Goal:
Integrating FLIm with AI for intraoperative surgical guidance in cancer resection. Utilizing multi-modal and biological information with supervised and semi-supervised models to address real-world clinical data challenges related to heterogeneity and out-of-distribution instances. Linked biological variables and treatment variability to enhance model accuracy, utilizing model interpretability techniques to discern risk factor influences within the models.
Enhanced Surgical Guidance in Head and Neck Cancer Through FLIm-Based Classification
Addressing the challenges of intraoperative identification in head and neck cancer surgeries, we harnessed the capabilities of fluorescence lifetime imaging (FLIm). Our research offers two pivotal advancements:
Anatomy-Specific Classification: By catering to the diverse anatomical regions of the head and neck, our FLIm-based classification models specifically targeted the “base of tongue,” “palatine tonsil,” and “oral tongue.” This anatomical specificity resulted in enhanced discrimination between healthy and cancerous tissues, achieving commendable ROC-AUC scores of up to 0.94.
Detection of Residual Cancer: Moving beyond traditional techniques like intraoperative frozen sections analysis, we introduced a FLIm-based model adept at identifying residual tumors intraoperatively, particularly during Transoral Robotic Surgery (TORS). Validated on a cohort of 22 patients, our approach not only pinpointed all instances of positive surgical margins but also demonstrated notable sensitivity and specificity metrics.
These innovations underscore FLIm's potential as a revolutionary tool for surgical guidance, promising more accurate interventions and better patient outcomes in head and neck cancer surgeries.
Reference
M. A. Hassan, et al., FLIm-based In Vivo Classification of Residual Cancer in the Surgical Cavity during Transoral Robotic Surgery, MICCAI, 2023
M. A. Hassan, et al., Anatomy-Specific Classification Model Using Label-free FLIm to Aid Intraoperative Surgical Guidance of Head and Neck Cancer, IEEE T. Bio. Eng., 2023
M. A. Hassan, et al. "Anatomy-specific classification models using FLIm for head & neck cancer surgical guidance." Imaging, Therapeutics, and Advanced Technology in Head and Neck Surgery and Otolaryngology 2023. Vol. 12354. SPIE, 2023.
Weyers, Brent W., et al. "Investigating sources of FLIm data variability in head & neck cancer." Advanced Biomedical and Clinical Diagnostic and Surgical Guidance Systems XX. SPIE, 2022.