On-site stereomicroscope quality evaluations to estimate white core cutoff lengths using EUS-FNA biopsy sampling with 22-gauge needles

Background and Aims: Although rapid on-site cytologic evaluation (ROSE) during EUS-guided FNA biopsy (EUS-FNAB) sampling may improve accuracy of pathologic analyses, cytopathologists are not widely available. We calculated the cutoff lengths required for accurate pathologic diagnoses from stereomicroscopically visible white cores (SVWCs) sampled using 22-gauge needles. Methods: Overall, 118 patients with mediastinal or upper abdominal solid masses requiring pathologic diagnoses were included. EUS-FNAB sampling was performed using 22-gauge needles. SVWCs were isolated and measured using stereomicroscopy, and the utility of calculated cutoff lengths in diagnosis was investigated. Results: The procedure success and SVWC sampling rates were both 100%, and the median SVWC length was 10 mm. Pathologic examination identified 75, 31, and 12 patients with pancreatic neoplasms (PNs), subepithelial lesions (SELs), and other lesions, respectively. The sensitivity, specificity, positive predictive value, negative predictive value, and accuracy for diagnosing malignancy using EUS-FNAB sampling were 93.1%, 100%, 100%, 69.6%, and 94%, respectively. The final diagnostic accuracy in the entire cohort, PNs, and SELs was 92.4%, 90.7%, and 93.5%, respectively. Receiver operating characteristic curves demonstrated the overall SVWC cutoff length to be 11 mm (11 mm for PNs, 3.5 mm for SELs). The overall sensitivity according to SVWC cutoff length was 91.4% (87.6% for PNs, 98.8% for SELs). Compared with cutoff length, multivariate analysis confirmed SVWC length to be a stronger independent factor for tissue diagnosis in both groups. Conclusions: Diagnosis improved significantly with SVWC cutoff lengths >= 11 mm. This may be a useful index for endoscopists, particularly where ROSE is unavailable.