Novel clinical radiomic nomogram method for differentiating malignant from non-malignant pleural effusions

Area of Science:

• Radiology
• Medical Imaging
• Oncology

Background:

• Pleural effusions are common clinical findings.
• Distinguishing malignant pleural effusion (MPE) from non-MPE (NMPE) is crucial for patient management.
• Accurate differentiation can be challenging using conventional methods.

Purpose of the Study:

• To develop and validate a clinical radiomics nomogram for differentiating malignant and non-malignant pleural effusions.
• To assess the diagnostic performance of a combined clinical and radiomics model.

Main Methods:

• Retrospective analysis of 146 MPE and 93 NMPE patients.
• Extraction of radiomics features from CT images of chest lesions.
• Development of a nomogram integrating selected radiomics features and clinical factors.
• Evaluation using receiver operating characteristic (ROC) curve and decision curve analysis (DCA).

Main Results:

• A nomogram was constructed using 19 radiomics features and 2 clinical factors.
• The combined model achieved high diagnostic accuracy with an AUC of 0.968 (training) and 0.873 (validation).
• The combined model significantly outperformed clinical and radiomics-only models in both cohorts.

Conclusions:

• CT-based radiomics offers superior diagnostic accuracy for MPE compared to clinical and radiological features alone.
• The combined radiomics nomogram demonstrates excellent diagnostic performance and clinical utility.
• The developed nomogram supports the clinical application of chest CT radiomics in diagnosing MPE.