• Users Online: 176
  • Print this page
  • Email this page

Table of Contents
Year : 2020  |  Volume : 3  |  Issue : 1  |  Page : 136-137

Authors reply to Pinninti et al., Niyogi and Baheti

1 Department of Radiology, Tata Memorial Centre, Homi Bhabha National Institute, Mumbai, Maharashtra, India
2 Department of Pathology, Tata Memorial Centre, Homi Bhabha National Institute, Mumbai, Maharashtra, India

Date of Submission27-Jan-2020
Date of Acceptance27-Jan-2020
Date of Web Publication24-Feb-2020

Correspondence Address:
Swapnil Ulhas Rane
Department of Pathology, Tata Memorial Hospital, Parel, Mumbai, Maharashtra
Login to access the Email id

Source of Support: None, Conflict of Interest: None

DOI: 10.4103/CRST.CRST_43_20

Get Permissions

How to cite this article:
Mahajan A, Vaidya T, Gupta A, Rane SU. Authors reply to Pinninti et al., Niyogi and Baheti. Cancer Res Stat Treat 2020;3:136-7

How to cite this URL:
Mahajan A, Vaidya T, Gupta A, Rane SU. Authors reply to Pinninti et al., Niyogi and Baheti. Cancer Res Stat Treat [serial online] 2020 [cited 2020 Jul 4];3:136-7. Available from: http://www.crstonline.com/text.asp?2020/3/1/136/279116

We thank Pinninti and Rajappa, Niyogi and Baheti for their interest in our article. As aptly pointed out by Pinninti and Rajappa,[1] artificial intelligence (AI) in healthcare is rapidly gaining importance, with its potential utility being demonstrated across several domains of medicine.[2],[3],[4] However, there are currently limited instances of such techniques being successfully translated into clinical practice. Several challenges need to be addressed before AI is successfully implemented in the healthcare sector.[5] One of the crucial issues highlighted by them is that the objective comparison of algorithms across studies poses a challenge due to each study's performance being reported using variable methodologies on different populations with unique sample distributions and characteristics. Unless this issue is overcome, it is difficult for physicians to determine which algorithm can be applied to obtain maximum benefit for their patient population. Another pertinent issue raised by them is the pressing need to develop regulatory frameworks to deploy AI algorithms in a secure and effective manner. Owing to the continuous advancements in the field of healthcare, it is necessary to develop performance monitoring tools/programs to continually calibrate models using human feedback to identify performance deficits and correct these for greater benefit.

With AI poised to establish itself at the forefront of the health sector, it is certainly natural for physicians to wonder: Is AI coming for my job? This point is rightly raised by Niyogi;[6] without a doubt, AI is in a prime position to alter clinical workflow, but its role would be to serve as more of a tool rather than a threat to health professionals, such that they would potentially derive benefit from its pattern recognition and predictive abilities. Thus, we would like to believe that the term 'augmented intelligence' is more apt considering its potential to improve physician accuracy, reduce errors in diagnosis and serve as an incredible asset to underserved regions with limited access to healthcare.

The comments by Baheti highlight the pressing concern of the entire scientific community regarding data ownership and privacy.[7] Often, data custodianship is mistaken for data ownership. While the organization and/or the practitioner who get access to the patient data as a part of routine care have definitely been given consent (written or implied) by the patient to 'hold' their data for the purpose of their care, extending this consent to using or sharing data (anonymized or otherwise) for any additional purpose is not discussed routinely with the patient. Furthermore, once acquired (through additional consent or through a consent waiver authorized by the institutional ethics committee), there is no effective check in place or mandate to make the data and results publicly available to the community at large.

From the researcher's perspective, having complex and time-consuming procedures to acquire additional consent for the use of anonymized data introduces hurdles in the scientific process. This results in the sense of ownership (rather than custodianship) of data in the mind of the researcher who goes through the process of acquiring the additional consent and converting the raw data into meaningful knowledge and wisdom. The solution to this problem should aim at combining these multiple viewpoints, creating a framework for data storage and access acceptable to both patients and researchers and possibly incentivize each party involved to follow the process. We do know of two such efforts by the Department of Science and Technology, where National policies and rules are being defined for data sharing, protection and toward a scientific social responsibility which were open for public consultation till mid-late 2019.[8],[9] Similar efforts are being made by private organizations such as Facebook which is supporting research in this domain to come up with a patient-centric framework for AI research in India and address other related regulatory and ethical issues involved in migrating to an AI-enabled environment in healthcare and other domains.[10]

While there are considerable efforts in recent times, these debates are still in their early stages in the general population and the media. More active interaction between the community at large and the scientific community is essential to bridge this gap.

Financial support and sponsorship


Conflicts of interest

There are no conflicts of interest.

  References Top

Pinninti R, Rajappa S. Artificial intelligence in health-care: How long to go? Cancer Res Stat Treat 2020;3:133-4.  Back to cited text no. 1
  [Full text]  
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.  Back to cited text no. 2
  [Full text]  
Vaidya T, Agrawal A, Mahajan S, Thakur MH, Mahajan A. The continuing evolution of molecular functional imaging in clinical oncology: The road to precision medicine and radiogenomics (Part I). Mol Diagn Ther 2019;23:1-26.  Back to cited text no. 3
Baid U, Talbar S, Rane S, Gupta S, Thakur MH, Moiyadi A, et al. Deep Learning Radiomics Algorithm for Gliomas (DRAG) Model: A Novel Approach Using 3D UNET Based Deep Convolutional Neural Network for Predicting Survival in Gliomas. International MICCAI Brainlesion Workshop. Springer, Cham; 2018. p. 369-79.  Back to cited text no. 4
Kelly CJ, Karthikesalingam A, Suleyman M, Corrado G, King D. Key challenges for delivering clinical impact with artificial intelligence. BMC Med 2019;17:195.  Back to cited text no. 5
Niyogi DM. AI in oncology. Cancer Res Stat Treat 2020;3:134-5.  Back to cited text no. 6
  [Full text]  
Baheti A. Artificial intelligence and its code and mode of conduct. Cancer Res Stat Treat 2020;3:135-6.  Back to cited text no. 7
  [Full text]  
Available from: https://dst.gov.in/sites/default/files/Draft%201%20-%20Biological%20Data%20Policy.pdf. [Last accessed on 2020 Jan 27].  Back to cited text no. 8
Available from: https://dst.gov.in/sites/default/files/Final%20SSR%20Policy%20Draft_2019.09.09_0.pdf. [Last accessed on 2020 Jan 27].  Back to cited text no. 9


    Similar in PUBMED
   Search Pubmed for
   Search in Google Scholar for
    Access Statistics
    Email Alert *
    Add to My List *
* Registration required (free)  

  In this article

 Article Access Statistics
    PDF Downloaded33    
    Comments [Add]    

Recommend this journal