Our study demonstrated no significant difference in malignancy risks between homogeneous vs. heterogeneous hypoechoic nodules in all subgroups and in homogenous vs. heterogeneous iso- or hyperechoic nodules in all subgroups except partially cystic nodules without suspicious features. Meanwhile, heterogeneous hypoechoic nodules showed significantly higher malignancy risk than heterogeneous isoechoic nodules in all subgroups except partially cystic nodules. Our study validated the concept of classifying nodules by their predominant echogenicity as a reasonable form of risk stratification. Regarding the degree of hypoechogenicity, nodules with moderate hypoechogenicity showed similar malignancy risks compared to markedly hypoechoic nodules. In contrast, moderately hypoechoic nodules showed significantly higher malignancy risks than mild hypoechoic nodules in all subgroups except partially cystic nodules without suspicious features. Based on our results, moderately hypoechoic nodules should be grouped with marked hypoechoic nodules for risk stratification.
For nodules with heterogeneous echogenicity, the EU-TIRADS suggested that nodules with any hypoechoic component should be regarded as hypoechoic nodules and classified as intermediate risk [10]. However, in our study, the malignancy risks of heterogeneous isoechoic nodules were not significantly different from their homogeneous counterparts except in the partially cystic nodules without suspicious features subgroup; this result aligned with the findings of our previous study [7]. Although the malignancy risks of heterogeneous isoechoic nodules were higher than homogeneous isoechoic nodules in overall nodules, the malignancy risks of heterogeneous isoechoic nodules were ranged within the low to intermediate risk categories, depending on concurrent suspicious US features. Therefore, our study’s results support the strategy provided by K-TIRADS and ACR TIRADS when assessing nodules with heterogeneous echogenicity.
In the EU-TIRADS [10] and ACR-TIRADS [11], moderate hypoechogenicity was classified as similar risk with mild hypochogenicity; however in this study, moderate hypochogenicity showed a similar malignancy risk to marked hypochogenicity. The results of this study suggest that the previous definition of marked hypoechogenicity should be revised as hypoechoic or similar echogenicity relative to the anterior neck muscles. The result of this study confirmed the validity of the revised definition of marked hypoechogenicity by 2021 K-TIRADS.
The results of our study are in line with those of our previous study in that the malignancy risks of moderately hypoechoic nodules are similar to that of markedly hypoechoic nodules [7]. However, the results in overall nodules were somewhat discrepant from the previous study [7], demonstrating that the malignancy risks of marked hypoechoic nodules were higher than that of moderately hypoechoic nodules [6]. In this cohort, concurrent suspicious US features occurred more frequently in marked hypoechoic nodules than moderately hypoechoic nodules (marked hypoechoic, 72.4% vs. moderate hypoechoic, 52.5%, P <.001). The higher prevalence of suspicious features in marked hypoechoic nodules might have caused confounding effects in the malignancy risks between these two groups.
In the partially cystic nodules without suspicious features subgroup, the malignancy risks of most nodules ranged within the low risk category, regardless of their predominant echogenicity. This contrasts partially cystic nodules with suspicious features’ risks, which fell within the intermediate risk category in most nodules. In contrast to solid nodules, partially cystic nodules without suspicious features showed no significant difference in malignancy risk between marked/moderate versus mild hypoechoic nodules, and the difference between various degrees of hypoechogenicity was diminished. Additionally, in partially cystic nodules, most partially cystic hypoechoic nodules showed mild hypoechogenicity (68.7-85.0%) and the incidence of marked hypoechogenicity was very rare. Based on our results, we assume that in partially cystic nodules, malignancy risk is mainly determined by the presence of suspicious features and the degree of hypoechogenicity had little impact.
Our study has several limitations. First, the reference standards for benign and malignant diagnoses were based on FNA results as well as surgical histology findings, which may inevitably cause false-negative or false-positive results, although the effect may have been small. Second, we retrospectively assessed the nodules’ US features. However, we believe our large sample size across a multicenter setting mitigated this shortcoming. Third, this study did not consider the interobserver agreement for US features described by different radiologists, possibly resulting in discrepant interpretations. However, we performed training sessions prior to formal analysis to mitigate for interobserver variability. Although a previous study demonstrated that this proposed classification of echogenicity showed improved reproducibility than the previous lexicon [7], future in-depth studies are needed to validate the reproducibility of this US lexicon in multiple readers.
In conclusion, the malignancy risk of nodules with heterogeneous echotexture can be stratified based on predominant echogenicity. Additionally, nodule hypoechogenicity can be classified as mild vs. moderate to marked hypoechogenicity for malignancy risk stratification.