Ultrasound Radiomics Features for Elderly Pancreatic Cancer Prognosis After HIFU Therapy

Additional file 2 of a study by Yumei Liu, last updated in April 2026, presents 1,561 radiomics features extracted from ultrasound region-of-interest (ROI) images. The dataset is a 7.7 MB CSV file shared under a CC-BY-4.0 license on figshare. It was developed to support the validation of a deep learning model for predicting prognosis in elderly patients with advanced pancreatic cancer following High-Intensity Focused Ultrasound (HIFU) therapy.

Use Cases

• Training machine learning models to predict patient prognosis based on 1,561 extracted radiomics features.
• Validating deep learning radiomics models for pancreatic cancer using ultrasound-derived features.
• Conducting feature importance analysis to identify key ultrasound radiomics signatures for cancer prognosis.
• Benchmarking new prognostic models against the features used in the original study.

Strengths

• Contains 1,561 distinct radiomics features per patient, providing a high-dimensional feature space for analysis.
• Dataset is openly shared under a permissive CC-BY-4.0 license.
• File size of 7.7 MB indicates a manageable download and processing footprint.

Limitations

• Row count is unknown, which may limit suitability assessment for model training.
• Column-level documentation is absent; field semantics must be inferred after download.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

Yumei Liu via figshare

Collection Method

Features were extracted from ultrasound region-of-interest (ROI) images of patients.

Freshness

Last updated 2026-04-11 03:38:36; freshness should be verified.