PTPRD/PTPRT mutation as a predictive biomarker of immune checkpoint inhibitors across multiple cancer types

BACKGROUND: Immune checkpoint inhibitors (ICIs) are dramatically changing the treatment landscape of a variety of cancers. Nevertheless, the variability in ICI responses highlight the importance in identifying predictive biomarkers. PTPRD and PTPRT (PTPRD/PTPRT) are the phosphatases of JAK-STAT signaling, a critical pathway in anti-cancer immunity regulation. However, the pan-cancer association between PTPRD/PTPRT mutation and the efficacy of ICIs remains unclear across pan-cancer patients. METHODS: We analyzed the association between PTPRD/PTPRT mutations and patient outcomes using clinical data and genomic mutations from TCGA pan-cancer cohort. Furthermore, the ICI-treatment cohort was used to evaluate the relationship between PTPRD/PTPRT mutation and the efficacy of ICIs. Another ICIs-treatment cohort was used to validate the findings. The TCGA pan-cancer dataset was analyzed to explore the correlation between PTPRD/PTPRT mutations and immune signatures. Moreover, we combined four factors to construct a nomogram model that could be used to predict the survival of pan-cancer patients receiving ICI treatment. The calibration curves and area under the curve were applied to assess the performance of the model. RESULTS: PTPRD/PTPRT mutations were shown to be associated with a worse prognosis in TCGA cohort (P < 0.05). In the Samstein cohort, prolonged overall survival (OS) was observed in PTPRD/PTPRT mutant cancers, compared with wild-type cancers (mOS: 40.00 vs 16.00 months, HR = 0.570, 95%CI: 0.479-0.679, P < 0.0001). In the validation cohort, significant OS advantage was observed in PTPRD/PTPRT mutant patients (mOS: 31.32 vs 15.53 months, HR = 0.658, 95%CI: 0.464-0.934, P = 0.0292). Furthermore, PTPRD/PTPRT mutations were associated with a higher tumor mutational burden, MSI score, and TCR score (P < 0.0001). Enhanced immune signatures were found in the PTPRD/PTPRT mutant cancers (P < 0.05). Finally, we successfully established a nomogram model that could be used to predict the survival of NSCLC patients who received ICI treatment. Based on the risk score of the model, patients in the low-risk group showed a better mOS than those in the high-risk group (mOS: 2.75 vs 1.08 years, HR = 0.567, 95%CI: 0.492-0.654; P < 0.001). CONCLUSIONS: PTPRD/PTPRT mutations may be a potential biomarker for predicting ICI treatment responsiveness in multiple cancer types.