The presence of a cytopathologist increases the diagnostic accuracy of endoscopic ultrasound-guided fine needle aspiration cytology for pancreatic adenocarcinoma: a meta-analysis

Abstract

Objective: A meta-analysis has not been previously performed to evaluate critically the diagnostic accuracy of endoscopic ultrasound-guided fine needle aspiration (EUS-FNA) of solely pancreatic ductal adenocarcinoma and address factors that have an impact on variability of accuracy. The aim of this study was to determine whether the presence of a cytopathologist, variability of the reference standard and other sources of heterogeneity significantly impacts diagnostic accuracy.

Methods: We conducted a comprehensive search to identify studies, in which the pooled sensitivity, specificity, likelihood ratios for a positive or negative test (LR+, LR-) and summary receiver-operating curves (SROC) could be determined for EUS-FNA of the pancreas for ductal adenocarcinoma using clinical follow-up, and/or surgical biopsy or excision as the reference standard.

Results: We included 34 distinct studies (3644 patients) in which EUS-FNA for a solid pancreatic mass was evaluated. The pooled sensitivity and specificity for EUS-FNA for pancreatic ductal adenocarcinoma was 88.6% [95% confidence interval (CI): 87.2-89.9] and 99.3% (95% CI: 98.7-99.7), respectively. The LR+ and LR- were 33.46 (95% CI: 20.76-53.91) and 0.11 (95% CI: 0.08-0.16), respectively. The meta-regression model showed rapid on-site evaluation (ROSE) (P = 0.001) remained a significant determinant of EUS-FNA accuracy after correcting for study population number and reference standard.

Conclusion: EUS-FNA is an effective modality for diagnosing pancreatic ductal adencarcinoma in solid pancreatic lesions, with an increased diagnostic accuracy when using on-site cytopathology evaluation.

Figure 1

Flow diagram demonstrating the process of identification and selection of studies for inclusion.

Figure 2

The funnel graph plots the log of diagnostic odds ratio (DOR) against the standard error of the log of DOR. Circles represent each study in the meta-analysis, the line, the mean; and curves, the 95% confidence intervals (CI) for the meta-analysis. The larger deviation from the funnel curve of each study suggests more asymmetry. Results from small studies will scatter widely at the bottom of the graph, with the spread narrowing among larger studies. This funnel plot suggests no publication bias for the meta-analysis.

Figure 3

Summary receiver-operator curve (SROC) showing the area under the curve (AUC) and the standard error (SE).

Figure 4

Forest plot of the sensitivity from the meta-analysis of diagnostic accuracy of EUS-FNA of pancreatic adenocarcinoma.

Figure 5

Summary receiver operator curve (SROC) plots for rapid on-site evaluation by the cytopathologist. The curves are the regression lines that summarize the overall diagnostic accuracy for each parameter. The black square indicates the absence of cytopathology and the blue square, the presence.