1. Introduction
Biomarkers are measurable biological parameters that distinguish between normal or pathological conditions [
1]. Cancer biomarkers comprise tumor cell features detected in tissue biopsies or body fluids that can be used to stratify prognosis and predict the benefit or toxicity of specific therapeutic interventions [
2].
In the era of precision oncology, biomarker research and implementation in routine clinical care contribute to the enhancement of tailored therapeutic approaches [
3]. Predictive and prognostic biomarkers are essential to define and optimize the therapeutic strategy for each patient, aiming for a maximal therapeutic response with minimal toxicity. A prognostic factor can be defined as a feature associated with clinical outcomes in the absence of therapy [
4]. The World Health Organization (WHO) classification of tumors focuses on pathological aspects with histological plus molecular categorizations, integrating established and investigational biomarkers [
5].
Tumor budding (TB) is defined as the presence of isolated single cancer cells or cell clusters of up to four cancer cells at the invasive tumor front and has emerged as a promising prognostic biomarker predicting disease progression across several tumor types [
6,
7,
8]. The association between tumor budding, disease progression, and worst outcomes in different solid tumors was first described in 1950 [
9].
Tumor buds are part of the tumor microenvironment (TME), being most prominent at the invasive front and associated with epithelial–mesenchymal transition (EMT) [
10]. EMT is characterized by cytoskeletal rearrangements, cell motility, increased cell-associated proteolytic activity, and invasion [
11]. TB is a dynamic process of tumor cell dissociation that shares some of these biological characteristics, with in vitro evidence showing a prominent role of E-cadherin as a regulator of EMT.
The role of TB as a prognostic cancer biomarker was first validated for colorectal cancer after the standardization of tumor budding assessment by the International Tumor Budding Consensus Conference (ITBCC) grading recommendations in 2016 [
12]. TB is currently included in international guidelines for colorectal cancer, influencing therapeutic decisions for patients with pT1 and stage II colon cancer [
13,
14,
15]. For patients with pT1c, intermediate-to-high-grade tumor budding correlates with node involvement, and it is used to guide complementary surgery following endoscopic resection. In stage II patients, high-grade tumor budding is a poor prognostic factor considered for adjuvant systemic therapy guidance [
16].
Emerging research suggests that TB has prognostic value for other tumor types, including breast, lung, head and neck, esophageal, gastric, and urogenital cancers. However, the lack of a validated disease-specific scoring system does not allow for wider adoption of tumor budding into the classifications of other cancers.
Despite evidence showing that tumor budding could have a prognostic impact on early-stage breast cancer, there is a lack of clinical evidence demonstrating the clinical applicability of tumor budding analysis and its utility as a prognostic factor to guide therapeutic decisions [
17]. This study aims to analyze the prognostic relevance of tumor budding determination for patients with early-stage ductal invasive breast carcinoma.
4. Discussion
Pathological changes associated with tumor budding represent the metastatic process in the initial stage, where cells acquire metastatic potential by detaching from the main tumor [
20]. The metastatic process relies on features collectively known as epithelial–mesenchymal plasticity that include two major processes: epithelial-to-mesenchymal transition and mesenchymal-to-epithelial transition [
21]. In the initial steps of the metastatic process, cells gain mesenchymal features that allow for cell motility plus invasion, and as tumor cells adhere to the metastatic site, they regain epithelial features. The EMT and MET empower the plasticity of stem cells, allowing tumor cells to alternate between epithelial and mesenchymal states and facilitating metastatic spread [
22].
Tumor buds are part of the TME and are associated with EMT, providing a known portrayal of the initial steps of the metastatic cascade [
23]. Histological and molecular subtypings of breast cancer have shown that EMT impacts prognosis, with basal-like breast cancers demonstrating a higher metastatic spread based on sustained mesenchymal features [
24]. Transcription factors, including Snail1/Snail, Snail2/Slug, Twist, and the ZEB family of transcription factors, are involved in the EMT process. The upregulation of ZEB is linked with invasive ductal breast cancer de-differentiation, increased vimentin plus N-cadherin expression, and downregulation of E-cadherin [
25]. Having a tumor suppressor role in breast cancer, partial or total loss of E-cadherin expression correlates with invasiveness, increased tumor grade, and poor prognosis [
26].
