| This ReadMe file was generated on 2024-06-13 by Diedre |
|
|
| ------------------------- |
| General Information |
| ------------------------- |
|
|
| Data Set Title: Long COVID Iowa-UNICAMP dataset |
|
|
| Data creator information: |
|
|
| Diedre S. Carmo[1][https://orcid.org/0000-0002-5922-9120], Alejandro A. Pezzulo[2][https://orcid.org/0000-0001-7544-5109], Raul A. Villacreses[2][https://orcid.org/0000-0002-9928-0182], McKenna L. Eisenbeisz[2][], |
| Rachel L. Anderson[2][], Sarah E. Van Dorin[2][https://orcid.org/0009-0005-7205-4045], Letıcia Rittner[1][https://orcid.org/0000-0001-8182-5554], Roberto A. Lotufo[1][https://orcid.org/0000-0002-5652-0852], Sarah E. |
| Gerard[3][https://orcid.org/0000-0001-8101-3150], Joseph M. Reinhardt[3][https://orcid.org/0000-0003-3526-3591], and Alejandro P. Comellas[2][https://orcid.org/0000-0003-1521-7520] |
|
|
| [1]School of Electrical and Computer Engineering, Universidade Estadual de Campinas, Campinas, 13083-852, Brazil |
| [2]Department of Internal Medicine, University of Iowa, Iowa City, 52242, USA |
| [3]Roy J. Carver Department of Biomedical Engineering, University of Iowa, Iowa City, 52242, USA |
|
|
|
|
| Date(s) of Data Collection: June 2020 to June 2022 |
|
|
|
|
| This work was supported by São Paulo Research Foundation (FAPESP) grant #2022/02344-8 |
|
|
| ------------------------------------------------ |
| SHARING/ACCESS INFORMATION |
| ------------------------------------------------ |
|
|
| Licenses/restrictions placed on the data, or limitations of reuse: Open Data Commons Attribution License (ODC-By) |
|
|
|
|
| Recommended citation for the data set: |
|
|
| Carmo, D. S., Pezzulo, A. A, Villacreses, P. A., Comellas A. P., Eisenbeisz, M. L., Anderson, R. L., Van Dorin, S. E., Rittner, L., Lotufo, R. A., Gerard, S. E., Reinhardt, J. M., |
| Comellas, A. P. (2024). Long COVID Iowa-UNICAMP [dataset]. University of Iowa. https://doi.org/10.25820/data.007301 |
|
|
| Citation for and links to publications that cite or use the data: |
|
|
| A paper describing this dataset and another that uses this dataset is under submission/review. This description will be updated to include its reference in the future. |
|
|
| ------------------------------------------------------- |
| ABSTRACT & PROJECT DESCRIPTION |
| ------------------------------------------------------ |
|
|
| The field of supervised automated medical imaging segmentation suffers from the lack of annotated groundtruth data. This problem is even more noticeable when dealing with the segmentation of multiple types of lung findings in computed tomography with uncertain borders, such as opacities and parenchymal consolidation resulting from pneumonia. In this work, we make available the first public dataset of ground glass opacity and consolidation in the lung of Long COVID patients. The Long COVID Iowa-UNICAMP dataset (LongCIU) was built by three independent expert annotators, blindly segmenting the same 90 selected axial slices manually, without using any automated initialization. We make available not only the final consensus segmentation, but also the individual segmentation from each annotator totaling 360 slices. This dataset can be used to train and validate new automated segmentation methods and to study interrater uncertainty in lung opacities' segmentation on computed tomography. |
|
|
|
|
| -------------------------------------------------- |
| METHODOLOGICAL INFORMATION |
| -------------------------------------------------- |
|
|
| CT scan data was originally collected at the University of Iowa Hospitals and Clinics, in an ongoing Long COVID study. Adults with history of COVID-19 infection confirmed by a positive antigen or reverse transcriptase-polymerase chain reaction that remained symptomatic 30 days or more following diagnosis were prospectively enrolled starting in June 2020. Patient data collected included age, sex, CT scan, days between acute COVID-19 diagnosis and CT scan, vaccination status, comorbidities, and other information. Chest CT scans were conducted with Siemens scanners. All scans employed tube current modulation. A standardized protocol for noncontrast chest CT imaging was followed, which entailed acquiring an inspiratory scan at total lung capacity (TLC) and an expiratory scan at residual volume (RV). Reconstruction of images was performed using iterative reconstruction techniques with a section thickness x interval of 1 x 0.5 mm. |
