American Cancer Society; Centers for Disease Control and Prevention; Swiss Re; Swiss Cancer Research foundation; Swiss Cancer League; Institut National du Cancer; La Ligue Contre le Cancer; Rossy Family Foundation; US National Cancer Institute; and the Susan G Komen Foundation.

Rates of PaC resection remain low in Europe and USA with great international variations. Further studies are warranted to explore reasons for these variations.

The Ki-67 labeling index has been found to bear prognostic significance in gastrointestinal neuroendocrine tumors (NETs), and it was recently incorporated in NET histological grading. Nevertheless, a reliable preoperative determination of NET grading could be useful in clinical practice. The aim of this study is to compare the results of Ki-67 labeling index, as measured on cytological samples and on surgical specimens of patients with pancreatic NETs (P-NETs). We also investigated whether concordance might be improved, using a 5 % (instead of 2 %) cutoff value for defining G2 tumors. We retrospectively identified 48 consecutive patients with 53 P-NETs, from our five institutions, and we measured Ki-67 labeling index on their cytological samples and surgical specimens. The traditional 2 % and the alternative 5 % cutoff values were used to classify G2 tumors. The concordance rate between cytological and histological grading was 46/53 (86.8 %; weighted κ statistic 0.77; 95 % confidence interval (95 % CI) 0.60-0.94). No cases of cytological G1-G2 NETs were upgraded to G3 neuroendocrine carcinoma (NEC) at histological grading. Cytology was found to be highly specific in the diagnosis of both G2 (94.1 %; 95 % CI 80.3-99.3) and G3 tumors (100.0 %; 95 % CI 92.8-100), but the sensitivity was poor for G2 NETs (66.7 %; 95 % CI 38.4-88.2) and high for the prediction of G3 NECs (100 %; 95 % CI 39.8-100.0). When the 5 % cutoff value was adopted, concordance rate was 49/53 (92.4 %; weighted κ 0.82; 95 % CI 0.64-1.00). In conclusion, Ki-67 cytological expression can distinguish well-differentiated (both G1 and G2) from poorly differentiated P-NETs, and it may be useful for their preoperative classification.

No abstract