Redefining QUAIDE: paving the way for better AI in preclinical endoscopy

Abstract

We read with great interest the recent article by Antonelli et al 1 on the quality assessment for artificial intelligence in digestive endoscopy (QUAIDE) framework, which enhances the quality and reproducibility of artificial intelligence (AI)-based gastrointestinal endoscopy research. While QUAIDE provides a solid foundation for pre-clinical research design, we see opportunities for further improvement in several key areas. First, preclinical endoscopy AI research often involves sensitive patient data, requiring strict adherence to data privacy and ethical standards.2 Currently, the QUAIDE framework does not adequately address data privacy protection, particularly in terms of anonymisation, de-identification, data-sharing agreements and informed consent. Including these standards is crucial for ensuring patient data safety, especially in multicentre collaborations, where varying anonymisation practices across centres increase the risk of data breaches.3 Establishing unified data management …