AI-Assisted Clinico-Quantitative Imaging Nomogram for Lung Nodule Malignancy Risk

951 patient records from a retrospective study of surgically resected pulmonary nodules ≤ 3 cm. The dataset was used to develop and internally validate a multivariable logistic regression model combining AI-extracted quantitative CT features with clinical variables. The model, created by Yingding Ruan and last updated in April 2026, achieved an AUC of 0.836 for preoperative malignancy risk assessment.

Use Cases

• Training machine learning models for malignancy risk prediction based on quantitative CT features like nodule size and attenuation.
• Benchmarking clinical prediction models against the reported AUC of 0.836.
• Developing interactive risk calculators for preoperative surgical decision support.
• Studying the integration of AI-extracted imaging features with clinical variables in a logistic regression framework.

Strengths

• Model performance is quantitatively reported with an AUC of 0.836 and a mean absolute error of 0.015.
• Based on data from 951 consecutive patients, providing a substantial sample for model development.
• Includes an interactive calculator for public access, facilitating practical application.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• The cohort is restricted to 951 surgically resected nodules, which may introduce selection bias and limit generalizability to screening populations.
• Row count for the underlying data is unknown, which may limit suitability assessment.

Provenance

Source

Retrospective study data analyzed by Yingding Ruan, shared on figshare.

Collection Method

Data from 951 consecutive patients who underwent surgical resection, with quantitative CT features automatically extracted using InferRead CT Lung AI software.

Time Range

null

Freshness

Last updated 2026-04-13 05:31:07.

Geography

null