Stereomicroscopic on-site evaluation in pancreatic endoscopic ultrasound-guided tissue acquisition: current evidence and future perspectives Authors: Ho Seung Lee, Jae Min Lee Published: 2026 The pancreas is anatomically difficult to access and is closely surrounded by critical vascular and ductal structures. Therefore, tissue acquisition remains inherently challenging. With the advent of endoscopic ultrasound (EUS) and the introduction of EUS-guided tissue acquisition (EUS-TA), endoscopic sampling of pancreatic lesions has become feasible. However, important limitations persist. EUS-TA specimens are often scant and blood-contaminated, limiting histologic adequacy. More importantly, the presence of core tissue is a key determinant of diagnostic accuracy. Various efforts have been made to overcome these limitations, improve diagnostic yield, and reduce the need for repeat biopsies. Technical modifications, such as the fanning technique, have been introduced,1 and alternative approaches using specimen weight as a surrogate marker for tissue adequacy and histologic core acquisition have also been proposed.2 Despite these efforts, substantial improvements in diagnostic accuracy remain difficult to achieve. To address this issue, on-site evaluations of specimen adequacy have been introduced, including rapid on-site evaluation (ROSE) and macroscopic on-site evaluation (MOSE). ROSE is the most established method and allows real-time cytopathological assessment and direct confirmation of diagnostic adequacy. However, its implementation requires the presence of a cytopathologist at the bedside, which limits its availability in many countries. Alternatively, MOSE can be performed without a cytopathologist. However, it is based on gross visual assessment and, therefore, depends heavily on operator experience and subjective judgment; consequently, its clinical utility remains controversial. In this context, stereomicroscopic on-site evaluation (SOSE) has emerged as a potential alternative that addresses some of the limitations of MOSE. Using a stereomicroscope, SOSE enables a more refined assessment of specimen characteristics. Its feasibility has been demonstrated not only in gastrointestinal subepithelial lesions3 but also more recently in resectable or borderline resectable pancreatic cancers.4 The potential of SOSE to improve specimen quality and diagnostic yield in settings where ROSE is not readily available is certainly of interest. Compared with conventional MOSE, SOSE enables stereomicroscopic visualization of white-core tissue, allowing a more objective assessment of specimen adequacy. Watanabe et al.5 demonstrated that SOSE provided ROSE, with a median assessment time of only 47 seconds, while achieving high histologic and overall diagnostic sensitivity. Similarly, Ishizaki et al.4 reported that approximately 73% of lesions met the stereomicroscopically visible white-core cutoff value after the first needle pass, allowing EUS-TA to be completed with a single puncture in most patients while maintaining high diagnostic accuracy. These findings suggest that the principal advantage of SOSE lies in facilitating white-core identification and optimizing the number of needle passes. However, whether SOSE can truly replace ROSE remains uncertain. While ROSE directly evaluates cytological features and determines diagnostic adequacy, MOSE and SOSE assess the presence of tissue components and, therefore, function as surrogate markers for adequacy. In other words, macroscopic adequacy does not necessarily translate into diagnostic adequacy. In clinical practice, ROSE remains the most established approach for on-site evaluation, although it is not feasible in many institutions because it requires the presence of an on-site cytopathologist. Under such circumstances, MOSE may serve as the simplest alternative, and SOSE may represent a more refined extension of MOSE. However, it remains premature to conclude that SOSE can completely replace ROSE. Despite its potential advantages, several practical issues should be considered before the widespread adoption of SOSE. The availability of stereomicroscopes may be limited in many institutions, and additional equipment, operator training, and familiarity with stereomicroscopic interpretation are required. Furthermore, the impact of SOSE on overall procedure time, interobserver variability, cost-effectiveness, and reproducibility across different centers has not yet been sufficiently established. Further validation through large-scale, multicenter prospective studies is required. Meanwhile, the application of artificial intelligence (AI) to on-site evaluation has emerged as a promising direction in EUS-TA. Recent developments have focused on applying AI to digital cytology and real-time image interpretation, with the goal of assisting specimen adequacy assessment during EUS-TA. AI-assisted ROSE, in which slide interpretation is supported or replaced by AI, has the potential to reduce operator-dependent bias and improve efficiency.6 Nevertheless, this approach remains in the early stages of development and requires substantial data accumulation, rigorous model training, and thorough validation. In conclusion, SOSE is a promising adjunct tool for on-site specimen evaluation during EUS-TA. However, further prospective multicenter studies are required to establish its clinical role and define its place alongside existing approaches.