NEW YORK – South Korean biotech firm Lunit said on Wednesday that it has obtained the CE-IVD mark for its artificial intelligence-based PD-L1 expression test, called Lunit Scope PD-L1 TPS.
The test uses machine learning to analyze whole slide images of immunohistochemistry-stained tumor tissue samples and detect PD-L1 expression in both tumor cells and inflammatory cells. By analyzing PD-L1 expression in tumor cells, the test calculates a tumor proportion score, and by assessing PD-L1 expression in tumor and inflammatory cells, it provides a combined positive score. The test also classifies patients into three groups according to their PD-L1 score: 50 percent or more, between 1 percent and 49 percent, and 0 percent.
Patients' PD-L1 expression status can be important for determining their eligibility for checkpoint inhibitor treatment. Lunit, based in Seoul, expects to launch Lunit Scope PD-L1 TPS in Europe in the second half of this year, making it available to pathology labs.
In a study presented at a medical meeting last year, researchers showed that Lunit's AI-powered PD-L1 test could accurately predict the PD-L1 combined positive score in urothelial carcinoma when compared to pathologists' consensus, manually determined PD-L1 readings.
"This CE mark certification is a meaningful milestone," Lunit CEO Brandon Suh said in a statement. "It demonstrates the expansion of Lunit's offerings to the cancer biopsy field beyond the scope of medical image analysis. As our AI technology gains reliability and recognition, we will do our utmost to pioneer the overseas market."
Lunit is also developing another AI-powered test that can analyze the presence of tumor infiltrating lymphocytes (TILs) in tumor tissue to predict checkpoint inhibitor response, called Lunit Scope IO. This test is currently available for research use, but the company aims to seek regulatory approval for the TIL test by 2024.