Pancreatic neuroendocrine tumours (PanNET) are a distinct tumour type with outcomes dependent, in part, upon grading by Ki67. Endoscopic ultrasound (EUS) guided fine needle aspiration (FNA) shows variable accuracy to determine Ki67 or grading. Our aim was to assess whether Ki67 and grade can be more accurately determined using fine needle biopsy (FNB) compared to FNA using surgical excision histology as the gold standard.Methods
Retrospective analysis of all pancreatic pathology for neuroendocrine tumours was performed for the period Jan 2009 – Jun 2017. Patients were included if they had undergone EUS guided sampling of the lesion prior to surgical resection. Patient demographics, lesion size and location were noted. FNA and FNB results were examined and Ki67 and grade recorded. Surgical histology reports were examined and time from EUS to surgery, operation performed, Ki67 and grade recorded and compared using correlation coefficient and Cohen’s Kappa.Results
162 patients were diagnosed with PanNET in our centre over the study period of which 57 underwent surgical resection (mean age 55.6, 30 males). 22 lesions (mean size 24.5 mm) were located in the head, 10 in the body and 25 in the tail of the pancreas. 35 lesions underwent FNA and 26 FNB (4 lesions underwent both) all of which confirmed PanNET on cytology or histology respectively. On surgical histology 33 lesions were grade 1, 22 were grade 2, 1 was grade 3 and 1 was mixed neuroendocrine-acinar. 23/35 FNA samples could report Ki67/grading compared to 26/26 FNB samples (p=0.0006). Ki67 on FNA showed a weak correlation with surgical pathology (R=-0.08, p=0.74) whereas Ki67 on FNB showed a moderate correlation (R=0.65, p=0.0004). With respect to tumour grading, FNA samples showed a poor correlation (kappa=0.026) and FNB samples showed a moderate correlation (kappa=0.474). Excluding cystic lesions gave similar results. 12 samples had been obtained using the Procore needle and 13 samples obtained using the Sharkcore needle. Procore correlation of Ki67 to surgical resection histology was moderate (r=0.521, 95% confidence interval −0.07–0.84, p=0.08). Sharkcore correlation of Ki67 to surgical resection histology was good (r=0.788, 95% confidence interval 0.42–0.93, p=0.0013). With respect to tumour grading, both Procore and Sharkcore showed moderate correlation (kappa=0.47 and 0.435 respectively).Conclusion
Both FNA and FNB can be used to confirm a diagnosis of PanNET. However, FNB samples were significantly more likely to provide adequate material for Ki67/grading and showed a closer match to Ki67/grading of the final surgical histology.