P1.211. Impact of Recovery Achievement and Timing of Esophagrams for Anastomotic Leak Detection After Robotic Esophagectomy

Abstract

Esophageal Cancer: Surgical Treatment of Esophageal Cancer – early outcomes and complications Esophagrams are routinely used after hybrid robotic assisted minimally invasive esophagectomy (hRAMIE) within enhanced recovery after surgery pathways (ERAS) to detect anastomotic leaks (AL). Because many patients discharge before the typical leak window, outpatient esophagram is performed shortly after discharge. This study evaluates optimal timing and utility of postoperative esophagram. A retrospective review of a prospectively maintained database was performed for patients undergoing hRAMIE between October 2018 and January 2025. Patients were stratified by recovery pathway using length of stay (LOS): accelerated (≤6 days), targeted (7–8 days), and delayed (≥9 days). Overall leak data was assessed across the full cohort. A focused outpatient analysis included patients discharged without a true inpatient leak who underwent outpatient esophagram within 30 postoperative days. The primary outcome was true AL detection across cohorts. Secondary outcomes included outpatient leak detection on esophagram, number needed to scan (NNS), and leak timing by postoperative day bins (6–10, >10–15, >15–20, >20). NNS was calculated as the number of outpatient esophagrams performed divided by true outpatient leaks detected among eligible patients. Categorical variables were compared using Fisher’s exact test, with Fisher–Freeman–Halton testing for overall three-group comparisons. Statistical analysis was performed using R studio. A total of 133 patients were analyzed (Accelerated n=87, Targeted n=27, Delayed n=16). True inpatient leak rates increased significantly across groups (0%, 3.7%, and 37.5%, respectively; p=5.0×10-6). Outpatient leak rates remained significantly different (7.5%, 20.8%, and 40.0%; p=0.009), driven primarily by Accelerated versus Delayed differences. Among patients undergoing outpatient esophagram within 30 days of their operation, leaks occurred in 7.8%, 15.8%, and 25.0% across groups (p=0.156). Median outpatient esophagram timing was POD 9.7 (IQR 7.6–23.0) overall and 9.1 (IQR 7.8–11.3) among detected leaks. Number needed to scan was 13.3, 4.2, and 2.25 by group (overall 7.3). Leak detection yield was 11.5% for POD 6–10, 15.0% for POD >10–15, and 25.0% for POD >15–20, while leaks were absent beyond POD 20. Recovery trajectory after hRAMIE was associated with leak incidence and diagnostic yield of postoperative esophagram. Targeted and delayed phenotypes identified higher-risk patients for subsequent outpatient leak and demonstrated lower NNS values. Leak detection was concentrated early postoperatively, with no leaks being detected being POD 20. Recovery- and timing-guided imaging may improve ERAS efficiency while preserving early detection of clinically significant anastomotic leak.