Professional Documents
Culture Documents
Academia Letters, August 2021 ©2021 by the authors — Open Access — Distributed under CC BY 4.0
Clinical Applications
Using machine learning in a clinal setting is not a novel concept. In medical imaging, ML is
used with computer vison tasks such as detection, segmentations and classification. For ex-
ample, in oncology, ML is applied to assist in detecting the presence or absence of metastases
on histological sections (Nichols et al., 2018). Using a color scale, computer-aided diagno-
sis can indicate the probability of each pixel in an image to be part of a metastatic region
(Nichols et al., 2018). In the past, detecting features in images such as edge detection of var-
ious boundaries using algorithms based on pixel intensity and gradients required inputs to be
set by the user. Now with improved deep learning methods and artificial neural networks,
the process can determine the settings through an iterative process to reduce the loss function
(Chartrand et al., 2017). Another benefit from the application of ML in imaging is the time
saved through automation. In the field of histopathology, automated segmentation using ML
has been successful in reducing the diagnostic workload (Litjens et al., 2016). Litjens et al.
also successfully identified all slides containing prostate cancer as well as micro- and macro-
metastases of breast cancer automatically, without the use of additional immunohistochemical
markers or human intervention (Litjens et al., 2016). Furthermore, with the increasing avail-
ability of large databases of raw data with attached diagnoses, improved training techniques,
network architecture and computer hardware, there has been significant breakthroughs in the
fields of dermatology, radiology, ophthalmology and cardiology (Chartrand et al., 2017).
Research Applications
Machine learning also plays a role in research. In the field of drug discovery, development
pipelines are often long, complex and depend on several factors. Using the computational
power of ML, scientists can mine through billions of potential drug compounds to quickly
discover treatments for currently incurable diseases. Additionally, ML can be applied in all
Academia Letters, August 2021 ©2021 by the authors — Open Access — Distributed under CC BY 4.0
Surgery Application
Machine learning application has a lot of potential in helping physicians predict major com-
plications after and in some cases during surgeries. Currently the ability to accurately predict
major complications prior to the surgery is limited. This is an important area of concern as
complications after surgeries can double the 30-day mortality and are associated with long-
term consequences (Bihorac et al., 2020). In the United States alone, 1.5 million patients de-
velop a medical complication and at least 150,000 patients die within thirty days after surgery
(Bihorac et al., 2020). Having the ability to accurately predict the risks and benefits of surgery
can help identify patients that could benefit from intraoperative strategies to offset the risk (Bi-
horac et al., 2020). In a study by Bihorac et al., the team used machine learning on existing
clinical data in electronic health records to forecast the risk probabilities for eight major post-
operative complications including sepsis, neurologic and cardiovascular complications, and
death up to 24 months after the surgery (Bihorac et al., 2020). This model was able to calculate
risk scores with high discriminatory ability using readily available health records data.
Another study focused on using ML to help anticipate hypoxaemia during surgical pro-
cedures. Hypoxaemia, or low arterial blood oxygen tension, can occur during general anaes-
thesia and is associated with serious patient complications including cardiac arrest, cardiac
arrhythmias and decreased cognitive function (Lundberg et al., 2018). Currently anaesthe-
Academia Letters, August 2021 ©2021 by the authors — Open Access — Distributed under CC BY 4.0
References
Bihorac, A., Ozrazgat-Baslanti, T., Ebadi, A., Motaei, A., Madkour, M., Pardalos, P. M.,
Lipori, G., Hogan, W. R., Efron, P. A., Moore, F., Moldawer, L. L., Wang, D. Z., Hob-
son, C. E., Rashidi, P., Li, X., & Momcilovic, P. (2019). MySurgeryRisk: Develop-
ment and Validation of a Machine-Learning Risk Algorithm for Major Complications and
Death after Surgery. Annals of Surgery, 269(4), 652–662. https://doi.org/10.1097/SLA.
0000000000002706
Chartrand, G., Cheng, P. M., Vorontsov, E., Drozdzal, M., Turcotte, S., Pal, C. J., Kadoury,
S., & Tang, A. (2017). Deep Learning: A Primer for Radiologists. Radiographics: A
Review Publication of the Radiological Society of North America, Inc, 37(7), 2113–2131.
https://doi.org/10.1148/rg.2017170077
Litjens, G., Sánchez, C. I., Timofeeva, N., Hermsen, M., Nagtegaal, I., Kovacs, I., Hulsbergen
- van de Kaa, C., Bult, P., van Ginneken, B., & van der Laak, J. (2016). Deep learning
Academia Letters, August 2021 ©2021 by the authors — Open Access — Distributed under CC BY 4.0
Lundberg, S. M., Nair, B., Vavilala, M. S., Horibe, M., Eisses, M. J., Adams, T., Liston, D.
E., Low, D. K.-W., Newman, S.-F., Kim, J., & Lee, S.-I. (2018). Explainable machine-
learning predictions for the prevention of hypoxaemia during surgery. Nature Biomedical
Engineering, 2(10), 749–760. https://doi.org/10.1038/s41551-018-0304-0
Nichols, J. A., Herbert Chan, H. W., & Baker, M. A. B. (2018). Machine learning: Ap-
plications of artificial intelligence to imaging and diagnosis. Biophysical Reviews, 11(1),
111–118. https://doi.org/10.1007/s12551-018-0449-9
Vamathevan, J., Clark, D., Czodrowski, P., Dunham, I., Ferran, E., Lee, G., Li, B., Mad-
abhushi, A., Shah, P., Spitzer, M., & Zhao, S. (2019). Applications of machine learning
in drug discovery and development. Nature Reviews Drug Discovery, 18(6), 463–477.
https://doi.org/10.1038/s41573-019-0024-5
Zhang, Y., Wang, Y., Zhou, W., Fan, Y., Zhao, J., Zhu, L., Lu, S., Lu, T., Chen, Y., & Liu,
H. (2019). A combined drug discovery strategy based on machine learning and molecular
docking. Chemical Biology & Drug Design, 93(5), 685–699. https://doi.org/10.1111/
cbdd.13494
Academia Letters, August 2021 ©2021 by the authors — Open Access — Distributed under CC BY 4.0