AI model improves prediction of immunotherapy response across multiple cancers

US researchers have developed an artificial intelligence (AI) model that could improve prediction of which cancer patients are likely to benefit from immune checkpoint inhibitors (ICIs), potentially supporting better patient selection in immunotherapy trials.

The model, COMPASS, uses bulk tumor RNA sequencing data to predict responses to ICIs, including anti-PD-1, anti-PD-L1 and anti-CTLA-4 therapies. In a Nature Medicine paper published on July 3, the researchers said COMPASS was pretrained on 10,184 tumors across 33 cancer types and assessed across 16 clinical cohorts covering 1,133 ICI-treated patients, seven cancers and six checkpoint inhibitors.