China Medical University Hospital leverages Google’s genAI for cancer care

China Medical University Hospital is leveraging Google Cloud’s generative AI to support its doctors in providing cancer care.

A formal collaboration between CMUH and Google Cloud was signed in December last year to use MedLM, a healthcare-specific gen AI based on the large language model, Med-PaLM 2. 

MedLM, which intelligence has been proven comparable to licensed doctors by scoring 85% on the MedQA-USMLE (United States Medical Licensing Examination), has become the foundation of the CMUH’s AI-assisted Physician system for making disease diagnosis, treatment planning, patient education, and medical research.

Based on a statement, their partnership aims to introduce assistive tools to support doctors in precision cancer treatment. So far, they have developed two genAI-based tools that help in developing cancer treatment plans, offering personalised treatment-related information to patients, and providing responses to patient health education inquiries:

  • Customised Cancer Treatment Guidelines: developed based on the National Comprehensive Cancer Network guidelines, it enables clinicians to quickly capture accurate information. It can also generate cancer treatment plans and provide personalised treatment recommendations for clinicians to review, edit, and use with patients.

  • Cancer Therapy Q&As: a system incorporated with CMUH’s cancer care-related health educational content. It helps healthcare professionals quickly search for information while also offering comprehensive treatment options.

Additionally, CMUH is using Google’s AI accelerators, TPU (Tensor Processing Unit), for the development of new drugs. Citing findings from preliminary tests, CMUH said the AI can help reduce the computation time related to protein folding “by over tenfold.”


CMUH is an early adopter of MedLM among university hospitals in Asia. The hospital aims to use it to establish industry-leading healthcare AI models for the Chinese-speaking market in the region while facilitating access to accurate medical information. 

CMUH superintendent Professor Der-Yang Cho noted how genAI can provide precise and rapid data analysis for drug guidelines, disease gene sequencing, and medical records, greatly accelerating the discovery of novel treatments for hard-to-treat diseases and the development of treatment plans for patients. 

By reducing workload, genAI can help ease the mental burden of staff and allow clinicians to spare more time with their patients, added Aashima Gupta, global director of Healthcare Strategy and Solutions at Google Cloud.

“[Our collaboration with Google Cloud] positions smart hospitals in Taiwan [like ours] at the front of providing more precision- and safety-oriented support to healthcare professionals and patients alike,” Prof Cho said. 


MedLM adds to the foundational technology of CMUH Artificial Intelligence Center for developing AI-powered medical services, which now run over 20. Among its noteworthy applications of AI is the Intelligent Anti-Microbial System (i.A.M.S.), a four-in-one platform that can predict the risk of sepsis, identify drug-resistant strains, make smart drug dosage recommendations and automatically compare drug-drug interactions and allergy history. 

i.A.M.S was cited by HIMSS as a commendable example of EMR use during CMUH’s revalidation for Stage 7 of the HIMSS Electronic Medical Record Adoption Model last year. In the same year, the hospital was also validated for Stage 7 of the Infrastructure Adoption Model and Stage 6 of the Adoption Model for Analytics Maturity. Proving its digital health capability partly through the use of AI, CMUH also ranked global third in the 2022 HIMSS Digital Health Indicator.


“Life sciences problems are fundamentally data problems, and this is where AI can fully unleash its potential. Generative AI solutions from Google Cloud can help healthcare and life sciences organisations operate more efficiently and improve patient outcomes,” said Kathy Lee, Managing Director of Google Cloud, North Asia.

Source link

This post originally appeared on TechToday.