28MO - Dynamic evolution of vascular features based on magnetic resonance imaging to predict pathological response and survival outcomes in breast cancer patients undergoing neoadjuvant chemotherapy

Background

Magnetic resonance imaging (MRI) has solidified its status as an indispensable instrument for the assessment of therapeutic responses to NAC in breast cancer patients. We endeavor to elucidate the nexus between the temporal evolution of vascular features extracted from MRI and disparate clinicopathological parameters, aiming to formulate nomograms for the prediction of pathologic response after NAC and survival outcomes.

Methods

This is an exploratory analysis based on two prospective neoadjuvant clinical trials. Patients with locally advanced breast cancer underwent breast MRI three times, respectively at baseline, after 2 cycles of NAC, and after final NAC. Vessel-through-lesion (VTL) was quantified as the number of vessels passing the lesion identified from the MRI, with ΔVTL₁₂ and ΔVTL₁₃ denoting the relative change of VTL post the second and third MRI comparing to the first MRI.

Results

182 patients were enrolled and were randomly divided (7:3) into training and validation sets. The breast pathological complete response (bpCR)-predictive nomogram, incorporating ΔVTL₁₂ and clinicopathological factors, demonstrated superior predictive performance compared to a model solely predicated on clinicopathological factors (validation Area Under the Curve (AUC) of 0.892 vs. 0.857). The RFS-predictive nomogram leveraging ΔVTL₁₃ and clinicopathological factors unveiled a robust concordance with RFS (validation AUCs of 0.918 in 1 year, 0.767 in 3 years, and 0.717 in 5 years), eclipsing the model utilizing total pCR (tpCR) alone. Upon stratification into dichotomous risk categories (high-risk vs. low-risk) based on an optimal risk score threshold (100 points), the low-risk group manifested a superior RFS (HR 0.25, 95% CI 0.12-0.5, p < 0.001).

Conclusions

Fusing the dynamic evolution of vascular features with clinicopathological characteristics, we have crafted nomograms that furnish both predictive and prognostic insights. These instruments are poised to bolster the precocious identification of individuals amenable to NAC precisely, enabling the tailoring of surgical interventions and the selection of adjuvant therapies.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

The authors.

Funding

Has not received any funding.

Disclosure

All authors have declared no conflicts of interest.