Quantitative detection and delineation of head and neck cancer using hyperspectral imaging and machine learning

Hyperspectral imaging;Head and neck cancer;Early cancer detection;Surgical margin assessment;Machine learning;Supervised classification;Discriminant analysis;Ensemble learning;Support vector machines;Preclinical animal models;Subcutanous cancer model;Chemically-induced carcinogenesis model;Ex vivo surgical specimen.

Lu, Guolan ; Fei, Baowei Biomedical Engineering (Joint GT/Emory Department) Chen, Georgia Z. Oshinski, John N. Pogue, Brian W. Wang, May D. ; Fei, Baowei

University:Georgia Institute of Technology

Department:Biomedical Engineering (Joint GT/Emory Department)

关键词: Hyperspectral imaging;    Head and neck cancer;    Early cancer detection;    Surgical margin assessment;    Machine learning;    Supervised classification;    Discriminant analysis;    Ensemble learning;    Support vector machines;    Preclinical animal models;    Subcutanous cancer model;    Chemically-induced carcinogenesis model;    Ex vivo surgical specimen.;   

Others  :  https://smartech.gatech.edu/bitstream/1853/60648/1/LU-DISSERTATION-2016.pdf

美国|英语

来源: SMARTech Repository

【 摘 要 】

Over 500,000 patients are diagnosed with head and neck squamous cell carcinoma worldwide each year. Most people who develop head and neck cancer have advanced disease at the time of diagnosis. Early cancer detection and subsequent surgical resection of tumor remain one of the most promising approaches to improve the survival and quality of life of patients. Hyperspectral imaging (HSI) is a promising optical modality for early cancer detection and image-guided surgery. The major advantage of HSI is that it is a noninvasive technology that does not require any contrast agent and that it combines wide-field imaging and spectroscopy to simultaneously attain both spatial and spectral information from an object in a non-contact way. Hyperspectral images contain high-dimensional spectral information that can be analyzed for visualization, characterization, and quantification of a disease state in biological tissue. Machine learning-based quantitative analysis is critical to exploit the rich spectral-spatial information provided by HSI for cancer detection. This dissertation investigated the potential of label-free HSI technology combined machine learning methods as a noninvasive diagnostic tool for quantitative detection and delineation of head and neck cancer. More specifically, this dissertation included four aims. The first two aims evaluated the diagnostic performance of HSI and machine learning algorithms to differentiate tumor from normal tissue in preclinical animal models, including a subcutaneous tumor model and a chemically-induced tongue carcinogenesis model. The last two aims investigated the detection and delineation of head and neck cancer in a surgical animal model and in ex vivo fresh surgical specimen of human patients. Overall, this work demonstrated that HSI combined with machine learning can provide a noninvasive and quantitative diagnostic tool for the detection and delineation of head and neck cancers, which has the potential to be translated into the clinic for early cancer detection and image-guided surgery in the future.