Relationship between PET response and pathologic response in distal esophageal/gastroesophageal junction (DE/GEJ) cancers.

Abstract

452 Background: The utility of PET scans (PETs) to predict outcomes after neoadjuvant treatment of DE/GEJ cancers is unclear. We aimed to explore the relationship between PET response and pathologic/clinical outcomes in a real-world setting. Methods: Patients (pts) with DE/GEJ cancer treated with curative intent perioperative chemotherapy or neoadjuvant chemoradiation followed by surgery in British Columbia from 2009-2018 were included. Retrospective chart review was conducted; pts were stratified into PET responder (R, ≥ 35% decrease in max SUV) and non-responder (NR, < 35%) groups. Chi-square and Kaplan Meier were used to test for associations between variables and outcomes. Results: Of 576 pts identified, 52% had pre- and post-induction PETs; 232 pts proceeded to surgery and were included for analysis. Treatment regimens comprised of CROSS (72%), MAGIC (24%) and FLOT (4%). Median age was 66 (IQR 57-72), 85% male, 91% ECOG 0/1, 62% GEJ involvement, and 81% adenocarcinoma histology. Characteristics and treatment regimens were balanced between the PET-R and PET-NR groups (all p>0.05). Median time from end of treatment to PET was 30 days (IQR 22-36); 67% were PET-R. Pathologic complete response (PCR) rates were similar for PET-R vs. PET-NR (14% vs. 13%, p=0.08). The discordance rate between PET vs. pathologic response was 34% (Table). Aborted surgery rate was higher in the PET-NR group (8% vs. 3%, p=0.03); 70% of aborted cases were due to peritoneal involvement. Median overall survival was similar between the two groups (PET-R 31.5 mo vs. PET-NR 36.1 mo, p=0.62). Conclusions: In our population-based cohort, PET response did not demonstrate prognostic utility and was associated with a significant pathology discordance rate. The role of PET/CT is evolving and the use of post-induction imaging for response assessment and prognostic value may be questionable. [Table: see text]