Purpose or Case Report: To evaluate diagnostic performance of quantitative gray-scale ultrasound as an objective method in evaluation of pediatric thyroiditis.


Methods & Materials: In this retrospective study of 37 children with tissue proven diagnosis, two radiologists independently reviewed thyroid ultrasounds twice and subjectively classified images according to presence or absence of thyroiditis. A consensus session was performed for patients for which there was disagreement. Unweighted kappa coefficients were calculated to assess intra- and inter-observer reliability. Pearson chi-square and Fisher’s exact tests were used to compare categorical measures by final pathology.
A third radiologist performed quantitative measurements of echo-intensity level of the thyroid and adjacent strap muscles from US images using a PACS-based tool. Thyroid /muscle ratio (TMR) was obtained by dividing thyroid mean by muscle mean values. Heterogeneity index (HI) for thyroid was calculated by dividing thyroid SD by thyroid mean values. Wilcoxon Rank Sum tests were used to assess distribution of continuous risk factors by final pathology.


Results: Patient group comprised 29 females and 8 males, with median age 17 years (interquartile range 15-18). By pathology, 19 (51.3%) patients had lymphocytic thyroiditis and 18 (48%) had normal thyroid. For subjective assessment, there was fair inter-observer agreement (kappa .36 (95% CI .14-.57), p=.004) and slight intra-observer agreement for each radiologist (kappa .13 and .17, p>.05). A large proportion of patients for whom consensus review indicated thyroiditis were confirmed with pathology (12/19 (63%), p=.03). For quantitative assessment, no significant difference between thyroiditis and normal thyroid groups was found for either TMR (1.51 and 1.62, respectively, p=.82) or HI (.23 and .23, respectively, p=.37).


Conclusions: Quantitative gray-scale ultrasound did not accurately diagnose thyroiditis. However, subjective consensus evaluation showed significant correlation with the condition, suggesting that multiple radiologists performing more than one review may be beneficial for accurate diagnosis of thyroiditis in children.