Tumor budding to investigate local invasion, metastasis, and prognosis of head and neck carcinoma: A systematic review

The aim of this systematic review is to shed light on the role of tumor budding (TB) in the biology, behavior, and prognosis of head and neck squamous cell carcinoma (HNSCC). A search was run in PubMed, Scopus, and Embase databases following PRISMA guidelines. After full‐text screening and application of inclusion/exclusion criteria, 36 articles were included. Several investigations support the prognostic role of TB, which might play a role in selecting rational treatment strategies. To achieve this goal, further research is needed for greater standardization in TB quantification. Although TB is not included as a negative prognostic factor in the current management guidelines, it might be reasonable to consider a closer follow‐up for HNSCC cases with high histopathological evidence of TB.


| INTRODUCTION
Tumor budding (TB) is a histological phenomenon encountered in various cancers, for which individual malignant cells and/or small clusters of malignant cells are seen in the tumor stroma.TB is the presence of single cancer cells or small clusters of less than five cancer cells outside the main part of the tumor.Budding was first described in 1954 by Imai 1 but entered the mainstream of pathology only in the last decade. 2Budding is associated with loss of cellular adhesion and has been postulated to Elisabetta Zanoletti and Antonio Daloiso contributed equally to this study.
be closely associated with epithelial to mesenchymal transition (EMT). 2,35][6][7] TB has been associated with poor prognosis in lung, 8,9 gastrointestinal, 10,11 breast, 12,13 colorectal, 14,15 and esophageal cancer. 16,17At present, TB has been accepted as being an independent prognostic factor in colorectal cancer. 18,191][22][23][24][25] Despite the promising prognostic significance of TB, its association with tumor local invasion, nodal and distant metastasis needs to be further studied in order to be used in clinical decision making.Moreover, there are several ways to analyze TB regarding staining techniques (Hematoxylin and Eosin [H&E] or immunohistochemistry [IHC]), qualitative or quantitative assessment methods, cutpoint values, and area of examination.However, an optimized and standardized evaluation method of TB still needs to be assessed. 26iagnosis and management of HNSCC improved significantly over the last decades, but long-term overall survival (OS) did not experience similar progresses, especially in advanced cases.HNSCCs are diverse and complex diseases showing high levels of intertumoral and intratumoral heterogeneity as well as disparities in therapeutic response irrespective of clinical stage. 27Investigating biomarkers beyond the conventional clinicopathological prognostic factors appears to be of paramount importance to identify the cases with unfavorable prognosis and individualize treatment.The main aim of this systematic review of the literature is to assess the current level of knowledge on the role of TB in the biology, behavior and prognosis of HNSCC, which could be crucial for the development of improved combined therapeutic approaches for HNSCC.

| Protocol registration
The protocol of this systematic review was registered on PROSPERO, an international database of prospectively registered systematic reviews in health and social care (Center for Reviews and Dissemination, University of York, York, UK), in May 2023 (registry number CRD42023417316).

| Search strategy
A systematic literature review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) recommendations. 28The electronic databases Scopus, Pubmed, and Embase were searched from database inception to May 2, 2023.A combination of MeSH terms and free-text words were used (Table S1, Supporting Information).The reference lists of all the included articles were thoroughly screened to find other relevant articles.References were exported to Zotero bibliography manager (v6.0.10,Center for History and New Media, George Mason University, Fairfax, VA).After duplicates removal, two reviewers (Antonio Daloiso and Tiziana Mondello) independently screened all titles and abstracts and then evaluated the full texts of the eligible articles based on the inclusion criteria.Any disagreement between the reviewers involved in the literature search was resolved through discussion with all authors to reach a consensus.

| Selection criteria
Studies were deemed eligible when the following inclusion criteria were met: (i) confirmed pathological diagnosis of HNSCC; (ii) presenting data on TB in HNSCC; (iii) definition and cut-off point of TB; (iv) tissue specimen analysis performed through H&E staining and/or IHC; (v) TB assessment based on the magnification field and/or high-power field (HPF).Exclusion criteria were as follows: (i) retrospective series with less than 50 cases; (ii) lack of relevant data; (iii) non-original studies (i.e., reviews, recommendations, letters, editorials, conference paper and book chapters); (iv) animal model studies, (v) studies not described in English.

