HANGZHOU (ANN/CHINA DAILY) – Artificial intelligence (AI)-enabled medical diagnostic tools are transforming cancer screening and diagnosis, significantly enhancing accuracy and treatment quality.
Zhejiang University recently unveiled OmniPT, an AI-powered universal pathology assistant that integrates advanced vision and language models for human-computer interaction. The tool, now in use at the First Affiliated Hospital of Zhejiang University School of Medicine in Hangzhou, focuses on high-incidence cancers such as gastric, colorectal, and cervical. The hospital is the first clinical institution in China to adopt an AI-powered pathology assistant.
OmniPT has achieved remarkable breakthroughs in cancer classification, grading, identification of vascular and neural invasion, and predicting disease progression. The hospital reported that its analyses and predictions are 80 to 90 per cent accurate across various cancer types, offering unprecedented support to pathologists.
Pathology, the cornerstone of accurate diagnosis, involves rigorous examination of tissue samples to understand disease progression. However, China faces a critical shortage of pathology professionals, with only 30,000 registered pathologists compared to the 150,000 to 200,000 needed nationwide. Regional imbalances exacerbate the issue, with remote areas particularly affected.
“Pathologists are even rarer than pediatricians,” said Zhang Jing, Chair of the Pathology Department and Vice-President of the Yuhang branch of the hospital. He explained that the long training periods and a lack of experienced professionals further complicate the issue.
OmniPT, developed by Professor Song Mingli’s team from Zhejiang University’s College of Computer Science and Technology, addresses these challenges by performing repetitive tasks, allowing pathologists to focus on final judgments. For instance, OmniPT can analyze pathology slides and count mitosis — a critical step in diagnosing gliomas — in under 10 seconds, a process that typically takes 30 minutes to an hour manually.
“It assists us but doesn’t drive us. We remain in control, using AI to solve complex problems in pathology. This is particularly beneficial in remote regions or for less experienced doctors,” Zhang said.
OmniPT not only improves diagnostic efficiency and reduces costs but also ensures greater accuracy by identifying details that might be overlooked by fatigued pathologists. By handling over 90 per cent of repetitive tasks, OmniPT allows medical professionals to allocate their expertise to critical decision-making, ultimately enhancing patient care and supporting underserved regions.
