E445: Neoadjuvant Chemotherapy for Early-Stage Gastric Cancer: Is There a Potential Survival Benefit?

National guidelines recommend upfront surgical resection for patients with clinical T1N0 gastric cancer. However, practice patterns deviating from these recommendations and the associated patient outcomes, have yet to be characterized. The objective of this study was to identify factors associated with neoadjuvant chemotherapy (NAC) use and their survival outcomes for patients with early-stage gastric cancer.

Methods:

The National Cancer Database was queried for patients with clinical T1N0 non-metastatic gastric cancer (2004-2020), identifying those treated with NAC followed by surgery. Patients who underwent palliative-intent resection or who received neoadjuvant radiation were excluded. Multivariable logistic regression assessed characteristics of patients treated with NAC compared to upfront surgery. To account for potential selection bias, 1:1 propensity score matching was performed based on age, race, comorbidity index, diagnosis year, tumor location, grade, and size. Kaplan-Meier methods and Cox proportional hazards regression assessed overall survival (OS).

Results:

Of 8,842 total patients with early-stage gastric cancer, 775 (9.3%) received NAC followed by resection. The median age was 68 (IQR 59-76) and most patients were male (62.6%), Non-Hispanic White (62.5%). On multivariable analysis, patients who received NAC were more likely treated at community hospitals, had moderate to poorly differentiated disease, larger tumor sizes, and tumors located in the cardia compared to those treated with upfront surgery (Table, all p< 0.05). In the unmatched cohort, the median OS for patients treated with NAC was 36.2 months (IQR 19.7-62.5) vs 45.7 months (IQR 22.3-76.9) for resection (p< 0.001). Treatment with NAC independently predicted worse OS on multivariable Cox regression as well (HR 1.14; 95%CI 1.03-1.25). After propensity score matching, 1,526 patients remained in which the median OS for patients receiving NAC was 36.0 months (IQR 19.6-61.15) vs 43.5 months (IQR 21.6-70.0) for resection (p< 0.001). In the matched cohort, treatment with NAC remained independently predictive of worse OS on multivariable cox regression (HR 1.15; 95%CI 1.03-1.29).

Conclusions:

Although patients who received NAC had worse prognostic features for early-stage gastric cancer, treatment with NAC demonstrated worse survival despite accounting for this selection bias. These results highlight the importance of adhering to national guidelines, regardless of differing patient and disease characteristics, which has significant implications on patient outcomes.