Reply to Tiwari et al.

10.1055/a-2715-5329

We appreciate the author’s thoughtful comments [1] on our recent article [2].

As outlined in the article introduction, this study was a pilot investigation evaluating an internally developed software (PANCRAIEUS) for real-time pancreatic mass detection during endoscopic ultrasound (EUS) examinations. The author raised four specific critiques. 1) Differentiation of exophytic versus infiltrative tumors. As noted in the article, the software is currently designed only to identify the pancreas, not adjacent organs or specific pathologies. Thus, differentiation of exophytic pancreatic tumors from infiltrating lesions of neighboring organs is beyond the present scope. 2) Use of contrast-enhanced EUS. We respectfully disagree that the absence of contrast-enhanced EUS limits the study’s applicability. At our institution, which performs over 4000 EUS procedures annually, the largest volume in the United States, contrast is not used in any patient. This practice has not impaired workflow or compromised outcomes, suggesting that contrast-enhanced EUS is not essential for routine application. 3) Identification of vascular landmarks. The current version of the software has not been trained to recognize vascular structures. At this early stage, our objective was limited to pancreas identification and detection of possible pathology – features of particular use to novice endosonographers. Standard EUS remains available for vascular assessment, and in our view, cross-sectional three-dimensional imaging provides more accurate and reliable vascular staging. 4) False-positive evaluations. We agree that broader validation is important to eliminate this limitation. As a first step, the system incorporates a persistence function to minimize false positives from erratic individual frames.

In summary, we value the author’s suggestions. However, such refinements must be incorporated in a stepwise manner; one cannot run before learning to walk.

Publication History

Article published online:
22 January 2026

• References

• 1 Tiwari A. Clinical and methodological considerations in artificial intelligence-assisted endoscopic ultrasonography for pancreatic mass detection. Endoscopy 2025;
• 2 Bang JY, Săftoiu A, Udriștoiu A. et al. Prospective clinical validation of a novel artificial intelligence system for real-time detection of solid pancreatic masses during endoscopic ultrasonography. Endoscopy 2025;