TI-RADS: Diagnostically valid, high reproducibility in ID’ing malignant thyroid nodules

The American College of Radiology (ACR) Thyroid Imaging Reporting and Data System (TI-RADS) has achieved high accuracy in categorizing the malignancy status of nearly 950 thyroid nodules detected on thyroid ultrasonography. Findings of a large, prospective multicenter study from Egypt, published in the August 2019 issue of the European Journal of Radiology, revealed that the combined TR4 and TR5 categorizations have a sensitivity of 98.3% and a 90.9% specificity in identifying malignant nodules.¹

TI-RADS, which is similar to the Breast Imaging Reporting and Data System (BI-RADS), was created to avoid unnecessary biopsies and to improve patient management.² It categorizes pattern recognition and standardizes report structure and content. It is intended to decrease inconsistency in interpretation and reporting of thyroid nodules. However, TI-RADS has not been adopted as widely as it could be, say the study authors, who assessed the diagnostic validity and reproducibility of TI-RADS in real-world conditions.

The patient cohort included 380 individuals with clinically suspected thyroid nodules. These individuals underwent thyroid ultrasonography at the hospitals of Zagazig University, Benha University, and Ain Shams University in Cairo, all performed with the same gray-scale ultrasound model. The thyroid and adjacent tissues were scanned transversely and longitudinally; color Doppler images were acquired of each target nodule.

Three radiologists independently evaluated nine morphological features for each of  948 nodules. They assigned TI-RADS categorization points to each one relating to composition, echogenicity, shape, margin, and echogenic foci, and analyzed the vascularity of each thyroid nodule and suspicious lymph nodes.

One hundred thirty-six nodules, or 14.3%, were malignant; they consisted of 111 papillary carcinomas, 2 medullary carcinomas, 2 anaplastic carcinomas, 2 metastatic carcinomas, and 10 poorly differentiated carcinomas. Sixty-six percent of the 812 benign nodules were follicular nodules, 11% were lymphocytic (Hashimoto) thyroiditis, and 9% were granulomatous (subacute) thyroiditis.

The investigators, led by radiologist Mohammad Abd Alkhalik Basha, MD, of Zagazig University, reported that 74% of 87 TR5  highly suspicious-rated nodules were malignant as well as 58% of the 120 TR4 moderately suspicious-rated nodules. Two nodules of the 171 TR3 mildly suspicious-rated nodules were also malignant, representing 1.2% of this category.

The TR4 and TR5 categories had 73 false-positive nodules. Forty-seven percent were benign follicular nodules, 37% were granulomatous (subacute) thyroiditis and 16% were lymphocytic (Hashimoto) thyroiditis. The two false negatives were papillary carcinomas.

“Our study had a 0, 1.2, 58.3, and 73.6% malignancy risk for TR2, TR3, TR4, and TR5, respectively,” wrote the authors. “We recommend the combination of TR4 and TR5 as a predictor for malignancy of thyroid nodules, because if we consider TR5 alone as conclusive for malignant thyroid nodules diagnosis, the ACR TI-RADS will miss a relevant number of malignant thyroid nodules.”

Overall agreement for assigning TI-RADS categories was good, the authors wrote, while inter-reader agreement was very good among the three radiologists with respect to nodule shape, diameter and echogenicity, as well as suspicious neck lymph nodes, a non-TI-RADS feature included in the analysis. Agreement was good for echogenic foci, and composition and margins of nodules.

Of 30 physicians queried on the usefulness of TI-RADS, 27 responded positively, while the remaining three said they would not  use the system as a clinical management protocol because felt that the system was too “novel,” the authors noted.

“Radiologists need to familiarize clinicians with TI-RADS, and explain why it was developed. Our large, multi-center, real-world study has established that the ACR TI-RADS improves diagnostic performance of ultrasound for predicting malignant thyroid nodules with high validity and high reproducibility,” said Dr. Basha, who is using TI-RADS in his private clinic.

He told Applied Radiology that the team is working on another study to demonstrate the added value of suspicious lymph node, halo sign, color Doppler examination, and elastography to the ACR TI-RADS.

REFERENCES

1. Basha MAA, Alnaggar AA, Refaat R, et al. The validity and reproducibility of the thyroid imaging reporting and data system (TI-RADS) in categorization of thyroid nodules: Multicentre prospective study. Eur J Radiol. 2019; 117:184-192.
2. Tessler FN, Middleton WD, Grant EG, et al. Thyroid imaging, Reporting, and Data System (TI-RADS): White paper of the ACR TI-RADS Committee. J Am Coll Radiol. 2017;14(5):587-595.