Improving diagnostic precision in peritoneal carcinomatosis via photodynamic diagnosis-AI guided laparoscopy in a rabbit model

Abstract

Achieving precise diagnostics in cancer care, especially for peritoneal surface malignancies, is critical due to the challenges of undetected metastatic lesions. This study evaluates the effectiveness of combining photodynamic diagnosis (PDD) with artificial intelligence (AI)-guided laparoscopy using Phonozen, a photosensitizer activated at 405 nm, in a rabbit model. Female New Zealand white rabbits developed peritoneal carcinomatosis through VX2 sarcoma cell inoculation, forming solid-tumor nodules. A dual-emission laparoscope, capable of both white light and 405 nm blue light, was employed. To reduce blue-light interference, a 550 nm long-pass filter was used. The PDD-enhanced diagnostic laparoscopy involved three stages: initial exploration under white light to mark suspicious nodules, blue light illumination for red fluorescence detection, and a confirmatory open laparotomy for additional nodules. Results from 398 nodules in 16 rabbits showed an increase in sensitivity from 84.2% with conventional laparoscopy to 93.9% with PDD. Further testing with 270 nodules in 10 rabbits using a real-time AI algorithm increased sensitivity from 75.7% to 100%. This combination of PDD and AI significantly improves metastatic peritoneal cancer detection during laparoscopy, offering promising implications for clinical practice. © 2025 SPIE.