Serum microRNAs to predict the response to nivolumab in patients with esophageal squamous cell carcinoma.

Abstract

e14511 Background: A phase II study demonstrated that nivolumab (Nivo), a PD-1 inhibitor, had a meaningful activity for patients with esophageal squamous cell carcinoma. Although several biomarkers, including PD-L1 expression levels in tumor tissue, are explored to predict the efficacy of Nivo, further investigation is needed. Here we evaluated whether serum levels of microRNA (miRNA) could be a candidate of predictive markers. Methods: Among 65 patients who participated in the phase II study (JapicCTINo.142422) and received Nivo 3 mg/kg IV Q2W, 19 patients were treated at our institution. Patients were classified into responder and non-responder by investigators based on modified ir-RECIST. Serum samples were stored before and during treatment in the National Cancer Center Biobank. Comprehensive miRNA microarray analyses were performed using a 3D-Gene Human miRNA Oligo Chip (Toray Industries, Inc.), which was designed to detect 2565 miRNA expression levels. miRNA were compared between responders and non-responders using Fisher’s linear discriminant analysis, and then we calculated the area under curve (AUC) values of receiver operating characteristic curve. We explored the miRNA that showed AUC values of more than 0.8 and difference by 0.5 in the average log2 value of miRNA levels (log2miRNA) between responders and non-responders, and investigated their relation to progression-free survival using Kaplan-Meier curves and log-rank tests. Results: Among 19 patients, 5 responded (CR/PR, 1/4) to Nivo. We identified 3 miRNAs in the serum before treatment that were related to response to Nivo with AUC of 0.84 (log2miRNA of non-responder/responder: 13.16/12.47), 0.90 (8.65/10.20) and 0.81 (7.18/6.57), respectively. In the serum after the first treatment, 5 miRNAs were related to response to Nivo with AUC of 0.93 (11.93/11.09), 0.93 (11.87/11.13), 0.85 (7.92/8.53), 0.92 (6.61/5.82) and 0.93 (7.77/ 7.09), respectively. Among these 8 miRNAs, 4 miRNAs were significantly associated with progression-free survival. Conclusions: We identified miRNA candidates that could predict the response to Nivo. Further validation is warranted to confirm the results of our explorative research.