Aims
Endoscopic ultrasound-guided fine needle biopsy (EUS-FNB) is the primary diagnostic tool for assessing pancreatic masses. The time taken to obtain an accurate diagnosis is critical; therefore, it is essential to acquire a high-quality histological sample. Techniques such as Rapid On-Site Evaluation (ROSE) and Macroscopic On-Site Evaluation (MOSE) may enhance the diagnostic accuracy of EUS-FNB and increase the proportion of histologically adequate samples. However, the necessity of these on-site evaluation techniques remains a topic of debate. This study aims to determine whether these techniques should be incorporated into routine practice.
Methods
Five medical databases (PubMed, EMBASE, CENTRAL, Scopus, WebOfScience) were searched comprehensively for studies comparing the diagnostic accuracy of EUS-FNB ± ROSE and/or MOSE in solid pancreatic lesions. In the meta-analysis, pooled proportion rates with 95% confidence intervals (CIs) were calculated, and then a multivariate meta-analysis model was constructed to calculate the differences.
PROSPERO registration number: CRD42024613021.
Results
Altogether, we included 4562 patients’ data from 31 articles in our quantitative and qualitative synthesis.
The sensitivity of ROSE (95% [92%; 97%]; p=0.0087) and MOSE (93% [92%; 95%]; p=0.0349) is significantly higher than stand-alone EUS-FNB (88% [83%; 92%]).
The specificity of both ROSE (94% [89%; 97%]; p=0.2368) and MOSE (90% [81%; 95%]; p=0.5854) does not differ significantly from stand-alone EUS-FNB (92% [85%; 96%]; p=0.0087).
Both the sensitivity (p=0.2535) and specificity (p=0.2535) of ROSE and MOSE are not significantly different.
Conclusions
MOSE is equally effective compared to ROSE, suggesting that routine incorporation of MOSE is sufficient in the diagnosis of pancreatic lesions.