Machine-Learning Assisted Discrimination of Precancerous and Cancerous from Healthy Oral Tissue Based on Multispectral Autofluorescence Lifetime Imaging Endoscopy

Complete resection of dysplastic and malignant tissue improves overall survival and delays cancer recurrence in oral cancer patients; however, intraoperative surgical margin assessment is limited to visual inspection and palpation, making it difficult to achieve total resection. There is currently no tool capable of providing real-time, accurate, and continuous margin-assessment guidance during oral cancer resection surgery. Multispectral autofluorescence lifetime imaging (maFLIM) is a label-free imaging modality that enables quantifying a plurality of metabolic and compositional autofluorescence biomarkers of oral dysplasia and cancer. We have developed and validated a machine-learning assisted computer aided detection (CAD) system for automated discrimination of dysplastic and cancerous from healthy oral tissue based on in vivo widefield maFLIM endoscopy data. This CAD system can be potentially embedded into maFLIM endoscopes to enable continuous in situ detection of positive margins during oral cancer resection surgery, thus facilitating maximal tumor resection and improving surgical outcomes for oral cancer patients. Multispectral autofluorescence lifetime imaging (maFLIM) can be used to clinically image a plurality of metabolic and biochemical autofluorescence biomarkers of oral epithelial dysplasia and cancer. This study tested the hypothesis that maFLIM-derived autofluorescence biomarkers can be used in machine-learning (ML) models to discriminate dysplastic and cancerous from healthy oral tissue. Clinical widefield maFLIM endoscopy imaging of cancerous and dysplastic oral lesions was performed at two clinical centers. Endoscopic maFLIM images from 34 patients acquired at one of the clinical centers were used to optimize ML models for automated discrimination of dysplastic and cancerous from healthy oral tissue. A computer-aided detection system was developed and applied to a set of endoscopic maFLIM images from 23 patients acquired at the other clinical center, and its performance was quantified in terms of the area under the receiver operating characteristic curve (ROC-AUC). Discrimination of dysplastic and cancerous from healthy oral tissue was achieved with an ROC-AUC of 0.81. This study demonstrates the capabilities of widefield maFLIM endoscopy to clinically image autofluorescence biomarkers that can be used in ML models to discriminate dysplastic and cancerous from healthy oral tissue. Widefield maFLIM endoscopy thus holds potential for automated in situ detection of oral dysplasia and cancer.

Automated summary

The combination of multispectral autofluorescence lifetime imaging (maFLIM) and machine learning allows for automated discrimination of dysplastic and cancerous oral tissue from healthy tissue, potentially improving outcomes for oral cancer patients by facilitating maximal tumor resection.