REPEATED ENDOSCOPIC ULTRASUND (EUS)-GUIDED FINE NEEDLE ASPIRATION (EUS-FNA) AFTER NON-DIAGNOSTIC OR INCONCLUSIVE RESULTS – A SYSTEMATIC REVIEW AND META-ANALYSIS

Aims:

EUS-guided fine needle aspiration (EUS-FNA) is the gold standard technique for the pathological diagnosis of solid pancreatic lesions. Several studies have been conducted to assess the role of repeat EUS-FNA after an inconclusive index study, reporting different outcomes. The aim of this study was to evaluate the pooled diagnostic accuracy of repeated EUS-FNA after previous non-diagnostic or inconclusive results on first attempt.

Methods:

We performed a systematic research on electronic databases (MEDLINE, PubMed, EMBASE) for relevant studies. Meta-analysis was performed to obtain pooled sensitivity, specificity, positive and negative likelihood ratio and diagnostic odd ratio. Summary ROC curve was used to calculate area under the curve. Meta-regression was used to assess the role of rapid on-site evaluation (ROSE).

Results:

12 studies (486 patients) were included in the analysis. Pooled sensitivity was 77.1% (72.4 – 81.4%) and pooled specificity was 89.4% (82.9 – 94.1%); significant heterogeneity among studies was found both in sensitivity and specificity; positive-likelihood ratio (LR) was 5.96 (2.38 – 14.90) and negative-LR 0.29 (0.19 – 0.45); pooled diagnostic odd ratio (DOR) was 25.0 (7.8 – 80.2). Summary of ROC curves showed a pooled area under curve (AUROC) of 0.882 with a standard error of 0.047. Meta-regression for potential source of heterogeneity identified a significant role of ROSE: relative DOR was 14.06 (95% C.I. 3.10 – 63.7; P = 0.003) for studies conducted with ROSE.

Conclusions:

These data provide strong evidence on the diagnostic accuracy of repeated EUS-FNA after first non-diagnostic or inconclusive results; the use of ROSE seems to be recommended in these cases.