Preoperative MRI features predict failed breast-conserving surgery: construction of a predictive model

Area of Science:

• Oncology
• Radiology
• Surgical Oncology

Background:

• Breast-conserving surgery (BCS) is the standard for early breast cancer.
• Accurate preoperative prediction of BCS feasibility optimizes surgical plans and patient outcomes.
• Magnetic resonance imaging (MRI) can identify features associated with BCS failure.

Purpose of the Study:

• To explore preoperative MRI features linked to failed BCS.
• To develop and validate an MRI-based predictive model for BCS outcomes.

Main Methods:

• Retrospective study of 120 patients (30 failed BCS, 90 successful BCS) with preoperative MRI.
• Comparison of MRI features (lesion type, volume, multifocality) between groups.
• Development and validation of a multivariate logistic regression model using ROC curve analysis.

Main Results:

• Failed BCS group showed more non-mass lesions, multifocality, larger lesion volume, and higher lesion-to-breast volume ratio.
• Lesion-to-breast volume ratio (OR=1.044) and multifocality (OR=11.161) were independent predictors of failed BCS.
• The MRI-based model achieved AUCs of 0.902 (training) and 0.821 (testing).

Conclusions:

• An MRI-based model can effectively predict failed breast-conserving surgery preoperatively.
• This predictive model aids clinicians in formulating accurate surgical strategies.
• Improved preoperative prediction enhances patient care and surgical decision-making for breast cancer.