Longitudinal contrast-enhanced cone beam breast computed tomography for monitoring treatment response in breast cancer neoadjuvant therapy: a comprehensive assessment.

OBJECTIVES

To validate cone-beam breast computed tomography (CBBCT) for evaluating NAT response using Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1, and to explore correlations between multidimensional CBBCT features and pathological complete response (pCR).

METHODS

We prospectively analyzed 73 patients who underwent pre- and post-NAT CBBCT, of whom 41 also received paired magnetic resonance imaging (MRI). Inter-modality and imaging-pathology concordance were assessed using Cohen's κ statistics. Determinants of CBBCT accuracy were identified through univariate and logistic regression analyses. Receiver operating characteristic (ROC) analysis with DeLong's test was used to determine the optimal quantitative parameter for discriminating pCR from non-pCR. Associations between pCR and changes in calcification, vascularity, and background parenchymal enhancement (BPE) were evaluated using Pearson's χ2 and Fisher's exact tests.

RESULTS

CBBCT showed excellent agreement with MRI (κ = 0.809) and good pathological concordance (κ = 0.618; MRI κ = 0.700). The area under the curve (AUC) for distinguishing responders from non-responders was 0.950 for CBBCT and 0.905 for MRI. For pCR assessment, the AUC values were 0.643 and 0.786, respectively. Targeted therapy (odds ratio [OR] = 0.059, p = 0.013) and linear/segmental calcifications (OR = 0.127, p = 0.042) were independent factors affecting CBBCT accuracy. The enhanced degree change (ΔE) on CBBCT demonstrated strong predictive performance for pCR (AUC = 0.941) with high computational efficiency. Significant associations with pCR were observed for reductions in calcification extent (p = 0.021), adjacent vascular sign (AVS) grade (p = 0.004), and BPE levels (p = 0.001).

CONCLUSION

CBBCT demonstrates high accuracy in assessing NAT response, with excellent agreement to MRI. ΔE is recommended as an optimal quantitative parameter for predicting pCR, supported by dynamic changes in calcification extent, AVS, and BPE grade as valuable markers, positioning CBBCT as a comprehensive tool for breast cancer management.

CLINICAL RELEVANCE STATEMENT: CBBCT enables a comprehensive and efficient tool for monitoring NAT response, showing excellent concordance with MRI, and its multidimensional features are also valuable for predicting pCR.