies, Management studies.
INTRODUCTION
The term artificial intelligence (AI) refers to a broad range of computing advancements that imitate the supporting systems of
human intelligence, such as cognition, deep learning, engagement, adaptation, and sensory perception [1,2].More and more, it
is clear that artificial intelligence has applications in clinical research and medical treatment. The study has shown the value
and potential of AI-enabled health solutions. Governments and innovative organizations are currently investing resources across
the board in using AI for clinical purposes [3,4].
The US Food and Drug Administration desires to expand the availability of medical devices assisted by AI. Clinical decision
support, patient follow-up, medical care intercessions, and medical services organization are the four areas where AI-powered
medical service delivery will likely have an impact.AI approaches have proven to be great at identifying important data patterns
after thorough testing as tools for clinical trials, particularly to help the decision-making in each stage for diagnoses and
subsequent treatments, as well as diagnosis and predictions [5].
If AI is to succeed, health organizations must overcome a number of obstacles. Here are a few of these difficulties: (1) a lack
of understanding of the capabilities and limitations of a certain type of AI technology; (2) a lack of clear plans for incorporating
different AI technologies into the current healthcare systems to successfully address the most important issues those industries
are now addressing; (3) a lack of employees with the training necessary to implement AI; (4) AI technologies' incompatibility
with old infrastructure; and (5) lack of availability of accurate and diverse data [6,7].Additionally, an AI system gathers helpful
data from a huge patient population to assist in decision-making. predictions of medical outcomes with real-time inferences and
health risk alerts [8].
Technology:
Machine Learning
Support vector machines (SVMs), a type of machine learning technique, are frequently used for the diagnosis or prognosis of
disease. This techniquepromises to address the health issue with precise computational capability. Regression, classification,
and data outlier detection are all used by SVM. According to some experts, SVM can employ medical data analytics to detect
a variety of health conditions, such as diabetes, high blood pressure, and cancer. SVM is going to undergo a major revolution
as it is used more frequently to address issues with global health[10].
An artificial neural network connects neurons in a computer model to represent how the human brain learns [11].The most
popular techniques are neural networks and predictive scores since they can be used to identify results that are nearly impossible
for humans to anticipate, like mortality rate, for example. The majority of the data used to train these models must be
preprocessed because it might be anything from electronic health records to imaging data from MRIs and CT scans, among
other things [12].Artificial neural networks and computer-aided diagnosis with deep learning are currently the focus of intense
study in the field of medical diagnosis using artificial intelligence (AI) systems, and it is predicted that their use in biomedical
systems will increase in the future[13,14].
Deep Learning
Since deep learning includes several hidden layers between its inputs and outputs, it is a type of artificial neural network that
can evaluate complex data with a variety of topologies [9].
The healthcare sector frequently uses unstructured textual data such as physician notes, test results, lab reports, and other forms.
medicines and instructions on how to leave the hospital. Natural language processing technologies can be used to extract
important information from such large descriptive data, such as patient information that can enhance diagnostic and therapy
suggestions[15-17].
1 Diagnosis The use of AI in healthcare for creating patient treatment regimens has been expanding. AI can analyze patient
data from past visits andprovide patients with better methods of treatmentand evaluating treatment strategies [18].Machine
learning is now widely used in the field of medical imaging, especially in the areas of computer-aided detection and image
analysis. Medical image analysis from CT scans, MRIs, X-rays, and ultrasounds can be used by AI to more quickly and reliably
detect symptoms of illness. Patients benefit from improved treatment options and speedier, more precise disease diagnosis.
Recent media coverage of IBM's Watson's capacity to concentrate on precision medicine, particularly cancer detection and
treatment, has been favorable. When it came to diagnosing diabetes and CVD, artificial neural networks (ANNs) outperformed
other AI techniques like support vector machines, decision trees, and neural networks[19,20].
2. Drug interaction
Drug interactions are a risk for patients taking many medications at once, and the risk rises as more prescriptions are filled.
Artificial intelligence (AI) has made it possible for computers to extract details regarding drug interactions and potential adverse
effects from medical literature, even if managing all drug interactions and their potential side effects can be challenging.
3. Dermatology
In the medical field of dermatology, imaging is crucial. Deep learning has significantly improved image processing. In
dermatology, there are three different forms of imaging: macro, micro, and contextual. AI can help dermatologists make clinical
decisions about both common skin conditions and specific tumors. The majority of the existing and future uses of AI in
dermatology focus on identifying and stopping the development of skin diseases. Convolutional neural networks have a 94%
accuracy rate in classifying skin lesions as skin cancer[22].
Because they allow for data analysis from the very distant past to the present, electronic health records are crucial to healthcare
because they improve various therapy modalities, medication use, and illness management. AI can be utilized to review the
records and give the doctors information. EHR can be used by algorithms to assess the probability of a condition based on past
data and family history [19].