| Search results and quality assessment
A total of 232 titles were collected from our literature search.After duplicates removal and exclusion of 47 records due to coherence with the inclusion/exclusion criteria, 71 articles relevant to the topic were examined.No records were unavailable for retrieving.Finally, 36 were included in the review. 25, A deailed flowchart of the search process is shown in Figure 1.
In accordance with the National Institutes of Health quality assessment tool for Observational Cohorts and Cross-Sectional Studies, 29 15 studies (42%) were deemed of Good quality, 21 (58%) Fair, and none as Poor (Table S2).

| DISCUSSION
TB represents the most non-cohesive pattern of neoplasm invasion. 65In fact, studies have shown that TB is composed of cells exhibiting typical features of EMT, with increasing invasiveness. 66

| Biological processes associated with tumor budding in HNSCC
To date, the mechanism by which the TB is induced and develops has not yet been fully clarified, but it has been  shown that the interaction between the development of TB and the EMT is very close.These events, with anoikis resistance and exosome production, are a crucial point in the progression and invasiveness of solid tumors, indeed they are a consequence of transcriptional reprogramming leading to the migration of the cancer cell detached from the tumor site, which preludes metastasis, and lymph node invasion. 67Anoikis is a programmed cell death process that manifests itself following the physiological detachment of epithelial cells; in tumors the system fails, a downregulation expression of E-cadherin and an upregulation of N-cadherin achieves the formation of TB, followed by the metastatic process. 68,69Even if most of the evidence reported in the literature mainly comes from the study of colorectal tumors (much more frequent), TB has also been well studied in HNSCC. 67he EMT is a typical process of embryogenesis occurring when mesenchymal transition is required, characterized by high migratory capacity of stem cells, whose main features are resistance to apoptosis and high extracellular matrix production. 37,70It is well known that the development of EMT is owing to an imbalance between the expression of epithelial and mesenchymal transcription factors.Among HNSCC, such dysregulation has been reported in oral squamous cell carcinoma (OSCC), where an increased expression was observed of zinc finger E-box binding homeobox 1 (ZEB1), paired related homeobox 1 (PRRX1), vimentin, nuclear β-catenin, as well as the activation of the transforming growth factor beta (TGFβ), and a decreased expression of mesenchymal-epithelial transcription factors, E-cadherin, and most members of the miR-200 family. 33,58,67,71To confirm the involvement of EMT in OSCC patients, it has also been shown that tumor budding is related to the nuclear expression of Snail and Twist transcriptors, known as EMT regulators, which result directly associated with poor cancer prognosis. 72n budding cells, it has been seen that as a consequence of increased expression of TGFβ activators (thrombombospondine-1 and integrin β6), a downstream activation occurs of genes promoters of the matrix deposition mediated by the fibronectin gene 1 (FN1), and a down-regulation of miR-200 transcription, the latter plays an important role in establishing epithelial identity and tumor progression. 67The FN1 gene encodes for fibronectin, a glycoprotein present in the plasma or cell surface and in the extracellular matrix, which is involved in the processes of both adhesion and cell migration, thus is implicated in anoikis resistance of cancer cells, 73 in cell aggregation promotion, and in tumor immune microenvironment.It is highly related with the activation of macrophages M0 and mast cells. 68Macrophages surrounding the tumor tissue induce the loss of cellular tight junction and consequently cell detachment. 45In tumor stroma an increased expression has been observed of the interleukin 17 (IL-17).This pro-inflammatory cytokine is primarily secreted by T-helper lymphocyte and neutrophils.The role of this protein in TB has not been well clarified, but some research groups suggested that it promoted the formation and progression of budding cells, other groups that it acted together with TB in tumor progression.Furthermore, it is well known that IL-17 promotes tumor dissemination through its effects on matrix metalloproteinases and angiogenesis. 64The former act proteolyzing the laminin-5 gamma 2 chain (laminin-5 c2), whose function is endorsing the static adhesion between basal epithelial cells and the adjacent basement membrane.The release of laminin-5 c2 (peptide obtained by the laminin-5 gamma 2 chain proteolysis) leads to disruption of the cell-basement membrane adhesion through the intracytoplasmic phosphorylation of integrin brought about by its binding to the epidermal growth factor receptor. 47Moreover, tumor microenvironments under hypoxic conditions induce the secretion of extracellular vesicles and the activity of fibroblast surrounding budding cells promoting cell invasion by reshaping the extracellular matrix and thus leading to angiogenesis and tumor progression. 45bout the correlation between TB and angiogenesis, some research groups reported that the angiogenetic markers CD34 (pan-expression in vascular proliferation) and CD105 (angiogenic marker induced by hypoxia 74 ) were both overexpressed in tumor stroma, promoting increased microvascular density but without any correlation with the number of budding cells, even if the CD34 resulted more highly represented in the budding area than outside, demonstrating high microvascular density in this area, even though the study examined biopsies. 38nother relevant feature is the correlation between cancer stem cells and TB.High expression of aldehyde dehydrogenase 1 (ALDH1) in TB cells has been observed in nasopharyngeal carcinoma (NPC) and OSCC. 45,46Luo et al. 45 reported that budding cells having fibroblast-like phenotype were barely identifiable within cancer stroma; thus, they tackled the issue by staining the tumor sections with this cytosolic enzyme.The ALDH1 is a cancer stem cells marker that promotes the oxidation of proteins involved in the early stages of cell differentiation (aldehyde and retinoic acid) and they show that it resulted directly correlated with tumor aggressiveness. 34,75To confirm the role of cancer cells stemness, in OSCC patients, the expression of CD44 resulted closely related to the EMT process and therefore to tumor aggressiveness. 76n summary, budding cells show characteristics between EMT and cancer stem cells, such as morphology and lack of E-cadherin, high expression of β-catenin, vimentin, and matrix metalloproteinase. 58Budding cells are influenced by fibroblastic and macrophage presence even if their prognostic value correlates independently.Overall, all these factors promote tumor aggressiveness enhancing lymphatic invasion and metastatic potential. 59,60

