Beyond Foundation Models: How Deciphex's CipherX Is Transforming Pathology AI

In this episode, Sidney Ocanagil-Tunstall speaks with Dr Pierre Moulin, Chief Scientific Officer at Deciphex, following the company's recent announcement of CipherX, a new pathology AI engine designed to move beyond traditional foundation models and address a much broader range of diagnostic and research questions. Building on Deciphex's official launch of CipherX, Pierre challenges the industry's ongoing race towards bigger models and higher benchmark scores, questioning whether these advances are translating into meaningful improvements for pathologists and pharmaceutical researchers. He introduces the technology behind CipherX, explaining how its semantic layer transforms complex AI outputs into recognisable histopathological features, making results more interpretable, auditable and adaptable across different applications. The conversation explores how CipherX is already powering Deciphex's clinical diagnostic service, Diagnexia, and pharmaceutical research business, Patholytix, supporting diagnostic quality assurance, biomarker discovery and companion diagnostic development. Looking ahead, Pierre discusses the potential to accelerate AI development for a wider range of pathology questions, including rare diseases and complex diagnostic challenges, and why the future of computational pathology may depend less on building bigger models and more on making AI genuinely useful in clinical practice. Read more about the announcement: https://www.pathologynews.com/deciphe...