1. AI in clinical practice
One of the medical specialties that have witnessed major advancements in AI recently is radiology. Imaging AI technologies
have the potential to help radiologists with the task of quantifying medical images. Deep network models, for instance, have
enabled segmentation with minimal human oversight by autonomously localizing and defining the boundaries of anatomical
structures or lesions [23,24].Recent research have shown that AI can predict outcomes when processing photos at a level of
competency comparable to a doctor[25].There are numerous theories surrounding the application of AI in radiography. The
gradual transition to AI that is expected to occur over the next ten years will have a substantial impact on radiologists' work[26].
Pathology
The popularity of digital pathology and the development of complete slide digital scanners may have an impact on the current
interest in artificial intelligence for pathology (the digitalization of histology slides to produce high-resolution images)[27,28].
For use with digital pathology, many organizations are creating pattern-recognition algorithms to aid in the interpretation of
tissue features and the prediction of disease development (such as metastasis and recurrence). AI software can recognize the
symptoms of cancer in terms of staging, grading, and differential diagnosis[29-32].
Oncology
In oncology, AI is becoming more prevalent, particularly in the field of cancer-focused cognition computer systems [33].
Today's AI provides a new and quickly developing paradigm for many scientific domains, including those concerned with the
care of cancer patients.34-37New methods for cancer screening, diagnosis, and classification, the characterization of cancer
genomics, the study of the tumour microenvironment, the evaluation of biomarkers for prognostic and predictive purposes, and
techniques for follow-up and drug discovery are a few of the expanding list of applications of AI[36-37].
Numerous of the above procedures, such as variant calling, genome explanation, variation labelling, and correspondence from
aggregates to genotypes, have been adjusted to determine the many developments pertaining to clinical genomic research.
Eventually, they might also be used to predict genotype-to-aggregate relationships. It is believed that AI may contribute to the
improvement of precision medicine based on genomics [38-39].With a greater interest in this field, machine learning is being
Mental health
AI has the ability to help mental health patients and lessen the negative consequences of a shortage of medical professionals
who specialize in treating mental health issues. In fact, several tools are being developed right now. These consist of interactive
chatbots, keyboard interface, speech, voice, facial recognition, sensors, and automated tracking of depression and mood[40-
42].
Neurology
The Radiology portion of this study made indications at the creation of several AI neurology algorithms to predict the evolution
of disease using neuroimaging. Beyond neuroimaging, advantages from machine learning are anticipated for the study of
neuroscience. It has been hypothesized that AI applications in neuroscience may be able to support the creation of numerous
hypotheses and illuminate the relationships, structures, and mechanisms underlying the functioning of the brain and
behavior[43-45].
Diabetes
A key component in the development of artificial pancreas systems is the computer programmer who links the data received by
the continuous glucose monitor with the insulin delivered by the insulin pump. Because AI can learn from data in ambiguous
contexts, researchers have looked into how it might enable these computer programs that personalize insulin distribution [46-
47].
Ophthalmology
A goal patient diagnosis may be made possible by AI advancements in ophthalmology, which have the potential to alter present
vision screening programmers. An area of great interest has emerged: using AI to check for diabetic retinopathy in persons with
diabetes [48-52]. The FDA approved the first artificial intelligence (AI)-enabled tool in April 2018 to identify diabetic
retinopathy that is "greater than mild" and that relevant individuals be referred to eye care specialists[53].
Surgery
In the field of surgery, decisions occasionally need to be taken quickly and when a patient's diagnosis and expected response to
therapy are not obvious. Lack of patient data (such as external hospital records or diagnostic tests) or dependable evidence to
support important management decisions may result in uncertainty. Clinicians may instead rely on cognitive shortcuts and snap
judgments employing pattern recognition and intuition under such time constraints and ambiguity [54-55].
Cardiology
Cardiologists must process complex spatiotemporal data from cardiac imaging modalities such cardiac ultrasound, cardiac
computer tomography, and cardiovascular magnetic resonance imaging carefully and meticulously. Innovative AI-driven
cardiac image processing algorithms have altered cardiac clinical practice by allowing cardiologists to assess patients more
quickly in their daily practice[56].
Clinical research
Recent developments have revealed novel applications of AI in clinical information retrieval, suggesting that biomedical
research benefits more from AI-derived solutions than clinical applications. Traditional medical knowledge resources, for
example, currently use machine learning (ML) approaches to rank search results, including algorithms that are constantly
learning from user search activity [57].
Randomized controlled trials are the most accurate method for evaluating the risks and benefits of any medical intervention
(RCT). However, it is not always possible to conduct an RCT. Poor patient selection, insufficient randomization, insufficient
sample size, and poor endpoint selection are common issues with ineffective RCTs. With the help of sophisticated statistical
techniques, Data-driven approaches can be used to train AI models to choose study participants and assess study endpoints with
greater accuracy. The application of AI will lead to more efficient execution and better statistical power than what is predicted
from standard RCTs [58].
