LungSimNet4D: Synthetic Thoracic Imaging Dataset for NSCLC Research

A 2.8 GB reproducibility package from 2026 includes synthetic thoracic phantom variants, CT-like imaging volumes, lesion masks, and trained model weights. It was authored by Poornima G and released on figshare under an MIT license. The dataset supports research in pulmonary nodule segmentation, classification, and digital twin workflows.

Use Cases

• Training segmentation models based on synthetic CT-like voxelized imaging volumes and lesion masks.
• Validating Monte Carlo simulations based on GATE-compatible voxel phantoms and validation outputs.
• Benchmarking classification algorithms based on dataset manifests and train/validation/test splits.
• Developing digital twin research workflows based on anatomy-aware synthetic thoracic phantom variants.

Strengths

• Includes a 2.8 GB package with multiple artifact types, such as voxelized volumes, phantoms, and model weights.
• Provides a reproducibility package with training scripts, configuration files, and environment specifications.
• Released under the permissive MIT license, facilitating open use and modification.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Freshness should be verified as the last update was 2026-06-03 17:45:46.

Provenance

Source

figshare

Collection Method

Generated by the LungSimNet4D framework using anatomy-aware mesh transformations and procedural lesion placement.

Freshness

Last updated 2026-06-03 17:45:46.