Differentiation of testicular seminomas from nonseminomas based on multiphase CT radiomics combined with machine...

Area of Science:

• Oncology
• Radiology
• Medical Imaging

Background:

• Accurate differentiation between seminomas and nonseminomas is critical for optimal treatment of testicular germ cell tumors (TGCTs).
• Current diagnostic methods may require invasive procedures.
• There is a need for noninvasive tools to aid in TGCT subtyping.

Purpose of the Study:

• To develop and validate a clinical-radiomics model for differentiating seminomas from nonseminomas in TGCT patients.
• To assess the diagnostic performance of the combined model compared to clinical and radiomics-only models.

Main Methods:

• A total of 221 TGCT patients were retrospectively analyzed across four hospitals.
• Radiomics features were extracted from CT images and combined with clinical data.
• Machine learning algorithms were employed to build and evaluate clinical, radiomics, and clinical-radiomics models.
• Model performance was assessed using receiver operating characteristic (ROC) curve analysis and decision curve analysis (DCA).

Main Results:

• The clinical-radiomics model exhibited superior discriminatory ability across training, validation, and test cohorts (AUCs: 0.918, 0.909, 0.839, respectively).
• The combined model outperformed both the clinical-only and radiomics-only models.
• DCA confirmed the clinical-radiomics model's greater net benefit in predicting tumor subtypes.

Conclusions:

• The developed clinical-radiomics model shows promise as a noninvasive tool for distinguishing between testicular seminomas and nonseminomas.
• This model can provide valuable guidance for clinical treatment strategies in TGCT management.