369 poor correlation between fdg-pet and pathologic results after experimental neoadjuvant immunochemotherapy for gastroesophageal adenocarcinoma: results of a clinical trial

Abstract A lack of FDG-PET response to neoadjuvant chemotherapy for gastroesophageal adenocarcinoma is considered as indicative of a poor prognosis, whereas pathologic complete or near-complete response (pCR/near-pCR) predicts a more favorable outcome. In a single-arm phase II trial of perioperative immunochemotherapy, we sought to establish a correlation between the results of an FDG-PET performed after the neoadjuvant part of the treatment and the pathologic results obtained after surgery. Methods Inclusion criteria: locally advanced gastroesophageal adenocarcinoma (T3 and/or N+); adequate organ function; ECOG PS 0–1; no contraindication to immunotherapy; Staging: CT, FDG-PET, EUS, diagnostic laparoscopy. Eligible patients receive docetaxel/cisplatin/5-FU (mDCF) on day 1 and avelumab on day 4 of each 2-week cycle and are restaged after 4 cycles. FDG-PET response is defined as reduction of >35% of pre-treatment SUV. Surgery is performed if no disease progression. Four cycles of mDCF/avelumab are administered after surgery. Tumor regression is scored from 0 (no tumor cells) to 3 (no tumor regression). The primary endpoint is the pCR/near-PCR rate (score 0–1). Results As of February 24, 2020, 28 patients have been enrolled and 22 have been operated. Paired FDG-PET response and pathologic response data are available for 21 patients. We observed 16/21 FDG-PET responses and 6/21 pCR/near-pCR. Of the latter, 4 were associated with an FDG-PET response and 2 with no FDG-PET response. Ten patients with score 3 tumor regression had had FDG-PET response. No statistical correlation could be found between FDG-PET response and likelihood of pCR/near-pCR. Conclusion Given the discrepancy observed between FDG-PET and pathologic results, the need for a repeat FDG-PET for gastroesophageal adenocarcinoma after neoadjuvant immunochemotherapy and before surgery should be questioned. Analysis of the tumor microenvironment could possibly explain these results. Future studies will look into correlation between the results presented and survival.