Surgical Optimization in Glioma Resection: Towards the Development of an Integrated Hyperspectral Fluorescence Imaging Software
Introduction
The evolving field of intraoperative detection of gliomas warrants further investigation due to the potential improvements it can provide to patients. Previous research has revealed that single-point spectrally resolved measurement of the fluorescent agent, protoporphyrin IX (PpIX) concentrations showed a statistically significant difference in all tumor groups when compared with normal brain tissue. The diagnostic performance of PpIX concentration for detecting neoplastic tissue was evaluated using receiver operating characteristic (ROC) curve analysis, which yielded a classification efficiency of 87% (AUC = 0.95, specificity = 92%, sensitivity = 84%). This was significantly higher compared to the conventional fluorescence imaging approach (AUC = 0.73, specificity = 100%, sensitivity = 47%, p < 0.0001). Interestingly, more than 81% (57 of 70) of the quantitative fluorescence measurements that were below the threshold of the surgeon’s visual perception were correctly classified in an analysis of all tumors. These results suggest that PpIX concentration measurements using a spectrally resolved system are highly accurate for detecting neoplastic tissue and have the potential to significantly improve the accuracy of both low- and high-grade glioma detection and resection rates. We expect that when integrating this spectrally resolved approach into a wide-field imaging system with the proposed software solution we will achieve similar accuracy as well as improve surgical workflow. The goal of this research is to develop a software solution integrating real-time imaging cameras for intraoperative visualization of gliomas. This software will facilitate hyperspectral data acquisition and enable visualization of spectrally-resolved fluorescence. The introduction of this tool in the operating room could significantly aid neurosurgeons in the accurate detection of tumor tissues. As a result, we anticipate an improvement in overall resection rates and, consequently, enhanced patient outcomes.
Methods
This research project sought to develop the system control and on-chip software to enable spectrally resolved wide-field imaging during brain tumor surgery. It involved developing custom-built software developed in Python that integrates multiple imaging cameras, including Cubert Ultris 5 Hyperspectral camera (Cubert) and the pco.panda 4.2 bi: Back Illuminated sCMOS camera (pco). The custom software will enable camera integration with illumination sources for seamless data acquisition and live on-chip data processing of complex spectrally resolved imaging data for display in an easy-to-use and intuitive graphical user interface (GUI).
Results
Developed a custom GUI allowing for control of the Cubert camera with the following features: Live View, Parameter Control (Set Integration Time, Distance, Save Location), Capture White & Dark Reference, and Capture Hyperspectral Cube. Using this system we were able to image implanted tumors in mice brains using both the Cubert and pco camera.
Conclusion
We expect that the development of a custom graphical user interface to allow for simultaneous multi-modal wide-field imaging of multiple fluorophores at various excitation wavelengths will allow for improved intraoperative imaging capabilities of both high and low-grade gliomas and ultimately improve tumor resection rates, surgical workflow, and patient outcomes. Future directions of this project include: 1) developing a custom Camera class module that will allow for easier integration of various cameras and their custom software development kits; 2) refactoring the current Python code into a Model-View-Controller architecture to allow for within-window histogram display, image manipulation, and data processing; 3) integrating the Alizé 1.7 high-end, scientific-grade InGaAs camera into the custom (GUI) and camera setup for use in a white matter tract imaging project; and 4) transitioning the camera setup for operating room functionality.