Keywords
Abstract
Accurate marking of tumour borders during glioma surgery remains problematic and critically shapes patient management. This study combines multispectral fluorescence navigation with molecular profiling, aiming to create an integrative framework for glioma infiltration zone detection in real time and with high precision. We developed a multiprobe strategy based on integrin αvβ3 conjugates together with some of the relevant metabolic markers IRDye800CW which permit spectral discrimination between tumour signals and background autofluorescence. Using our custom-built multispectral imaging system with synchronised illumination sources and ML boundary detection, we achieved 92.3% sensitivity and 88.7% specificity for tumour detection, which are both major improvements compared to conventional 5-ALA fluorescence. Most importantly the system revealed infiltrative tumour cells up to 3.6±0.5mm beyond MRI-defined margins which correlates with better extent of resection (92.3±4.5% vs. 84.7±6.2% in controls). Clinical validation of 42 patients with high grade glioma showed significant improvement in PFS (14.2 vs 9.8 months) without higher rate of neurological deficits, thus confirming the potential of the integrated approach for precision-guided neurosurgery.