A Pilot Study of Neoadjuvant FOLFIRINOX followed by Chemoradiation for Gastric and Gastroesophageal Cancer: Preliminary Results and Prognostic Implications of ctDNA

Abstract

We performed a single-arm pilot study of neoadjuvant FOLFIRINOX and chemoradiation (CRT) with concurrent carboplatin/paclitaxel followed by surgery in patients with locally advanced gastric or gastroesophageal (GEJ) cancer. Patients were enrolled on an NCI sponsored, prospective, single arm study (NCT03279237). Key eligibility criteria included: histologically confirmed T3/4 or lymph node (LN) positive gastric or GEJ adenocarcinoma, ECOG PS ≤1, and life expectancy > 3 months. Extensive LN disease beyond the surgical field (supraclavicular or para-aortic) was permitted if deemed able to be encompassed within the RT field. Pts received neoadjuvant FOLFIRINOX x 8, restaging, CRT (45 Gy for gastric, 50.4 Gy for GEJ) with concurrent C/T, restaging, followed by resection. Serum ctDNA was measured at baseline, prior to CRT, and preoperatively. Tumor genotyping was performed on all pre-treatment specimens to identify mutations in the primary tumor, and mutation-specific droplet digital PCR was used to detect mutation fraction in ctDNA. The primary objective was to determine the rate of completion of FOLFIRINOX x 8 followed by CRT delivered in the preoperative setting. Secondary endpoints included: 1) acute toxicity and 2) pathologic complete response (pCR), and 3) correlation of ctDNA to treatment response. Fisher’s exact test was applied to determine predictive value of ctDNA to treatment response. From Oct 2017 to June 2018, 25 pts were enrolled. Median age was 60 (range:30-76), and 17 pts were male (68%). All pts started FOLFIRINOX, and 23 pts were able to complete all 8 planned cycles. Two pts did not complete the planned 8 cycles due to metastatic progression. Rates of grade 3+ toxicity for overall, gastrointestinal, and hematologic were 28, 12, and 28% respectively. Of the entire cohort, 23 pts started chemoRT and 22 pts completed chemoRT (1 pt died during CRT due to PEA arrest). All 22 pts who completed CRT went for surgical exploration. Of the 22 pts (88%) who went for surgical exploration, 2 pts were found with intraoperative metastases. Thus, in total, 20 (80%) pts underwent surgical resection all with R0 resection. In total, 7 pts had a pCR (35% in resected cohort, 28% in ITT cohort). In the entire cohort, tumor specific mutations were identified in 21 patients (84%) for ctDNA. Of the 21 patients with evaluable ctDNA levels, 3 patients had undetectable ctDNA at baseline (defined as 0) and 18 had detectable ctDNA. Although not significant due to low power, undetectable ctDNA was associated consistently with higher rate of pCR compared to detectable ctDNA at all-time points (67% vs. 28% at c1D1, p=0.24, 50% vs. 11% at c1D8, p=0.16, and 46% vs. 20% at post-CRT, p=0.6). Neoadjuvant FOLFIRINOX followed by CRT is feasible with acceptable grade 3+ toxicity. In our small series, the rate of pCR is promising and a follow-up study is currently planned. Although our current study had limited events, ctDNA appears to be a promising predictor of pCR.