TY - GEN
T1 - Accuracy of 3D Ultrasound in the Measurement of Thyroid Nodules Compared to 2D Ultrasound
AU - Li, Guodong
AU - Köhler, Melanie
AU - Petzold, Julian
AU - Krönke, Markus
AU - De Benetti, Francesca
AU - Wendler, Thomas
AU - Eilers, Christine
AU - Weber, Wolfgang
N1 - Publisher Copyright:
© 2025 SPIE.
PY - 2025
Y1 - 2025
N2 - Aim In patients with multinodular goiter, precise identification and measurement of thyroid nodules using traditional 2D ultrasound (US) can be challenging. This study aimed to assess the accuracy of size measurements for thyroid nodules using 3D ultrasound volumes acquired with the PIUR tUS system, comparing results with routine 2D measurements. Methods A total of 273 thyroid nodules identified on routine 2D US were analyzed. Transversal (T), anteroposterior (AP), and longitudinal (cranio-caudal) diameters (L) were measured in both datasets. Linear measurement volumes (LV) were calculated using the formula T × AP × L × 0.524. Volume measurement volumes (VV) from 3D US were obtained directly. Results Eight nodules were not fully covered, and 41 nodules could not be measured by 3D US. Among the remaining 224 nodules, Bland-Altman analysis showed low biases (T: 0.079 mm, AP: -0.526 mm), with high ICCs (T: 0.979, AP: 0.980). For longitudinal diameter, precise measurement was not possible in 35 nodules. In 189 nodules, Bland-Altman analysis showed minimal biases (L: -0.242 mm, LV: -2.415 mm3, VV: 98.607 mm3), with high ICCs (L: 0.961, LV: 0.976, VV: 0.941). ICCs for different echographic characteristics ranged from 0.940 to 0.990, except for hyperechoic nodules (LV: 0.779, VV: 0.797). ICCs indicated high reliability across different nodule outlines: regular (LV: 0.973, VV: 0.974), moderately irregular (LV: 0.960, VV: 0.908), and substantially irregular (LV: 0.974, VV: 0.936). Quantile regression analysis shows that the discrepancy between 2D and 3D ultrasound measurements increases with nodule size, with larger nodules exhibiting greater differences. Conclusion Size measurements of thyroid nodules using 3D ultrasound in routine clinical practice show good agreement with 2D ultrasound measurements, as evidenced by Bland-Altman analysis and ICCs. These findings support the integration of 3D ultrasound into clinical workflows for improved accuracy in thyroid nodule assessment.
AB - Aim In patients with multinodular goiter, precise identification and measurement of thyroid nodules using traditional 2D ultrasound (US) can be challenging. This study aimed to assess the accuracy of size measurements for thyroid nodules using 3D ultrasound volumes acquired with the PIUR tUS system, comparing results with routine 2D measurements. Methods A total of 273 thyroid nodules identified on routine 2D US were analyzed. Transversal (T), anteroposterior (AP), and longitudinal (cranio-caudal) diameters (L) were measured in both datasets. Linear measurement volumes (LV) were calculated using the formula T × AP × L × 0.524. Volume measurement volumes (VV) from 3D US were obtained directly. Results Eight nodules were not fully covered, and 41 nodules could not be measured by 3D US. Among the remaining 224 nodules, Bland-Altman analysis showed low biases (T: 0.079 mm, AP: -0.526 mm), with high ICCs (T: 0.979, AP: 0.980). For longitudinal diameter, precise measurement was not possible in 35 nodules. In 189 nodules, Bland-Altman analysis showed minimal biases (L: -0.242 mm, LV: -2.415 mm3, VV: 98.607 mm3), with high ICCs (L: 0.961, LV: 0.976, VV: 0.941). ICCs for different echographic characteristics ranged from 0.940 to 0.990, except for hyperechoic nodules (LV: 0.779, VV: 0.797). ICCs indicated high reliability across different nodule outlines: regular (LV: 0.973, VV: 0.974), moderately irregular (LV: 0.960, VV: 0.908), and substantially irregular (LV: 0.974, VV: 0.936). Quantile regression analysis shows that the discrepancy between 2D and 3D ultrasound measurements increases with nodule size, with larger nodules exhibiting greater differences. Conclusion Size measurements of thyroid nodules using 3D ultrasound in routine clinical practice show good agreement with 2D ultrasound measurements, as evidenced by Bland-Altman analysis and ICCs. These findings support the integration of 3D ultrasound into clinical workflows for improved accuracy in thyroid nodule assessment.
KW - 3D ultrasound
KW - measurement accuracy
KW - thyroid nodules
UR - https://www.scopus.com/pages/publications/105004554455
U2 - 10.1117/12.3039267
DO - 10.1117/12.3039267
M3 - Conference contribution
AN - SCOPUS:105004554455
T3 - Progress in Biomedical Optics and Imaging - Proceedings of SPIE
BT - Medical Imaging 2025
A2 - Gimi, Barjor S.
A2 - Krol, Andrzej
PB - SPIE
T2 - Medical Imaging 2025: Clinical and Biomedical Imaging
Y2 - 18 February 2025 through 21 February 2025
ER -