1138 Surveillance Is Associated With a Lower Tumor Stage and Increased Survival in Barrett's Esophagus Patients Diagnosed With Esophageal Adenocarcinoma

Abstract

Surveillance Is Associated With a Lower Tumor Stage and Increased Survival in Barrett’s Esophagus Patients Diagnosed With Esophageal Adenocarcinoma Romy E. Verbeek*, Max B. Leenders, Martijn G. Van Oijen, Fiebo J. Ten Kate, Frank P. Vleggaar, Jantine W. Van Baal, Peter D. Siersema Gastroenterology and Hepatology, University Medical Center Utrecht, Utrecht, Netherlands; Pathology, University Medical Center Utrecht, Utrecht, Netherlands Introduction: The (cost-)effectiveness of surveillance in patients with Barrett’s esophagus (BE) is still being discussed, particularly whether it is able to detect esophageal adenocarcinoma (EAC) at an early stage and to improve survival of BE patients. We investigated whether tumor stage and long term outcome were better in EAC patients with BE that were included in a surveillance program compared to those not undergoing surveillance. AIM: To determine differences in tumor stage at diagnosis, treatment and survival between patients with EAC and a prior diagnosis of BE participating in a surveillance program and patients not participating in such a program. Methods: All BE patients diagnosed with EAC between 1999 and 2009 in the Netherlands were identified by linking the Dutch cancer registry with the Dutch nationwide histopathology registry (PALGA). A prior BE diagnosis was defined as a BE diagnosis being available in PALGA at least one year before EAC diagnosis. Surveillance participation was defined as at least one additional biopsy sampling episode between the first BE diagnosis and EAC detection. Multivariate logistic regression and Cox proportional hazards regression analyses were performed to identify differences between surveillance and non-surveillance participants. Odds ratios (OR) and hazard ratios (HR) were corrected for age, gender and year of EAC diagnosis. Results: In total, 997 patients with a prior diagnosis of BE and EAC were included, of which 774 (78%) participated in a surveillance program according to the definition. Median time between first BE diagnosis and EAC detection was 6.0 (IQR1.5-11.1) years for surveillance participants and 5.0 (IQR2.5-8.5) years for those not participating (p 0.8). Patients with BE included in a surveillance program had a higher likelihood of having a well-differentiated EAC (OR3.4, 95%CI1.3-9.1), tumor stage 0 (OR4.3, 95%CI1.2-15.6), stage I (OR2.7, 95%CI1.54.9) or stage II (OR1.9, 95%CI1.1-3.2), were more often diagnosed in a university hospital (OR1.7, 95%CI1.0-2.7) and more frequently underwent surgical treatment (OR1.7, 95%CI1.1-2.7) than patients not undergoing surveillance. Surveillance participation was associated with a reduced mortality (HR0.6, 95%CI 0.5-0.7), which can be explained by a difference in tumor stage and differentiation grade between participants and non-participants, as including these characteristics in the analysis abolished this association (HR0.8, 95%CI0.6-1.0). Conclusion: In this large population-based cohort of BE patients diagnosed with EAC, participating in a surveillance program was associated with a lower tumor stage, a higher likelihood of undergoing surgical treatment and a survival advantage when compared to those not undergoing surveillance. The more favorable outcome in surveillance participants supports its use when a diagnosis of BE is established.