System for the detection of brain tumours in real time using hyperspectral imaging

This Spanish University has been researching new ways for detection of brain tumours. Unlike many tumours, where their identification is relatively simple, the detection of brain tumours during surgical operations continues to be a great challenge. The work carried out allows exploiting the characteristics of hyperspectral images, developing a demonstrator capable of accurately locating malignant tumours during brain surgical procedures. A very precise delineation of the tumour boundaries is achieved, which improves the results of surgery. As a proof of concept, the demonstrator developed is capable of generating thematic maps of the exposed brain surface using spectral information from the range between 400 and 1000 nm. These thematic maps distinguish between four different classes: normal tissue, tumour tissue, hypervascularized tissue (blood vessels) and elements of the background. Hyperspectral imaging allows the acquisition of large numbers of spectral bands throughout the electromagnetic spectrum (within and beyond the visual range) with respect to the surface of scenes captured by sensors. Using this information and a set of complex classification algorithms is possible to determine which material or substance is located in each pixel. One of the major benefits of this technology is that it can be used as a guidance tool during brain tumor resections. Unlike other tumors, brain tumor infiltrates the surrounding normal brain tissue and thus their borders are indistinct and extremely difficult to identify to the surgeon's naked eye. The surrounding normal brain tissue is critical and there is no redundancy, as in many other organs, where the tumor is commonly resected together with an ample surrounding block of normal tissue. This is not possible in the brain, where it is essential to accurately identify the margins of the tumor to resect as less healthy tissue as possible. In this sense, the work performed aims to exploit the characteristics of hyperspectral imaging to develop an intraoperative demonstrator capable of performing a precise localization of malignant tumors during brain surgical procedures. A precise delineation of tumor boundaries is expected to improve the outcomes of surgery. In these maps, the boundaries of the tumour can be easily identified, providing the result in less than 10 seconds when using acceleration on high-computing power GPUs. This work has achieved excellent results in discriminating between tumour and normal brain tissue in a non-invasive, thus improving the results of neurosurgical procedures. As a proof-of-concept, the demonstrator developed is able to generate thematic maps of the exposed brain surface using spectral information of the range comprised between 400 and 1000 nm. These thematic maps distinguish between four different classes previously established: normal tissue, tumor tissue, blood vessels/hypervascularized tissue, and background. In these maps, the tumor boundaries can be easily identifiable. A hyperspectral brain cancer detection algorithm, based on a mix of unsupervised and supervised machine learning approaches, was developed and implemented into the system. he results demonstrate that the system did not introduce false positives in the parenchymal area when no tumor was present and it was able to identify low-grade tumors that were not used to train the brain cancer detection algorithm, resulting in a robust and generalized classification algorithm. The system is protected by granted patent. This initiative has been selected by the European Union to be part of the Innovation Radar, a platform that brings together the European projects with the greatest impact and excellence for developing innovations with high potential to reach the market. The University seeks collaborations for developing a fully working prototype, to be employed in a large clinical trial, by technical cooperation agreement and private or public contact for license agreement.