Diagnostic accuracy of EUS-guided fine-needle biopsy sampling by macroscopic on-site evaluation: a systematic review and meta-analysis

Background and Aims: Assessment of EUS-guided fine-needle tissue acquisition by macroscopic on-site evaluation (MOSE) is gathering attention. Studies report good diagnostic parameters with MOSE; however, the overall data are limited. We conducted this systematic review and meta-analysis to report on the pooled diagnostic assessment parameters of EUS-guided tissue acquisition by MOSE using fine-needle biopsy sampling (FNB). Methods: Multiple databases were searched (from inception to December 2021), and studies that reported on the diagnostic assessment of EUS-guided tissue acquisition by MOSE were selected. Pooled diagnostic accuracy, sensitivity, specificity, and positive and negative predictive values were calculated by standard meta-analysis methods following the random-effects model. Heterogeneity was assessed by I² statistics. Results: Fourteen studies were included in the analysis, and 1508 lesions were biopsy sampled in 1489 patients undergoing EUS-guided tissue acquisition. MOSE definition included a visible core of tissue with opacity and “wormlike” features of adequate size and length (≥4 mm). The pooled accuracy of FNA and/or FNB specimens in yielding a pathologic diagnosis by MOSE was 91.3% (95% confidence interval [CI], 88.6-93.3; I² = 66%), pooled sensitivity was 91.5% (95% CI, 88.6-93.6; I² = 66%), pooled specificity was 98.9% (95% CI, 96.6-99.7; I² = 80%), pooled positive predictive value was 98.8% (95% CI, 97.4-99.5; I² = 33%), and pooled negative predictive value was 55.5% (95% CI, 46.9-63.9; I² = 95%). Subgroup analyses by newer-generation FNB needles demonstrated similar pooled rates, with minimal adverse events (2.5%; 95% CI, 1.5-3.9; I² = 21%). Conclusions: Excellent pooled diagnostic accuracy parameters were demonstrated in EUS-guided tissue acquisition by FNB using the MOSE method.