Is the 2-cm size cutoff relevant for small nonfunctioning pancreatic neuroendocrine tumors: A French multicenter study

Background. Nonfunctioning pancreatic neuroendocrine tumors (NF-PNETs) are often discovered at a small size. No clear consensus exists on the management of NF-PNETs <= 2 cm. The aim of our study was to determine the prognostic value of indicators of malignancy in sporadic NF-PNETs <= 2 cm. Methods. Eighty patients were evaluated retrospectively in 7 French University Hospital Centers. Patients were managed by operative resection (operative group [OG]) or observational follow-up (non-OG [NOG]). Pathologic characteristics and outcomes were analyzed. Results. Sixty-six patients (58% women) were in the OG (mean age, 59 years; 95 % CI, 56.0-62.3; mean tumor size, 1.6 cm; 95 % CI, 1.5-1.7); 14 (72% women, n = 10) were in the NOG (mean age, 63 years; 95 % CI, 56-70; mean tumor size, 1.4 cm; 95 % CI, 1.0-1.7). All PNETs were ranked using the European Neuroendocrine Tumor Society grading system. Fifteen patients (19%) had malignant tumors defined by node or liver metastasis (synchronous or metachronous). The median disease -free survival was different between malignant and nonmalignant PNETs, respectively: 16 (range, 4-72) versus 30 months (range, 1-156; P = .03). On a receiver operating characteristic (ROC) curve, tumor size had a significant impact on malignancy (area under the curve [AUG], 0.75; P = .03), but not Ki-67 (AUC, 0.59; P = .31). A tumor size cutoff was found on the ROC curve at 1.7 cm (odd ratio, 10.8; 95 % CI; 2.2-53.2; P = .003) with a sensitivity of 92% and a specificity of 75 % to predict malignancy. Conclusion. Based on our retrospective study, the cutoff of 2 cm of malignancy used for small NF-PNETs could be decreased to 1.7 cm to select patients more accurately.