Factors associated with the diagnostic yield of endoscopic ultrasound (EUS)-guided tissue sampling for the cyto-histological diagnosis of solid pancreatic tumors: analysis of a large prospective registry

Aims EUS-guided sampling is considered the gold standard for the cyto-histological diagnosis of solid pancreatic tumors (SPT). The development of new needles over the last few years has improved the diagnostic accuracy of EUS sampling. Our study aimed to evaluate factors associated with the diagnostic yield of EUS sampling for the cyto-histological diagnosis of SPT in clinical practice.

Methods Retrospective analysis of a large prospective registry of SPT evaluated by EUS from January 2008 to December 2021. Patients who underwent EUS sampling (FNA or FNB) of the SPT were identified and included in the study. EUS was performed with linear Pentax echoendoscopes and Hitachi systems. Tissue acquisition was performed with standard cytology needles and core needles [Procore, Franseen, and Fork-Tip]. Samples were collected in a cytological solution (Cytolit) or smeared and processed for cytological or histological evaluation. Diagnostic yield was analyzed using the histopathological evaluation of the surgical specimens, and the clinical and radiological long-term follow-up in non-operated patients, as gold standard. Results are shown as mean±standard deviation or percentages as appropriate. Multivariate logistic regression was used to identify the factors significantly and independently associated with the diagnostic yield of EUS sampling, and the results are expressed as OR and 95% CI. Needle type, rapid on-site evaluation (ROSE), number of needle passes, and size of the lesion sampled were considered independent variables The correct diagnosis reached by EUS-sampling (yes or no) compared to the gold standard was used as the dependent variable.

Results 1072 patients were included (mean age 67.6 years, range 17-92, 597 male). The size of SPT was 34.3±14.7 mm. 615 (57.4%) tumors were located in the head of the pancreas, 341 (31.8%) in the body, 83 (7.7%) in the tail, and 33 (3.1%) in the uncinate process. Cytology needles were used in 542 patients (50.6%) and core needles in 530 (49.4%). The mean number of needle passes was 1.7±0.8 in both groups. ROSE was done in 224 (20.9%) cases. Global sensitivity, specificity, and overall accuracy for malignancy were 86.1%, 100%, and 88.1%, respectively. Sensitivity and overall accuracy were higher with core needles compared with cytology needles (89.10% vs 83.10%, and 90.40% vs 85.80%, respectively, p<0.05). In the multivariate analysis, the tumor size (OR 1.04 [1.01-1.04], p=0.001), ROSE (OR 6.56 [2.81-15.32], p<0.001), and the use of core-needles (OR 1.89 [1.28-2.78], p=0.001), but not number of needle passes, were factors associated with a correct diagnosis.

Conclusions EUS-guided sampling is an accurate technique for the cyto-histological diagnosis of solid pancreatic tumors. A higher diagnostic yield is obtained with core needles and with ROSE if cytology needles are used.

Publikationsverlauf

Artikel online veröffentlicht:
15. April 2024