Automatic Identification of Parathyroid in Optical Coherence Tomography Images.

Fang Hou,Yang Yu,Yanmei Liang

DOI: https://doi.org/10.1002/lsm.22622

2017-01-01

Lasers in Surgery and Medicine

Abstract:Background and ObjectiveThe identification and preservation of parathyroid is a major problem in thyroid surgery. In order to solve this problem, optical coherence tomography was involved as a real‐time, non‐invasive high‐resolution imaging technique. This study demonstrated an effective and fast method to distinguish parathyroid tissue from thyroid, lymph node, and adipose tissue in their ex vivo optical coherence tomography (OCT) images automatically.MethodsOCT images were obtained from parathyroid, thyroid, lymph node, and adipose tissue, respectively. A classification and an identification system based on texture features analysis and back propagation artificial neural network (BP‐ANN) were established to classify the four types of tissue and identify each of the four types automatically.ResultsA total of 248 OCT images were taken from 16 patients undergoing thyroidectomy. The accuracy of classification for parathyroid, thyroid, lymph node, and adipose were 99.21, 98.43, 97.65, and 98.43%, respectively.ConclusionThe proposed automatic identification method is capable of distinguishing among parathyroid, thyroid, lymph, and adipose automatically and effectively. Compared with the identification results of human, it has a better accuracy and reliability. For identifying parathyroid from the other entities, it has a satisfying performance. Lasers Surg. Med. 49:305–311, 2017. © 2017 Wiley Periodicals, Inc.