ChirurgixAI/Chirurgix_CT_Dataset

Chirurgix CT Dataset

Dataset Overview

The Chirurgix CT Dataset is a collection of 270 patient cases containing medical CT scans and corresponding segmentation labels. Each case includes original CT scan images (.nii.gz format) and segmentation masks (.nii.gz files) for various anatomical structures and pathologies. The dataset is accompanied by a metadata file (meta_raw.csv) that provides detailed patient information, scan types, pathology descriptions, and more.

Dataset Structure

The dataset is organized as follows:

Chirurgix_dataset_270
├── meta_raw.csv
├── patient_10201
│   ├── ct_upper_leg_left.nii.gz
│   └── seg_tags
│       ├── femur_left.nii.gz
│       ├── gluteus_maximus_right.nii.gz
│       ├── hip_left.nii.gz
│       ├── hip_right.nii.gz
│       └── scapula_right.nii.gz
├── patient_10372
│   ├── ct_angiography_abdomen-pelvis-leg.nii.gz
│   └── seg_tags
│       ├── adrenal_gland_left.nii.gz
│       ├── adrenal_gland_right.nii.gz
│       ├── aorta.nii.gz
│       ├── atrial_appendage_left.nii.gz
│       ├── autochthon_left.nii.gz
│       ├── autochthon_right.nii.gz
│       ├── brachiocephalic_trunk.nii.gz
│       ├── brachiocephalic_vein_left.nii.gz
│       ├── brachiocephalic_vein_right.nii.gz
│       ├── clavicula_left.nii.gz
│       ├── clavicula_right.nii.gz
│       ├── colon.nii.gz
│       ├── common_carotid_artery_left.nii.gz
│       ├── common_carotid_artery_right.nii.gz
│       ├── costal_cartilages.nii.gz
│       ├── duodenum.nii.gz
│       ├── esophagus.nii.gz
│       ├── femur_left.nii.gz
│       ├── femur_right.nii.gz
│       ├── gallbladder.nii.gz
│       ├── gluteus_maximus_left.nii.gz
│       ├── gluteus_maximus_right.nii.gz
│       ├── gluteus_medius_left.nii.gz
│       ├── gluteus_medius_right.nii.gz
│       ├── gluteus_minimus_left.nii.gz
│       ├── gluteus_minimus_right.nii.gz
│       ├── heart.nii.gz
│       ├── hip_left.nii.gz
│       ├── hip_right.nii.gz
...

File Descriptions

• meta_raw.csv: Metadata file containing detailed patient information with the following fields:

  • index: Unique index number.
  • image_id: Patient ID.
  • age: Patient age.
  • gender: Patient gender.
  • study_type: Type of CT scan (e.g., ct abdomen-pelvis, ct angiography abdomen-pelvis-leg).
  • kvp: Scan voltage (kVp).
  • pathology: Pathology description (e.g., no_pathology, vascular, tumor).
  • pathology_location: Location of pathology (e.g., abdomen, thorax, throat).
  • date: Scan date.
  • notes: Additional notes (e.g., Routine checkup, Scheduled follow-up, Emergency case).

• patient_xxxxx: Directory for each patient, containing:

  • ct_xxxxx.nii.gz: Original CT scan in .nii.gz format, where xxxxx describes the scan region (e.g., upper_leg_left, angiography_abdomen-pelvis-leg).
  • seg_tags: Directory containing segmentation masks for anatomical structures (e.g., femur_left, aorta, heart).

Dataset Details

• Total Samples: 270 patients.
• Samples per Category: 10 samples per category (e.g., specific anatomical structure or pathology type).
• Data Format:
  • CT Images: .nii.gz (NIfTI format, compressed).
  • Segmentation Labels: .nii.gz (NIfTI format, compressed).
  • Metadata: .csv.
• Data Collection Period: February 2024 to January 2025.

Intended Use

This dataset is designed for:

• Medical imaging analysis.
• Anatomical structure segmentation.
• Pathology detection and diagnosis.
• Training and evaluation of machine learning and deep learning models.

Usage Example

1. Loading CT Images:

   import nibabel as nib
   ct_image = nib.load('path/to/ct_upper_leg_left.nii.gz').get_fdata()
   

2. Loading Metadata:

   import pandas as pd
   metadata = pd.read_csv('path/to/meta_raw.csv')
   

3. Dataset Splitting: Use the study_type or pathology fields to categorize or split the dataset.

License

**Usage Restrictions**: For academic and non-commercial use only. Commercial use is prohibited.

Citation

If you use this dataset in your research, please cite it as follows:

Chirurgix CT Dataset 270, ChirurgiAI, 2025, https://www.chirurgixai.pro/

Contact

About Chirurgix: At Chirurgix, we merge AI-powered diagnostics, 3D-printed surgical solutions, and robotic-assisted therapies to revolutionize cancer care. Our mission is to enhance treatment accuracy, personalize patient outcomes, and push the boundaries of medical innovation. Our Innovations:

• Advanced algorithms analyze MR & CT scans for precision diagnostics.
• Custom anatomical models optimize surgical accuracy and treatment planning.
• AI-powered robotics enhance precision in minimally invasive cancer therapies.

For questions or feedback, please contact:

• Email: info@chirurgixai.pro
• Github: https://github.com/ChirurgixAI/Chirurgix_AnatoGen
• Website: https://www.chirurgixai.pro
• Hugging Face Dataset Page: https://huggingface.co/datasets/ChirurgixAI/Chirurgix_CT_Dataset