| The pathological evaluation of tumor budding in HNSCC
As stated by the International Tumor Budding Consensus Conference (ITBCC), TB is defined as a single tumor cell or a cluster of up to four tumor cells. 18This definition was primarily proposed for colorectal cancer and then adopted for other tumors including HNSCC. 20In accordance with previous studies and with the ITBCC recommendations, TB should be assessed on H&E-stained slides (Figure 2), while the use of immunoreaction for cytokeratins should be limited to increase evaluation accuracy only in challenging cases. 77he ITBCC recommends assessing TB in a single field measuring 0.785 mm 2 on the invasive front, which should be selected where the tumor shows the highest density of tumor buds ("Hotspot method"). 18Different scoring systems for TB have been used in HNSCC, as highlighted in our review.The Hotspot method was the most frequently adopted (about 75% of the studies), whereas 22% of studies used a multiple fields evaluation method, meaning that examiners provided the mean value of buds counted in multiple fields along the tumor invasion front (e.g., "5 high power field" and "10 high power field" methods).Interestingly, one study applied a different scoring system that considered the spatial extension of the buds along the front of tumor invasion. 53Despite the differences in TB detection methods, its clinical relevance was maintained across the studies, endorsing the concept that the prognostic significance of TB results independent of the scoring system used.However, the presence of a single adopted standard is crucial for reproducibility in routine diagnostic practice and for research purposes, and the definition of shared recommendations for TB evaluation in HNSCC is highly desirable.Moreover, despite ITBCC recommended adopting a three-tier score system (low risk ≤4 buds; intermediate risk 5-9 buds; high risk ≥10 buds) to perform risk stratification, most of the studies discussed herein adopted a two-tier score system with a cut-off point of five buds (low risk <5 buds vs. high risk ≥5 buds), while only four studies maintained the three-tier score system. 34,44,56,60enerally, the 5 buds cut-off seemed to be an appropriate discriminator to stratify patients; however, further studies are needed to validate these data. 20he evaluation of TB in preoperative biopsies is an aspect less investigated in HNSCC.TB in biopsy material is better defined as intratumoral budding as it is not possible to consider a biopsy as representative of the entire tumor invasion front. 18Few studies 24,[78][79][80] assessed TB in preoperative biopsies, and in all the cases the TB resulted a reliable marker for tumor aggressiveness, suggesting that TB seemed to maintain in biopsies the same clinical relevance observed in surgical specimens.In any case, the real impact of intratumoral budding in preoperative biopsies deserves further investigations.

