|LETTER TO EDITOR
|Year : 2020 | Volume
| Issue : 1 | Page : 133-134
Artificial intelligence in health-care: How long to go?
Rakesh Pinninti, Senthil Rajappa
Department of Medical Oncology, Basavatarakam Indo-American Cancer Hospital and Research Institute, Hyderabad, Telangana, India
|Date of Submission||05-Jan-2020|
|Date of Acceptance||05-Jan-2020|
|Date of Web Publication||24-Feb-2020|
Department of Medical Oncology, Basavatarakam Indo-American Cancer Hospital and Research Institute, Banjara Hills, Hyderabad, Telangana
Source of Support: None, Conflict of Interest: None
|How to cite this article:|
Pinninti R, Rajappa S. Artificial intelligence in health-care: How long to go?. Cancer Res Stat Treat 2020;3:133-4
We read with interest the recent article regarding artificial intelligence (AI) in health-care by Mahajan et al. AI and machine learning (ML) algorithms promise to deliver faster and more consistent diagnoses and ultimately improve patient care. This exciting technological innovation has the potential to disrupt the current diagnostic and management paradigms in medicine. With advanced computational algorithms and deep neural network statistical analytics, AI can automate radiographic assessment of neoplastic lesions, aid molecular subtyping, and predict biological behavior based on qualitative phenotypic descriptions., Similarly, a supervised-learning process using histopathology slides can automate cancer diagnosis based on predefined characteristics of cancer cells, tumor microenvironment, and relationship between various biomarkers. AI can integrate inputs of multiple data streams and varied diagnostic systems spanning radiographic images, genomics, pathology, electronic health records, and epidemiological trends in the intended population improving the accessibility, efficiency, reliability, and quality of health-care services.
With only one physician for 1445 populations, India consistently ranks among the least performing countries across the world in health-care access to masses. There is a significant crunch in human resources and expertise in health-care sectors. AI systems have the potential to overcome the handicap of lack of infrastructure of training and increasing competence of health-care professionals. Several decision support tools could be developed to aid clinicians and patients in remote locations. Moreover, AI is expected to help deliver better health-care at lower costs.
However, there are several inherent shortcomings of AI in health-care that need to be addressed before its wider application. There is a concern regarding the uniformly acceptable reference gold standard or 'semantic features' used for ML. Most available databases are not robust in terms of quality to be used for AI algorithms. The core strength of speed and consistency can also become a concern, as this may result in consistent overdiagnosis and may amplify the biases in clinical decisions. ML algorithms can generate data that may be based on race, gender, age, and religion, resulting in discrimination and unfair results which might be better for some people than others. This may be more relevant in a country like India which is populated by people of multiple ethnicities. To ultimately replace human beings, one needs to prove that AI is at least as good if not better in efficiency and accuracy. As with any other new technology in health-care, AI too should be subject to regulatory approvals relying on clinical trials and evidence-based improvements in clinical outcomes in the intended populations.
As with any innovation, AI will make health-care more efficient and cheaper, though there is quite some way to go before that is a reality.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Mahajan A, Vaidya T, Gupta A, Rane S, Gupta S. Artificial intelligence in healthcare in developing nations: The beginning of a transformative journey. Cancer Res Stat Treat 2019;2:182-9. [Full text]
Young JD, Cai C, Lu X. Unsupervised deep learning reveals prognostically relevant subtypes of glioblastoma. BMC Bioinformatics 2017;18:381.
Wadhwa SM, Krishna BG, Malhotra M, Prabhash K, Noronha V, Joshi A, et al
. Radiogenomic signatures of NSCLC brain metastases: A potential noninvasive imaging marker for ALK mutation. Ann Oncol 2019;30 Suppl 2:Ii20-1.
GBD 2017 SDG Collaborators. Measuring progress from 1990 to 2017 and projecting attainment to 2030 of the health-related Sustainable Development Goals for 195 countries and territories: A systematic analysis for the global burden of disease study 2017. Lancet 2018;392:2091-138.
Philip CC, Mathew A, John MJ. Cancer care: Challenges in the developing world. Cancer Res Stat Treat 2018;1:58-62. [Full text]
Noronha V. Making a case for cancer research in India. Cancer Res Stat Treat 2018;1:71-4. [Full text]