AI-assisted decision support for hematology.

The customer

Universitätsklinikum Leipzig is a prominent academic hospital located in Leipzig, Germany. As a major health care provider, it offers comprehensive medical services across all major specialties, integrating patient care with research and teaching. This project was sponsored by Janssen.

The challenge

Hematologists are continually searching for new treatments for rare blood cancers. They rely on the latest research publications to identify the best possible treatments for their patients. However, the sheer volume of publications and the rapid pace at which new studies are released present significant challenges. Hematologists must sift through vast amounts of data to find relevant publications for making personalized therapy decisions. The primary difficulty lies in accurately matching detailed patient data - such as demographics, previous treatments, and health conditions - with the information described in these publications, which is often presented in complex medical language.

“Within our KAIT project, Merantix Momentum developed an NLP solution in which relevant clinical publications are automatically selected and ranked according to their evidence-based value for a given patient case. The team did an awesome job and created a unique technology that will allow hematologists to find the right publication faster and spend more time with patients.”

Dr. Anne-Sophie Kubasch

Specialist Physician for Internal Medicine, Hematology and Oncology at Leipzig University Hospital

Solution

Our developed NLP solution is specifically designed to assist hematologists in finding the most relevant publications for their patients, thereby aiding in making personalized therapy decisions. This system selects and ranks clinical publications based on their relevance to specific medical cases. By integrating the expertise of hematologists with artificial intelligence, this approach enhances the diagnostic process and the formulation of personalized treatment decisions, making cancer care more patient-centered, targeted, and effective. The NLP component is being developed in collaboration with the Clinic for Hematology, Cell Therapy, and Hemostaseology at the University Hospital Leipzig and the Innovation Center Computer Assisted Surgery (ICCAS). It will be incorporated as another key component into the KAIT platform, which provides various innovative and data-driven methods to optimize hematological treatment while promoting collaboration and exchange between clinical professionals.

Results & effects

The solution has enhanced support for personalized therapy decisions by providing an intelligent search function for medical publications. This has streamlined the process of finding and utilizing the most relevant research to inform treatment plans. This comprehensive approach not only saves time but also increases the accuracy and effectiveness of patient care.

Official website

Official KAIT Whitepaper