Software-Based Analysis of the Taller-Than-Wide Feature of High-Risk Thyroid Nodules

The diagnostic value of a new ultrasonographic method for the measurement of a taller-than-wide shape of benign and malignant thyroid nodules

PURPOSE

To evaluate the diagnostic value of a new ultrasonographic method in the measurement of thyroid nodules with a taller-than-wide (TTW) shape.

METHODS

A total of 982 thyroid nodules were analysed, 571 of which were malignant thyroid nodules and 411 of which were benign nodules. Nodules were divided into two groups by size: Group A (<1 cm, n = 627) and Group B (≥1 cm, n = 355). The angle between the maximum diameter of each nodule and the horizontal axis was measured by image analysis software. The best cut-off value for distinguishing between benign and malignant thyroid nodules and the diagnostic accuracy of the new ultrasonographic method in the measurement of TTW shapes were determined by receiver operating characteristic (ROC) curve analysis.

RESULTS

The cut-off value for distinguishing benign and malignant thyroid nodules was 44.5 degrees. The area under the ROC curve (AUC) was 0.849 (95% CI: 0.822~0.875), and the sensitivity and specificity of the diagnosis of malignant thyroid nodules were 86.9% and 84.4%, respectively. Regarding the angle between the maximum diameter and the transverse axis of the thyroid nodules, an angle greater than 45 degrees was a significant indicator of a diagnosis of malignant thyroid nodules. The AUC for distinguishing malignant from benign thyroid nodules with the new ultrasonographic method in the measurement of TTW shapes was higher than that with the first method (FM) in the whole group, Group A and Group B (respectively, 0.849 vs. 0.812, 0.853 vs. 0.808, 0.852 vs. 0.828). The diagnostic sensitivity of a TTW shape measured by the new ultrasonographic method for predicting thyroid malignancy was significantly higher than that measured by the FM in the whole group, Group A and Group B (respectively, 0.858 vs. 0.760, 0.764 vs. 0.669, 0.890 vs. 0.815).

CONCLUSION

A TTW shape measured by our new ultrasonographic method showed superior performance for predicting thyroid malignancy.

Assessing Detection Accuracy of Computerized Sonographic Features and Computer-Assisted Reading Performance in Differentiating Thyroid Cancers

For ultrasound imaging of thyroid nodules, medical guidelines are all based on findings of sonographic features to provide clinicians management recommendations. Due to the recent development of artificial intelligence and machine learning (AI/ML) technologies, there have been computer-assisted detection (CAD) software devices available for clinical use to detect and quantify the sonographic features of thyroid nodules. This study is to validate the accuracy of the computerized sonographic features (CSF) by a CAD software device, namely, AmCAD-UT, and then to assess how the reading performance of clinicians (readers) can be improved providing the computerized features. The feature detection accuracy is tested against the ground truth established by a panel of thyroid specialists and a multiple-reader multiple-case (MRMC) study is performed to assess the sequential reading performance with the assistance of the CSF. Five computerized features, including anechoic area, hyperechoic foci, hypoechoic pattern, heterogeneous texture, and indistinct margin, were tested, with AUCs ranging from 0.888~0.946, 0.825~0.913, 0.812~0.847, 0.627~0.77, and 0.676~0.766, respectively. With the five CSFs, the sequential reading performance of 18 clinicians is found significantly improved, with the AUC increasing from 0.720 without CSF to 0.776 with CSF. Our studies show that the computerized features are consistent with the clinicians’ findings and provide additional value in assisting sonographic diagnosis.

A Closer Look at "Taller-Than-Wide" Thyroid Nodules: Examining Dimension Ratio to Predict Malignancy

OBJECTIVE

To evaluate nodule height-to-width ratio as a continuous variable predicting likelihood of thyroid malignancy.

STUDY DESIGN

Retrospective cohort study.

SETTING

All study information was collected from a single academic tertiary care hospital.

METHODS

Subjects included adult patients with thyroid nodules who underwent thyroid surgery between 2010 and 2020. The following variables were collected: patient demographics, nodule dimensions via ultrasound, fine-needle aspiration biopsy results, and surgical pathology results. Statistical analysis included logistic regression modeling malignancy with variables of interest. We used a receiver operating characteristic curve to assess the discriminatory value of variables.

RESULTS

Height-to-width ratio, as a continuous variable, was associated with malignancy (with each 0.1 increase in ratio; odds ratio [OR], 1.25; 95% CI, 1.14-1.37). The same relationship was true for height-to-length ratio (OR, 1.36; 95% CI, 1.24-1.56). The area under the receiver operating characteristic curve for height-to-width ratio was 63.7%. In line with current emphasis on the transverse ultrasound view, we determined 4 different height-to-width ratio intervals: <0.8, 0.8 to <1.0, 1.0 to <1.5, and ≥1.5. Likelihood ratios of malignancy for each interval were 0.6, 1.0, 2.3, and 4.9, respectively.

CONCLUSION

Our results support the association between greater height-to-width ratio and malignancy but suggest that a multilevel rather than binary variable improves prediction. The likelihood ratios at different intervals give a more nuanced view of how height-to-width ratio predicts malignancy. With continuing review of guidelines for thyroid nodule biopsy, it is important to consider these data for any point total attributed to shape.