The Diagnostic Efficiency of Ultrasound Computer–Aided Diagnosis in Differentiating Thyroid Nodules: A Systematic Review and Narrative Synthesis

Abstract

Computer-aided diagnosis (CAD) techniques have emerged to complement qualitative assessment in the diagnosis of benign and malignant thyroid nodules. The aim of this review was to summarize the current evidence on the diagnostic performance of various ultrasound CAD in characterizing thyroid nodules. PUBMED, EMBASE and Cochrane databases were searched for studies published until August 2019. The Quality Assessment of Studies of Diagnostic Accuracy included in Systematic Review 2 (QUADAS-2) tool was used to assess the methodological quality of the studies. Reported diagnostic performance data were analyzed and discussed. Fourteen studies with 2232 patients and 2675 thyroid nodules met the inclusion criteria. The study quality based on QUADAS-2 assessment was moderate. At best performance, grey scale CAD had a sensitivity of 96.7% while Doppler CAD was 90%. Combined techniques of qualitative grey scale features and Doppler CAD assessment resulted in overall increased sensitivity (92%) and optimal specificity (85.1%). The experience of the CAD user, nodule size and the thyroid malignancy risk stratification system used for interpretation were the main potential factors affecting diagnostic performance outcomes. The diagnostic performance of CAD of thyroid ultrasound is comparable to that of qualitative visual assessment; however, combined techniques have the potential for better optimized diagnostic accuracy.

S-Detect Software vs. EU-TIRADS Classification: A Dual-Center Validation of Diagnostic Performance in Differentiation of Thyroid Nodules

“…Differentiation between benign and malignant TNs is challenging, even for experienced specialists. Accordingly, some recent studies have focused on S-Detect system efficacy in comparison to evaluations performed by experienced medical physicians or other types of evaluations using CAD [14,15]. A recent meta-analysis performed by Zhao et al including 723 lesions from five studies revealed that the sensitivity of the CAD was comparable to the assessment provided by experienced radiologists (0.87 vs. 0.88) but presented lower specificity (0.79 vs. 0.92).…”

“…To overcome this difficulty, computer-aided diagnosis (CAD) systems are gaining interest for US image analysis and are developed on the basis of statistical data mining algorithms, collected from medical center databases. It enables non-invasive judgement on benignity or malignancy of TNs on the basis of US image analysis [14,15]. The purpose of CAD is to increase the diagnostic confidence, achieve interpretation constancy of US features, and eliminate the inter-observer variability in order to increase diagnostic accuracy, especially when the examination is performed by ultrasonographers outside referenced centers for thyroid cancer diagnostics.…”

“…The purpose of CAD is to increase the diagnostic confidence, achieve interpretation constancy of US features, and eliminate the inter-observer variability in order to increase diagnostic accuracy, especially when the examination is performed by ultrasonographers outside referenced centers for thyroid cancer diagnostics. The practical potential of CAD has been suggested in a few recent studies [8,9,14,16]. Kwak TIRADS has high sensitivity and low specificity.…”

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

“…It will help to substantially alleviate the financial burden on the health care system and the anxiety and financial burden on patients. In comparison to the previous articles that use computer-aided diagnostic methods to simply discriminate malignant from benign nodules, our AI methods are designed to identify which nodules warrant FNAB (25). The central aim of this study was to determine which approach reduces the unnecessary FNAB of benign nodules.…”

Visual Interpretability in Computer-Assisted Diagnosis of Thyroid Nodules Using Ultrasound Images

“…The diagnostic accuracy of deep learning can currently reach more than 92%, which is better than the performance of doctors [ 35 – 37 ]. This study found that the datasets used in some studies on deep learning in medical image-assisted diagnosis are based on full images as inputs, which may result in the model learning interference information, leading to poor robustness of the deep learning model without good generalizability among experiments.…”