| Prognostic role
For OSCCs with higher TB scores, the existing literature agrees in describing significantly less favorable outcomes compared to those with lower TB, in terms of OS, 48,57- 60,62-64 disease-specific survival (DSS), 30,39 and diseasefree survival (DFS). 33,42,48,55,57,63The summary of findings of the included studies has been reported in Table 1.Generally speaking, such prognostic associations were significant regardless of the considered TB scoring system, mostly including either two-tier (<5 vs. ≥5 buds) 30,33,35,39,58,64 or three-tier (<5, 5-9, ≥10 buds) 34,42,48,55,60 classifications.Also, research groups who used other TB classification methods (e.g., TB absence vs. presence) obtained results in line with this general tendency.Ho et al. 43 found significantly lower locoregional recurrence-free survival in cases with the presence of TB at histopathological examination, compared to those with no evidence of TB.However, such an association did not reach statistical signification for DSS.
Recently, clinicopathological research has focused on the prognostic role of TB in the specific setting of earlystage OSCC.Retrospectively considering a Brazilian cohort of 254 OSCC patients, Dourado et al. 39 concluded that a TB count higher than 5 was independently associated with reduced DSS.Dividing such a cohort into two subgroups (early [I-II] vs. advanced [III-IV] stages), the authors found that TB was not significantly associated with prognosis in the advanced cases, whereas it was a significant predictor of DSS in the early stages. 39In line with such results, also the Almangush et al. 30 data seemed to support the prognostic role of TB in early oral cancers.In fact, in their Finnish cohort of 233 T1-2N0 OSCC cases, this research group found that TB ≥5 was significantly associated with reduced DSS at both univariate and multivariate analysis. 30Similar results were reported by Attramadal et al. 33 in early-stage OSCCs.In their retrospective analysis of cT1-2N0 OSCCs (all treated with upfront surgery followed by adjuvant radiation therapy in the case of close or positive margins), these authors found that a high number of tumor buds (≥5) was significantly associated with a poor prognosis in terms of reduced DFS. 33Moreover, in that investigation, a significantly higher recurrence rate (33%) was found among the high-budding cases, compared to the lowbudding ones (14%). 33Results in line with those described above were also found by Shimuzu et al., 55 Xie et al., 59 and Xu et al. 60 The prognostic value of TB has also been tested also in OSCC occurring in young adults.In a retrospective study on 150 OSCC patients younger than 40 years, Mneimneh et al. 48found that high-risk TB (≥10 buds) was associated with significantly reduced OS and distant metastasis-free survival, thus substantially confirming the prognostic value of such a histological feature also in this specific cohort.

