Prediction of lymphovascular invasion in breast cancer with non-mass enhancement by DCE-MRI and clinicopathological features: a retrospective case-control study

Background: Lymphovascular invasion (LVI) is one of the significant prognostic risk factors for breast cancer patients. However, research on factors affecting LVI in non-mass enhancement (NME) breast cancer remains limited. This study aims to investigate the relationship between Dynamic Contrast-Enhanced Magnetic Resonance Imaging (DCE-MRI) features, clinicopathological characteristics, and LVI in NME-type invasive breast cancer. Methods: A retrospective analysis was conducted on 192 adult female patients with non-mass enhancement (NME) who underwent breast DCE-MRI at our hospital from January 2019 to December 2023, including 192 invasive carcinoma lesions. Based on routine postoperative pathology results, patients were divided into the LVI-positive [LVI(+)] group (N=50) and the LVI-negative [LVI(-)] group (N=142). Multivariate logistic regression was used to analyze risk factors for LVI formation. The performance of the combined risk factor score (Distribution + ADC value + Molecular subtype) was evaluated using receiver operating characteristic (ROC) curves and areas under the curve (AUC). The net reclassification improvement (NRI) and integrated discrimination improvement (IDI) were employed to assess the predictive ability of the NME distribution pattern when added to LVI prediction. Findings: Age ≥45 years (OR=0.406, 95% CI: 0.191-0.844, P=0.017) and higher ADC values (OR=0.133, 95% CI: 0.017-0.825, P=0.041) were identified as independent protective factors against LVI formation. Among the NME distribution patterns, patients with linear distribution had a higher risk of LVI positivity compared to those with focal distribution (OR=13.540, 95% CI: 1.390-172.644, P=0.030). Regarding molecular subtypes, patients with Luminal B (OR=5.081, 95% CI: 1.326-33.859, P=0.039), HER-2 overexpression (OR=9.378, 95% CI: 1.922-71.460, P=0.012), and triple-negative subtypes (OR=11.599, 95% CI: 2.043-96.710, P=0.010) were at a higher risk of LVI formation compared to those with the Luminal A subtype. The combined risk factors provided a good prediction of LVI status, with an AUC of 0.723 (95% CI: 0.645-0.801). Additionally, the inclusion of distribution patterns significantly improved the predictive ability of the combined risk factors for LVI, as indicated by the NRI (0.389; 95% CI: 0.083-0.696) and IDI (0.047; 95% CI: 0.012-0.083). Interpretation: The combined risk factors (Distribution + ADC value + Molecular subtype) provided a good prediction of LVI status, with an AUC of 0.723. The inclusion of distribution patterns significantly enhanced the predictive ability of the combined risk factors for LVI. However, since our study is a single-center retrospective analysis, further studies with larger sample sizes are needed to validate our conclusions. Funding: None.