Size Criteria Is Not Sufficient in Selecting Patients for Local Excision Versus Radical Excision for Rectal Neuroendocrine Tumors >2 cm: A National Cancer Database Analysis

Background: Patients with rectal neuroendocrine tumors >2 cm often undergo radical surgery, despite limited data supporting this practice. Five- and 10-year survival rates for these patients have been reported previously as 74.8% and 58.6%. Objective: Overall survival was compared between local excision and radical surgery and pN0 and pN1 within the radical surgery subgroup for rectal neuroendocrine tumors >2 cm. Factors independently associated with survival were identified. Design: A retrospective, nationwide, multivariate regression analysis was performed. SETTINGS: Data are from the National Cancer Database (2004-2013). Patients: Patients with rectal neuroendocrine tumors >2 cm, excluding stages T4 and M1, were included. Main Outcome Measures: Outcome measures were overall survival and independent risk factors for overall survival based on multivariate regression analysis. Results: Each group had 178 patients. After local excision, 5- and 10-year overall survival rates were 88% and 72% vs 51% and 42% after radical surgery (p < 0.001). A multivariate Cox proportional hazards model showed similar survival (p = 0.96). Tumor factors independently associated with survival were nodal metastasis (HR = 2.01 (95% CI, 1.01-3.97)), poorly differentiated tumors (HR = 4.82 (95% CI, 1.65-14.01)), and undifferentiated tumors (HR = 9.91 (95% CI, 2.77-35.49)). After radical surgery, patients with and without nodal metastasis had 5-year survival rates of 44% vs 59% (unadjusted p = 0.09; adjusted p = 0.11), with insufficient 10-year survival data. Limitations: The study is a retrospective analysis and includes only Commission on Cancer-accredited hospitals. Long-term follow-up was limited. Lymphovascular invasion was missing for a majority of patients analyzed. Conclusions: Local excision for select patients with rectal neuroendocrine tumors >2 cm is a viable alternative to radical surgery. Nodal status and tumor grade independently predict survival and should be factored into surgical intervention selection. In higher-risk patients selected for radical surgery, survival was similar between the pN0 and pN1 groups, possibly indicating a benefit of radical surgery for these patients.