Protocol for CLASSICA software as medical device trial

Area of Science:

• Oncology
• Medical Imaging
• Artificial Intelligence

Background:

• Current methods for detecting rectal neoplasia before transanal excision are suboptimal.
• Fluorescence angiography (FA) combined with artificial intelligence (AI) classification may improve diagnostic accuracy.
• The CLASSICA Project is validating this AI-driven FA approach as a medical device software.

Purpose of the Study:

• To validate a real-time AI-driven FA method for in-situ rectal cancer detection.
• To assess the efficacy of AI-guided FA for endoscopic biopsy guidance.
• To compare AI-driven FA with traditional biopsy and excision pathology standards.

Main Methods:

• A multi-centre, prospective study enrolling up to 127 patients across seven surgical cancer centers in five countries.
• Utilizing a real-time AI-driven fluorescence angiography method for digital detection of rectal cancer.
• Employing traditional endoscopic biopsies and excision specimen pathology as the comparative standard.

Main Results:

• Prior collaboration has recruited 130 patients, demonstrating feasibility and physician capability.
• An enlarged training dataset of 200 FA videos has been compiled.
• Development of a secure online data-sharing platform and clinical-grade medical device software is underway.

Conclusions:

• The AI-driven FA method shows promise for enhancing rectal cancer detection and guiding biopsies.
• Further optimization and accuracy determination are key objectives of the ongoing trial.
• This technology has the potential to improve pre-operative cancer staging and treatment planning.