A Multiparametric Ultrasound Diagnostic Model for Adenomyosis by Integrating Cervical Ultra-Microangiography and Shear Wave Elastography: A Prospective Study.
Authors
Affiliations (6)
Affiliations (6)
- Department of Medical Ultrasound, The First Affiliated Hospital of Guangzhou Medical University, Guangzhou, China; Department of Ultrasound, Zhongshan Hospital of Traditional Chinese Medicine, Zhongshan, China.
- Department of Clinical Laboratory, Zhongshan Hospital of Traditional Chinese Medicine, Zhongshan, China.
- Department of Medical Ultrasound, The First Affiliated Hospital of Guangzhou Medical University, Guangzhou, China.
- Department of Ultrasound, Hubei Cancer Hospital, Wuhan, China.
- Department of Ultrasound, Guangzhou Women And Children's Medical Center, Guangzhou, China.
- Department of Medical Ultrasound, The First Affiliated Hospital of Guangzhou Medical University, Guangzhou, China. Electronic address: [email protected].
Abstract
To develop and validate a multiparametric diagnostic model for adenomyosis utilizing standardized cervical anatomical landmarks. By integrating cervical shear wave elastography (SWE) and ultra-microangiography (UMA), this study aims to provide an objective and quantitative tool to complement conventional corpus-based ultrasound assessments. This prospective study enrolled 422 women (222 with adenomyosis and 200 controls) between June and October 2025. Participants were randomly assigned to training and validation cohorts (6:4 ratio). Multiparametric ultrasound was performed to extract acoustic stiffness (SWE) and hemodynamic (UMA) parameters from four specific cervical regions with verified measurement reproducibility. Optimal features were identified using a machine learning ensemble comprising least absolute shrinkage and selection operator (LASSO) regression, support vector machine recursive feature elimination (SVM-RFE), and random forest (RF) algorithms. A multivariable logistic regression-based nomogram was subsequently developed and validated. Four final cervical predictors were identified through the algorithmic ensemble to construct the diagnostic model: the mean Young's modulus of the cervical internal os anterior lip(CIO_AL_SW_E<sub>mean</sub>), the resistance index of the cervical internal os anterior lip (CIO_AL_RI), the mean Young's modulus of the cervical internal os posterior lip (CIO_PL_SW_E<sub>mean</sub>), and the mean Young's modulus of the cervical mid-cervix posterior lip (MC_PL_SW_E<sub>mean</sub>), Individually, the CIO_AL_SW_E<sub>mean</sub> and CIO_PL_SW_E<sub>mean</sub> demonstrated strong diagnostic performance, with areas under the curve (AUCs) of 0.88 and 0.86, respectively. The integrated nomogram incorporating these four optimized parameters achieved an AUC of 0.94 in the training cohort and 0.96 in the validation cohort. A multiparametric model based on cervical Young's modulus and microvascular parameters provides high diagnostic accuracy for adenomyosis. By utilizing the cervix as an accessible diagnostic window, this noninvasive approach offers a standardized method that may serve as a valuable diagnostic complement to conventional myometrial ultrasound evaluations.