E-cadherin expression differs between breast cancer subtypes, with lobular subtypes showing a high loss of E-cadherin expression and invasive ductal carcinomas having a low frequency of E-cadherin expression loss [
27]. The level of E-cadherin expression influences the invasion pattern, with lobular carcinomas having a single-cell infiltration pattern and invasive ductal carcinomas presenting with solid layers or ductal structures. These patterns justify the reason for the reported tumor budding assessments in the literature being performed exclusively on invasive ductal carcinomas. The solid layer or ductal invasion pattern allows for an easy morphological assessment of TB in H&E slides without the need for additional immunohistochemical characterization [
28,
29].
Evidence implies the role of TB as a prognostic factor for several solid cancers [
23], being validated as an independent risk factor for colorectal cancer after the standardization provided by the ITBCC recommendations [
13]. Despite some evidence pointing to a similar prognostic role of TB in breast cancer, no validated assessment method exists, and no cut-off values are defined to guide clinical practice. Our study explored the association between TB and clinicopathologic features of early breast cancer, aiming to assess TB’s clinical value as a prognostic biomarker.
Lymphovascular (LVI) and perineural invasion, defined as the presence of tumor emboli within an endothelium-lined space, has been associated with worst outcomes and increased probability of distant metastases in breast cancer [
30]. Tumor budding has been established as a predictor of lymphovascular invasion for colorectal cancer [
31]. However, the association between TB and LVI for breast cancer is not entirely established, with some research showing that higher scores of budding correlate to a higher extent of LVI [
22,
29]. Analysis of the surgical specimens presenting with LVI in our population demonstrated a statistically significant association between the TB score and the occurrence of LVI. Of the 30 patients with LVI, 24 (80%) had a TB score greater than 5 buds, and 9 (30%) had a TB score greater or equal to 10 buds (
p < 0.001). These findings support the current literature on the association between TB score and LVI for breast cancer, with higher scores meaning higher LVI extension and more aggressive biological behavior.
The histological grade (HG), representing the morphological degree of tumor differentiation, is one of the best-established prognostic factors in breast cancer [
32]. Evidence shows that HG can predict tumor behavior, particularly in early small tumors, with high-grade breast cancers tending to recur and metastasize early following diagnosis [
33]. Despite representing two different biological processes, evidence points to an association between the histological grade and tumor budding, with higher budding scores associated with higher histological grades and tumor de-differentiation in colorectal cancer [
34]. The relation between the histological grade and TB for invasive ductal breast carcinoma is controversial, with some studies showing a trend for higher de-differentiation with higher TB but without statistical significance [
35]. Analyzing the distribution of the histological grade via TB group in our study, we found that 93% (
n = 39) of low-grade tumors (grade 1) presented with a low TB score (0–4) and 75% (
n = 18) of high-grade tumors presented with an intermediate-to-high TB score (>5 buds) (
p < 0.001). These findings demonstrate a significant association between the histological grade and TB score, supporting the current evidence for colorectal cancer.
Tumor size is part of the classical prognostic factors for breast cancer and one of the most important factors in the American Joint Committee on Cancer’s staging system [
36]. The association between tumor size and degree of tumor budding has been explored for different solid tumors, with the literature reporting a transversal association between both features [
23,
37,
38,
39]. The assessment of mean tumor size across our TB groups showed a significant difference between low and high TB scores, with an increase in tumor size for higher TB scores (low-TB median size: 15.00 (3–49) mm; high-TB median size: 22.50 (8–45) mm; and
p = 0.012).
Immunohistochemical evaluation of the Ki-67 proliferation index is critical for the St. Gallen Consensus on differentiating Luminal A and Luminal B molecular subtypes [
40]. Being strongly associated with tumor proliferation and aggressive tumor biology, Ki-67 is accepted as a prognostic biomarker for breast cancer and plays a role in therapeutic decisions for breast cancer [
41]. Our results are in line with the current literature, showing a significant relation between the Ki-67 index and the degree of tumor budding with an increase in the median Ki-67 index from low to high TB scoring groups (low-TB median Ki-67 index (%): 33.14 (3–90); high-TB median Ki-67 index (%): 51.57 (15–90); and
p = 0.011).
