As the sentinel lymph node (SLN) is the primary site of metastasis from the primary cancer lesion, assessing its metastatic status helps in the clinical selection of the optimal treatment strategy to maximize the survival benefit for patients [1]. Lymph node metastases play a key role in patient treatment failure and shortened survival. A phase 3 clinical trial noted that SLN-negative patients who underwent SLN resection alone without axillary lymph node dissection (ALND) had similar overall survival and disease-free survival to those who underwent ALND [2]. Subsequently, Donker et al. [3] reached similar conclusions. From this, it can be assumed that not all SLN-positive individuals will benefit from ALND. Currently, there is still some controversy regarding the management of SLN in early-stage breast cancer patients with negative axillary lymph nodes, but pathology suggests the presence of one to two SLN-positive lesions, and patients may not receive ALND for those with postoperative adjuvant radiotherapy or other combination treatment options [4, 5]. Therefore, finding reliable protocols to screen for nonsentinel lymph node (NSLN) metastases at high risk from SLN-positive patients effectively prevents patients from receiving unnecessary ALND.

Vascular endothelial growth factor (VEGF) promotes the production of blood vessels and lymphatic vessels. It also has various functions, such as promoting epithelial cell division and regulating vascular permeability [6]. VEGF-C and -D both belong to the VEGF family, which is widely expressed in early embryos and various organs and tissues. In recent years, studies related to malignant tumours such as oesophageal cancer [7] and squamous lung cancer [8] have pointed to the association of VEGF-C and D with lymph node metastasis. However, their roles in lymph node generation and metastasis of breast cancer are not well defined. This study aimed to investigate the predictive value of VEGF-C and D combined with ultrasound and pathological features for NSLN metastasis to guide the choice of clinical ALND treatment.

The VEGF-C- and D-positive rates in breast cancer tissues were 78.24% and 61.76%, respectively, which were higher than the positive rates in paraneoplastic tissues (36.47% and 18.82%), indicating that VEGF-C and D in breast cancer lesions were abnormally highly expressed, while they were expressed at low levels in paraneoplastic tissues. The VEGF-C- and D-positive rates showed differences in the presence or absence of vascular infiltration, the number of positive SLNs $>$2 or not, and the percentage of positive SLNs $>$0.5, suggesting that the positive expression of both proteins was related to the above pathological features, which may be related to the proangiogenic function of the VEGF family [11]. Recent studies have found that VEGF-C and VEGF-D can induce lymphangiogenesis, and both can activate signalling pathways related to endothelial cell migration and tubulogenesis by binding to VEGF receptor 3 (VEGFR-3) in the endothelium [12, 13]. Lymphatic vessels are important conduits for distant metastasis of tumour cells, and VEGF-C and VEGF-D facilitate lymphatic vessel expansion.

In this study, the logistic analysis revealed that the risk of NSLN metastasis increased 2.332-fold in those with vascular infiltration compared to those without vascular infiltration, and the risk of NSLN metastasis increased 4.372-fold for each increase in the number of positive SLNs. SLN positive percentage $>$0.5 increased 5.353-fold compared to $\le$0.5, suggesting a strong association between vascular infiltration, the number of SNL positivity $>$2 and its proportion $>$0.5 and NSLN metastases. Independent risk factors for metastasis, consistent with the findings of several previous studies [14, 15, 16]. SLN is the first drainage area of primary tumour metastasis. Toberer et al. [17] found that VEGF and its receptors 2 and 3 were expressed more intensely in SLN-positive cutaneous melanoma patients than in SLN-negative patients, and VEGFR-3 was confirmed to be associated with SNL status. The burr sign is a typical imaging signs of malignant lesions. Previous study has shown that the marginal burr sign is more likely to be found in triple-negative breast cancer patients and can predict prognostic regression [18]. This study revealed that NSLN metastasis risk was increased 2.724-fold in patients with burr signs compared to those without, suggesting that burr signs increase the risk of NSLN metastasis.