Personalized medicine
AI techniques may accelerate the development of personalized medicine by assessing the therapeutic value of various research
methodologies and a variety of data sources. Precision medicine is a new area of therapy that is based on knowledge about a
person's genetic makeup, personal lifestyle, genes, and environment. By determining the characteristics that predispose a person
to a particular disease and characterizing the underlying biological pathways that cause the ailment, Precision medicine allows
us to customize a patient's preventative and treatment strategies [59,60].
Drug discovery
To create new medications and extract chemical data from vast compound databases, drug designers commonly use ML
algorithms. The creation of AI methods to implement creative modeling based on the sizeable nature of pharmacological
databases is at the core of this transition. Since large data modeling and analysis is the basis for the efficacy and safety evaluation
of prospective medications, freshly developed AI techniques offer fresh perspectives. techniques and tools for integrating,
deciphering, and analyzing the vast volumes of data. To discover eight antibacterial chemicals that were structurally distinct
from well-known antibiotics, the programmer screened more than one billion molecules and digitally tested more than 107
million[61].
Public health
AI can be used to pinpoint particular demographics or geographic regions with high disease frequency or high-risk behavior.
There is also a wide variety of AI systems that help enhance disease surveillance. To study methods for concern verification,
digital epidemiological surveillance combines case- and event-based monitoring with news and online media, sensors, digital
traces, mobile devices, social media, and clinical reporting. Using this, air pollution and medication side effects early warning
systems have been developed [62, 63].
Global health
AI might make it possible to solve health issues in low- and middle-income nations. These difficulties include a severe lack of
health professionals and inadequate public health surveillance systems. These difficulties aren't specific to these nations, but
given their link to morbidity and death in low- and middle-income settings, they are particularly important there[64].
4. AI in healthcare administration
AI and related technologies are starting to be employed in healthcare as well as other sectors of business and society. Numerous
aspects of patient care as well as operational procedures within provider, payer, and pharmaceutical industry could be
transformed by these technologies[65].
AI is able to gather all of a person's data in one location and use it to gain insight into both past and present health issues. A
comparison of disease information in this way helps doctors diagnose patients more precisely. The database of health
applications has gathered millions of symptoms, diagnoses, and specific cases that enable forecasting future health issues that
a person may experience.
AI streamlines processes, automates activities, shares data instantaneously, and organizes operations, all of which free up
medical personnel from having to juggle too many tasks.
AI might reduce costs and clinician stress, but it might also replace some jobs. Due to this factor, professionals who spent time
and money on their healthcare education may be out of work, creating equality issues.
Medical AI heavily relies on diagnosis information gleaned from millions of instances that have been cataloged. When there is
little knowledge about specific illnesses, demography, or environmental factors, a misdiagnosis is quite possible. This element
becomes crucially important when recommending a specific treatment.
Human monitoring is still essential in medicine despite the advancements AI has achieved. For instance, robots used in surgery
operate logically rather than sympathetically. Health experts can identify significant behavioral indications that can aid in the
diagnosis or prevention of medical disorders.
Cyberattacks will use AI to become smarter with each success and failure, making them more difficult to spot and prevent, just
as AI uses data to make systems smarter and more accurate. Once serious threats are able to bypass security measures, it will
be far more difficult to stop the attacks.
Artificial intelligence is a growing field of study. There are many uses of AI outside of healthcare, including data management,
drug research, diabetes treatment, digital consultation, etc. Studies show that the market for AI in healthcare is fundamentally
growing. There are various instances that show how medical AI can help practitioners deliver care to patients in the twenty-
first century much more effectively.The field of medicine has benefited greatly from AI. The main uses of AI in the medical
industry include disease diagnosis, therapy planning, the detection of mental problems, etc. For identifying cancer, strokes, and
other diseases, there are a number of AI methods available.
CONFLICT OF INTEREST
FINANCIAL SUPPORT
None
REFERENCES
1. Tran BX, Vu GT, Ha GH, et al. Global evolution of research in artificial intelligence in health and medicine: a bibliometric study. J Clin Med. 2019;
8(3):360.
2. Kaplan A, Haenlein M. Siri, Siri, in my hand: Who’s the fairest in the land? On the interpretations, illustrations, and implications of artificial intelligence.
Bus Horiz. 2019;62(1):15–25.
3. Gursoy D, Chi OH, Lu L, et al. Consumers acceptance of artificially intelligent (AI) device use in service delivery. Int J Inf Manag. 2019; 49:157– 69.
