MRNet
MRNet: A Dataset of Knee MRIs
1,370 clinical knee MRI exams from Stanford University Medical Center (2001-2012) with report-derived labels for abnormality, ACL tear and meniscal tear; 120 exams form a hidden test set. Non-commercial research use only.
Overview
MRNet is a collection of 1,370 clinical knee MRI exams from Stanford University Medical Center, acquired between 2001 and 2012. The Stanford Machine Learning Group assembled it to train and evaluate MRNet, a convolutional network that classifies whole knee exams, and released it together with a public benchmark. Each exam carries three binary labels: any abnormality, anterior cruciate ligament (ACL) tear and meniscal tear. The data are now distributed by the Stanford AIMI center through Redivis.
Composition
Of the 1,370 exams, 1,104 are abnormal, 319 show an ACL tear and 508 a meniscal tear; 194 have both tears. The exams are divided into a training set (1,130 exams, 1,088 patients with an identifier), a validation set (120 exams, 111 patients) and a test set (120 exams, 113 patients). The paper calls the validation set the tuning set and the test set the validation set. All exams of one patient fall into the same split, and the two smaller sets were sampled so that each holds at least 50 positive cases per label. The test set is held back for the benchmark and is not part of the download. The mean patient age is 38.0 years, and 569 exams belong to female patients.
Acquisition
All exams were done on GE scanners with a standard knee coil and a routine protocol without contrast: coronal T1-weighted, coronal T2-weighted with fat saturation, sagittal proton density, sagittal T2-weighted with fat saturation and axial proton density with fat saturation. 775 exams were acquired at 3 T and the rest at 1.5 T. The paper used three series per exam (sagittal T2-weighted, coronal T1-weighted and axial proton density), resampled to 256 × 256 pixels.
Annotations
Training and validation labels were extracted by hand from the clinical radiology reports. For the test set, the reference labels are the majority vote of three musculoskeletal radiologists who had the images, reports, clinical history and follow-up exams.
Known limitations
Released labels come from reports, not surgery, and are exam-level only, with no localization. The data come from one institution and one scanner vendor. The research use agreement forbids commercial use, redistribution and derivative works.
Cohort
Aggregate numbers from the sources below. Bars are relative to the largest value.
Age
mean 38No age bins reported.
Condition
studies, values can overlap
- Meniscal tear 508 37%
- Anterior cruciate ligament tear 319 23%
Scanner vendor
studies
- GE HealthCare 1,370 100%
Field strength
studies
- 3 T 775 57%
- 1.5 T 595 43%
Split
subjects
- Training 1,088
- Test 113
- Validation 111
License and access
Our reading of the license, not legal advice. Before you use the data, read the original license and confirm that your use is allowed. We take no responsibility for how you use a dataset. Full disclaimer
Create an account and accept terms online
Register and accept the Stanford research use agreement. The data are downloaded through Redivis. The test set is not released.
Stanford University School of Medicine Dataset Research Use Agreement (AIMI)
Free for personal, non-commercial research only. You may not share the data or the download link, may not create derivative works, and must not try to re-identify patients. Stanford sells a separate commercial license.
The MRNet page publishes the Stanford University School of Medicine Research Use Agreement with "MRNet Dataset" as the dataset name. It also states that the dataset is for non-clinical research use only and must not be relied upon in patient care.
What you can do
- No
- Not stated
- No
- Not stated
What you can share
- No
- No
- Share trained models Not stated
What you must do
- Yes
- Share alike No
- Yes
- Ethics approval No
- Manuscript review No
- Release code No
- Return results No
- No
Limits
- Yes
- Location limits No
Commercial license: Yes Companies can apply for a one-year license per dataset, reviewed by a Stanford committee. The page lists an annual fee of USD 70,000 per dataset for agreements contracted in FY25. Details
Citation
Bien N, Rajpurkar P, Ball RL, et al. Deep-learning-assisted diagnosis for knee magnetic resonance imaging: Development and retrospective validation of MRNet. PLoS Med 15(11): e1002699 (2018). https://doi.org/10.1371/journal.pmed.1002699. Dataset DOI: https://doi.org/10.71718/rcbp-8c35
All numbers
Every number on this page, as stored in stats.csv, with its source.
| Measure | Breakdown | Value | Source |
|---|---|---|---|
| Studies | total knee MRI exams performed 2001-2012 | 1,370 | bien2018 Methods - Dataset |
| Studies | split=train | 1,130 | mrnet-page Splits |
| Studies | split=validation called tuning set in the paper | 120 | mrnet-page Splits |
| Studies | split=test hidden test set (validation set in the paper); not released | 120 | mrnet-page Splits |
| Subjects | split=train for the 1114 training exams with a patient identifier | 1,088 | bien2018 Table 1 |
| Subjects | split=validation | 111 | mrnet-page Splits |
| Subjects | split=test | 113 | mrnet-page Splits |
| Studies | sex=female reported as female patients; the percentage (41.5%) is of the 1,370 exams | 569 | bien2018 Abstract |
| Studies | split=train;sex=female reported as female patients; the percentage is of exams | 480 | bien2018 Table 1 |
| Studies | split=validation;sex=female reported as female patients; the percentage is of exams | 50 | bien2018 Table 1 |
| Studies | split=test;sex=female reported as female patients; the percentage is of exams | 39 | bien2018 Table 1 |
| Mean age | total | 38 | bien2018 Abstract |
| Studies | condition=acl_tear labels from clinical reports | 319 | bien2018 Methods - Dataset |
| Studies | condition=meniscal_tear labels from clinical reports | 508 | bien2018 Methods - Dataset |
| Studies | split=train;condition=acl_tear labels from clinical reports | 208 | bien2018 Table 1 |
| Studies | split=validation;condition=acl_tear labels from clinical reports | 54 | bien2018 Table 1 |
| Studies | split=test;condition=acl_tear labels from clinical reports; 58 by the radiologist reference standard | 57 | bien2018 Table 1 |
| Studies | split=train;condition=meniscal_tear labels from clinical reports | 397 | bien2018 Table 1 |
| Studies | split=validation;condition=meniscal_tear labels from clinical reports | 52 | bien2018 Table 1 |
| Studies | split=test;condition=meniscal_tear labels from clinical reports; 65 by the radiologist reference standard | 59 | bien2018 Table 1 |
| Studies | field_strength=3 | 775 | bien2018 Methods - Dataset |
| Studies | field_strength=1.5 1,370 minus the 775 exams at 3 T; the paper states the remaining exams used 1.5 T | 595 | bien2018 Methods - Dataset |
| Studies | vendor=ge all exams on GE scanners | 1,370 | bien2018 Methods - Dataset |
Sources
The keys used in the table above.
- bien2018 Bien et al. 2018, Deep-learning-assisted diagnosis for knee MRI - development and retrospective validation of MRNet (PLoS Med 15(11)) paper
- mrnet-page Stanford ML Group MRNet dataset and competition page website
- datacite-mrnet DataCite record of the MRNet dataset DOI (publication year 2018) website