|
|
| The same selected CT slices were annotated by three pairs of trainee and expert annotator. A STAPPLE[2] final consensus is provided as the final annotation, but the individual results from each pair of annotators is also provided for uncertainty studies. More details are present on the accompanying paper (UNDER SUBMISSION). |
|
|
| This study was performed in line with the principles of the Declaration of Helsinki. Study protocols were approved by the institutional review board and were Health Insurance Portability and Protection Act–compliant. Participants were required to sign written informed consent before inclusion, including the possibility of sharing anonymized images. |
|
|
|
|
| ----------------------------------- |
| DATA & FILE OVERVIEW |
| ----------------------------------- |
|
|
|
|
| # Expected Directory tree |
|
|
| .\ |
| ├── longciu_1_tgt.nii.gz\ |
| ├── longciu_2_tgt.nii.gz\ |
| ├── longciu_3_tgt.nii.gz\ |
| ├── longciu_img.nii.gz\ |
| ├── longciu_medpseg_output.nii.gz\ |
| ├── longciu_splits.json\ |
| ├── longciu_STAPLE_tgt.nii.gz\ |
| └── staple_pub_stats.json |
|
|
| File tree for the dataset, including images, annotations per annotator and consensus, MEDPSeg output, optional proposed training splits, and STAPLE statistics. .nii.gz formats can be read using libraries such as SimpleITK or NiBabel, or visualization softwares such as 3D Slicer and ITKSnap. Following is the description of each file. |
|
|
| |
| ## longciu_img.nii.gz |
| Anonymized CT scan slices used for annotation, in original Hounsfield Unit intensities. |
|
|
| ## longciu_STAPLE_tgt.nii.gz |
| Final consensus segmentation for all slices, using STAPLE[2]. Masks are aligned with "img". |
|
|
| ## longciu_1_tgt.nii.gz, longciu_2_tgt.nii.gz, longciu_3_tgt.nii.gz |
| Results from blind annotations from annotator 1, 2, and 3 for all slices. Masks are aligned with "img". |
|
|
| ## longciu_medpseg_output.nii.gz |
| Output from the MEDPSeg[1] model, which was not trained in this dataset. Masks are aligned with "img". |
|
|
| ## longciu_splits.json |
| Optional, proposed data splits for supervised or semi-supervised machine learning applications. |
|
|
| ## staple_pub_stats.json |
| Output statistics from the SimpleITK STAPLE implementation, for reproducibility. These should match attempts at reproducing our provided code for STAPLE consensus creation. |
|
|
| ---------------------------------------------------------------------- |
| SOFTWARE, PREREQUISITES & DEPENDENCIES |
| ---------------------------------------------------------------------- |
|
|
| Users can either read the image and mask arrays (.nii.gz) with libraries such as SimpleITK ans NiBabel or visualize them with medical imaging visualization software such as 3D Slicer or ITK-Snap. |
|
|
| --------------------------------------------------------------------------------------------------- |
| STAPLE CREATION AND TECHNICAL VALIDATION REPRODUCIBILITY |
| --------------------------------------------------------------------------------------------------- |
|
|
| Code related to the creation and validation of this dataset is in: https://github.com/MICLab-Unicamp/LongCIU, with the version corresponding to this initial data release at the following link: https://github.com/MICLab-Unicamp/LongCIU/releases/tag/v1.0. Note that this code is not necessary for using the dataset. It is provided for dataset creation and technical validation reproducibility purposes. This code can be used to generate all the figures included in the submitted paper. |
|
|
| -------------------- |
| REFERENCES |
| -------------------- |
|
|
| [1] Carmo, Diedre S., et al. "MEDPSeg: Hierarchical polymorphic multitask learning for the segmentation of ground-glass opacities, consolidation, and pulmonary structures on computed tomography" arXiv preprint arXiv:2312.02365 (2024). |
|
|
| [2] Warfield, Simon K., Kelly H. Zou, and William M. Wells. "Simultaneous truth and performance level estimation (STAPLE): an algorithm for the validation of image segmentation." IEEE transactions on medical imaging 23.7 (2004): 903-921. |
|
|
