Towards machine learning-based quantitative hyperspectral image guidance for brain tumor resection
David Black1, Declan Byrne1, Anna Walke2
1Department of Electrical and Computer Engineering, University of British Columbia, Vancouver, BC, Canada.
Background:
Complete resection of malignant gliomas is hampered by the difficulty in distinguishing tumor cells at the infiltration zone. Fluorescence guidance with 5-ALA assists in reaching this goal. Using hyperspectral imaging, previous work characterized five fluorophores' emission spectra in most human brain tumors.
Methods:
In this paper, the effectiveness of these five spectra was explored for different tumor and tissue classification tasks in 184 patients (891 hyperspectral measurements) harboring low- (n = 30) and high-grade gliomas (n = 115), non-glial primary brain tumors (n = 19), radiation necrosis (n = 2), miscellaneous (n = 10) and metastases (n = 8). Four machine-learning models were trained to classify tumor type, grade, glioma margins, and IDH mutation.
Results:
Using random forests and multilayer perceptrons, the classifiers achieve average test accuracies of 84-87%, 96.1%, 86%, and 91% respectively. All five fluorophore abundances vary between tumor margin types and tumor grades (p < 0.01). For tissue type, at least four of the five fluorophore abundances are significantly different (p < 0.01) between all classes.
Conclusions:
These results demonstrate the fluorophores' differing abundances in different tissue classes and the value of the five fluorophores as potential optical biomarkers, opening new opportunities for intraoperative classification systems in fluorescence-guided neurosurgery.
Insights
Five fluorophores identified using hyperspectral imaging show promise as optical biomarkers for distinguishing brain tumor types and grades during surgery. This aids in achieving complete tumor resection in fluorescence-guided neurosurgery.
Area of Science:
- Neurosurgery
- Medical Imaging
- Biomarkers
Background:
- Distinguishing malignant gliomas at infiltration zones is crucial for complete resection.
- 5-aminolevulinic acid (5-ALA) fluorescence guidance aids tumor visualization.
- Hyperspectral imaging has characterized five fluorophores in brain tumors.
Purpose of the Study:
- To evaluate the effectiveness of five fluorophore emission spectra for classifying brain tumors and tissues.
- To develop machine-learning models for intraoperative tumor classification.
- To assess the potential of these fluorophores as optical biomarkers.
Main Methods:
- 184 patients with various brain tumors and tissues were analyzed using 891 hyperspectral measurements.
- Four machine-learning models were trained to classify tumor type, grade, glioma margins, and IDH mutation.
- Random forests and multilayer perceptrons were employed for classification.
Main Results:
- Classifiers achieved average test accuracies of 84-87% for tumor type, 96.1% for grade, 86% for margins, and 91% for IDH mutation.
- Fluorophore abundances significantly varied between tumor margin types and grades (p < 0.01).
- At least four fluorophore abundances differed significantly between tissue types (p < 0.01).
Conclusions:
- The study demonstrates distinct fluorophore abundances across different tissue classes.
- The five fluorophores show potential as optical biomarkers for intraoperative classification.
- This research opens avenues for advanced fluorescence-guided neurosurgery systems.