| Predictive role
As TB has been widely established as an adverse prognostic factor for colorectal carcinoma, 81 there is growing interested in confirming such a predictive role in OSCC, as well.61]63 Table 2 reports the findings of the studies dealing with predictive role of TB in OSCC.In a retrospective study which included 186 surgically resected OSCC, Noda et al. 49 found in their bivariate and multivariate analyses that high TB and pathological depth of invasion >10 mm in resections were independent factors for the presence of extranodal extension (ENE+).Moreover, their combination showed high specificity (91%) and accuracy (83%) in predicting ENE+.In 2022, a clinicopathological research focused on the predictive role of TB in the specific setting of early-stage OSCC. 51Retrospectively considering a cohort of 70 early-staged OSCC patients, Sakai et al. 51 concluded that a TB count higher than 5 was independently associated with neck node metastasis in a univariate setting, while in a multivariate one tumor-stroma ratio correlated more strongly with neck node metastasis than TB grade.Similar conclusions were proposed by Sakata et al., 52 who demonstrated that the evaluation of TB was effective for pinpointing patients with cT2N0 OSCC at high risk of occult neck metastasis.Previously, Xie et al. 59 reported that TB independently predicted prognosis of patients with T1/2 classification oral tongue SCC and might be used for routing pathological diagnosis and the decision for elective lymph node dissection.Some patients with early-stage oral cancer have a poor prognosis due to delayed neck metastasis occurrence.In this setting, Yamakawa et al. 61 underlined the role of TB as additional to conventional features to predict delayed neck metastasis in early oral tongue SCC.Recently, Hamada et al. 41 analyzed a sample of 99 patients diagnosed with early-stage OSCC of the tongue identifying lymphatic vessel invasion and podoplanin expression as predictors for neck lymph node metastasis.These authors further defined a cluster of 77 patients with low-TB scores, where podoplanin expression in TB cells was an independent prognostic factor for neck metastasis.In conclusion, they suggested that patients with T1-2 tongue carcinoma, with lowgrade TB and no podoplanin expression might be candidates for a less aggressive treatment of the neck.
In 2017, Arora et al. 32 developed a new outcome prediction model in early-stage OSCC based on histopathological parameters.At first, they found TB as an independent predictor of neck node metastasis, then they graded it with other significant independent factors (depth of invasion, pattern of invasion, perineural invasion, lympho-vascular invasion, lymphoid response) to design a scoring system that allowed accurate evaluation of the risk of metastasis with accuracy independent of the traditional TNM system.
The predictive role of TB could be of clinical importance for OSCC patients where the clinicians have refrained from neck dissection.A high degree of budding might indicate a higher risk of neck node metastases, implying that a tighter follow-up is warranted and elective neck dissection could probably be recommended.

