Data Sheet 1_Integrating MRI habitat heterogeneity and peritumoral radiomics into a nomogr

A multicenter retrospective cohort of 290 patients with locally advanced rectal cancer (LARC) was analyzed by Liang Zhang, with data last updated in April 2026. The dataset likely contains features derived from high-resolution MRI scans, including intratumoral and peritumoral radiomic features, habitat heterogeneity metrics, and clinical variables used to build a predictive nomogram. It was created to improve risk stratification and personalize treatment decisions for LARC patients.

Use Cases

• Training machine learning classifiers for cancer risk prediction based on MRI-derived radiomic features.
• Developing integrated nomograms for clinical decision support based on tumor microenvironment heterogeneity.
• Comparing the performance of logistic regression and support vector machine models for medical imaging data.
• Analyzing the relationship between peritumoral region margins (1mm, 2mm, 3mm) and predictive performance.

Strengths

• Multicenter study design involving 290 patients, with a defined training set (n=178) and external test set (n=112).
• Model performance metrics are provided, including an AUC of 0.928 in the training cohort and 0.817 in the test cohort.
• The dataset is shared under a permissive CC-BY-4.0 license, facilitating reuse and modification.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The dataset is very small at 15.1 KB, indicating limited scope.

Provenance

Source

figshare

Collection Method

Retrospective analysis of MRI scans from a multicenter cohort; radiomic features were extracted and filtered using LASSO regression.

Freshness

Last updated 2026-04-21 04:21:57; freshness should be verified.