Previously mentioned features are the basis for defining breast cancer molecular subtype classification, which has an established prognostic and therapeutic impact [
41]. Xiang et al. performed a molecular analysis on 240 tumor tissue microarrays and assessed TB scores for each molecular subtype according to the ITBCC recommendations. Results showed that HER-2-positive and triple-negative breast cancers were associated with higher levels of TB and worse outcomes [
17,
42]. To provide insight into the relevance of molecular subtyping for TB expression, we grouped patients according to their molecular profile in Luminal A (
n = 23), Luminal B (
n = 35), HER-2 low (
n = 19), HER-2 positive (
n = 15), and TNBC (
n = 8). Tumor budding analysis stated a significant difference in molecular distribution across TB groups, with a lower number of Luminal tumors (A and B) in the high-TB group (
n = 6; 10%) compared to HER-2 positive or TNBC tumors (
n = 6; 25%) (
p = 0.019). The asymmetric distribution with a higher percentage of HER-2 positive/TNBC in the high-TB group seems to reflect the intrinsic aggressive biological behavior associated with these subtypes of breast cancer.
Regarding prognostic biomarkers for TNBC, tumor-infiltrating lymphocytes (TILs) have a level 1B of evidence to predict clinical outcomes in early TNBC. TILs represent a surrogate biomarker of lymphocyte-mediated immunity, with higher TIL scores associated with higher responses and improved survival outcomes [
29,
43]. TILs can be characterized according to their distribution in the tumor microenvironment into stroma TILs (sTILs) and intra-epithelial compartment TILs. Recently, the spatial heterogeneity of the tumor microenvironment has been investigated, as subpopulations of tumor cells are unevenly spatially distributed and responsible for different immune microenvironments. Current evidence demonstrates a negative relationship between TB and TILs, with low levels of TB being a surrogate for higher levels of TILs [
44]. Therefore, high levels of TILs seem to represent low TB scores and convey a better prognosis.
Our results point to a relation between higher TB scores and clinicopathological features known to be associated with aggressive biological behavior. This is also supported by the relation found between higher levels of the TB score and the indication for systemic adjuvant chemotherapy. Analyzing the distribution of patients that underwent adjuvant systemic chemotherapy, there was a significant difference between the high TB (n = 9, 64%) and intermediate TB (n = 11; 65%) groups compared to the low TB (n = 23; 33%) (p = 0.014).
These findings illustrate that the subset of high-TB patients has clinicopathological features of greater biological aggressivity that led multidisciplinary tumor boards to propose adjuvant chemotherapy to reduce the relapse rate. We can extrapolate that if no adjuvant strategies were implemented, the difference in disease-free survival of 7 months observed between the low-TB and high-TB groups would be even higher.
Regarding the impact of tumor budding on survival, the most structured evidence is provided by a post hoc analysis of the IDEA-France phase III (PRODIGE-GERCOR) clinical trial. In post hoc analysis, intermediate and high tumor budding scores strongly correlate with poor disease-free (HR: 1.41, 95% CI: 1.12–1.77;
p = 0.003) and overall survival (HR 1.65, 95% CI:1.22–2.22;
p = 0.001) for colorectal cancer. The 3-year DFS and the 5-year OS rates for low TB versus intermediate-to-high TB were 79.4% versus 67.2% (
p = 0.001) and 89.2% versus 80.8% (
p = 0.001) [
45]. The impact of tumor budding on breast cancer-specific mortality and relapse is controversial, with studies suggesting that there is a decrease in overall and disease-free survival and others not finding any relevant correlation [
38,
42,
46]. Our results showed a significant difference in the number of relapses between the low TB and the intermediate-to-high TB groups, with only one (1.5%) relapse in the first group and eight (26%) relapses in the combined groups (
p < 0.001). A significant difference was also found for disease-free survival (DFS), with a shorter time for the intermediate (96.00 (8–111) months) and high TB groups (94.00 (21–111) months) versus the low TB group (101 (10–112) months) (
p = 0.00). No statistically significant differences were seen for overall survival between groups, but data remain immature regarding the follow-up time.
These findings are in line with the current literature, suggesting a possible impact of TB on disease-free survival for breast cancer patients.
Despite the results described, the present study has limitations that must be considered when extrapolating our findings to a broader population of breast cancer patients. The study was designed as a retrospective, single-center study enrolling a small number of patients. Furthermore, the tumor budding assessment was performed by two pathologists using ITBCC 2016 guidance, which is not validated for breast cancer, and different results could be found if a different methodology was applied. Finally, despite a median follow-up time of 101 (8–112) months or 8 (0.7–9.3) years, few relapses were documented, which could have an impact on survival analysis.