Malignancy risk score in thyroid nodules with atypia of undetermined significance cytology: an approach based on cytological subgrouping and ultrasonographic data.
Thyroid nodules diagnosed as atypia of undetermined significance (AUS) are one of the most challenging categories in thyroid cytopathology due to histopathological ambiguity and complexities in clinical management. Therefore, this study aimed to develop and evaluate the diagnostic accuracy of a novel malignancy risk scoring system integrating cytological subclassification and key ultrasonographic features.
A retrospective analysis was conducted on 357 patients who underwent thyroidectomy following an AUS cytology diagnosis. Cytological samples were subclassified as AUS-N or AUS-O, based on the presence of nuclear atypia, and fundamental sonographic parameters were recorded. Independent predictors of malignancy were identified using multivariable logistic regression, and the AUS-malignancy risk score (AUS-MRS) was developed accordingly. Diagnostic performance was compared with existing classification systems: ACR-TIRADS, EU-TIRADS, and K-TIRADS.
The AUS-MRS demonstrated the highest diagnostic performance, with a sensitivity of 84.6%, specificity of 79.6%, and a diagnostic odds ratio (DOR) of 21.8. Among the conventional systems, ACR-TIRADS showed the highest sensitivity (89.1%) but had lower specificity (69.7%) and a DOR of 18.4. EU-TIRADS and K-TIRADS showed more balanced metrics, with sensitivities of 82.7% and 81.4%, specificities of 73.6% and 74.1%, and DORs of 18.2 and 16.7, respectively. The inclusion of nuclear atypia increased specificity across TIRADS models (>85%); however, this was accompanied by reduced sensitivity and did not yield substantial gains in diagnostic performance.
The AUS-MRS may offer improved diagnostic utility compared to existing ultrasound-based classification systems for assessing malignancy risk in AUS nodules.
A retrospective analysis was conducted on 357 patients who underwent thyroidectomy following an AUS cytology diagnosis. Cytological samples were subclassified as AUS-N or AUS-O, based on the presence of nuclear atypia, and fundamental sonographic parameters were recorded. Independent predictors of malignancy were identified using multivariable logistic regression, and the AUS-malignancy risk score (AUS-MRS) was developed accordingly. Diagnostic performance was compared with existing classification systems: ACR-TIRADS, EU-TIRADS, and K-TIRADS.
The AUS-MRS demonstrated the highest diagnostic performance, with a sensitivity of 84.6%, specificity of 79.6%, and a diagnostic odds ratio (DOR) of 21.8. Among the conventional systems, ACR-TIRADS showed the highest sensitivity (89.1%) but had lower specificity (69.7%) and a DOR of 18.4. EU-TIRADS and K-TIRADS showed more balanced metrics, with sensitivities of 82.7% and 81.4%, specificities of 73.6% and 74.1%, and DORs of 18.2 and 16.7, respectively. The inclusion of nuclear atypia increased specificity across TIRADS models (>85%); however, this was accompanied by reduced sensitivity and did not yield substantial gains in diagnostic performance.
The AUS-MRS may offer improved diagnostic utility compared to existing ultrasound-based classification systems for assessing malignancy risk in AUS nodules.
Authors
Öztürk Öztürk, Kökbudak Kökbudak, Kocabaş Kocabaş, Karaköse Karaköse, Kulaksizoğlu Kulaksizoğlu, Karakurt Karakurt
View on Pubmed