Phase II randomized trial of first-line pembrolizumab and vorinostat in patients with metastatic NSCLC (mNSCLC).

Andreas Saltos, Tawee Tanvetyanon, Ben Creelan, Michael Shafique, Scott Antonia, Eric B. Haura, Hong Zheng, Margaret L. Barlow, James J. Saller, Anna Castellano-Fornelli,

Journal of Clinical Oncology · 2020 · 13 citations · 0 references

Concepts

Abstract

9567 Background: Histone deacetylase inhibitors may enhance tumor immunogenicity through various mechanisms including induced expression of T cell chemokines. A previous phase I trial demonstrated the combination of pembrolizumab (P) with vorinostat (V) in mNSCLC was well tolerated with signals of activity in ICI-pretreated pts. We initiated a randomized trial in the first-line setting with the primary objective to determine if the combination had superior ORR compared to pembrolizumab monotherapy. Methods: Pts with treatment-naïve mNSCLC and PD-L1 expression ≥ 1% were eligible. Pts were randomized open-label 1:1 to receive P 200 mg IV q3 wk as monotherapy [Arm A] or P 200 mg IV q3 wk plus V 400 mg PO daily [Arm B]. The primary endpoint was overall response rate (ORR). Secondary endpoints included DOR, PFS and OS. Tumor biopsies were collected both pre- and on-treatment (day 15-21) for analysis of CD8+ TIL, scored using a 0-3 scale in tumor beds. Here we report results after a preplanned interim analysis for efficacy, with accrual ongoing to a planned total of 39 patients per arm. Results: Between 7/2017 – 1/2019, 49 pts were enrolled, with 47 pts evaluable for response (24 in Arm A and 23 in Arm B). Median age was 69 (range 47 - 87), 49% female, ECOG PS 0/1 in 11%/89%. PD-L1 TPS was ≥50% in 13/24 (54%) of pts in Arm A, and in 13/23 (57%) of pts in Arm B. The most common TRAEs in Arm A included diarrhea (13%), fatigue (8%), and pruritus (8%). 3 pts in Arm A experienced grade ≥ 3 irAEs (including 1 each of grade 3 hepatitis, pneumonitis, and rash). The most common TRAEs in Arm B included anorexia (43%), fatigue (43%), nausea (35%) and increased creatinine (35%). 1 pt in Arm B experienced grade ≥ 3 irAE (1 grade 3 pneumonitis). Pre-treatment CD8+ TIL were not significantly different between Arm A and Arm B (p = 0.85) with the majority of tumors in both arms having a low TIL score of 1 (65% Arm A and 73.7% Arm B). A significant increase from pre-treatment to on-treatment TIL scores was seen in both Arm A (p = 0.001) and Arm B (p = 0.002). The ORR in Arm B pts with low pre-treatment TIL (score = 1) pts was substantially higher (66.7%) than in Arm A (33.3%), suggesting the combination may be especially beneficial against low TIL tumors. Conclusions: The combination arm had a considerably higher ORR compared to pembrolizumab monotherapy, with a manageable toxicity profile. The combination of pembrolizumab plus vorinostat in mNSCLC warrants further investigation. Clinical trial information: NCT02638090 . [Table: see text]