Classification of Salivary Gland Tumors Based on Quantitative Optical Coherence Tomography.
Zihan Yang,Jianwei Shang,Chenlu Liu,Jun Zhang,Yanmei Liang
DOI: https://doi.org/10.1002/lsm.23370
2021-01-01
Lasers in Surgery and Medicine
Abstract:Background and Objectives: Visual inspection is the primary diagnostic method for oral diseases, and its accuracy of diagnosis mainly depends on surgeons' experience. Histological examination is still the golden standard, but it is invasive and time-consuming. In order to address these issues, as a noninvasive imaging technique, optical coherence tomography (OCT) can differentiate oral tissue with advantages of real-time, in situ, and high resolution. The aim of this study is to explore optimal quantitative parameters in OCT images to distinguish different salivary gland tumors. Study Design/Materials and Methods: OCT images of four salivary gland tumors were obtained from 14 patients, including mucoepidermoid carcinoma (MC), adenoid cystic carcinoma (ACC), basal cell adenoma (BCA), and pleomorphic adenoma (PA). Two parameters of optical attenuation coefficient (OAC) and standard deviation (SD) along the depth of OCT signal were combined to create a computational model of classification, and sensitivity/specificity of classification was calculated statistically to evaluate their results. Results: A total of 5,919 two-dimensional (2D) OCT images were used for quantitative analysis. The classification sensitivities of 89.6%, 95.0%, 89.5%, 97.8%, and specificities of 97.6%, 99.0%, 98.0%, 98.2%, respectively, were obtained for MC, ACC, BCA, and PA, with the thresholds of 3.6 mm(-1) based on OAC and 0.22/0.18 based on SD. Conclusion: It was demonstrated that OAC and SD could be considered as important parameters in quantitative analysis of OCT images for salivary gland tissue characterization and intraoperative diagnosis. It is of great potential value in promoting the application of this method based on OCT in clinical practice. Lasers Surg. (c) 2020 Wiley Periodicals LLC