• DMP ID: 10.48321/D1J31B
  • Version: 15 Oct 2022

This page describes a data management plan written for the Federation of American Societies for Experimental Biology (faseb.org) using the DMPTool.

FAIR annotated dataset of stroke MRIs, CTs, and metadata

Contributors to this project

Project details

  • Research domain: Health sciences
  • Project Start: October 03, 2022
  • Project End: October 01, 2027
  • Created: October 14, 2022
  • Modified: October 15, 2022
  • Ethical issues related to data that this DMP describes? no

Citation

Funding status and sources for this project

Project description

  • To extract meaningful and reproducible models of brain function from stroke images, for both clinical and research proposes, is a daunting task severely hindered by the great variability of lesion frequency and patterns. Large datasets are therefore imperative, as well as fully automated image post-processing tools to analyze them. The development of such tools, particularly with artificial intelligence, is highly dependent on the availability of large datasets to model training and testing. We will create and share a public dataset of 4,000 multimodal clinical MRIs and CTs of patients with acute and early subacute stroke, with manual lesion segmentation, and metadata. The dataset provides high quality, large scale, human-supervised knowledge to feed artificial intelligence models and enable further development of tools to automate several tasks that currently rely on human labor, such as lesion segmentation, labeling, calculation of disease-relevant scores. It also represents a valuable training and testing resource for translational research relating lesion features to risk factors, brain functions and patients’ outcomes.

Planned outputs

ANNOTATED CLINICAL MRIS AND LINKED METADATA OF PATIENTS WITH ACUTE STROKE, BALTIMORE, MARYLAND, 2009-2019

A dataset with 2,888 annotated multimodal clinical MRIs and metadata, deposited in ICPSR (https://doi.org/10.3886/ICPSR38464). The public release, by curation completion, is expected by the end of 2022

ACUTE-STROKE DETECTION AND SEGMENTATION

This a free, public, user-friend tool developed with this dataset, to automatically detect, segment and quantify acute ischemic strokes. It converts images in CDO (computable data objects) enabling direct AI modeling. It runs in real time, in regular local computers, with a single command line, facilitating the use by non-experts and researchers in diverse fields and enabling easy reproducible and replicable research. Recently added features are the output of radiological reports and the stroke score ASPECTS. Currently this resource has more than 400 downloads in Nitrc. Accessible at https://www.nitrc.org/projects/ads

  • Format:Software
  • Anticipated volume:unspecified
  • Release timeline:May 05, 2021
  • Intended repository:NITRC

ARTERIAL ATLAS

This is the first digital 3D atlas of brain arterial territories, developed with this dataset, currently with more than 700 dowloads all around the world. Accessible at https://www.nitrc.org/projects/arterialatlas

  • Format:Interactive Resource
  • Anticipated volume:unspecified
  • Release timeline:May 05, 2021
  • Intended repository:NITRC

A LARGE PUBLIC DATASET OF ANNOTATED CLINICAL MRIS OF PATIENTS WITH ACUTE STROKE AND LINKED METADATA

Description of the StrokeFAIR dataset. Pre-print available in Research Square https://www.researchsquare.com/article/rs-1705779/v1. Submitted to Nature Scientific Data.

  • Format:Data Paper
  • Release timeline:January 14, 2023

DEEP LEARNING-BASED DETECTION AND SEGMENTATION OF DIFFUSION ABNORMALITIES IN ACUTE ISCHEMIC STROKE

Liu, CF., Hsu, J., Xu, X. et al. Deep learning-based detection and segmentation of diffusion abnormalities in acute ischemic stroke. Commun Med 1, 61 (2021). https://doi.org/10.1038/s43856-021-00062-8. https://www.nature.com/articles/s43856-021-00062-8. This paper describes a tool, created with this dataset, for automated detection, segmentation, and quantification of acute strokes. It outputs the 3D mask of the stroke core and the ratio of brain regions affected by the stroke in two parcelation schemes (classical anatomy and arterial territories). This toll is free and publicly available, runs in local computers, is user-friendly to non-experts and people in diverse field, facilitating reproducible and replicable research

