127 Digital pathology prognostic biomarkers- time for clinical application in melanoma

Résumé fourni par la source

Background Melanoma is a highly immunotherapy sensitive cancer, and there is an urgent need for immune based biomarkers for clinical trial stratification and to guide clinical management. Digital pathology correlates of a favorable tumor immune micro-environment (TIME) have been proposed, but none have yet been validated for successful application to clinical practice. We validated three previously published prognostic immune biomarkers in stage II-III melanoma - NanoString based Melanoma Immune Profile (MIP), CD8/CD68 ratio using quantitative immunofluorescence (qIF), and electronic TILS (eTILS). Methods We tested all three biomarkers (MIP, CD8/CD68 and eTILs) in a new independent cohort of 160 patients from Roswell Park Comprehensive Cancer Center (RPCC). Receiver operating curves (ROC) and Kaplan Meier (KM) curves were generated using previously published methods. Demographics are shown in table 1. Success rates for tissue processing are shown in table 2. Results All three immune biomarkers correlated with distant metastatic recurrence (DMR), the original endpoint used in prior publications. Area under the curve (AUC) was 0.745, p<0.001, for NanoString immune signature (MIP), 0.681, p<0.001 for CD8/CD68 ratio and and 0.722, p<0.001 for eTILs (figure 1A). Next, because recurrence free survival (RFS) is the primary endpoint used in adjuvant trials, we tested whether the biomarkers correlated with RFS. Again, all three biomarkers correlated with recurrence within 36 months (AUC=0.755, p< 0.001, AUC=0.619, p<0.001, and AUC=0.706, p<0.001) (figure 1B). All three immune biomarkers high and low risk RFS groups as shown in figure 2 (p=0.00039, p=0.014, and p=0.0074). Finally, stage IIB-IIIA tumors were analyzed separately as there is urgent clinical need for these lower risk patients. Both MIP and eTILs significantly distinguished high and low risk populations (p<0.0001 and p=0.0047, figure 3), and contribute to Cox survival models (<0.001 and 0.012 table 2). Conclusions We propose that immune biomarkers should be further tested in prospective setting for near term application as stratification tools for clinical trials in early-stage melanoma. This may accelerate development of successful therapeutics by allowing for smaller and more conclusive trials enrolling high risk patients. Further, patients at low risk fo recurrence could be managed more conservatively avoiding toxic immunotherapy combinations while higher risk patients would be prioritized for clinical studies. Applying pathomics to clinical care is an urgent need and early- stage melanoma is an ideal case use.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.