| Tumor budding in non-oral HNSCC
The role of TB in non-oral HNSCCs has been investigated, too (Table 3).Considering the pathological assessment of TB, several studies showed IHC of cytokeratin to have higher detection rates of single cells at the invasive tumor front than the H&E method. 82,83A retrospective case series by Gupta et al. 40 confirmed a high efficacy of IHC in TB assessment in LSCC.Indeed, IHC was able to detect scattered single tumoral cells mixed within the inflammatory cells at the invasive front. 40On H&E sections, low-grade TB and high-grade TB were found in 30% and 42% of the specimens, respectively.On the other hand, IHC showed low-grade TB in 17% of tumors and high-grade TB in 59% of them.Therefore, Gupta et al. 40 defined IHC as a better method for TB assessment and confirmation in LSCC.
Several studies focused on the formulation of new grading systems based on easily reproducible morpho-pathological findings with potential prognostic role. 30,84,85In 2016, Weichert et al. 86 suggested that in pulmonary SCC, tumor cell nest size (CNS) and TB had an impact on the prognosis of patients and proposed a grading scheme.Similarly, for HNSCC, Boxberg et al. 35 suggested a grading system based on CNS and TB scores in OSCC, while Caruntu et al. 36 found significant correlations between TB and tumor necrosis and the clinical staging in cutaneous and mucosal HNSCC series.In addition, TB was significantly correlated with perineural invasion. 36More recently, the same group adapted their grading system in patients with laryngeal and hypopharyngeal squamous cell carcinomas (LSCC, HSCC). 25In addition to TB and CNS, other morphopathological parameters were taken into consideration, such as degree of keratinization, mitotic count, nuclear size, and stromal content. 25TB resulted to be prognostic for OS, DSS, and DFS, in particular the OS of patients with high TB was the worst.The keratinizing histotype significantly correlated both with a higher frequency of TB and small CNS/single-cell infiltration compared with nonkeratinizing SCC.High TB levels significantly correlated with moderate/high stromal content, intermediate/large nuclear size, and CNS.In addition, the presence of TB and small CNS/single-cell infiltration was both significantly associated with lymph-angio-invasion and perineural invasion and were also correlated with each other.Both TB and CNS, were significantly correlated with pT classification, pN classification, and UICC classification.Therefore, the new grading system proposed by Boxberg et al. 25 was able to successfully stratify patients from a prognostic viewpoint into three groups (nG1, nG2 and nG3) in both LSCC and HSCC as well as in OSCC.
A recent investigation by Stögbauer et al. 56 suggested an improved histopathological grading system that for the first time was comprehensive of both HPV positive and negative HNSCC.In particular, they analyzed the prognostic significance of TB and CNS and stratified tumors in two subgroups of cellular dissociation grading according to budding levels.The prognostic value of TB in HPV-positive HNSCCs was confirmed in this study. 56ubsequently, they compared this new grading system to the well-established WHO classification and Brandwein-Gensler (BG) risk models. 5,87In a multivariate analysis, the new grading system seemed to be the only one with a significant prognostic role. 56Moreover, they studied the correlation between TB and neck lymph node metastasis in cN0 patients, thus suggesting the role of TB in identifying lymph nodes with an elevated risk of micro-metastasis and highlighting the importance of this finding in the decision making for surgical treatment planning.In addition, they hypothesized that small tissue areas that can be analyzed for TB under 1 HPF could be transferable to an excisional biopsy sample. 56he prognostic role of TB in LSCC was initially studied in 2010 by Sarioglu et al. 53 who analyzed a sample of 64 patients with LSCC.In this series, N+ status and higher TB scores were significantly associated with distant metastasis in LSCC. 53Analogously, in the Ozturk et al. 50series, higher TB levels were associated with reduced OS and DFS.Moreover, the lymphocytic host response was found to be significantly reduced in highgrade TB tumors, thus resulting in a poorer prognosis. 50he prognostic role of TB was further evaluated by Karpathiou et al. 44 who focused their attention on tumor stroma parameters in a large series of LSCC and pharyngeal SCC biopsies and surgical specimens.In particular, TB and abundant stroma tumors were associated with a worse prognosis and a more aggressive tumor behavior.Indeed, TB levels, stroma-rich tumors, and fibroblastic stroma type were all associated with advanced T classification, whereas abundant stroma tumors were only marginally correlated with N classification.Response to neoadjuvant chemotherapy was evaluated in association with tumor related variables; low TB, high CNS, and stroma-poor tumors were predicting factors for a good response to chemotherapy.On the other hand, high TB, stroma rich tumors, and a low CNS seemed to be negatively correlated with a poorer lymphocytic host response, thus highlighting the fundamental role of immune surveillance to reduce tumor aggressiveness. 44imilarly, a Romanian group studied the prognostic potential of tumor/stroma ratio, local immune response, TB activity and tumor necrosis using standard (H&E) staining, in SCCs involving distinct areas of the head and neck.They suggested that TB could provide additional information on the aggressiveness of SCCs.In particular, both TB and tumor necrosis were significantly related with tumor dimensions and advanced clinical stages of the disease.Moreover, perineural invasion was more frequently present in tumors with high TB activity.
EMT markers such as Twist and Snail are known to be associated with tumor size, distant metastasis, patient survival, and tumor bud formation in the context of lymph node metastasis in several tumors, for instance breast cancer and colorectal cancer. 88,89Some studies suggest that TB may play a role in EMT in several types of cancer as well as in HNSCCs. 2,33,58,79A recent study by a Korean research group investigated the relationship between TB and both p16 expression and EMT markers, namely Twist and Snail, in a cohort of 121 oropharyngeal SCC and 270 cases of non-oropharyngeal HNSCC as a control group. 37A marked expression of Twist and Snail/ Slug was observed in p16-positive cases among all HNSCCs.Focusing on oropharyngeal SCC, p16-positive tumors showed significantly higher Twist and Snail/Slug expression than the p16-negative ones.Moreover, high TB count was significantly associated with p16 positivity in oropharyngeal SCCs, whereas at IHC, p16 expression was not related to TB in the control group.Compared with non-oropharyngeal SCCs, oropharyngeal SCCs were significantly associated with lympho-vascular invasion, lymph node metastasis, and larger maximal diameter of metastatic foci in lymph nodes, as well as with p16 positivity.These findings suggested that EMT markers changes, as well as morphological changes, can be related to the lymph node metastasis in oropharyngeal SCCs.Cho et al. 37 reported that p16 expression in oropharyngeal SCC was correlated with EMT markers expression and TB formation, thus suggesting the important role HPV infection played in altering such markers, leading to an increase in tumor aggressiveness and consequently to metastasis formation, despite the relatively small tumor size.
NPC is an uncommon and aggressive tumor, also because it tends to have cervical lymph node metastases even at an early stage. 90In 2012, Luo et al. 45 studied the intensity of TB levels and its prognostic significance in NPC biopsy samples.Furthermore, they investigated similar features to cancer stem cells in TB cells at the invasive tumor front using IHC for ALDH1.A total of 122 NPC samples were analyzed and TB was stained with pancytokeratin antibodies.They suggested that TB could play a significant role in NPC invasion and malignant progression.Indeed, a significant association was found between the degree of budding and advanced stage tumors, highlighting a possible correlation between TB and tumor invasion's depth, as well as an increase in metastatic potential when a higher degree of TB is present.Moreover, Luo et al. 45 reported high TB levels were contributing for lymphatic and vascular tumor invasion.Subsequently, they found a positive correlation between ALDH1 expression and the grade of budding, in terms of staging, lymph node metastasis and both lymphatic and vascular invasion.TB with a high level of ALDH1 expression were reported to be negatively associated with OS, thus suggesting the prognostic relevance of this factor in NPC. 45