  • Format:Data Paper
  • Release timeline:December 06, 2021

DIGITAL 3D BRAIN MRI ARTERIAL TERRITORIES ATLAS

Liu CF, Hsu J, Xu X, Kim NG, Sheppard SM, Meier EL, Miller M, Hillis AE, Faria AV. Digital 3D Brain MRI Arterial Territories Atlas. https://www.biorxiv.org/content/10.1101/2021.05.03.442478v2.full.pdf. In Press in Sci Data. This paper describes the fisrt digital 3D atlas of brian arterial territories, created with this dataset

  • Format:Data Paper
  • Release timeline:December 02, 2022

AUTOMATIC COMPREHENSIVE ASPECTS REPORTS IN CLINICAL ACUTE STROKE MRIS

Faria, A., Liu, C. F., Li, A., Kim, G., Miller, M., & Hillis, A. (2022). Automatic Comprehensive Aspects Reports in Clinical Acute Stroke MRIs. Pre-print in Research Square. This paper describes our software to calculate ASPCTS in patients with ischemic strokes, using this dataset. Under revision in Sci Reports

  • Format:Data Paper
  • Release timeline:December 15, 2022

AUTOMATIC COMPREHENSIVE RADIOLOGICAL REPORTS FOR CLINICAL ACUTE STROKE MRIS

Liu, C. F., Zhao, Y., Miller, M. I., Hillis, A. E., & Faria, A. (2022). Automatic comprehensive radiological reports for clinical acute stroke MRIs. Available at Research Square. This manuscript describes our free and public software to generate automated reports, developed with this dataset. Under revision in Nat Comm Med

  • Format:Data Paper
  • Release timeline:December 14, 2022

DETECTION AND SEGMENTATION OF PERFUSION DEFICITS IN CLINICAL ACUTE STROKE MRIS

This tool will detect, segment and quantify perfusion deficits in perfusion-weighted images. It will output not only volumetric information but also the 3D masks of perfusion deficits, to be used on clinical research, particurly to correlate with function or to predcit outcomes

  • Format:Software
  • Anticipated volume:unspecified
  • Release timeline:May 21, 2021
  • Intended repository:NITRC

MRI-BASED PREDICTION OF TIME TO STROKE ONSET

This tool will automatically calculate time to stroke onset based on brian MRIs. The time to onset is a crucial informaiton for acute treatment, but is unknown in about 25% of patients with ischemic stroke, which prevents them to be treated. As our other tools, this will be free and publicly available, and accessible to expert and on-expert users

  • Format:Software
  • Anticipated volume:unspecified
  • Release timeline:June 14, 2023
  • Intended repository:NITRC

CONTENT-BASED IMAGE RETRIEVAL (CBIR) FOR ACUTE STROKE MRIs

Using our "ADS" system already in place, this tool will automatically calculate transform the original data in a computational data object and search in our dataset (which will be then a library of thousands of computational data objects) for the clusterof similar cases. Then, it will output frequence reports of relevant information in this cluster, for instance, 90-days follow up scores and response to acute treatment. It will also be useful to identify specific risk-factors, as it is linked to the metadata. This will enable population stratification and personalized medical approach. As our other tools, it will be free and publicly available, and accessible to expert and on-expert users

  • Format:Software
  • Anticipated volume:unspecified
  • Release timeline:September 14, 2023
  • Intended repository:NITRC

SYNTHETIC CTs FOR QUANTIFICATION OF ACUTE ISCHEMIC STROKE BASED ON DIFFUSION MRISSYNTHETIC CTs FOR QUANTIFICATION OF ACUTE ISCHEMIC STROKE BASED ON DIFFUSION MRIS

By cost reasons, computed tomography (CT) is still the first image acquired for most of patients with acute stroke. However, if it less sensitive than MRIs in the hyper acute stage, which leads to misdiagnosis or underestimation of the stroke volume. Using our dataset, we will create an artificial intelligence tool (likely based on deep-learning) to estimate the volume of the ischemic stroke in CTs, based on the information provided by the diffusion weighted images. By submitting a CT showing a questionable stroke core, the user will receive what the MRI would look-like and the predicted area and volume of the stroke core in the CT. As our other tools, this will be free and publicly available, and accessible to expert and on-expert users.

