Tumor budding (TB) has emerged as a promising independent prognostic biomarker in colorectal cancer (CRC). The prognostic role of TB has been extensively studied and currently affects clinical decision making in patients with stage I and II CRC. However, existing prognostic studies on TB in stage III CRC have been confined to small retrospective cohort studies. Consequently, this study investigated the correlation among TB categories, clinicopathological features, and prognosis in stage III-IV CRC to further enhance the precision and individualization of treatment through refined prognostic risk stratification.

To analyze the relationship between TB categories and clinicopathological characteristics and assess their prognostic value in stage III-IV CRC to further refine the prognostic risk stratification of stage III-IV CRC.

The clinical data of 547 CRC patients were collected for this retrospective study. Infiltration at the front edge of the tumor buds was counted according to the 2016 International Tumor Budding Consensus Conference guidelines.

Multivariate Cox proportional hazards regression analysis demonstrated that chemotherapy (P = 0.004), clinical stage IV (P < 0.001), ≥ 4 regional lymph node metastases (P = 0.004), left-sided colonic cancer (P = 0.040), and Bd 2-3 (P = 0.002) were independent prognostic factors in patients with stage III-IV CRC. Moreover, the density of tumor infiltrating lymphocytes was higher in Bd 1 than in Bd 2-3, both in the tumor stroma and its invasive margin.

TB has an independent predictive prognostic value in patients with stage III-IV CRC. It is recommended to complete the TB report of stage III-IV CRC cases in the standardized pathological report to further refine risk stratification.

Colorectal cancer (CRC) is the third most common malignant tumor in the world[1]. The occurrence and development of CRC represent a complex multifactorial process[2]. Despite significant advancements in medical technology both domestically and internationally, the incidence of CRC has continued to rise in recent years. This trend can be attributed to improvements in people's living standards as well as changes in lifestyle and dietary habits. There were an estimated 1.93 million new CRC cases diagnosed, and 0.94 million CRC-related deaths in 2020 worldwide. China is expected to have the highest estimated number of new CRC cases over the next 20 years. The number of new CRC cases is expected to increase from 0.56 million (2020) to 0.91 million (2040) in China[3].The tumor-node-metastasis (TNM) staging system published simultaneously by the Union for International Cancer Control and The American Joint Committee on Cancer (AJCC) remains a widely accepted standard for classifying malignant tumors[4]. The prognosis of patients with CRC is often influenced by the time gap between cancer detection and treatment initiation. Notably, even patients with identical stage of disease may exhibit diverse clinical outcomes, suggesting the limited predictive value of TNM stage in estimating CRC prognosis. Tumor budding (TB), defined as a single cancer cell of up to four cancer cells at the tumor invasive margin, has emerged as a promising independent prognostic biomarker in CRC[5].

TB reflects an invasive growth pattern with metastatic potential that plays a key role in the tumor microenvironment (TME) and epithelial-mesenchymal transition (EMT). The prognostic role of TB has been extensively studied and currently affects clinical decision making in patients with stage I and II CRC[6,7]. However, existing prognostic studies on TB in stage III CRC have been confined to small retrospective cohort studies. Consequently, this study investigated the correlation among TB categories, clinicopathological features, and prognosis in stage III-IV CRC to further enhance the precision and individualization of treatment through refined prognostic risk stratification.

Tumor-infiltrating lymphocytes (TILs) are a heterogeneous group of cells that leave the blood and migrate to the tumor regions with antigenic effects. They are also an important part of the TME and are generally considered to protect the host against tumor development. It also plays a vital role in inhibiting cancer cell invasion and distant metastasis[8]. The interaction between TB and the immune system is known as the attack-defense model[9]. TB reflects an aggressive disease phenotype; however, TILs can regulate immune function and improve the body's ability to kill tumors. Therefore, in our study, we also discussed the relationship between TB and TILs based on the idea that the combination of TB and TILs can effectively predict the prognosis of stage III-IV CRC patients and analyzed the changes in TILs in cases of different TB levels to preliminarily explore the correlation between TB and the immune microenvironment.

We retrospectively analyzed the medical records and clinical data of 838 patients with stage III-IV resectable CRC pathologically diagnosed at Union Hospital, Tongji Medical College of Huazhong University of Science and Technology (Wuhan City, Hubei Province, China) from January 1, 2015, to December 31, 2018 (Figure 1). A total of 547 patients with CRC were enrolled, and their TB categories and clinicopathological features were analyzed.

