Association between surveillance imaging and survival outcomes in small bowel neuroendocrine tumors.

BACKGROUND: Surveillance guidelines following the resection of small bowel neuroendocrine tumors (SB-NETs) are inconsistent. We evaluated the impact of surveillance imaging on SB-NET recurrence and overall survival (OS). METHODS: Patients with completely resected SB-NETs referred to a provincial cancer center (2004-2015) were reviewed. Associations between imaging frequency, recurrence, post-recurrence treatment, and OS were determined using univariate and Cox-regression analyses. RESULTS: Among 195 completely resected SB-NET patients, 31% were ≥70 years, 43% were female, and 80% had grade 1 disease. Imaging frequency was predictive of recurrence (hazard ratio 2.52, 95% confidence interval 1.84-3.46, p < 0.001). 72% underwent interventions for recurrent disease. Patients who were treated for the recurrent disease had comparable OS to those who did not recur (median 152 vs. 164 months; p = 0.25). Imaging frequency was not associated with OS in those with treated recurrent disease (p = 0.65). Patients who recurred underwent more computerized tomography (CT) scans than those who did not recur (CT: 1.47 ± 0.89 vs. 1.02 ± 0.81 scans/year, p < 0.001). Detection of disease recurrence was 5%-7% per year during the first 6 years of surveillance and peaked at 17% in Year 9. CONCLUSION: Less frequent imaging over a longer duration should be emphasized to capture clinically relevant recurrences that can be treated to improve OS.

Risk of lymph node metastasis in early gastric cancer for a Western population.

INTRODUCTION: Endoscopic resection (ER) of early gastric cancer (EGC) is increasingly used in Eastern countries due to their low rates of lymph node metastasis (LNM); however, there is a paucity of evidence in Western countries. We investigated LNM and its effect on overall survival (OS) in Western patients with EGC. METHODS: Patients diagnosed with T1 gastric cancer between 2000 and 2017 were retrospectively evaluated. Univariate Kaplan-Meier, multivariate logistic and Cox-regression models were used to assess the associations between clinical characteristics, LNM, and OS. RESULTS: Among 86 patients, median age was 68 years and 72% were male. Node positivity was 30%. Two percent of patients met the classical guidelines for ER and all were node-negative, while 16% met expanded criteria of which 14% were node-positive. T1b disease (odds ratio [OR] 41.2 [95% confidence interval [CI] 1.62-1048], p = 0.02) and lymphovascular/perineural invasion (OR 18.0 [95% CI 2.41-134], p = 0.01) were predictive of node positivity. The 5-year OS for node-negative and node-positive patients was 84% and 53% (p = 0.004), respectively. CONCLUSIONS: The risk of LNM in Western patients with EGC is higher; therefore, generalizability of the expanded criteria for ER should be interpreted with caution.