Picture: A web application for decision support in glioma surgery

Background and Objective: Patients with glioma, the most common primary malignant brain tumor, often undergo surgery, aiming to remove as much tumor as possible while maintaining functional integrity. However, there is large variation in surgical decisions. This study aims to provide a data-driven approach to surgery planning and evaluation, estimating personalized potential extent of resection, based on a large multicenter MRI database. Methods: We developed an interactive web-application (PICTURE tool), that uses segmented MRI scans from prior surgeries to create resection probability maps. The maps depict the chance of tumor tissue resection based on decisions in prior surgeries. Results: The PICTURE tool enables uploading scans of a new patient and comparing these with the resection probability map of previous patients. This map can then be filtered for clinical characteristics to compare with similar patients and can be interactively explored to determine which parts of the tumor are more or less likely to be resected in a particular patient. Additionally, tumor characteristics and expected extent of resection are reported. Conclusions: The PICTURE tool can enable data-driven glioma surgery planning through interactive generation of resection probability maps.