CT-Based Model for Pulmonary Ground-Glass Nodule Classification

1,067 patient records from Shandong First Medical University Cancer Hospital were used to develop a model for classifying pulmonary ground-glass nodules. The integrated model, combining clinical features, radiomics, and deep learning, achieved a validation AUC of 0.871 for malignancy classification. This research dataset, authored by Jian Zhang and shared under CC-BY-4.0, is documented in a 36.7 KB file.

Use Cases

• Training machine learning classifiers to differentiate benign from malignant pulmonary nodules based on clinical and imaging features.
• Benchmarking integrated models that combine clinical variables, radiomics, and deep learning for medical image analysis.
• Developing non-invasive decision support tools for pathological subtype classification of ground-glass nodules.

Strengths

• Dataset is based on a study of 1,067 patients, providing a substantial sample size for model development.
• The described integrated model achieved a validation AUC of 0.871, indicating strong diagnostic performance for the primary task.
• The research methodology is detailed, including the use of multiple classifiers like SVM, Random Forest, and XGBoost for comparison.

Limitations

• The underlying data file is a 36.7 KB DOCX document, suggesting the dataset itself is likely a summary or results table, not the primary imaging or patient data.
• Row count and column-level documentation for any potential tabular data are unknown, limiting suitability assessment.
• Data reflects a single-center, retrospective study from one hospital, which may introduce geographic or institutional bias.

Provenance

Source

Shandong First Medical University Cancer Hospital

Collection Method

Retrospective study collecting clinical and imaging data from patient records.

Freshness

Last updated 2026-05-26 10:15:53; freshness should be verified.

Geography

China (inferred from hospital location)