IMG-137. OncoLight: Optical fingerprinting of brain tumors with snapshot hyperspectral imaging for accurate, real-time fluorescence-guided neurosurgery

Abstract Effective resection of brain tumors is a challenge due to poor contrast between tumor tissue and normal brain. Fluorescence-guided surgery (FGS) has emerged as an effective real-time surgical adjunct that enhances contrast using fluorescent biomarkers (e.g., protoporphyrin IX [PpIX]) that accumulate in tumors. While existing FGS systems effectively delineate tumor bulk, they cannot detect infiltrative margins due to low fluorescence signal and tissue autofluorescence. Moreover, the imaged fluorescence intensity does not reflect accurate biomarker concentrations due to tissue scattering and absorption. Therefore, there is a critical need for advancing FGS technologies for optimal surgical outcomes. Our objective is to develop OncoLight, a real-time hyperspectral imaging (HSI) system that can “optically fingerprint” tissue during FGS, enabling accurate visualization of tumor tissue and infiltrative margins. OncoLight is an add-on module to a Zeiss Kinevo surgical microscope with two cameras: RGB and HSI. OncoLight captures real-time HSI cubes for fluorescence under violet-blue light, and reflectance under white light. OncoLight uses fluorescence HSI as an optical fingerprint to isolate PpIX fluorescence. Further, it uses reflectance to correct PpIX fluorescence for the distorting effects of tissue optical properties to enable accurate biomarker estimation in real-time across the surgical field of view. OncoLight is able to accurately isolate and estimate PpIX in phantom studies compared to current methods. In animal studies, OncoLight reveals infiltrative tumor and tumor hidden by hypervascularized absorption. Finally, in a first-in-human clinical study, we show the superiority of OncoLight over spectral filtering and conventional systems like the Zeiss Kinevo, in revealing infiltrative tumor tissue. OncoLight therefore provides quantitative, real-time tumor visualization including infiltrative margins superior to current FGS microscopes. Its design allows seamless integration into surgical workflows, with the potential for better surgical outcomes.