Surveillance and outcomes after curative resection for gastroesophageal adenocarcinoma

BACKGROUND: The goal of surveillance testing is to enable curative salvage therapy through early disease detection, however supporting evidence in gastroesophageal adenocarcinoma is limited. We evaluated frequency of successful salvage therapy and outcomes in patients who underwent surveillance. METHODS: A single‐site, retrospective cohort study was conducted to identify all patients who received curative resection for gastroesophageal adenocarcinoma. Surveillance testing were those investigations not triggered by abnormal symptoms, physical examination, or blood tests. Successful salvage therapy was any potentially curative therapy for disease recurrence which resulted in postrecurrence disease‐free survival ≥2 years. Time‐to‐event data were analyzed using the Kaplan‐Meier method and log rank tests. RESULTS: Between 2011 and 2016, 210 consecutive patients were reviewed. Esophageal (14%), gastroesophageal junction (40%), and gastric adenocarcinomas (45%) were treated with surgery alone (29%) or multimodality therapy (71%). Adjuvant therapy was administered in 35%. At median follow‐up of 38.3 months, 5‐year overall survival (OS) rate was 56%. Among 97 recurrences, 53% were surveillance‐detected, and 46% were symptomatic. None was detected by surveillance endoscopy. Median time‐to‐recurrence (TTR) was 14.8 months. Recurrences included locoregional only (4%), distant (86%), and both (10%). Salvage therapy was attempted in 15 patients, 4 were successful. Compared to symptomatic recurrences, patients with surveillance‐detected recurrences had longer median OS (36.2 vs 23.7 months, P = .004) and postrecurrence survival (PRS, 16.5 vs 4.6 months, P < .001), but similar TTR (16.2 vs 13.3 months, P = .40) and duration of palliative chemotherapy (3.9 vs 3.3 months, P = .64). CONCLUSIONS: Among patients surveyed, 96% of recurrences were distant, and salvage therapy was successful in only 1.9% of patients. Longer OS in patients with surveillance‐detected compared to symptomatic recurrences was not associated with significant earlier disease detection, and may be contributed by differences in disease biology. Further prospective data are warranted to establish the benefit of surveillance testing in gastroesophageal adenocarcinoma.