Evidence on urgent endoscopic ultrasound (EUS) for subepithelial lesions (SELs) is limited. ESGE guidelines recommend EUS-guided fine-needle biopsy (FNB) or mucosal-incision-assisted biopsy (MIAB) for SELs ≥20 mm or when histology is required to guide therapy, while allowing imaging-based management for clearly benign lesions. We evaluated the feasibility, diagnostic yield and impact on management of urgent EUS for unclear lesions detected during urgent endoscopy in a tertiary emergency hospital.

We retrospectively reviewed all EUS procedures performed between 01.01.2023 and 20.11.2025 (524 total) and selected only urgent EUS performed after emergency endoscopy failed to clarify the diagnosis (incidentally detected SELs, mucosal thickening, suspected masses). For each case we collected lesion location, size (largest diameter), EUS features, real-time elastography, contrast-enhanced EUS (CEH-EUS), performance of EUS-FNB, final histology/diagnosis, management decision and follow-up. Primary endpoints were: (1) proportion of lesions characterized by multimodal EUS and (2) rate of management change after EUS and/or tissue diagnosis. Secondary endpoints were diagnostic yield of FNB and procedural complications.

Thirty-five urgent EUS procedures were analysed (57.1% males; median age was 65 years, IQR: 56–70). Lesions were predominantly gastric and duodenal (24/35 = 68.6%). Additional cases included esophageal, jejunal and mediastinal lesions. Median lesion size was 28 mm (IQR: 18–40) and 23/35 (65.7 %) lesions were ≥20 mm. EUS allowed characterization of layer of origin and echogenic pattern in all cases, and real-time elastography supported the diagnosis of soft lesions such as lipomas, while CEH-EUS highlighted tissue vascularization by contrast uptake pattern. EUS-FNB was performed in 9/35 (25.7 %) cases, mainly in suspected GIST/NET or indeterminate solid masses (7/23 lesions ≥20 mm and 2/12 <20 mm). Of these, 5/9 (55.6 %) FNBs were diagnostic, while 4/9 (44.4 %) were inconclusive due to insufficient material. Final diagnoses included 15 GISTs, 3 neuroendocrine tumors, 3 lipomas, 3 cystic lesions, and several other benign or inflammatory conditions; five cases remained histologically or etiologically inconclusive. EUS findings and/or histology led to a change in management in 85 % (30/35 of cases), with surgery planned mainly for GIST, NET, carcinoma, lymphoma and significant stenosis, and surveillance / conservative management for lipomas, cysts, focal nodular hyperplasia, inflammatory lesions and extrinsic compressions. No EUS-related complications or deaths were recorded. 

In this 3-year emergency hospital registry, urgent EUS with selective use of FNB, in lesions where histology could change management, was feasible and safe, frequently clarified uncertain findings after emergency endoscopy and supported tailored decisions between surgery or conservative management. The selective use of FNB by avoiding random sampling of clearly benign or non-mass lesions, appears consistent with ESGE size and risk-based recommendations. Larger prospective series with standardized data collection will be useful to refine predictors of which SELs benefit most from urgent tissue diagnosis.