A team from Vall d’Hebron Institute of Oncology and Bellvitge University Hospital has developed a user-friendly, open-access AI tool called DISCERN. This tool utilizes deep learning from standard MRI scans to accurately diagnose brain tumors non-invasively, surpassing current methods.
The study, led by VHIO and published in Cell Reports Medicine, highlights DISCERN’s effectiveness in distinguishing between common brain malignancies, such as glioblastoma multiforme, brain metastasis, and primary CNS lymphoma. These tumors often require different treatment approaches, making accurate diagnosis crucial.
Traditionally, brain tumor diagnosis involves neurosurgical interventions after contrast-enhanced MRI assessments. However, this new AI tool automates the diagnostic classification process with high precision. DISCERN was trained on 50,000 voxels, three-dimensional display elements representing MRI scans, to learn distinct characteristics of various brain tumors.
Results from testing on over 500 additional cases demonstrated a remarkable 78% accuracy, outperforming conventional methods. The tool serves as a diagnostic support system, aiding multidisciplinary medical teams in decision-making regarding the necessity and type of surgery required to confirm the diagnosis.
To ensure widespread adoption, the VHIO’s Radiomics Group collaborated closely with the Bellvitge University Hospital’s Neuroradiology Unit, implementing DISCERN as an open-access application. The user-friendly interface enhances study reproducibility, making it a valuable tool for clinicians.
The collaborative effort included researchers from the Neuroradiology Unit of the Bellvitge Biomedical Research Institute, Clínic Hospital in Barcelona, Spain, HT Medica in Andalucia, Spain, and the Radiation Medicine and Applied Sciences and Bioengineering Departments of the University of California in San Diego, U.S. Access the DISCERN app .
Source: Medicalexpress
