Positive Association Between Location of Melanoma, Ultraviolet Signature, Tumor Mutational Burden, and Response to Anti–PD-1 Therapy

PURPOSE Emerging evidence suggests a correlation between the tumor mutational burden (TMB) and the response to programmed cell death-1 protein (PD-1) monotherapy across multiple cancer types. In skin cancers, as high TMB is mostly because of ultraviolet (UV) exposure, we hypothesized a correlation between the primary melanoma cutaneous location according to sun exposure and response to anti–PD-1 monotherapy. METHODS The aim of this study was to analyze, in advanced melanoma, the relationship between TMB, locations according to sun exposure, and response to PD-1 inhibitors. We conducted a prospective multicentric analysis, by sequencing the most recent metastatic sample before PD-1 inhibitors using FoundationOne assay. RESULTS One hundred two patients were included, with TMB available for 94 cases. In univariate and multivariate linear regression, TMB was significantly associated with sun-exposed areas of the primary melanoma location and with age (coefficients of the association with log-TMB: non-UV location, –1.05; chronic sun-exposed area, 1.12; P value for the location, < 10–5; age, 0.021 per year, P value for age, .002). Molecular UV signature present on the metastatic site was associated with higher TMB ( P = .003). Melanomas bearing a high TMB had a higher probability of response to PD-1 inhibitors compared with melanomas with a low TMB, with a dose-dependent effect following an exponential curve and a negative odds ratio of 0.40 (95% CI, 0.20 to 0.72, P = .004) between log-TMB and 6-month progression. CONCLUSION Cumulative sun exposure related to skin location and molecular UV signature present on the metastatic site appear to be relevant biomarkers directly linked to TMB. Because TMB is not yet available to all for routine clinical use, the location of the primary melanoma in a sun-exposed area may play an important role in clinical decisions regarding therapeutic choice.

mTORC1 activation assessed in metastatic sample to predict outcome in patients with metastatic breast cancer treated with everolimus-exemestan: Results from the SAFIRTOR study.

1024 Background: Using samples from TAMRAD study (Treilleux, Ann Oncol, 2015), we previously reported that p4EBP1, a downstream protein of mTOR, was associated with higher benefit to everolimus (eve). SAFIRTOR study was designed to validate clinical utility of this biomarker. Methods: Patients (pts) with ER+, HER2 negative, AI resistant MBC were prospectively included (NCT02444390). All pts had a biopsy of a metastatic site and were then treated with standard eve + exemestane (exe) combination. The primary end point was to validate that p4EBP1 expression is associated with longer PFS in patients treated with eve. 120 evaluable pts were needed for the pre planed statistical analysis. All samples were collected and processed in a standardized procedure in order to allow phophoproteins IHC staining. In addition to p4EBP1, we explored prognostic value of pS6K, pAkt, PTEN and LKB1, together with genomic alterations assessed by NGS and CGH arrays. Results: 150 pts were included, 30 pts had no adequate sample, and further 13 had missing clinical data, 107 were evaluable for primary objective. Median age was 62, they had previously progressed on AI treatment, either in the adjuvant (22 pts) or the metastatic setting (83 pts). 20 were considered as primary hormone resistant, 87 as secondary resistant. The median Allread score for p4EBP1 was 5.5 (range: 0-6.5). Analysis of the primary endpoint showed that p4EBP1 staining above the median is associated with a longer PFS on eve+exe. (median PFS: 9.3 months, 95CI 6.3-13.1 for high p4EBP1 versus 5.8 months, 95CI 3.7-7.8 for low p4EBP1, p = 0.02). Prognostic value of high pEBP1 remained significant when assessed in a multivariate analysis along classical clinico-biological prognostic factors for MBC (HR 0.57, 95%CI 0.38-0.88, p = 0.01). In this AI resistant population, the tumor of 42 (46%), 33 (35%) and 5 (5.3%) pts carried an activating mutation for ESR1, PIK3CA and AKT1, respectively. None of these mutational statuses were correlated to outcome. Conclusions: This prospective study validates p4EBP1 expression analysis to select patients most likely to benefit from everolimus + exemestane. Clinical trial information: NCT02444390.