PRATIK SHAH

Artificial intelligence for drug discovery and clinical trials

Phase 3 clinical outcome trials to evaluate new drugs, therapies, and vaccines are among the most complex experiments performed in medicine. Around 50% of Phase 3 trials fail. The U.S. Food and Drug Administration’s recently published white paper, “22 case studies where Phase 2 and Phase 3 trials had divergent results” demonstrates that a common theme is the difficulty of predicting clinical results in a wide patient base, even with the backing of solid data. More importantly, the barriers to this cost healthcare industries, government, and academic research hospitals millions of dollars each year, drive up costs, delay life-saving treatments to patients, and in some cases lead to adverse events. The crucial obstacle is a limited knowledge of the key parameters that need to be considered in order to test candidate molecules, eliminate adverse events, and select optimal IC50. Consequently, we are developing new machine learning architectures with automated learning and predictions gleaned from the past experimental successes and failures of drugs leading to the designing of faster, safer, and more efficacious clinical trials. This work has a significant impact on the ethical and regulatory decisions facing patients, the pharmaceutical industry, and the FDA and EMA.

Publications

  1. 2018. Reinforcement learning with action-derived rewards for chemotherapy and clinical trial dosing regimen selection [Abstract] [Full Paper]

    Gregory Yauney, Shah P*

    (*Senior author supervising research)

    Machine Learning for Healthcare Conference

  2. 2018. Machine learning and Artificial Intelligence in clinical development: a translational perspective

    Pratik Shah* et al.

    (*Corresponding author)

    Nature Reviews Drug Discovery-(Under review)

Other MIT Research Areas