Thyroid nodules ultrasound classification and the importance of the endocrinologist clinical feeling.

Background and aim of the study Several ultrasound (US) classifications for estimating thyroid nodules risk have been proposed. Since most of them are hardly applicable in clinical practice, we created a local tool, named Modena classification (MC), considering US characteristics and clinician subjective impression. The aim is to verify the diagnostic accuracy of MC and to compare it to US classifications of American Thyroid Association(ATA) (1) and British Thyroid Association(BTA) (2). Methods We prospectively enrolled 111 patients (33M, 78F; age 19–75; total 457 nodules) with an indeterminate, suspicious for malignancy or malignant cytology. All the patients underwent neck US before surgery and a score risk was assigned, according to MC: low (not certainly nodular or not suspect); intermediate (indeterminate); high (suspect or very suspect). Then, we retrospectively classified nodules according to ATA and BTA. The US pattern was related to hystology. Results All the classifications had low sensitivity and positive predictive value (PPV), and high specificity and negative predictive value (NPV) for low risk categories. For the intermediate risk category, BTA had the highest accuracy (68%). For higher risk categories, MC had good sensitivity (62%), high specificity (89%) and accuracy (81%); ATA had high sensitivity (83%), low specificity (48%), accuracy 58%; BTA had high sensitivity (88%), low specificity (44%), accuracy 57%. Conclusions A classification that considers the subjective impression of the clinician, in addition to the known US characteristics, has highest accuracy and specificity compared to guidelines classifications, particularly if the nodule has suspect US features. References (1) Haugen et al. Thyroid. 2016, 26: 1–133. (2) Perros et al. Clin Endocrinol (Oxf). 2014; 81 (Suppl 1):1–122. DOI: 10.1530/endoabs.49.EP1383