Evaluating the Intermediate Suspicion Category in Thyroid Nodules with AI Assistance: A Comparative Analysis of C-TIRADS, ACR-TIRADS, and ATA Guidelines.
Authors
Affiliations (4)
Affiliations (4)
- Department of Medical Ultrasonics, Institute of Diagnostic and Interventional Ultrasound, The First Affiliated Hospital of Sun Yat-sen University, Guangzhou, China.
- Department of Ultrasonography, The Sixth Affiliated Hospital, Sun Yat-sen University, Guangzhou, China.
- Department of Medical Ultrasonics, Institute of Diagnostic and Interventional Ultrasound, The First Affiliated Hospital of Sun Yat-sen University, Guangzhou, China. Electronic address: [email protected].
- Department of Medical Ultrasonics, Institute of Diagnostic and Interventional Ultrasound, The First Affiliated Hospital of Sun Yat-sen University, Guangzhou, China. Electronic address: [email protected].
Abstract
To compare diagnostic performance and fine-needle aspiration (FNA) decision-making for thyroid nodules classified as "Intermediate Suspicion" across three major ultrasound-based risk stratification systems (2017 ACR-TIRADS, 2015 ATA guidelines, and 2020 C-TIRADS), with and without AI assistance. This retrospective study analyzed 1,911 ultrasound images of thyroid nodules from 1,040 patients (mean age, 46.1 ± 13.1 years) collected between 2021 and 2022. The Intermediate Suspicion category was defined as ACR-TR4, ATA-4, and C-TR 4B. Seven radiologists independently categorized nodules according to each guideline, and an artificial intelligence (AI) model independently evaluated all nodules. AI-assisted diagnostic and FNA strategies were implemented and compared across the three guidelines. Diagnostic performance and 95% confidence intervals (CIs) were assessed using patient-level clustered logistic generalized estimating equations. In the Intermediate Suspicion category, diagnostic accuracy was lower for both radiologists and the AI model than in the overall cohort. Radiologists' overall accuracy increased from 66.4%-71.4% without AI to 74.1%-79.4% with AI across all three guidelines (P < 0.001). With AI, C-TIRADS achieved the highest accuracy (79.4%), specificity (66.7%), and PPV (77.4%) (all P < 0.05), with no significant differences between ACR-TIRADS and ATA guidelines. Without AI, FNA rates were lowest with ACR TI-RADS (37.7%) and highest with C-TIRADS (52.2%). AI assistance reduced FNA rates by 20.3%-34.7%, increased malignant detection rate (MDR) by 15.0%-23.4%, and decreased missed malignancy rate (MMR) by 23.7%-36.8% (all P < 0.001). With AI assistance, ATA-4 had the lowest FNA rate (14.7%), whereas C-TR 4B had the lowest MMR (29.8%) and the numerically highest MDR (70.0%). AI-assisted interpretation improves diagnostic accuracy and FNA decision-making for Intermediate Suspicion nodules. C-TIRADS combined with AI shows superior performance among the three systems.