Herein, we present an anatomy-specific classification model using FLIm to differentiate between benign tissue, dysplasia, and cancer within the oral cavity and oropharynx. A total of 54 features, comprising both time-resolved and spectral intensity features, were used to train and test the classification model. This anatomy-specific classifier improves on our previous classification approach, now yielding an overall ROC-AUC of 0.94 during binary benign vs. cancer classification, and 0.92 while discriminating between healthy, cancer, and dysplasia. The proposed classification model demonstrates that FLIm has the potential to be used as an adjunctive diagnostic tool to facilitate head and neck cancer surgical guidance.
|