[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"health-study-detail:100631419":3},{"organization":4,"armGroups":7,"interventions":22,"overallOfficials":28,"centralContacts":32,"locations":41,"responsibleParty":56,"collaborators":59,"id":62,"slug":63,"hasResults":64,"nctId":65,"briefTitle":66,"officialTitle":67,"acronym":68,"eligibilityCriteria":69,"healthyVolunteers":70,"sex":71,"minAge":72,"maxAge":73,"enrollmentInfo":74,"targetDuration":10,"studyType":77,"phases":10,"briefSummary":78,"conditions":79,"keywords":83,"overallStatus":43,"whyStopped":10,"lastUpdateSubmitDate":90,"lastUpdatePostDateStruct":91,"startDateStruct":94,"completionDateStruct":96,"leadSponsor":98,"locationsCount":99},{"fullName":5,"class":6},"Peking Union Medical College Hospital","OTHER",[8,14,18],{"label":9,"type":10,"description":11,"interventionNames":12},"Normal Breast",null,"Breast ultrasound images showing normal glandular tissue across different tissue composition types, with no focal lesions identified. Confirmed by senior radiologist review.",[13],"Diagnostic Test: Multimodal AI Model Diagnostic Evaluation",{"label":15,"type":10,"description":16,"interventionNames":17},"Benign Lesion","Breast ultrasound images containing pathologically confirmed benign lesions (BI-RADS 2-4B), including fibroadenoma, cyst, lipoma, sclerosing adenosis, intraductal papilloma, and selected non-mass lesions (NML).",[13],{"label":19,"type":10,"description":20,"interventionNames":21},"Malignant Lesion","Breast ultrasound images containing pathologically confirmed malignant lesions (BI-RADS 3-5), including invasive ductal carcinoma, invasive lobular carcinoma, mucinous carcinoma, and selected non-mass lesions (NML).",[13],[23],{"type":24,"name":25,"description":26,"armGroupLabels":27,"otherNames":10},"DIAGNOSTIC_TEST","Multimodal AI Model Diagnostic Evaluation","Retrospective evaluation of de-identified breast ultrasound images by multiple AI systems, including baseline deep learning models (ResNet-50, USFM) and multimodal large language models, using standardized BI-RADS-guided chain-of-thought prompts via API. No patient contact or clinical decision-making is involved.",[15,19,9],[29],{"name":30,"affiliation":5,"role":31},"Qingli Zhu, MD","PRINCIPAL_INVESTIGATOR",[33,37],{"name":30,"role":34,"phone":35,"phoneExt":10,"email":36},"CONTACT","+86 13621376699","zqlpumch@126.com",{"name":38,"role":34,"phone":39,"phoneExt":10,"email":40},"Yinglan Wu, MD","+86 15626121076","wuylan7@gmail.com",[42],{"facility":5,"status":43,"city":44,"state":10,"zip":45,"country":46,"countryCode":47,"cosmosGeoPoint":48,"geoPoint":53,"contacts":54},"RECRUITING","Beijing","100730","China","CN",{"type":49,"coordinates":50},"Point",[51,52],116.39723,39.9075,{"lat":52,"lon":51},[55],{"name":30,"role":34,"phone":35,"phoneExt":10,"email":36},{"type":31,"investigatorFullName":57,"investigatorTitle":58,"investigatorAffiliation":5,"oldNameTitle":10,"oldOrganization":10},"Qingli Zhu","Professor, Department of Ultrasound, Peking Union Medical College Hospital",[60],{"name":61,"class":6},"Chinese Academy of Medical Sciences","100631419","construction-of-a-benchmark-for-breast-ultrasound-ai-interpretation-and-performance-evaluation-of-multimodal-ai-models-100631419",false,"NCT07500428","Construction of a Benchmark for Breast Ultrasound AI Interpretation and Performance Evaluation of Multimodal AI Models","Construction of a Standardized Benchmark Evaluation System for Intelligent Breast Ultrasound Image Interpretation and Systematic Performance Assessment of Multimodal Artificial Intelligence Models Based on ACR BI-RADS v2025 Criteria","BUST-AI Bench","Inclusion Criteria:\n\n* B-mode breast ultrasound grayscale images from the institutional PACS database or from published open-access breast ultrasound datasets with documented original institutional ethics approval\n* Image quality adequate for clinical diagnosis with clear visualization of the region of interest\n* Pathological diagnosis confirmed (for benign and malignant lesion groups), or normal breast status confirmed by a senior radiologist with \\>15 years of breast ultrasound experience (for the normal group)\n* Complete de-identification with removal of all personally identifiable information\n\nExclusion Criteria:\n\n* Severely degraded image quality precluding meaningful BI-RADS assessment\n* Duplicate images from the same patient (only the most representative image retained per lesion)\n* Images with residual personally identifiable information after de-identification processing\n* Cases with ambiguous, disputed, or unavailable pathological results\n* Non-B-mode ultrasound images, including elastography, contrast-enhanced ultrasound, and Doppler imaging",true,"FEMALE","18 Years","75 Years",{"count":75,"type":76},1380,"ESTIMATED","OBSERVATIONAL","This single-center, retrospective, observational study aims to construct a standardized benchmark evaluation system for intelligent breast ultrasound image interpretation and to systematically assess the diagnostic performance of current mainstream multimodal artificial intelligence (AI) models.\n\nDe-identified B-mode breast ultrasound images with confirmed pathological diagnoses will be retrospectively collected from the institutional archive (2018-2025) and supplemented with images from published open-access datasets. Expert radiologists with varying experience levels will independently annotate all images according to the American College of Radiology (ACR) Breast Imaging Reporting and Data System (BI-RADS) v2025 criteria, including glandular tissue composition, lesion characterization (mass vs. non-mass lesion), morphological descriptors, and final BI-RADS classification.\n\nBaseline deep learning models (CNN-based ResNet-50 and Transformer-based USFM) will be trained to establish performance baselines and to stratify cases by diagnostic difficulty through cross-architecture consensus. Multiple multimodal large language models (MLLMs), including both general-purpose and medical-domain models, will then be evaluated via standardized API calls using BI-RADS-guided chain-of-thought prompts at temperature 0 for reproducibility.\n\nPrimary endpoints include BI-RADS classification accuracy and diagnostic AUC for benign-malignant differentiation. Model robustness and safety will be assessed through out-of-distribution rejection testing, temperature-stability experiments, and thinking-mode ablation studies. This study adheres to the FLAIR and TRIPOD-LLM reporting guidelines.",[80,81,82],"Breast Neoplasms","Breast Diseases","Ultrasonography",[84,85,86,87,88,89],"Breast Ultrasound","BI-RADS","Artificial Intelligence","Multimodal Large Language Model","Benchmark","Computer-Aided Diagnosis","2026-03-24",{"date":92,"type":93},"2026-03-30","ACTUAL",{"date":95,"type":93},"2026-03-12",{"date":97,"type":76},"2027-03-01",{"name":5,"class":6},1]