| Future perspectives
It is necessary to investigate the role of TB in SCC arising from different anatomic sites of the head and neck region, as well as HPV-negative and HPV-positive tumors. 23There is a paucity of data specifically relating to predisposing factors for TB, which could contribute to our understanding of TB.From a pathological viewpoint, TB has been framed as a marker that could reflect EMT in different malignancies including HNSCC.A clearer understanding of the molecular background of TB in HNSCC necessarily requires more studies.
Moreover, an issue exists in current literature regarding the relationship between TB and the worst pattern of invasion (WPOI).WPOI was firstly proposed by Brandwein-Gensler et al. 5 Unlike the TB, the WOPI grading system is focused on the pattern of invasion, and provides a qualitative assessment stratified in five invasion patterns: Type 1, pushing borders; Type 2, finger-like growth; Type 3, large separate islands, >15 cells per island; Type 4, small tumor islands, ≤15 cells per island; and Type 5, tumor satellites, ≥1 mm from the main tumor or next closest satellite. 5The last AJCC staging system endorsed the use of WPOI method for OSCC risk stratification 91 ; however, the complementary role of TB and WOPI in stratify patients with HNSCC is an aspect still under investigation. 5,60,92t has been proposed that in the future routine reporting of surgical pathological specimens, "TB intensity" will have an essential role. 20,93In HNSCC, the prognostic and predictive role of TB need not necessarily be considered in isolation but also in combination or in panels of other clinicopathological variables as the WOPI score.Routine assessment of TB on biopsy should be considered with caution and further investigated, as it can be affected by section effects and compression artifacts.

| CONCLUSIONS
TB represents groups of cancer cells that are more invasive than that in the tumor core, leading to an increased risk of lymphovascular spread.There is a substantial body of data indicating that TB may be an important sign of invasive growth and metastasis in HNSCC.Consequently, several studies supported the clinicopathological relevance and prognostic role of TB.
Evidence suggests that TB may play a role in future staging systems for HNSCC, 94 potentially influencing treatment decisions. 87However, to facilitate this, further research (particularly prospective studies) is needed on biopsy and surgical material to standardize TB quantification and implement it into the pathology reports.Although TB is not included as a negative prognostic factor in the current guidelines for the management of HNSCC, it is reasonable to assume that cases with high histopathological evidence of TB can be addressed for closer follow-up.
of Padova.The authors thank Alison Garside for correcting the English version of this paper.