In this study, the NSLN metastasis rate in VEGF-C- and D-positive patients was higher than that in VEGF-negative patients, and logistic analysis showed that VEGF-C- and D-positive patients increased 4.464- and 3.604-fold, respectively, compared with negative patients, indicating that both positive proteins were risk factors for NSLN metastasis. Li et al. [19] concluded that VEGF-C and its receptor 3 can promote lymph node metastasis in renal cell carcinoma. One study noted that VEGF-C and D transcript and protein expression levels were increased in different grades of endometrial cancer [20], suggesting that high expression of both increases the metastatic intensity of endometrial cancer. This result suggested that the possible mechanism by which VEGF-C and D promote lymphatic metastasis is that VEGF-C and D can bind to VEGFR-3 in lymphatic vessel endothelial cells and activate multiple signalling pathways, such as phosphatidylinositol 3-kinase regulatory subunit 1 (PIK3R1) [21] and wingless-type mouse mammary tumor virus (MMTV) integration site family (WNT5A) [22], to promote lymphatic vessel expansion, thus increasing lymph node metastasis risk. In addition, this study showed by ROC analysis that the AUC (95% CI) of VEGF-C and D combined with ultrasound and pathological features to predict NSLN metastasis was 0.859 (0.796–0.907), which was higher than the predictive efficacy of VEGF-C and D alone and their combined ultrasound or pathological features, respectively, indicating that these two proteins alone and in combination with ultrasound and pathological features help to improve the early identification of NSLN metastasis and reduce the risk of unnecessary surgery in people with a low risk of NSLN metastasis.

The analysis of this study as a single-centre retrospective study still has some limitations, the sample size that can be included in this time is limited, and the measurements of VEGF-C and VEGF-D have some differences between different testing institutes, which still needs to be improved in the future work, and can be explored by multicentre prospective large-sample data analysis implementation.

The ultrasound and pathological features of SLN-positive breast cancer, such as vascular infiltration, a high number of positive SLNs, an SLN positivity ratio $>$0.5, burr sign, and VEGF-C and D positivity, are independent risk factors for NSLN metastasis, and VEGF-C and D alone and in combination with the above ultrasound and pathological features have high predictive efficacy for NSLN metastasis. This study can be used as a reference for whether SLN-positive patients should receive ALND. It’s unclear whether the study’s findings are applicable to other types of cancer or patient groups. Further research may be needed to generalize the findings.

The simulation experiment data used to support the findings of this study are available from the corresponding author upon request.

All the authors have contributed to the document retrieval, conception and design of this study. Material preparation, data collection and analysis, and patient follow-up were conducted by JC, QL, WL, XT, ZW and LX. The first draft of the manuscript was written by JC and QL, and all authors commented on the first few versions of the manuscript. All authors contributed to editorial changes in the manuscript. All authors read and approved the final manuscript. All authors have participated sufficiently in the work to take public responsibility for appropriate portions of the content and agreed to be accountable for all aspects of the work in ensuring that questions related to its accuracy or integrity.

All subjects gave their informed consent for inclusion before they participated in the study. The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the Ethics Committee of The Third Hospital of Nanchang (approval number: 201526).

We would like to express our gratitude to all those who helped us during the writing of this manuscript. Thanks to Shegan Gao and Yiwen Liu for advice on the research concept and design, and Yiwen Liu for professional editing and language revision, and Yiwen Liu, Xiang Yuan, Jinyu Kong, Wei Sun, Yijun Qi, Hong Yang for development of methodology, and Shegan Gao, Fuyou Zhou, Kuisheng Chen, Haijun Yang for acquisition of data, and Jinyu Kong, Yiwen Liu, Wei Sun for analysis and interpretation of data.

This study was supported by the project of Jiangxi Provincial Health Commission (project No. 20164016).

The authors declare no conflict of interest.