The work entitled “Towards a Multicentric Open Digital Pathology Assistant Benchmark: Initial Results from the DALPHIN Study”, led by the Computational Pathology Group at Radboud University Medical Center in the Netherlands and involving several international partners, including IMP Diagnostics, was presented on March 25 at the USCAP (United States and Canadian Academy of Pathology) congress in Boston by researcher Carlijn Lems.
This groundbreaking study compares various AI-powered virtual assistants applied to pathology cases. Two general-purpose models (GPT-4 and Gemini 1.5 Pro) and one pathology-specific model (PathChat 1.0) were analyzed. The results indicate that PathChat 1.0 outperforms the general-purpose models, although it still shows limitations in more complex cases. Nevertheless, these models demonstrate promising potential as support tools for pathologists.
You May Also Like: Interview with Our Pathologists Diana Montezuma and Domingos Oliveira on the ARABESC Project and the Impact of AI in Healthcare
It is an honor for IMP Diagnostics to take part in this project. We are proud to continue advancing our efforts and exploring new digital frontiers that aim to pave the way for the future of medicine and anatomical pathology.