Roche has claimed a breakthrough designation from the FDA for a cancer test that combines an antibody assay with artificial-intelligence-powered analysis to provide what the company says could become the agency’s first AI-driven companion diagnostic for non-small cell lung cancer.
Described as a computational pathology system, the Ventana TROP2 RxDx Device includes the immunohistochemistry staining kit and instrument, along with the company’s digital pathology scanners, image management platform and AI algorithm.
The test processes whole-slide images of non-small cell lung cancer tissue to provide a quantitative TROP2 score, which Roche says provides a level of precision beyond manual methods.
TROP2 has become a major target for antibody-drug conjugates, with over a dozen candidates in clinical development. The transmembrane protein is overexpressed in several solid cancers and has been linked to more aggressive tumors.
Roche’s companion diagnostic test aims to identify patients with advanced or metastatic NSCLC who may benefit from treatment with Datroway (datopotamab deruxtecan-dlnk), an ADC developed by AstraZeneca and Daiichi Sankyo that received its first U.S. approval in January for HR-positive, HER2-negative breast cancer.
That same month, Datroway was granted priority review by the FDA for EGFR-mutated NSCLC—which analysts have described as the larger market opportunity, with a potential approval decision expected in mid-July.
With regards to its first-quarter debut as a breast cancer therapy, Daiichi Sankyo executives said this week that Datroway has had a “smooth market launch” in the U.S. and Japan, collecting about $10 million in sales through the end of March.
In NSCLC, however, AstraZeneca and Daiichi Sankyo have seen Datroway—pitched as a follow-up to their blockbuster ADC Enhertu (fam-trastuzumab deruxtecan-nxki)—stumble in a phase 3 trial last year, where it missed its target in improving overall survival.
The drugmakers pursued an AI-powered diagnostic strategy after a retrospective analysis of the study showed Datroway performed better in some patients versus others. Roche’s test incorporates AstraZeneca’s proprietary algorithm platform, known as quantitative continuous scoring, or QCS.
Compared to previous companion diagnostics, which focus on protein expression on the surface of cancer cells, the Ventana TROP2 assay measures the ratio between its appearances on the membranes of tumor cells as well as within their inner cytoplasm. This normalized membrane ratio score, dubbed QCS-NMR, can then classify patients with non-squamous NSCLC as either positive or negative based on a predefined cutoff level.