Objectives: Assessment of proliferation by the Ki-67 labelling index (Ki67-LI) is an important parameter of pancreatic neuroendocrine tumour (pNET) prognosis on resection specimens. Ki67-LI values for grading are not fully established on endoscopic ultrasound-guided fine needle aspiration (EUS-FNA). The aim of the study was to determine the accuracy of Ki67-LI on EUS-FNA to predict a final grade of pNET and to analyse the relationship between cytological grading and progression-free survival (PFS).
Methods: Between 1996 and 2010, 46 pNETs (33 were resected) from 45 patients were diagnosed by EUS-FNA. Ki67-LI was evaluated on cytological and histological material for each tumour and classified according to the 2010 WHO grading system.
Results: A very good inter-observer agreement for Ki67-LI on EUS-FNA and surgical specimens, respectively, were obtained. Discrepancies were observed between histology and cytology, especially in grade 2 (G2) tumours, where cytology underestimated grading owing to tumour heterogeneity. Still, EUS-FNA was able to distinguish a poor prognostic group, as the actuarial PFS of cytological (c) G3 tumours was 10 ± 4 months versus 29 ± 7 and 68 ± 10 for cG2 and cG1 tumours, respectively (P < 0.0001).
Conclusion: This study attests the reproducibility of Ki67-LI of pNETs whether counted on cytology or histology with a very good inter-observer correlation. Determination of Ki67-LI on EUS-FNA of pNETs should be included systematically in their prognostic work-up.
Keywords: Ki-67; Ki-67 labelling index; endoscopic ultrasound-guided fine needle aspiration; grading; neuroendocrine tumour; pancreas.
© 2013 John Wiley & Sons Ltd.