A Comparative Analysis of Two Machine Learning-Based Diagnostic Patterns with Thyroid Imaging Reporting and Data System for Thyroid Nodules: Diagnostic Performance and Unnecessary Biopsy Rate

Abstract

Background: The risk stratification system of the American College of Radiology Thyroid Imaging Reporting and Data System (ACR TI-RADS) for thyroid nodules is affected by low diagnostic specificity. Machine learning (ML) methods can optimize the diagnostic performance in medical image analysis. However, it is unknown which ML-based diagnostic pattern is more effective in improving diagnostic performance for thyroid nodules and reducing nodule biopsies. Therefore, we compared ML-assisted visual approaches and radiomics approaches with ACR TI-RADS in diagnostic performance and unnecessary fine-needle aspiration biopsy (FNAB) rate for thyroid nodules. Methods: This retrospective study evaluated a data set of ultrasound (US) and shear wave elastography (SWE) images in patients with biopsy-proven thyroid nodules (‡1 cm) from the Shanghai Tenth People's Hospital (743 nodules in 720 patients from September 2017 to January 2019) and an independent test data set from the Ma'anshan People's Hospital (106 nodules in 102 patients from February 2019 to April 2019). Six US features and five SWE parameters from the radiologists' interpretation were used for building the ML-assisted visual approaches. The radiomics features extracted from the US and SWE images were used with ML methods for developing the radiomics approaches. The diagnostic performance for differentiating thyroid nodules and the unnecessary FNAB rate of the ML-assisted visual approaches and the radiomics approaches were compared with ACR TI-RADS. Results: The ML-assisted US visual approach had the best diagnostic performance than the US radiomics approach and ACR TI-RADS (area under the curve [AUC]: 0.900 vs. 0.789 vs. 0.689 for the validation data set, 0.917 vs. 0.770 vs. 0.681 for the test data set). After adding SWE, the ML-assisted visual approach had a better diagnostic performance than US alone (AUC: 0.951 vs. 0.900 for the validation data set, 0.953 vs. 0.917 for the test data set). When applying the ML-assisted US+SWE visual approach, the unnecessary FNAB rate decreased from 30.0% to 4.5% in the validation data set and from 37.7% to 4.7% in the test data set in comparison to ACR TI-RADS. Conclusions: The ML-assisted dual modalities visual approach can assist radiologists to diagnose thyroid nodules more effectively and considerably reduce the unnecessary FNAB rate in the clinical management of thyroid nodules.

Radiomics in Differentiated Thyroid Cancer and Nodules: Explorations, Application, and Limitations

“…Furthermore, Yoon et al [ 87 ] also applied texture analysis and the LASSO method in US images to predict malignant thyroid nodules with indeterminate cytology, demonstrating good predictive performance. Zhao et al [ 88 ] compared the diagnostic performance and unnecessary FNAB rate for thyroid nodules of assisted visual-based and radiomic-based machine learning approaches in ultrasound images. In this study, ten machine learning classifiers, including decision tree, naïve Bayes, k nearest neighbors (KNN), logistics regression, SVM, KNN-based bagging, random forest, extremely randomized trees (XGBoost), multilayer perception, and gradient boosting tree classifiers, were verified.…”

Artificial Intelligence in Thyroid Field—A Comprehensive Review

“…Subsequently, this method was used as a step of radiomics analysis. In this section we describe the workflow of the ML algorithm with classification task frequently encountered in the CAD framework [ 10 , 21 , 22 ] ( Figure 1 ). A supervised ML model is composed of two phases, i.e., training and application phase ( Figure 1 a).…”

Pre-Operative Prediction of Mediastinal Node Metastasis Using Radiomics Model Based on 18F-FDG PET/CT of the Primary Tumor in Non-Small Cell Lung Cancer Patients

“…Data pre-processing: to ensure that the features were comparable, training/testing cohort division, missing-value filling, and data standardization were performed. First, to maintain the distribution of the original data, a stratified sampling method was applied to identify the training (501 samples, 70%) and testing cohorts (215 samples, 30%) (19). Moreover, the missing values (0 and 5) were filled with the median in the training and testing cohorts, respectively, and then, the same normalization was used for the data.…”