F
I G U R E 1 PRISMA diagram resembling electronic database search and inclusion/exclusion process of the review.Legend: date of last search May 2, 2023 [Color figure can be viewed at wileyonlinelibrary.com]T A B L E 1 Summary of studies dealing with prognostic performance of tumor budding in oral squamous cell carcinoma included in the systematic review

F
I G U R E 2 Representative Hematoxylin and Eosin images of the different degree of tumor budding in three different moderately differentiated squamous cell carcinomas of the oral cavity (panoramic overview of the corresponding digital slides in the upper-left corner of each box).Low tumor budding (A), intermediate tumor budding (B), and high tumor budding associated to high peritumoral inflammation (C1); black arrows indicate tumoral buds spreading into the peritumoral stroma.Due to the inflammatory context in C1, budding detection is challenging, and it is better recognized by cytokeratin stain (black arrows).Original magnification 200Â (A, B, C1, C2); immunostaining with cytokeratin AE1-AE3 in C2 [Color figure can be viewed at wileyonlinelibrary.com] Summary of studies on non-oral squamous cell carcinomas included in this systematic review Abbreviations: CNS, cell nest size; DFS, disease-free survival; DSS, disease-specific survival; EMT, epithelial-mesenchymal transition; H&E, Hematoxylin and Eosin stain; HM, hotspot method = number of buds in a single 200Â field (0.785 mm 2 ) at the tumor front of invasion area with the highest budding intensity); HNSCC, head and neck squamous cell carcinoma; HSCC, hyphopharyngeal squamous cell carcinoma; IHC for CK, immunohistochemistry for cytokeratin; LSCC, laryngeal squamous cell carcinoma; MFE, multiple field evaluation = number of buds in 10 HPF selected within the tissue area with highest budding density); NPC, nasopharyngeal carcinoma; NR, not reported; OPS, observational prospective study; ORS, observational retrospective study; OS, overall survival; OSCC, oral squamous cell carcinoma; OTSCC, oral tongue squamous cell carcinoma; TB, tumor budding.T A B L E 2 Summary of studies dealing with predictive performance of tumor budding in oral squamous cell carcinoma included in the systematic reviewT A B L E 2 (Continued)Abbreviations: CNS, cell nest size; DFS, disease-free survival; DSS, disease-specific survival; EMT, epithelial-mesenchymal transition; H&E, Hematoxylin and Eosin stain; HM, hotspot method = number of buds in a single 200Â field (0.785 mm 2 ) at the tumor front of invasion area with the highest budding intensity); HNSCC, head and neck squamous cell carcinoma; HSCC, hyphopharyngeal squamous cell carcinoma; IHC for CK, immunohistochemistry for cytokeratin; LSCC, laryngeal squamous cell carcinoma; MFE, multiple field evaluation = number of buds in 10 HPF selected within the tissue area with highest budding density); NPC, nasopharyngeal carcinoma; NR, not reported; OPS, observational prospective study; ORS, observational retrospective study; OS, overall survival; OSCC, oral squamous cell carcinoma; OTSCC, oral tongue squamous cell carcinoma; TB, tumor budding.aBuddingassessed at 400Â.T A B L E 3Abbreviations: CNS, cell nest size; DFS, disease-free survival; DSS, disease-specific survival; EMT, epithelial-mesenchymal transition; H&E, Hematoxylin and Eosin stain; HM, hotspot method ) at the tumor front of invasion area with the highest budding intensity); HNSCC, head and neck squamous cell carcinoma; HSCC, hyphopharyngeal squamous cell carcinoma; IHC for CK, immunohistochemistry for cytokeratin; LSCC, laryngeal squamous cell carcinoma; MFE, multiple field evaluation = number of buds in 10 HPF selected within the tissue area with highest budding density); NPC, nasopharyngeal carcinoma; NR, not reported; OPS, observational prospective study; ORS, observational retrospective study; OS, overall survival; OSCC, oral squamous cell carcinoma; OTSCC, oral tongue squamous cell carcinoma; TB, tumor budding.aBudding assessed at 400Â.