QSM radiomics in multiple sclerosis (Bologna)
MRI dataset for susceptibility-based radiomic feature extraction in healthy controls and patients with multiple sclerosis
Brain MRI with quantitative susceptibility mapping (QSM) from adults with multiple sclerosis and healthy controls, collected to extract and test radiomic features of normal-appearing white matter and six white matter tracts.
Overview
Brain MRI from 100 adults with multiple sclerosis (MS) and 50 healthy controls, scanned at the IRCCS Istituto delle Scienze Neurologiche di Bologna, Italy. The dataset was built to study whether radiomic features computed from quantitative susceptibility mapping (QSM) of normal-appearing white matter are robust, and it ships the images, the volumes of interest, the extracted features and the processing scripts. It is described in Fiscone et al., Scientific Data 2024.
Composition
The cohort has 100 MS patients (58 women, 42 men, mean age 47.6 years) and 50 controls (31 women, 19 men, mean age 57.1 years). Patients had relapsing-remitting, primary progressive or secondary progressive MS under the 2017 McDonald criteria and were treated with intravenous anti-CD20 antibodies. Each of the 150 subject folders, organised in BIDS, holds a T1-weighted MPRAGE, a 3D FLAIR, the raw magnitude and phase of a five-echo gradient-echo scan, the QSM map registered to T1w space, and preprocessed diffusion-weighted images with b-values, b-vectors and registration matrices. A participants file gives age, sex, scan date, group and whether the diffusion scan was single-shell or multi-shell.
Acquisition
All scans come from one Siemens Magnetom Skyra 3 T scanner with a 64-channel head and neck coil, acquired between November 2020 and August 2023. The diffusion protocol changed during the study from single-shell (b = 2000) to multi-shell; the shared diffusion files contain the unweighted and b = 2000 volumes. Images were defaced and de-identified.
Annotations
Volumes of interest were produced automatically: white matter from MRtrix3 5ttgen, MS lesions from the LST lesion prediction algorithm (patients only), and six tracts (arcuate, corticospinal, frontal aslant, inferior fronto-occipital, optic radiation, uncinate) from tractography, each split by hemisphere with lesions removed. For each of the 14 volumes, 107 PyRadiomics features from the QSM map are provided. No manual labels are included.
Known limitations
Single site and single scanner. Controls are on average about ten years older than patients. For three controls the FLAIR comes from an earlier exam, up to 42 months before the other sequences. All masks are automatic; the lesion masks used to define normal-appearing white matter are not listed among the shared files.
Cohort
Aggregate numbers from the sources below. Bars are relative to the 150 subjects.
Sex
- Female 89 59%
- Male 61 41%
Scanner vendor
- Siemens Healthineers 150 100%
MRI data acquisition in Fiscone et al. 2024, Multiparametric MRI dataset for susceptibility-based radiomic feature extraction and analysis, Scientific Data
Field strength
- 3 T 150 100%
MRI data acquisition in Fiscone et al. 2024, Multiparametric MRI dataset for susceptibility-based radiomic feature extraction and analysis, Scientific Data
Condition by sex
Reported cross table. A dot marks a cell the source does not give.
| Female | Male | |
|---|---|---|
| Multiple sclerosis | 58 | 42 |
| Healthy control | 31 | 19 |
Age
range 23 to 86
No age bins reported.
Condition
Groups can overlap
- Multiple sclerosis 100 67%Age mean 47.6 ± 11.4, range 23 to 73
- Healthy control 50 33%Age mean 57.1 ± 16.7, range 24 to 86
Country
- Italy 150 100%
Participants in Fiscone et al. 2024, Multiparametric MRI dataset for susceptibility-based radiomic feature extraction and analysis, Scientific Data
Contrast combinations
How many subjects have exactly each set of contrasts.
| T1w | FLAIR | Chimap | MEGRE | dwi | Subjects with exactly this set |
|---|---|---|---|---|---|
150 |
Data Records in Fiscone et al. 2024, Multiparametric MRI dataset for susceptibility-based radiomic feature extraction and analysis, Scientific Data
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
Sign a data use agreement, often reviewed by the provider
The Zenodo record is restricted. Users create a Zenodo account with a public username or email, email the signed Data Use Agreement to the data holders and are then added as readers of the Zenodo community that holds the data.
MRI dataset for susceptibility-based radiomic feature extraction: Data Use Agreement
Investigators who want to use the data for research or education sign the agreement and email it to the data holders, who then grant read access on Zenodo. The agreement forbids re-identification, contacting participants and passing the data to anyone else, including colleagues at the same institution, and requires acknowledging the data source and citing the data paper.
What you can do
- Not stated
- Not stated
- Not stated
- Not stated
- Conditional
What you can share
- No
- Not stated
- Share trained models Not stated
What you must do
- Yes
- Share alike No
- Yes
- Conditional
- Manuscript review No
- Release code No
- Return results No
- Conditional
Limits
- Yes
- Location limits No
Citation
Fiscone C, Sighinolfi G, Manners DN, Motta L, Venturi G, Panzera I, Zaccagna F, Rundo L, Lugaresi A, Lodi R, Tonon C, Castelli M. Multiparametric MRI dataset for susceptibility-based radiomic feature extraction and analysis. Scientific Data 11, 575 (2024). doi:10.1038/s41597-024-03418-6
Sources
Every number on this page comes from one of these documents. Each chart names the table or page it is taken from. The raw numbers are in stats.csv.