4614 - Predictors of Response to Checkpoint Inhibitors in Naïve and Ipilimumab-Refractory Melanoma

Background

Response to checkpoint inhibitors (CI) is governed by the tumor immune environment and understanding this immune contexture can predict response. Therapeutic intervention can change this environment even in the absence of clinical response. Patients failing initial immunotherapy may respond to a second line of CI; however, these cohorts show lower overall response rates (ORR). This study identifies transcriptional signatures associated with response to first- and second-line CI monotherapy in melanoma.

Methods

CI-naïve or ipilimumab-refractory patients were treated with ipilimumab, nivolumab or pembrolizumab at the Instituto Nazionale Tumori and clinical response was evaluated by irRECIST 1.1 criteria. Pretreatment tumor biopsies (n = 82) from metastatic lesions were collected and RNA was profiled with the NanoString® IO360 gene expression panel.

Results

Compared to CI-naïve cohorts, ipilimumab-refractory cohorts had reduced ORR to nivolumab (naïve: 35%, n = 6; refractory: 20%, n = 10) or pembrolizumab (naive: 67%, n = 6; refractory: 20%, n = 10) with multiple genes differentially expressed between groups. The Tumor Inflammation Signature, an investigational 18 gene signature of suppressed adaptive immune response enriching for pembrolizumab response, was higher in responders versus non-responders in first-line (log2 fold change: 1.56, p = 0.21), but not second-line pembrolizumab (log2 fold change: 0.41, p = 0.60). First-line pembrolizumab responders had elevated MHC2 (log2 fold change: 1.35, p = 0.02) and B cell (log2 fold change: 2.14, p = 0.02) signatures. Upon stratifying the CI-naïve cohort between no prior treatment versus prior targeted/chemotherapy, the latter had increased immune expression suggesting these therapies prime the tumor immune environment.

Conclusions

Correlating patterns of tumor gene expression with clinical response can lead to the development of biomarkers enriching for CI response in both first-line and CI-refractory patients. Utilization of a clinical grade platform such as the NanoString nCounter® may speed the development of diagnostic assays used to predict and monitor patient response to immunotherapy.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

The authors.

Funding

NanoString Technologies.

Disclosure

S. Ong: Full / Part-time employment: NanoString Technologies. S. Warren: Shareholder / Stockholder / Stock options, Full / Part-time employment: NanoString Technologies. A. Cesano: Shareholder / Stockholder / Stock options, Full / Part-time employment: NanoString Technologies. J.M. Beechem: Shareholder / Stockholder / Stock options, Full / Part-time employment: NanoString Technologies. P.A. Ascierto: Advisory / Consultancy: Amgen; Advisory / Consultancy: Array; Advisory / Consultancy: BMS; Advisory / Consultancy: Incyte; Advisory / Consultancy: Immunocore; Advisory / Consultancy: MedImmune; Advisory / Consultancy: IDERA; Advisory / Consultancy: Genmab; Advisory / Consultancy: Merck; Advisory / Consultancy: Roche; Advisory / Consultancy: Genentech; Advisory / Consultancy: Sandoz; Advisory / Consultancy: Syndax; Advisory / Consultancy: Sun Pharma; Advisory / Consultancy: Ultimovacs; Advisory / Consultancy: Pierre Fabre; Advisory / Consultancy: Novartis. All other authors have declared no conflicts of interest.