Case Study: Memorial Sloan Kettering Cancer Center achieves 93% HER-2 patient classification accuracy with Snorkel AI

How Memorial Sloan Kettering Cancer Center (MSKCC) used Snorkel Flow to scale clinical trial screening

Memorial Sloan Kettering Cancer Center (MSKCC) sought to use AI to identify patients for clinical trials by classifying the HER-2 protein in complex medical records. Their progress was bottlenecked by the slow and expensive process of manually labeling training data, which required highly skilled clinicians, and the resulting labels were often inconsistent. They partnered with Snorkel AI to overcome this challenge.

Using Snorkel AI's proprietary technology, the team created a programmatic labeling solution. They developed just a few labeling functions to auto-label thousands of patient records and trained a model that achieved 93% accuracy. This solution, built in weeks instead of months, now powers an AI-driven screening system that accelerates clinical trial recruitment and advances treatment research.

Memorial Sloan Kettering Cancer Center

John Cadley

Bioinformatics Engineer