Real-world clinical characteristics and outcomes of gastroesophageal cancer treated at a tertiary academic center (2021–2025).

Abstract

e16085 Background: Gastroesophageal cancer (GEC) remains an aggressive malignancy with high mortality despite therapeutic advances. Real-world outcomes often differ from clinical trial data due to comorbidities, sequencing variability, and access barriers. We characterized contemporary real-world GEC patients treated at a high-volume academic center and evaluated selected clinical and treatment-associated outcomes. Methods: We conducted a retrospective cohort study of adults with GEC at Siteman Cancer Center (2021–2025). Outcomes included stage at diagnosis (stage I–III vs IV), overall survival (OS), and recurrence. OS was summarized using Kaplan–Meier. Additional analyses: (1) association of Esophageal Squamous Cell Carcinoma (ESCC) with underweight at diagnosis (BMI < 18.5) using logistic regression and (2) an exploratory comparative analysis among stage I–III patients, comparing definitive chemotherapy + radiation (CRT) versus trimodality (CRT→surgery) using a 6-month landmark Cox model adjusted for stage and age. Results: Among 220 patients, the mean age was 65.3 years; 77.7% were male; 87.7% White, and 10.9% Black. Stage at diagnosis was stage I 5.0%, stage II–III 38.2%, and stage IV 44.5%. Biomarker availability varied: PD-L1 positive 38.6%, HER2 positive 20.5%, MMR deficient 3.6%, and CLDN18.2 positive 4.5%. Treatments included chemotherapy 86.4%, radiation 49.1%, immunotherapy 50.5%, and surgery 30.0%. At cutoff, recurrence was 13.2% and vital status was 48.6% alive, 32.7% deceased, and 18.6% loss to follow-up. In adjusted stage IV models, Black race and female sex were not associated with stage IV [aOR 1.31 (95% CI 0.46–3.80; p = 0.618) and aOR 1.09 (95% CI 0.49–2.42; p = 0.832) respectively] while older age was associated with lower odds of stage IV [aOR 0.95 (95% CI 0.92–0.98; p < 0.001)]. Median follow-up was 12.0 months. ESCC was associated with underweight at diagnosis (unadjusted OR 10.34 [95% CI 3.19–39.84], p = 0.00020; adjusted OR 6.81 [1.92–27.76], p = 0.0039). In stage I–III CRT patients (N = 59), trimodality versus CRT-only was not associated with significantly different OS in a 6-month landmark model (adjusted HR 0.54 [0.14–2.01], p = 0.355). Conclusions: In this real-world GEC cohort, nearly half of the patients presented with stage IV disease, underscoring persistent gaps in early detection and referral. ESCC was strongly associated with underweight at diagnosis, even after adjustment, supporting the need for early nutritional risk screening and supportive care integration. Our small landmark analysis did not show an OS difference between trimodality and CRT alone, likely reflecting limited power and selection in a real-world setting. These findings support prioritizing multimodality and biomarker-directed approaches when feasible while also targeting modifiable factors to improve real-world GEC outcomes.