Tumor detection of the thyroid and salivary glands using hyperspectral imaging and deep learning

• Type of Material: Article
• Authors: Martin Halicek, University of Texas Dallas; James D. Dormer, University of Texas Dallas; James Little III, Emory University; Amy Chen, Emory University; Baowei Fei, Emory University
• Language: English
• Date: 2020-03-01
• Publisher: Optica Publishing Group
• Publication Version: Final Published Version
• Copyright Statement: © 2020 Optical Society of America.
• License: Creative Commons Attribution-NonCommercial 4.0 International
• Final Published Version (URL): http://dx.doi.org/10.1364/BOE.381257
• Title of Journal or Parent Work: Biomedical Optics Express
• Volume: 11
• Issue: 3
• Start Page: 1383
• End Page: 1400
• Grant/Funding Information: National Institutes of Health10.13039/100000002 (CA156775;, CA176684;, CA204254); Cancer Prevention and Research Institute of Texas10.13039/100004917 (RP190588).
• Keywords: Physical Sciences; Biochemical Research Methods; Optics; Radiology, Nuclear Medicine & Medical Imaging; Frozen section; Biopsy; Science & Technology; Life Sciences & Biomedicine; Management; Biochemistry & Molecular Biology
• Research Categories: Physics, Optics; Health Sciences, Radiology; Health Sciences, Opthamology; Engineering, Biomedical

Abstract

The performance of hyperspectral imaging (HSI) for tumor detection is investigated in ex-vivo specimens from the thyroid (N = 200) and salivary glands (N = 16) from 82 patients. Tissues were imaged with HSI in broadband reflectance and autofluorescence modes. For comparison, the tissues were imaged with two fluorescent dyes. Additionally, HSI was used to synthesize three-band RGB multiplex images to represent the human-eye response and Gaussian RGBs, which are referred to as HSI-synthesized RGB images. Using histological ground truths, deep learning algorithms were developed for tumor detection. For the classification of thyroid tumors, HSI-synthesized RGB images achieved the best performance with an AUC score of 0.90. In salivary glands, HSI had the best performance with 0.92 AUC score. This study demonstrates that HSI could aid surgeons and pathologists in detecting tumors of the thyroid and salivary glands.