Endoscopic ultrasonography (EUS) is most sensitive modality for accurately establishing tissue diagnosis in patients with solid pancreatic mass lesions. However, small pancreatic masses can sometimes be challenging to detect, particularly for less experienced endosonographers. Therefore, outcomes of EUS are operator dependent. We validated performance of novel artificial intelligence-enhanced EUS to detect solid pancreatic mass lesions. In this single-center, prospective, non-randomized, comparative study, high-risk patients aged ≥18 years referred for pancreatic cancer screening or with suspected (solid and cystic) pancreatic lesions due to clinical symptoms, radiological or laboratory findings were evaluated in real-time using AI-EUS software (PANCRAIEUS). Model included 32,713 EUS frames (training/testing phases) of normal, solid and >10mm cystic pancreatic lesions from 202 patients. Clinical validation was conducted prospectively when EUS findings were evaluated concurrently in real-time by two independent expert examiners, one using conventional EUS and another with AI-EUS, both blinded to alternate assessments. Primary outcome was detection of solid pancreatic masses. Between January and July 2024, 308 patients were evaluated. Performance of AI-EUS for detection of solid pancreatic masses was not significantly different to conventional EUS performed by experts (97.1 vs. 100%; risk difference 2.9%, 95% CI -1.2 to 6.8, p=0.246). Final pathology of 105 pancreatic solid masses revealed neoplasia in 93 (88.6%), benign lesion in 12 (11.4%) patients. Performance of AI-EUS was not significantly different to experienced endosonographers for detection and segmentation of solid pancreatic masses. By standardizing performance, AI-EUS may have potential to optimize clinical outcomes in pancreatic cancer.
Bang et al. (Mon,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: