Multicenter phase II study of neoadjuvant nivolumab or nivolumab plus relatlimab (anti-LAG3 antibody) plus chemoradiotherapy in stage II/III esophageal/gastroesophageal junction (E/GEJ) carcinoma.

321 Background: The phase III CheckMate 577 study resulted in adjuvant nivolumab becoming a new standard of care for patients with completely resected E/GEJ cancer with residual pathologic disease post neoadjuvant chemoradiotherapy. We evaluated if neoadjuvant nivolumab (N) or nivolumab/relatlimab (N/R) combined with chemoradiotherapy (CRT) can further improve patient outcomes. Methods: Patients with stage II/III E/GEJ carcinoma eligible for curative resection were treated with standard of care regimen of carboplatin (AUC2), paclitaxel (50mg/m2), RT 41.1Gy in 23 fractions and an Ivor-Lewis esophagectomy (E/MIE) 6-10 weeks after last CRT/IO dose. Patients on arm A (n=16) received 2 cycles of induction N (240mg q2 wks) plus three additional cycles of N on week 1, 3 and 5 of CRT. After safety and feasibility evaluation of arm A, patients on arm B (n=16) received N (240mg q 2 wks) plus R (80mg q 2 wks) following the same schedule. The primary endpoints of the study were safety and feasibility. Secondary endpoints include pCR, MPR (<10% residual cancer cells), DFS and OS. We also evaluated pathologic response and molecular ctDNA responses via longitudinal targeted error correction sequencing. Results: From August 2017 to July 2021, 32 patients were enrolled. Median age 65 (39 to 73), male 81%, adeno/SCC (87.5%, 12.5%). CRT combined with N on arm A was well tolerated with 4/16 (25%, 95% CI: 9.3-52.6%) reporting grade 3 AEs. Dual IO inhibition targeting PD1 and LAG3 combined with CRT on arm B demonstrated unacceptable toxicities as per predefined early stopping rule after 9 patients which resulted in a protocol amendment; 6/9 patients (66%) in arm B developed G3 or higher IO-related toxicities including pericarditis (2/9, 22%) and adrenal insufficiency (2/9, 22%). The amended arm B (n=7) involved induction N + R for 2 cycles prior to standard CRT and was well tolerated. In 31 evaluable patients to date, the pCR rate is 29.0% (95% CI: 14.9-48.2%), (arm A, pCR 6/16 (37.5%, 95% CI: 16.3-64.1%) and MPR 8/16 (50.0%, 95% CI: 28.0-72.0%) (arm B, pCR 3/15 (20.0%, 95% CI: 5.3-48.6% and MPR 8/15 (53.3%, 95% CI: 27.4-77.7%). With a median follow-up time of 30m, the median DFS is 35.4m (95% CI: 24.7-NA) and 1 year DFS rate of 79.1% (95% CI: 65.5-95.6%). For patients in Arm A (arm B data pending), ctDNA clearance was associated with pathologic response, while ctDNA persistence was linked with disease recurrence. Conclusions: The addition of N to preoperative CRT is safe and is associated with a higher MPR rate and pCR in Arm A compared to historical controls. In this study, neoadjuvant anti PD-1/LAG3 combination IO-IO strategies with CRT were challenging due to enhanced IR toxicities. In depth immune correlates and liquid biopsy analyses will be presented. Clinical trial information: NCT03044613.