At the present, identifying head and neck squamous cell carcinoma (HNSCC) patients for immune checkpoint inhibitor therapy (ICIT) is achieved through the determination of Tumor Proportion Score (TPS) or the percentage of tumor cells positively labeled for PD-L1. Estimation of TPS is largely done in a manual fashion by a trained pathologist. In the case of HNSCC, the histological section can be over 1 cm in size in which over 100,000 cancer cells need to be evaluated for PD-L1 expression. To expedite the TPS evaluation process for such large specimens, we have developed a platform in which artificial intelligence (AI) is used for TPS determination. With additional development, this approach may be used in the clinical setting to assist pathologists in TPS evaluation.
|