CT radiomics based model for differentiating malignant and benign small (≤20mm) solid pulmonary nodules.

Area of Science:

• Radiology
• Medical Imaging
• Oncology

Background:

• Limited models exist for evaluating small (≤ 20 mm) solid pulmonary nodules (SPNs) using computed tomography (CT) radiomics.
• Accurate differentiation of benign and malignant SPNs is crucial for patient management.

Purpose of the Study:

• To develop and validate a CT radiomics-based model for distinguishing benign from malignant small SPNs.

Main Methods:

• A cohort of 210 patients with small SPNs was randomly divided into training (168) and testing (42) groups.
• CT images were analyzed to extract radiomics features.
• A radiomics scoring model was developed using logistic regression and validated in the testing cohort.

Main Results:

• A model incorporating 9 CT radiomics features, lobulation, spiculation, and diameter demonstrated high diagnostic accuracy.
• The CT radiomics-based model achieved an area under the curve (AUC) of 0.957 (training) and 0.943 (testing).
• Significant predictors included lobulation (P < 0.001), spiculation (P < 0.001), and diameter (P < 0.001).

Conclusions:

• The developed CT radiomics-based model provides valuable diagnostic accuracy for differentiating benign and malignant small SPNs.
• This model can potentially improve the diagnostic process for small pulmonary nodules.