Using Large-Scale Clinico-Genomics Data for in silico Clinical Trials and Precision Oncology
summary
A computational platform called Trial Pathfinder uses Flatiron EHR plus Foundation Medicine tumor panels to run millions of in silico trials, matching cohorts with propensity/optimal-transport methods, and finds data-driven eligibility that on average more than doubles eligible patients while lowering hazard ratios; a separate analysis scans linked clinico-genomics to report 400+ mutation–treatment interactions, releases PrecisionCancer.org, validates key findings against randomized trials, and is adopted by Roche/Genentech.
AI-written summary · about
chapters (14)
- 0:00Intro
- 1:08Clinical trial circa 600BC
- 3:26In silico trials using large real-world clinico-genomics data
- 5:06Real-world data + Al to make clinical trials more inclusive
- 8:01Eligibility design often anecdotal
- 9:48Trials for similar drugs have different exclusions
- 12:46Using EHR to emulate a trial
- 14:33Using EHR to evaluate trial eligibility
- 18:35Trial Pathfinder uses EHR + Al to design eligibility criteria
- 19:37Example: safely relaxing bilirubin threshold
- 20:33Data-driven design of trial eligibility Broadening eligibility thresholds for lab values (e.g. bilirubin, platelets, hemoglobin).
- 22:46Enabling precision oncology w/ real-world data
- 24:11Example:APC mutation → better immunotherapy response in bladder cancer patients
- 30:46Gene-gene interactions modify the effects of targeted therapies
related talks
20mTranslating a Trillion Pointo Diseases of Data into Therapies
36mReimagining the Future of Clinical Trials
53mThe Algorithm for Precision Medicine
50mThe algorithm for precision medicine
17mBig data in personalized medicine
37mHow Developers can treat Ovarian Cancer
10mGenesis Final Round - Drylab AI
17mPrecision Medicine for All: Using Tidymodels to Validate PRS in Brazil