Brain tumor surgery presents complex challenges due to a delicate tradeoff between removing as much neoplastic tissue as possible while minimising the loss of brain function. This is usually obtained using intraoperative direct cortical stimulation (DCS) which is considered the gold standard technique for functional mapping of the cortex. Nevertheless, DCS may trigger seizures and can only be performed intraoperatively, thus extending the duration of the procedure and preventing detailed preplanning of the intervention.

In the last two decades, functional magnetic resonance imaging (fMRI) has emerged as a valuable tool for non-invasive assessment of human brain activity and is now used to determine brain regions that should be spared to prevent functional impairment after surgery. In this regard fMRI, complemented with other MRI modalities (e.g. MR-angiography and white matter tractography), offers a unique opportunity to optimise treatment planning with a significant reduction of surgical time.

However, despite the scientific bases of fMRI and the ensuing assisted surgery are well established, there are two major technological limitations in its current implementation: i) a specialised software that calculate, integrates and promptly outputs the imaging results to surgical navigation systems is still lacking. Indeed, a complex cascade of operations is performed manually to transfer, analyse, reformat and output the MRI data to a neuronavigator. These operations require different software packages, mainly developed for research purposes and thus often difficult to use in a clinical setting, preventing large scale diffusion of presurgical mapping; ii) despite neuronavigation systems offer advantages in terms of preoperative planning, its current implementation is not optimal, generally showing only a 2-dimensional representation of preoperative images and virtual surgical instruments. This requires to the surgeon a significant effort to mentally integrate the different informations, with prolonged intervention duration and increased error risk.

Thus, the main objectives of this project are:

1) to develop a specialised software able to implement the analysis of multimodal MRI in a single application and to transfer the results to the neuronavigator, minimising manual procedures. This will bring the diffusion of presurgical mapping procedures at a completely new level, including hospitals outside the academic community.

2) to develop a specialised application that integrates the MRI data in a commercially available wearable device, using the emerging mixed-reality imaging technology. In particular, 3-dimensional holograms obtained from MRI will be anchored with the physical head of the patient and the surgical microscope/exoscope view, allowing the surgeon to virtually explore deeper tissue layers highlighting critical brain structures that need to be preserved. This is expected to significantly improve the surgeon accuracy and decrease mistakes.