22.03.2024

Researchers at Chalmers University of Technology unveil an AI model capable of detecting lymphatic cancer with 90 percent accuracy, improving medical imaging analysis. This new technology promises to enhance diagnostic capabilities and streamline patient care.

In the realm of medical image analysis, AI-driven solutions have emerged as powerful tools for detecting and diagnosing various conditions, including cancer. By leveraging deep learning algorithms, these systems provide valuable insights and assist healthcare professionals in making informed decisions.

Image: An image in white and green visualizes the procedure of this new AI model; Copyright: Ida Häggström

An image from positron emission tomography (PET) is entered and analyzed by the AI model. It is trained to find patterns and features in the image to make the best possible prediction of whether the image is positive or negative, i.e. whether it contains lymphoma or not.

Products and exhibitors related to this topic

Exhibitors and products related to this topic can be found in the database of MEDICA:

Key findings from extensive study

Collaborating with leading institutions like Sahlgrenska Academy and Memorial Sloan Kettering Cancer Center, researchers conducted a comprehensive study involving over 17,000 images from 5,000 lymphoma patients. Their findings, published in The Lancet Digital Health, showcase the effectiveness of the AI model in identifying visual signs of lymph node cancer.

The Lymphoma Artificial Reader System

Developed by Associate Professor Ida Häggström, the AI model, named Lars (Lymphoma Artificial Reader System), employs deep learning techniques to analyze positron emission tomography (PET) images. Trained through supervised learning, Lars can accurately distinguish between positive and negative scans, aiding in the early detection of lymphoma.

Empowering radiologists with advanced support

By assisting radiologists in interpreting complex images, Lars serves as a valuable second opinion tool, particularly in cases where diagnoses are challenging. With an estimated accuracy rate of 90 percent, the AI model offers reliable support to healthcare professionals, enhancing diagnostic accuracy and efficiency.

Challenges and future prospects

While the development of Lars marks a significant milestone, further validation and clinical testing are essential before its widespread adoption in clinical practice. Challenges, such as differentiating cancer from treatment-related changes, underscore the need for continued research and refinement of AI-based diagnostic tools. The AI model developed by Chalmers University heralds a new era in cancer diagnosis and treatment. As researchers continue to refine and validate this technology, the potential for improving patient outcomes and advancing personalized healthcare remains promising.

MEDICA-tradefair.com; Source: Chalmers University of Technology

More about Chalmers University of Technology at: www.chalmers.se/en