PRATIK SHAH

Pratik Shah leads special session on Regulatory Science with the FDA for Diversity, Equity, and Inclusion in Biomedical Imaging at IEEE ISBI

Pratik Shah leads special session on Regulatory Science with the FDA for Diversity, Equity, and Inclusion in Biomedical Imaging at IEEE ISBI

March 31, 2022
8:00am - 10:00am ET

Dr. Pratik Shah conceptualizes and leads regulatory science special session on the value of diversity, equity and inclusion of races, genders and ethnicities in biomedical images for uncertainty quantification and fair and unbiased deep learning.

Conference: IEEE International Symposium on Biomedical Imaging 2022.

Abstract and Motivation
Tools, models and statistical methods for signal processing and medical image analysis and training deep learning models are of special interest to the biomedical community to create research prototypes for eventual clinical applications. Dr. Shah will lecture on the recent research from his lab published in Cell Reports Methods, that outlined strategies and methods for training of unbiased, high-performance locked and adaptive deep learning models for uncertainty quantifications. Experts from University of California at San Francisco will emphasize the value of integration of unbiased imaging data (race and ethnicity for dermatology as examples) and their impact on deep learning model performance. Speakers from the US Food and Drug Administration (FDA) will then discuss regulatory pathways for real-world performance of medical devices using medical imaging within the context of AI/ML-enabled software and deep learning. Finally, the session will explore benchmarking of devices, software and medical image data to improve safety and patient outcomes with notions of diversity, fairness and inclusion.

Target Audience and Value Proposition
This session is suitable for community members interested in modern topics in acquisition of medical images and their real-world clinical applications in deep learning, and challenges such as bias and inequity and emerging regulatory framework. Topics include digital image processing, deep learning, tissue structure and anatomy, uncertainty estimations and statistical modeling. Newer concepts in using regulatory science methods for Software as a Medical Device (SaMD), and equity and fairness for benchmarking data and models for medical imaging will also be covered.