More Power to Clinicians

The easiest, fastest and possibly most impactful way to address the shortage of radiologists in rural India is by giving clinicians and surgeons the ability to read scans on their own using assistance provided by Al. General physicians and pulmonologists in rural India often order chest X-rays for their patients but seldom get reports for them - an app on their smartphone linked to a cloud-based Al system can help them read these scans faster and more accurately. Abdominal surgeons in India are already reading CT scans - imagine a situation where Al directs them to the abnormality and confirms the provisional diagnosis made by them. Clinicians often complain that they find it hard to read MRI scans because of the different ‘contrasts’ that MRI presents - Al can potentially combine all MRI sequences to make a ‘CT-like’ image which is easily readable by them. In such a scenario, one can assume that most routine, mundane, day-to-day reporting tasks would be handled by clinicians in conjunction with Al, and radiologists could instead focus on the more difficult rare cases, thereby optimally utilizing their time. That said, just as how chess amateurs team up with computers to beat experts (Centaur chess, 2019), could one envision a day when primary care clinicians team up with Al to become better than radiologists?

In summary, Al will surely address the shortage of radiologists in rural India in the short term but the extent to which it would affect impact on-ground would depend on use-cases chosen by developers, the accuracy of the algorithms and economics involved in development and deployment of such solutions. But for the successful deployment of algorithms, we need a systematic methodology to validate and audit these algorithms.

References

Bhaktwani, A. (2012) The PC-PNDT act in a nutshell. Indian Journal of Radiology Imaging. 22(2), 133-134. doi: 10.4103/0971-3026.101114.

Bien, N., Rajpurkar, P., Ball, R. L., et al. (2018) Deep-learning-assisted diagnosis for knee magnetic resonance imaging: development and retrospective validation of MRNet. PLoS Medicine. 15, el002699-

Centaur chess shows power of teaming human and machine. Retrieved from https://www.huffpost.com/entry/centaur-chess-shows-power_b_6383606 (accessed on April 2, 2019).

ChestX-ray8: hospital-scale chest X-ray database and benchmarks on weakly-supervised classification and localization of common thorax diseases. (n.d.). Retrieved from https://arxiv.org/abs/1705.02315 (accessed on November 4, 2019).

Chilamkurthy, S., Ghosh, R., Tanamala, S., et al. (2018). Deep learning algorithms for detection of critical findings in head CT scans: a retrospective study. The Lancet. 392, 2388-2396.

Dhanlaxmi Bank. (2019). In-focus: rural India - where is it? Retrieved from https:// www.dhanbank.com/pdf/reports/InFocus-December%201,%202010.pdf (accessed on April 2, 2019).

Fischer, D. (2017). Need a CT scan in India? You might have to look around.

Global Healthcare Insights Magazine. Retrieved from https://globalhealthi. com/2017/04/20/medical-imaging-india/ (accessed on November 4, 2019).

Funding analysis of companies developing machine learning solutions for medical imaging; Simon Harris, Signify Research, June 2018.

Guidelines for competency based postgraduate training programme for MS in Orthopedics. Retrieved from https://old.mciindia.org/PG-Curricula/ MS-Orthopedics.pdf (accessed on April 2, 2019).

Healthtech startup 5C network raises funding from Unitus Ventures, Axilor and CIIE-IIMA (2019). Retrieved from https://yourstory.com/2019/01/ healthtech-startup-5c-network-funding.

Hwang, E. J., Park, S., Jin, K. N., et al. (2019). Development and validation of a deep learning-based automated detection algorithm for major thoracic diseases on chest radiographs. JAMA Network Open. 2, el91095.

Irvin, J., Rajpurkar, P., Ko, M., et al. (2019). CheXpert: a large chest radiograph dataset with uncertainty labels and expert comparison. arXiv:1901.07031 [cs, eess]. Retrieved from http://arxiv.org/abs/1901.07031.

IvanStajduhar, Mihaela Mamula, Damir Miletic, and GozdeUnaL Semi-automated detection of anterior cruciate ligament injury from MRI. Computer Methods and Programs in Biomedicine, 140:151-164, Mar 2017. doi: 10.1016/j. cmpb.2016.12.006.

Johnson, A. E. W„ Pollard, T. J., Greenbaum, N. R„ et al. (2019). MIMIC-CXR: a large publicly available database of labeled chest radiographs. arXiv:1901.07042 [cs, eess]. Retrieved from http://arxiv.org/abs/1901.07042.

Kim, D. W„ Jang, H. Y, Kim, K. W„ Shin, Y. & Park, S. H. (2019). Design characteristics of studies reporting the performance of artificial intelligence algorithms for diagnostic analysis of medical images: results from recently published papers. Korean Jornal of Radiology. 20, 405.

Malik, P. S., & Raina, V. (2015). Lung cancer: prevalent trends & emerging concepts. Indian Journal of Medical Research. 141(1), 5-7.

Nagarajan, R. (2019). Services, PSU hospitals can churn out 1,100 more specialists.

Times News Network. Retrieved from https://timesofindia.indiatimes.com/india/ services-psu-hospitals-can-churn-out-1100-more-specialist-docs-niti-aayog/ articleshow/68441374.cms (accessed on April 2, 2019).

NHA. (2019). About Pradhan Mantri Jan Arogya Yojana (PM-JAY) | Ayushman Bharat. Retrieved from https://pmjay.gov.in/about-pmjay (accessed on November 4, 2019).

Putha, P., Tadepalli, M., Jain, S., et al. (2018). Efficacy of deep learning for screening pulmonary tuberculosis. European Congress of Radiology 2018, February 28-March 4, 2018, Vienna. SS 1805 - Deep Learning.

Rajpurkar, P., Irvin, J., Ball, R. L., et al. (2018). Deep learning for chest radiograph diagnosis: a retrospective comparison of the CheXNeXt algorithm to practicing radiologists. PLoS Medicine. 15, el002686.

Sahu, A. K., Aggarwal, B., Dhar, A., Kapoor, G., & Rao, P. (2018). Evaluating the use of a deep learning algorithm for radiology quality assurance in out-patient chest X-ray reporting. Radiological Society of North America 2018 Scientific Assembly and Annual Meeting, November 25-November 30, 2018, Chicago IL. Retrieved from http://archive.rsna.org/2018/18015032.html.

Singh, R., Kalra, M. K., Nitiwarangkul, C., et al. (2018). Deep learning in chest radiography: detection of findings and presence of change. PLoS One. 13, e0204155.

The National Lung Screening Trial Research Team. (2011). Reduced lung-cancer mortality with low-dose computed tomographic screening. New England Journal of Medicine. 365, 395-409.

Venugopal, V., Manoj, T. D. L., Bhargava, R., et al. (2019a). Automated classification of chest X-rays as normal/abnormal using a high sensitivity deep learning algorithm. European Congress of Radiology 2019, February 27-March 3, 2019, Vienna. SS1905a - Machine Learning in chest radiology.

Venugopal, V, Vaidya, A., Ahuja, A., et al. (2019b). Towards radiologist-level malignancy detection on chest CT scans: a comparative study of the performance of convolutional neural networks and four thoracic radiologists. European Congress of Radiology 2019, February 27-March 3, 2019, Vienna. Scientific Exhibit, doi: 10.26044/ecr2019/C-2065.