A Multimodal Diagnostic Model for Breast Cancer Invasiveness Based on Ultrasound Imaging and Serum Biomarkers

Area of Science:

• Oncology
• Medical Imaging
• Biomarkers

Background:

• Breast cancer invasiveness is critical for treatment and prognosis.
• Current diagnostic methods can be improved by integrating imaging and biomarkers.
• Accurate assessment of invasiveness is needed for optimal patient management.

Purpose of the Study:

• To develop and validate a multimodal diagnostic model for breast cancer invasiveness.
• The model integrates ultrasound B-mode, Doppler imaging, and serum biomarkers.
• To enhance the accuracy of diagnosing invasive versus non-invasive breast cancer.

Main Methods:

• Developed a multimodal model using ultrasound data and serum biomarkers (CA125, CA15-3, CEA, CA19-9).
• Applied machine learning algorithms (Logistic Regression, Random Forest, XGBoost) for invasiveness prediction.
• Evaluated model performance using accuracy, precision, recall, F1-score, and AUC.

Main Results:

• The multimodal model significantly outperformed single-modality approaches.
• XGBoost algorithm achieved the highest accuracy (88.90%) and AUC (0.930).
• Inclusion of specific serum biomarkers notably improved diagnostic accuracy for invasiveness.

Conclusions:

• A multimodal diagnostic model integrating ultrasound and serum biomarkers is highly accurate for assessing breast cancer invasiveness.
• This model shows potential for improving clinical decision-making.
• Enhanced diagnostic accuracy can lead to better patient outcomes in breast cancer care.