3D Intratumoral Heterogeneity Scores for Predicting Lung Cancer Pleural Invasion

A multicenter study of 1,301 patients developed a stacking ensemble model for preoperative prediction of visceral pleural invasion in lung adenocarcinoma. The model integrates 3D intratumoral heterogeneity scores from chest CT scans with clinicoradiologic features, achieving an AUC of 0.878. The dataset, authored by Qunzhi Ouyang and last updated in May 2026, contains the quantitative scores and features used in this research.

Use Cases

• Training machine learning models for preoperative risk stratification based on radiomic and clinical features.
• Benchmarking new predictive models against the reported stacking ensemble classifier with an AUC of 0.878.
• Analyzing feature importance for lung cancer prognosis using the SHAP values mentioned in the description.
• Developing adjunct tools for individualized treatment planning in lung adenocarcinoma.

Strengths

• Data originates from a multicenter study involving 1,301 patients, which likely improves generalizability.
• Model performance is quantitatively reported with an AUC of 0.878 for the final ensemble classifier.
• The study includes a fixed external validation cohort from a third medical center.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The 17.5 KB file size suggests a very limited scope, likely containing summary scores or model parameters rather than raw patient data.

Provenance

Source

figshare

Collection Method

Retrospective analysis of patient data from three medical centers.

Freshness

Last updated 2026-05-12 04:19:51; freshness should be verified.