Automated 3D brain tumor segmentation in MRI — University Hospital Tübingen × statworx

Title

Automated 3D brain tumor segmentation in MRI — University Hospital Tübingen × statworx

Content

University Hospital Tübingen (Uniklinik Tübingen) partnered with statworx on a pro bono computer vision project to automatically determine the three-dimensional extent of brain tumors in MRI images, replacing the time-consuming manual drawing of tumor boundaries by radiologists. "What impressed me most was that, despite the heterogeneity of the data due to anatomical variants, different tumor sizes, and different imaging modalities, good segmentation accuracy was achieved," said PD Dr. med. Florian Grimm of University Hospital Tübingen. On next steps, Grimm added: "We would very much like to continue this exciting development collaboration — preferably in a third-party-funded collaboration project." In brain tumors, an accurate understanding of growth behavior is critical for treatment decisions. Doctors typically had to manually outline the extent of tumors in MRI images — a time-consuming process. The interdisciplinary team at Uniklinik Tübingen worked with computer vision and biomedicine experts from statworx. The model uses computer vision to recognize the spatial extension of tumors in MRI images across three dimensions, allowing medical professionals to assess growth trajectory and establish appropriate therapy. Marcel Plaschke, Head of Strategy, Sales & Marketing at statworx, presented the project as a pro bono engagement. The model's robustness across anatomical variants, tumor sizes, and imaging modalities is what produces the segmentation accuracy Grimm singled out. Per Grimm, the team plans to extend the work into a third-party-funded research project.

URL