Skip to content

Research on artificial intelligence based breast mammography combined with multi-source diagnosis and treatment data for lesion identification, pathological analysis, and prognosis prediction

Research on artificial intelligence based breast mammography combined with multi-source diagnosis and treatment data for lesion identification, pathological analysis, and prognosis prediction

Status
Active, not recruiting
Phases
Unknown
Study type
Observational
Source
ChiCTR
Registry ID
ChiCTR2600128322
Enrollment
Unknown
Registered
2026-07-17
Start date
2026-07-19
Completion date
Unknown
Last updated
2026-07-20

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

Breast lesions (including benign breast lesions, suspicious breast lesions, and breast cancer)

Interventions

Retrospective Cohort:None

Sponsors

Beijing Friendship Hospital ,Capital Medical University
Lead Sponsor

Eligibility

Sex/Gender
Female

Inclusion criteria

Inclusion criteria: This study intends to include the previous breast related diagnosis, treatment or examination records of Beijing Friendship Hospital Affiliated to Capital Medical University, which should have mammography imaging data of the breast, as well as breast ultrasound examination and examination reports, and also have pathology Results and case data of pathological reports. The molybdenum target images, ultrasound examination reports and pathological results of the included cases should be able to be reasonably matched by the hospital information management department based on the in-hospital data association rules, combined with information such as the examination time, examination side and lesion location. The research team only accepted the research data that had been desensitized by the platform for the analysis of this study.

Exclusion criteria

Exclusion criteria: Exclude those with missing mammography images or severely poor image quality. Those who are missing key materials such as ultrasound examination reports, pathological results, and pathological reports; Those whose molybdenum target images, ultrasound reports and pathological results cannot be effectively matched; Image data with severe artifacts, image damage or abnormal format that cannot be used for model analysis; Those whose basic information or key diagnostic and therapeutic information is seriously lacking and cannot meet the requirements of research and analysis; And in the past, they have explicitly refused to use their medical records for scientific researchers.

Design outcomes

Primary

MeasureTime frame
Performance of Breast Lesion Detection Model;Performance of Imaging–Pathology Association Model;Performance of Breast Lesion Risk Prediction Model;

Countries

China

Contacts

Public ContactLv Han

Beijing Friendship Hospital ,Capital Medical University

chrislvhan@126.com+86 10 63139013

Outcome results

None listed

Source: ChiCTR (via WHO ICTRP) · Data processed: Jul 23, 2026