Pre- and Post-Surgery Primary Brain Tumour Volume Estimation
Overview
The volume of brain tumour and especially the pre- and post-surgery difference in tumour volume is an important clinical parameter in primary brain cancer patients. It helps clinicians predict complications and survival. Similar techniques can be of use also in non-primary (metastatic) brain tumours and even other brain lesions. Deep learning is used, focused particularly on matching pairs (before and after surgery) of cranial MR images with several modalities (T1, T2, FLAIR). While initial training uses publicly available images (with the vast majority of non-matching images), the main training is facilitated by data from the continuously updated GlioBank repository curated by the National Institute of Biology.
Outcome
Near real-time assessment of change in brain tumour volume after neurosurgery. The model computes the difference in brain tumour volume before and after surgery. The volume is computed for three different important regions: tumour core, tumour rim, and surrounding hematoma.
Service Card
| Category | AI products |
| Domain | Healthcare, Radiology |
| Vertical | Health and Biotechnology |
| Target users | Radiologists |
| Service type | Model |
| Outputs | Brain tumour volume estimation service with pre- and post-surgery volume difference computation across three regions. |
| Requirements | Access to own (brain) MR images (pre- and post-surgery). |
| Limitations | |
| Availability | Q1/2027 (initial), Q3/2027 (final) |
| Licensing | |
| Visual content | |
| Visual caption |