ECR Study: Adjunctive AI Facilitates 38 Percent Reduction in Ultrasound Assessment Time for Thyroid Nodules

Adjunctive AI may lead to significant improved efficiency and accuracy with ultrasound evaluation of thyroid nodules, according to new research presented at the European Congress of Radiology (ECR).

For the blinded multi-reader study, researchers compared the use of AI assistant software (Thyroid Suite, DeepHealth) with unaided ultrasound interpretation for 600 cases involving thyroid nodules. The cohort included 104 biopsy-confirmed malignancies, according to the study.

The study authors found that the use of adjunctive AI reduced thyroid ultrasound interpretation time by 38 percent (44.8 seconds vs. 72.4 seconds). Noting that the study involved 18 radiologists, the researchers also pointed out an 8.4 percent increase in inter-radiologist agreement for ACR TI-RADS evaluations for adjunctive AI (62.3 percent vs. 53.9 percent).

Study findings revealed that the use of adjunctive AI improved accuracy across the board for multiple IT-RADS descriptors. The researchers noted accuracy improvements as follows:

• a 4.5 percent increase in nodule composition evaluation (84.9 percent vs. 80.4 percent);

• a 7.4 percent increase in echogenicity assessment (77.4 percent vs. 70 percent);

• a 4.1 percent increase in echogenic foci evaluation (75.2 percent vs. 71.1 percent); and

• a 16.2 percent increase in nodule margin assessment (73.5 percent vs. 57.3 percent).

“An AI-enhanced thyroid ultrasound assistant can improve diagnostic efficiency, consistency, and accuracy and thus enhance overall quality of care,” noted lead study author Sadaf Monajemi, Ph.D., the head of AI and clinical science for ultrasound at DeepHealth, and colleagues.,

(Editor’s note: For related content, see “Study: AI Boosts Ultrasound AUC for Predicting Thyroid Malignancy Risk by 34 Percent Over TI-RADS,” “Ultrasound-Based Nomogram May Enhance Risk Stratification for Thyroid Nodules with Peripheral Calcifications” and “Can PET/CT Bolster Malignancy Detection with Thyroid Incidentalmoas?”)

Reference

1. Monajemi S, Louis LD, Ng A, et al. AI assistant improves interpretation efficiency, consistency, and accuracy for thyroid ultrasound. Presented at the European Congress of Radiology (ECR), March 4-8, 2026, Vienna, Austria.