Datasets:

andreasenz
/

TRUTH-LUNG

license: cc-by-4.0
task_categories:
  - image-classification
  - image-segmentation
  - image-to-3d
language:
  - en
tags:
  - lung
  - lung nodule
  - nodule
  - nodule segmentation
  - nodule detection
  - histological
  - histopatological
  - histological markers
  - biomarkers
pretty_name: truth-lung

CT Lung Nodule Dataset (Case–Control)

Overview

This dataset contains chest CT scans organized in a case–control setting for lung nodule analysis.

Case group: CT scans with lung nodules and corresponding segmentation masks.
Control group: CT scans without lung nodules (no segmentations available).

The dataset is intended for research purposes, including but not limited to:

lung nodule detection and classification,
segmentation,
case–control studies,
radiomics and machine learning experiments.

All data have been anonymized.

Dataset Structure

dataset_root/
│
├── case/
│ ├── <study_id_1>/
│ │ ├── volume.nii.gz
│ │ └── segmentation.nii.gz
│ ├── <study_id_2>/
│ │ ├── volume.nii.gz
│ │ └── segmentation.nii.gz
│ └── ...
│
├── control/
│ ├── <study_id_3>/
│ │ └── volume.nii.gz
│ ├── <study_id_4>/
│ │ └── volume.nii.gz
│ └── ...
│
├── labels.csv
├── pathology_features_1.csv
├── pathology_features_2.csv
└── README.md

Case and Control Folders

`case/`

Contains CT scans with lung nodules.

Each subfolder corresponds to a single imaging study, identified by a unique study_id.

Contents of each case/<study_id>/ folder:

volume.nii.gz
Preprocessed chest CT volume (NIfTI format).
segmentation.nii.gz
Binary segmentation mask of lung nodules, aligned with the CT volume.

`control/`

Contains CT scans without lung nodules.

Each subfolder corresponds to a single imaging study, identified by a unique study_id.

Contents of each control/<study_id>/ folder:

volume.nii.gz
Preprocessed chest CT volume (NIfTI format).

⚠️ Control cases do not include segmentation masks, as no nodules are present.

Study Identifier

Each subfolder name (<study_id>) uniquely identifies one imaging study.
The same study_id is consistently used across:
- folder names,
- labels.csv,
- pathology-related CSV files.

Labels File

`labels.csv`

This file provides the malignancy label for each study.

Format:

study_id	malignancy
XXXXXXX	1
YYYYYYY	0

Label definition:

1 → malignant nodule
0 → benign nodule

Control studies are not included in this file while the label is always malignancy = 0.

Pathology Data

`histo_dataset_wide.csv`

`histo_markers_long.csv`

These files contain anatomical pathology (histopathological) information associated with the studies in the case group.

Each row corresponds to a study_id.
Columns represent pathology-derived variables (e.g., histological findings, grading, molecular or morphological features).
The exact meaning of each column is described in the corresponding data dictionary or publication.

⚠️ Important note:
Control studies do not have pathology data, as no biopsy was performed due to the absence of lung nodules.

Data Modalities Summary

Group	CT Volume	Segmentation	Pathology	Malignancy Label
Case	✔️	✔️	✔️	✔️
Control	✔️	❌	❌	✔️ (benign)

File Formats

CT volumes and segmentations:
- NIfTI (.nii.gz)
Metadata and labels:
- CSV (.csv)

All volumes are stored in a consistent orientation and can be directly loaded using standard medical imaging libraries (e.g., SimpleITK, nibabel).

Notes and Limitations

Control cases did not undergo biopsy, as no lung nodules were detected.
Pathology data are therefore available only for case studies.
The dataset is intended for research use only and not for clinical decision-making.

Citation

If you use this dataset in your research, please cite the associated publication (to be added).

Contact

For questions, clarifications, or additional information, please contact the dataset authors.

License

This dataset is released under the Creative Commons Attribution 4.0 International (CC BY 4.0).

If you use this dataset, please cite:

[Paper / Author / Institution]

License details: https://creativecommons.org/licenses/by/4.0/