Radiomics-Based Non-Invasive MRI Differentiation of Uterine Sarcomas and Fibroids

• Conditions: Uterine Fibroid; Uterine Sarcoma; Diagnose Disease; AI (Artificial Intelligence)

• Registration Number: NCT07129005
• Lead Sponsor: Tongji Hospital

Brief Summary

This retrospective case-control study aims to develop and validate a diagnostic model based on multimodal big data and artificial intelligence to differentiate uterine leiomyoma from uterine sarcoma. Investigators will extract historical case data from existing inpatient and outpatient records, including medical history, physical and gynecological examination findings, MRI imaging data, laboratory results, and pathological records. The study seeks to address the question of whether integrating diverse retrospective clinical data with advanced AI techniques can accurately classify uterine tumors as benign leiomyomas or malignant sarcomas, thereby supporting clinical decision-making and optimizing diagnostic workflows.

Detailed Description

Uterine fibroids are the most common benign gynecological tumors among women of reproductive age in China, with an incidence that has been increasing annually. Statistics show that the prevalence of uterine fibroids among women over 30 years old in China has reached 20%-30%, and the onset age is trending younger. During the "12th Five-Year Plan" period, significant progress was made in the minimally invasive and pharmacological treatment of uterine fibroids through enhanced allocation of medical resources, advancement of clinical research, and improvement of diagnostic and treatment guidelines. However, with the rapid economic and social development in China, changes in environmental factors, lifestyle shifts, and delayed childbearing associated with improved living standards have contributed to a continued rise in the incidence of uterine fibroids. Uterine fibroids have now become a major public health issue affecting women's health in China.

Elucidating the mechanisms underlying the onset, recurrence, and malignant transformation of uterine fibroids, developing individualized treatment plans based on fertility preservation, and identifying high-risk populations to reduce disease progression and recurrence have become critical challenges in the field of reproductive health and women's and children's health research in China. Solving these issues is not only essential for improving women's health and well-being but also for enhancing population quality and reducing the healthcare burden.

In collaboration with the National Clinical Research Center for Obstetrics and Gynecology and regional medical centers (under construction), participating institutions will collect clinical, imaging, pathological, laboratory, and molecular testing data to establish a multicenter, systematic database. Machine learning algorithms will be used to develop early-warning models for malignant transformation and prognostic risk prediction models. Internal validation and optimization will be performed using different grouped datasets from this database, while large-scale data accumulated in Project 1 will be used for both internal and external validation, ultimately resulting in the construction of accurate and efficient early-warning and risk prediction models.

This multicenter retrospective observational study is led by Tongji Hospital in collaboration with several tertiary hospitals, including Zhongnan Hospital of Wuhan University, The Second Hospital of Shandong University, Shenzhen Second People's Hospital, West China Second University Hospital of Sichuan University, and The Third Affiliated Hospital of Zhengzhou University. The study protocol, including the use of existing inpatient and outpatient medical records, has been reviewed and approved by the Ethics Committee of Tongji Hospital (serving as the central IRB). Participating centers have either obtained approval from their local institutional review boards (IRBs) or formally accepted the central IRB approval. All procedures strictly adhere to the Declaration of Helsinki and relevant national ethical guidelines to ensure the protection of patient privacy and data confidentiality.

Study Timeline

• Study Start: 2025-01-01
• Status Verified: 2025-07
• Primary Completion: 2025-07-30
• Study First Submitted: 2025-08-03
• Study First Submitted QC: 2025-08-18
• Last Update Submitted: 2025-08-18
• Study First Posted: 2025-08-19
• Last Update Posted: 2025-08-19
• Study Completion: 2025-12-30

Recruitment & Eligibility

• Status: ENROLLING_BY_INVITATION
• Sex: Female
• Target Recruitment: 520

Inclusion Criteria

1. Histopathological confirmation of uterine sarcoma or leiomyoma.
2. Availability of preoperative MRI, includingT2WI and DWI, performed within 2 months of the surgery.

Exclusion Criteria

1. Tumors smaller than 2 cm. Small tumors may be difficult to accurately perform segmentation and feature extraction, which may affect the accuracy and reliability of the model.
2. Non-primary uterine sarcomas. Sarcomas from other sites with metastasis to the uterus were excluded because the biological characteristics and imaging findings of these tumors may differ from those of primary uterine sarcomas and may lead to bias in the diagnostic model.
3. Concurrent pelvic malignancies. To avoid the influence of other types of tumors on the imaging features of uterine sarcoma and leiomyoma, and to ensure the pertinence and accuracy of the model.

Study & Design

• Study Type: OBSERVATIONAL
• Study Design: Not specified

Primary Outcome Measures

Name	Time	Method
Sensitivity	through study completion, about July.2025	Ability of the test to correctly identify those with uterine sarcoma (true positive rate)
Specificity	through study completion, about July.2025	Ability of the test to correctly identify those without uterine sarcoma (true negative rate）
Negative Predictive Value (NPV)	through study completion, about July.2025	Probability that subjects with a negative test truly don't have the uterine fibroids
AUC	through study completion, about July.2025	AUC stands for Area Under the Curve, specifically under the ROC (Receiver Operating Characteristic) curve
Positive Predictive Value (PPV)	through study completion, about July.2025	Probability that subjects with a positive test truly have uterine sarcoma

Secondary Outcome Measures

Name	Time	Method
Intraclass Correlation Coefficient	Immediately after VOI delineation on baseline MRI
SHapley Additive exPlanations	through study completion, about July.2025
Comparative Performance of the Intratumoral, Peritumoral, and Combined Models	At model performance evaluation (following baseline imaging analysis),about August，2025

Trial Locations

Locations (1)

Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology; 🇨🇳 Wuhan, Hubei, China