Merck & Co. has tapped the artificial intelligence developer Owkin to help assemble a new slate of digital diagnostics for cancer testing.
The collaboration will expand upon Owkin’s machine learning-powered test for determining microsatellite instability status among people with colorectal cancer—using it as a platform to build similar tests in endometrial, gastric, small intestine and bile duct cancers.
The French company’s MSIntuit CRC program—which has received a CE Mark approval in Europe—is used to pre-select scanned pathology slides, prioritizing those that should undergo full immunohistochemistry testing, PCR screening or next-generation sequencing to capture the genomic biomarker.
In patients with colorectal cancer, finding out a tumor’s level of microsatellite instability can be essential in guiding treatment. Representing defects in a cell’s ability to correct errors in its DNA, those with high levels can see fewer benefits from chemotherapy, but better outcomes when matched with targeted immunotherapies—such as Merck’s mainstay immune checkpoint inhibitor Keytruda.
Testing for high status, known as MSI-H, is now recommended by international cancer treatment guidelines—with about 15% of colorectal cancer patients carrying the biomarker—however, according to Owkin, that screening is not typically performed in the collaboration’s four cancer targets.
“It is clear that there is a need for AI diagnostics that can both ease bottlenecks and resource pressures while also ramping up biomarker testing to match patients with optimal treatments,” Owkin’s chief diagnostics officer, Meriem Sefta, said in a statement.
Keytruda previously received a groundbreaking approval from the FDA in 2017 that cleared its use against tumors based on MSI-H status, regardless of the cancer’s location in the body. Owkin’s planned digital tests for endometrial, gastric, small intestine and bile duct cancers are bound first for Europe, where Merck trades under the name MSD.