Combined Biomarkers for Prediction of Immune Checkpoint Inhibitor Response in Patients With Triple‐negative Breast Cancer

Background/Aim: Triple-negative breast cancer (TNBC) is an aggressive malignancy with few available targeted therapies. In this study, we examined the predictive power of several biomarkers, comprising the IMmunotherapy Against GastrIc Cancer (IMAGiC) model, PD-L1 combined positive score (CPS), intra-tumoral tumor-infiltrating lymphocytes (iTILs), and stromal TILs (sTILs), in an Asian population of patients with metastatic TNBC treated with immunotherapy. Patients and Methods: Thirty-one metastatic TNBC patients receiving immunotherapy were analyzed. For measuring expression levels of the mRNA of four immune-related genes in the IMAGiC test, quantitative real-time polymerase chain reaction was used. TIL detection and quantification were conducted using Lunit SCOPE IO, which is an AI powered spatial TIL analyzer. Results: Patients were classified into IMAGiC responder and non-responder groups according to IMAGiC score cut off value. There were significantly more clinical responders [complete (CR) or partial (PR) response] in the IMAGiC responder group than in the IMAGiC non-responder group (50% vs. 15.3%, p=0.05). Area under the curve (AUC) values were calculated to examine the predictive value of the IMAGiC score, PD-L1 CPS, iTILs, and sTILs, for response to immunotherapy. The AUC values of the IMAGiC group and score were 0.684 and 0.632, respectively. When the IMAGiC group and iTIL level were combined, the highest AUC value of 0.755 was obtained.