  • Format:Software
  • Anticipated volume:unspecified
  • Release timeline:December 14, 2023
  • Intended repository:NITRC

IMAGE AND NON-IMAGE DETERTMINANTS OF NIHSS IN ACUTE ISCHEMIC STROKES

This paper is an analysis of image and metadata in our database to identify the factors that most influence the calculation of NIHSS scores in patients with acute stroke, an index of stroke severity that is relevant for treatment and outcome prediction.

  • Format:Data Paper
  • Release timeline:June 14, 2023

VARIOUS TESTS OF LEFT NEGLECT ARE ASSOCIATED WITH DISTINCT TERRITORIES OF HYPOPERFUSION IN ACUTE STROKE

Stein, C., Bunker, L., Chu, B., Leigh, R., Faria, A., & Hillis, A. E. (2022). Various tests of left neglect are associated with distinct territories of hypoperfusion in acute stroke. Brain communications, 4(2), fcac064. This paper exemplifies the use of our dataset, and of the tools for automated analysis derived from it, to stablish anatomic-functional correlation in stroke patients.

  • Format:Data Paper
  • Release timeline:August 14, 2022

LEFT HEMISPHERE BIAS OF NIH STROKE SCALE IS MOST SEVERE FOR MIDDLE CEREBRAL ARTERY STROKES

https://www.frontiersin.org/articles/10.3389/fneur.2022.912782/full. Using this dataset, we found a hemisphere-related bias in a largely used clinical indice for strokes, the NIHSS.

  • Format:Data Paper
  • Release timeline:August 14, 2022

ASSOCIATION OF INFERIOR DIVISION MCA STROKE LOCATION WITH POPULATIONS WITH ATRIAL FIBRILLATION INCIDENCE

Kim, G., Vitti, E., Stockbridge, M. D., Hillis, A. E., & Faria, A. V. (2021). Association of inferior division MCA stroke location with populations with atrial fibrillation incidence. Available at https://www.medrxiv.org/content/10.1101/2021.12.06.21267371v1. This paper describes a demographic association of the stroke location in our dataset. Under revision in Heliyon

  • Format:Data Paper
  • Release timeline:December 14, 2022

POOR GLYCEMIC CONTROL IS ASSOCIATED WITH WORSE BLOOD-BRAIN BARRIER DISRUPTION IN ISCHEMIC STROKE PATIENTS

We describe the association between metabolic profile and an MRI indice of blood perfusion, found in an external small dataset and confirmed in our dataset. This proves its value as an external test set for replication studies. The abstract will be presetned in the ANA Oct 2022, Chicago

  • Format:Data Paper
  • Release timeline:October 28, 2022

WORSE BLOOD-BRAIN BARRIER DISRUPTION IN ISCHEMIC STROKE IS ASSOCIATED WITH LONGER HOSPITAL LENGTH OF STAY

This manuscript describes an association between metabolic profile and an MRI indice of blood perfusion, found in an external small dataset and confirmed in our dataset. This proves its value as an external test set for replication studies. The abstract will be presented in the WSC Oct 2022, Singapore

  • Format:Data Paper
  • Release timeline:October 28, 2022

AUTOMATIC RADIOLOGICAL REPORTS FOR ACUTE ISCHEMIC STROKES

https://www.nitrc.org/projects/ads. The second version of "ADS" (ADSv1) additionally offers imaging mapping to different standard spaces, the generation of automated radiological reports. It outputs the feature vectors used and comprehensive reports of the classification results. It runs in real time, in regular local computers, with a single command line, facilitating the use by non-experts and researchers in diverse fields and enabling easy reproducible and replicable research.

  • Format:Software
  • Anticipated volume:unspecified
  • Release timeline:August 12, 2022
  • Intended repository:NITRC

AUTOMATIC CALCULATION OF ASPECTS FOR ACUTE ISCHEMIC STROKES IN BRIAN MRIs

https://www.nitrc.org/projects/ads. The second version of "ADS" (ADSv1) additionally offers the calculation of "ASPECTS" for brain MRIs, which is an important index for acute treatment and prognosis. It outputs the feature vectors used and comprehensive reports of the classification results. It runs in real time, in regular local computers, with a single command line, facilitating the use by non-experts and researchers in diverse fields and enabling easy reproducible and replicable research.

  • Format:Software
  • Anticipated volume:unspecified
  • Release timeline:August 12, 2022
  • Intended repository:NITRC

Other works associated with this research project