Precise identification of suspected cancerous gliomas through MRI imaging — Joint development of new AI technology by FUJIFILM and the National Cancer Center

FUJIFILM Corp. and the National Cancer Center Japan have jointly developed AI technology that accurately identifies, from an MRI image, suspected gliomas (a type of tumor). The group aims to enhance the precision of image evaluation prior to the treatment of these cancers. This AI technology was created using FUJIFILM's AI technology development service, "SYNAPSE Creative Space" (which is currently only available in Japan). FUJIFILM plans to implement the technology in society for the first time and aims to quickly introduce products equipped with this technology to the market.

Gliocytoma, a type of malignant brain tumor, originates from neuroepithelial cells and is commonly referred to as "glioma" due to its derivation from neuroglia cells. Although glioma is a rare cancer, accounting for a small percentage of all cancer cases, it is the most prevalent and clinically significant among primary malignant brain tumors. As glioma grows, cerebral edema, or brain swelling, occurs around it due to changes in blood flow and inflammation. This tumor and brain edema affect brain function, leading to various symptoms, including headaches, nausea, and impaired motor and language functions.

In the current ordinary treatment protocol for glioma, MRI images are initially evaluated through an MRI scan conducted before treatment. The tumor is then surgically removed as thoroughly as possible, followed by radiotherapy and chemotherapy to preserve functions such as limb movement and speech. Recent years have seen an increased use of AI technology to identify tumor regions in MRI images and to measure their volumes for cancer observation. However, there is no AI technology specifically tailored to glioma, which affects only a limited number of patients and is characterized by insufficient clinical data, making it challenging to accurately determine the area and size of the tumor.

This AI technology was developed by a team comprising Researcher Kazuma Kobayashi, Director Ryuji Hamamoto (Division of Medical AI Research and Development, National Cancer Center Research Institute), Medical Director Masamichi Takahashi (Department of Neurosurgery and Neuro-Oncology) and Medical Director Mototaka Miyake (Department of Diagnostic Radiology), both at the National Cancer Center Hospital.

The AI technique was developed using FUJIFILM's "SYNAPSE Creative Space" to efficiently conduct annotation work for extracting glioma areas from head MRI images, followed by training the AI with the created data. As a result, it is now possible to precisely extract areas suspected of gliomas in MRI images and measure the volume of these extracted areas. Consequently, it is possible to conduct pre-treatment image evaluations of glioma with higher accuracy. This advancement is expected to contribute not only to the early detection of glioma and enhancement of diagnostic accuracy in the future, but also to the optimization of treatment plans for radiotherapy and surgery.

"SYNAPSE Creative Space," which was used in the development, is a service created by FUJIFILM to support developing AI technology for image diagnosis support at medical and research institutions. This support service utilizes the research infrastructure system for developing AI technology, which was jointly developed by FUJIFILM and the National Cancer Center in 2021. The service enables the performance of a series of process involved in the development, from project management and annotation (tags and explanations for various types of information, including text and images) to learning and testing of AI technology.

Multiple learning models specialized for medical images are available, enabling doctors and researchers to develop their own AI imaging diagnoses without the need for specific knowledge, such as programming.

Director Hiroyuki Mano of the National Cancer Center Research Institute positively stated, "The medical AI technology, conceived by clinicians and researchers at the National Cancer Center, led to a joint development with FUJIFILM and will be actually implemented in society. In the future, we hope that many innovative medical AI technologies are created to address various clinical challenges and be utilized in patient care."

Toshiyuki Nabeta (Corporate Vice President, and General Manager of the Medical Systems R&D Center at FUJIFILM), stated, "The knowledge and AI technical ideas from the National Cancer Center, which advances the treatment and research of rare cancers, were combined with the knowhow and AI technology from the 3D image analysis system of our company that has been built for integrating physicians' knowledge into annotation information at high precision. This combination produced a valuable result. We will continue to utilize "SYNAPSE Creative Space" to support the image diagnosis of various illnesses, including rare diseases, and to promote the social implementation of AI technology that benefits the entire workflow."