Overview

Schema Version

https://atlas.rsna.org/schemas/2025-11/dataset.json

Name

Brain MRI scans with radiology reports from five Chinese tertiary hospitals (2018–2023) used in ReportGuidedNet study

Link

https://dx.doi.org/10.1148/ryai.230520

Indexing

Keywords: Brain MRI, Radiology reports, Lesion detection, Interpretable deep learning, Multicenter, Noncontrast MRI

Content: MR, NR

RadLex: RID35976, RID10312, RID45946, RID28768, RID43359, RID13060

Author(s)

Lisong Dai

Jiayu Lei

Fenglong Ma

Zheng Sun

Haiyan Du

Houwang Zhang

Jingxuan Jiang

Jianyong Wei

Dan Wang

Guang Tan

Xinyu Song

Jinyu Zhu

Qianqian Zhao

Songtao Ai

Ai Shang

Zhaohui Li

Ya Zhang

Yuehua Li

Organization(s)

Institute of Diagnostic and Interventional Radiology, Shanghai Sixth People’s Hospital Affiliated to Shanghai Jiao Tong University School of Medicine

Clinical Research Center, Shanghai Sixth People’s Hospital Affiliated to Shanghai Jiao Tong University School of Medicine

Shanghai AI Laboratory

School of Computer Science and Technology, University of Science and Technology of China

The Pennsylvania State University College of Information Sciences and Technology

Department of Electrical Engineering, City University of Hong Kong

Department of Radiology, Affiliated Hospital of Nantong University

Department of Radiology, Shanghai Ninth People’s Hospital Affiliated to Shanghai Jiao Tong University School of Medicine

Department of Radiology, Shanghai Public Health Clinical Center

Department of Radiology, Wuhan Hankou Hospital

Cooperative Medianet Innovation Center, Shanghai Jiao Tong University

License

Text: CC BY 4.0

URL: https://creativecommons.org/licenses/by/4.0/

Contact

Yuehua Li, email as shown in article: moc.361@9250auheuyil

Funding

National Natural Science Foundation of China (8225024); National Key Research and Development Program of China (2022ZD0160702); Shanghai Science and Technology Innovation Action Plan (20S31907300).

Ethical review

Institutional review board approval (approval no. 2023-KY-082 [K]) with waiver of informed consent.

Comments

Retrospective, multicenter dataset of noncontrast brain MRI scans paired with de-identified radiology reports for developing and evaluating a deep learning model (ReportGuidedNet) for interpretable lesion detection across 14 diseases plus normal brain.

Date

Published: 2024-10-09

References

[1] Dai L, Lei J, Ma F, et al.. "Boosting Deep Learning for Interpretable Brain MRI Lesion Detection through the Integration of Radiology Report Information". Radiology: Artificial Intelligence. 2024-11-01. doi:10.1148/ryai.230520. PMID: 39377669. PMCID: PMC11605145.

Dataset

Motivation

Use radiology report–derived textual features as weak annotations to guide model attention for interpretable brain lesion detection without manual image segmentation.

Sampling

Adult (≥18 years) outpatients and inpatients undergoing noncontrast brain MRI. Exclusions: prior brain surgery, low image quality, incomplete data, or uncertain diagnosis after expert review.

Partitioning scheme

Center 1 (2018–06/2023) used for training/validation/internal testing; Centers 2–5 (01–12/2022) reserved for external testing.

Relationships between instances

If multiple brain imaging exams existed for a patient during the study period, only the earliest scan was included (one scan per patient).

Noise

Low-quality scans (artifact/improper settings compromising diagnosis) were excluded per radiologist assessment.

Confidentiality

Radiology reports were de-identified with personal identifiers removed and replaced by unique IDs to match MR images.

Sensitive data

Medical images and associated clinical radiology reports from hospital PACS/records.