4. Jarrahi MH. Artificial intelligence and the future of work: human-AI symbiosis in organizational decision making. Bus Horiz. 2018; 61(4):577–86.
5. Ostherr K. Artificial Intelligence and Medical Humanities. J Med Humanit. 2022;43: 211–232. https://doi.org/10.1007/s10912-020-09636-4.
6. JMckinsey and Company. AI adoption advances, but foundational barriers remain. Survey Report. 2018; (11). Available at: https://
www.mckinsey.com/featured-insights/artificial-intelligence/aiadoption- advances-but-foundational-barriers-remain. Accessed June 6, 2019.
7. Chen M, Decary M. AI in healthcare: from hype to impact. Workshop presented at ITCH 2019: Improving Usability, Safety and Patient Outcomes with
Health Information Technology, Victoria, British Columbia, Canada. Available at: https:// www.slideshare.net/MeiChen39/ai-in-healthcarefrom-hype-
toimpact. Accessed June 6, 2019.
8. Murdoch TB, Detsky AS. The inevitable application of big data to health care, JAMA. 2013;309,1351–2.
9. Jiang F, Jiang Y, Zhi H, et al. Artificial intelligence in healthcare: past, present and future. Stroke Vasc Neurol. 2017; 2(4):230- 243.
10. Crown WH. Potential application of machine learning in health outcomes research and some statistical cautions. Value Health.2015;18(2):137-140.
11. Standing Senate Committee on Social Affairs, Science and Technology. Challenge ahead : integrating robotics, artificial intelligence and 3D printing
technologies into Canada’s healthcare systems. Ottawa (ON): Canada Senate; 2017: http://publications.gc.ca/site/eng/9.846477/publication. html.
Accessed 2018 May 24.
12. A. Pandit and A. Garg, 2021. Artificial Neural Networks in Healthcare: A Systematic Review.11th International Conference on Cloud Computing, Data
Science & Engineering (Confluence) 1-6, doi: 10.1109/Confluence51648.2021.9377086.
13. Beam AL, Kohane IS. Translating artificial intelligence into clinical care. J Am Med Assoc. 2016; 316(22),2368-2369.
14. Mieloszyk RJ, Bhargava P. Convolutional neural networks: The possibilities are almost endless. Curr Probl Diagn Radiol. 2018; 47(3):129-130.
15. Mills T. What is natural language processing and what is it used for? 2018. https://www.forbes.com/sites/forbestechcouncil/2018/07/02/what-is-natural-
language-processing-and-what-is-it-used-for/#722389015d71. Accessed 2018 Sep 10.
16. SAS Canada. Natural language processing: What it is and why it matters. 2018; https://www.sas.com/en_ca/insights/analytics/what-is-natural-language-
processing-nlp.html. Accessed 2018 Sep 10.
17. S. Alshahrani, E.Metais, F. Meziane, et al. Are deep learning approaches suitable for natural language processing.Natural language processing and
information systems, Springer, Cham. 2016;343-349.
18. C. I. F. N. I. F. B. Digital healthcare s.l.gpbullhound. 2013:91643080
19. Eren A, Subasi A, Coskun O, 2008. A decision support system for telemedicine through the mobile telecommunications platform. Journal of Medical
Systems.2008;32(1):31–5.
20. Suzuki K. Overview of deep learning in medical imaging. Radiol Phys Technol. 2017;10(3):257-273.
21. Ravi Manne. Machine learning techniques in drug discovery and development. International Journal of Applied Research.2021; 7(4):21-28.
22. Sennaar K, 2018. Machine learning for dermatology – 5 current applications. https://www.techemergence.com/machine-learning-
dermatologyapplications/. Accessed 2018 May 24.
23. Gillies RJ, Kinahan PE, Hricak H. Radiomics: Images are more than pictures, they are data. Radiology. 2016;278(2):563-577.
24. Hatt M, Tixier F, Visvikis Det et al. Radiomics in PET/CT: More than meets the eye? J Nucl Med. 2017;58(3):365-366.
25. Krittanawong C, Tunhasiriwet A, Zhang H, et al. Deep learning with unsupervised feature in echocardiographic imaging. J Am Coll Cardiol.2017;
69(16):2100-2101.
26. Paiva OA, Prevedello LM. The potential impact of artificial intelligence in radiology. Radiologiabrasileira.2017; 50(5), V-vi.
27. Molteni M, 2017. If you look at x-rays or moles for a living, AI is coming for your job. https://www.wired.com/2017/01/look-x-rays-moles-living-
aicoming- job/. Accessed 2018 May 24.
28. Janowczyk A, Madabhushi A, 2016. Deep learning for digital pathology image analysis: A comprehensive tutorial with selected use cases. J Pathol
Inform. 2016 Jul 26;7:29. doi: 10.4103/2153-3539.186902. PMID: 27563488; PMCID: PMC4977982.