br Discussion br We proposed a novel nomogram
We proposed a novel nomogram based on clinical, pathologic and radiological features to predict VI in breast cancer. The nomogram in-corporated 10 most strong predictors including breast density, tumor size (maximum diameter), location, margins, lobulation sign, ALNM, Stearamide enhancement patterns, DWI appearance, TIC patterns, and pathological type. Our results showed the nomogram had excellent discrimination in both the training dataset and the validation dataset. We developed an easy-to-use, repeated and relatively objective nomo-gram that facilitated the individualized prediction of VI.
For many years, patient’s age, tumor size, axillar lymph node status, histological grade of malignancy, ER, PR and HER2 represented prin-cipal factors used for the stratification of breast cancer patients for the
Fig. 2. The nomogram was developed in the training dataset, with the breast density, tumor size (maximum diameter), location, margins, lobulation sign, ALNM, contrast enhancement patterns, DWI appearance, TIC patterns, and pathological type incorporated.
purposes of evaluating the prognosis and determining the appropriate strategy of treatment [22–25]. Lymphovascular invasion (LVI) also has been reported as a strong prognostic factor in patients with breast cancer. Tumor cell invasion of blood vessels or lymphatic vessels is the critical step of tumor cell dissemination and metastasis for predicting disease recurrence or progression. However, most previous studies did not separate LVI into VI and lymphatic invasion (LI) [26–29]. Fujii T et al found the presence of VI, instead of LI, could be an indicator of high biological aggressiveness and may be a valid prognostic factor for breast cancer . Compared to LI, VI may better represent systemic disease due to the strong association between VI and LI . Therefore, to predict systemic disease, it would be helpful to identify the subset of patients with VI among breast cancer patients with or without LI. Pa-tients with VI may require for stronger adjuvant therapies because of
the high risk of local, regional, and distant recurrence. However, to our best of knowledge, there was no any study predicted VI in patients with breast cancer before treatment based on available data.
Some studies had showed VI was associated with lower mean age, body weight, breast density, tumor size, positive lymph nodes, and histologic grade [30–32]. Our study found the risk for VI increased with the tumor size. The mean tumor size in patients with VI was larger than that in patients without VI. Some studies have examined the association between breast density and breast cancer characteristics and found breast density was positively associated with tumor size, lymph node status, and VI among women with screen-detected cancers [31,33,34]. We also identified breast density as a predictor of VI, those patients with fatty breast density had higher risk for VI. We observed breast tumor with VI usually had morphological characteristics such as ill-
Fig. 3. The ROC curves of the nomogram in the training dataset and validation dataset.
defined margins, lobulation sign, and spiculation sign. In this study, VI was more frequently found in the lower-inner quadrant than other quadrants, which was inconsistent with a previous study. It said may because breast cancer was most likely to occur in the upper-outer quadrant, which was the quadrant with highest breast area and dense area . Analyses from larger trials are warranted to identify this finding. A review has shown that the presence of VI correlates closely with locoregional lymph node involvement . Our study observed VI was associated with ALNM. Multiple ALNM was more common in breast cancer with VI than those without VI (56% vs. 17%, p < 0.001). We also noticed pathological type was associated with breast cancer VI. Breast tumors with different pathological types have different biolo-gical behavior and therefore have different possibilities of invasion of surrounding tissues. The results showed grade 3 IDC had the highest risk for VI.