PanTrack: Longitudinal CT Benchmark for Pancreatic Cancer Lesion Tracking

PanTrack is a longitudinal CT benchmark dataset for tumor lesion tracking and segmentation. It was introduced alongside a MICCAI 2026 paper and was used exclusively as a held-out out-of-distribution test set. The dataset was authored by mrokuss and last updated on Hugging Face in May 2026.

Use Cases

• Benchmarking lesion tracking algorithms based on longitudinal CT scans
• Evaluating segmentation model performance on out-of-distribution medical imaging data
• Developing interactive tools for clinician-verified lesion tracking mentioned in the description

Strengths

• Designed as a held-out out-of-distribution test set, ensuring unbiased evaluation.
• Associated with a peer-reviewed MICCAI 2026 paper, indicating academic validation.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and file formats are unknown, which may limit suitability assessment.

Provenance

Source

Hugging Face dataset authored by mrokuss.

Collection Method

Introduced as a benchmark alongside a research paper; collection method not specified.

Freshness

Last updated 2026-05-21 07:48:32; freshness should be verified.