Computer-Assisted Fine-Needle Aspiration Cytology of Thyroid Using Two-Stage Refined Convolutional Neural Network

• DOI: 10.3390/electronics11244089
• Publication Date: 2022-12-09T07:50:51Z
• Authors: Wensi Duan; Lili Gao; Juan Liu; Cheng Li; Peng Jiang; Lang Wang; Hua Chen; Xiaorong Sun; Dehua Cao; Baochuan Pang; Rong Li; Sai Liu
• Keywords: 03 medical and health sciences; 0302 clinical medicine; two-stage CAD system; fine-needle aspiration cytology (FNAC); thyroid cytopathology; deep learning; object detection; 3. Good health

Fine-needle aspiration cytology (FNAC) is regarded as one of the most important preoperative diagnostic tests for thyroid nodules. However, traditional process FNAC time-consuming, and its accuracy highly related to experience cytopathologist. Computer-aided (CAD) systems are rapidly evolving provide objective recommendations. So far, studies have used fixed-size patches usually hand-select model training. In this study, we develop a CAD system address these challenges. order be consistent with working mode cytopathologists, mainly composed two task modules: detecting module that responsible regions interest (ROIs) from whole slide image FNAC, classification identifies ROIs having positive lesions. The can then output top-k highest probabilities cytopathologists review. obtain overall good performance system, compared different object detection models, combination YOLOV4 EfficientNet networks in our system.