Open in New Tab   Full Size Figure   Download Figure

Figure 1 Study design and flow diagram. CRC: Colorectal cancer.

The primary objective of this study was to investigate the relationship between TB and progression-free survival (PFS), defined as the time from the beginning of radical surgery to relapse or death, whichever occurred first. The secondary objective was to determine the association between TB and overall survival (OS), defined as the time from the start of radical surgery until death from any cause. The survival time and prognosis of the patients were followed up until the end of the study. The follow-up period was 61 mo. This study was approved by the review committee of the affiliated institution of Tongji Medical College of Huazhong University of Science and Technology (2018-S377), and consent was obtained from all patients participating in the study.

Statistical analysis

All statistical analyses were performed using the statistical software SPSS version 27.0, with two-sided statistical testing, and P < 0.05 was considered statistically significant. Clinicopathological features and KRAS, NRAS, BRAF mutation status of the patients were descriptively analyzed. The χ² test (or Fisher’s exact test, if appropriate) was used to evaluate the relationship between the TB categories and clinicopathological features. Spearman correlation analysis was used to evaluate the consistency between the TB categories evaluated by the two pathologists (Figure 2). Kaplan-Meier survival analysis was used to compare PFS and OS between patients with TB and those with other clinicopathological features. Cox proportional hazard regression models were used to estimate hazard ratios (HRs) and 95% confidence interval (CI). The association of baseline parameters with PFS and OS was first examined using univariate Cox analysis, and variables with P < 0.05 were entered into a Cox regression multivariate model. After normal distribution detection of TILs in the tumor stroma and at the TB of the tumor invasion front was conducted, it was found that both groups of data were biased. The Mann-Whitney U test was used to evaluate the relationship between TILs in the tumor stroma and TILs at TB of the tumor invasion front with Bd 1 and Bd 2-3, respectively.

Open in New Tab   Full Size Figure   Download Figure

Figure 2 Spearman correlation analysis was used to evaluate the consistency between tumor budding categories evaluated by two pathologists. R = 0.865, P < 0.001.

As early as 1954, Imai first reported the phenomenon of TB in various solid tumors and its correlation with prognosis of disease[10]. Subsequent studies have further demonstrated that TB serves as an independent prognostic biomarker in CRC and is observed in a variety of solid tumors, including tongue, larynx, esophagus, stomach, breast, and intrahepatic cholangiocarcinomas. It has been shown to be predictive of cancer progression[11-14].

Currently, high-grade TB is an independent predictor of LNM to determine subsequent surgical treatment for stage I CRC[6,15]. In particular, Bd 3 is a strong adverse prognostic factor used to identify those stage II CRC patients with high recurrence rates and high mortality to guide adjuvant treatment[16]. In the new 2017 TNM and 2019 World Health Organization classifications, TB is regarded as an important complementary biological marker for predicting the prognosis of CRC[17]. Although the prognostic impact of TB on stage I and II CRC is clear, the application of this criterion in clinical practice for treatment decision-making in stage III-IV CRC is limited. This is mainly due to the lack of relevant research data in this context[18]. In this study, 547 patients with stage III-IV CRC were evaluated for TB and TILs according to the ITBCC 2016 scoring system. Combined with patient history, clinicopathological data and patient prognosis, the association between TB and various clinicopathological features and TILs, as well as the prognostic significance, were analyzed. These results suggest that TB independently affects the prognostic outcome of stage III-IV CRC.

The results of our study are consistent with the conclusion reached by Akabane et al[18] who retrospectively collected clinicopathological data from 237 patients with stage III CRC in a single center. This study showed that Bd 1 patients showed significantly better disease-free survival. The importance of TB in adjuvant treatment decision-making for stage III CRC has been emphasized. Notably, our study showed that TB was an independent adverse prognostic biomarker for PFS and OS in stage III-IV CRC. A single-center retrospective study of 589 patients with stage III CRC reported by Yamadera et al[19] used a binary classification system similar to that of Akabane et al[18] but the cut-off value was different (0-9 buds: Low; ≥ 10 buds: High). The results showed that among patients with stage III CRC, the Bd1 group had better survival after adjuvant chemotherapy. Rogers et al[16] conducted a meta-analysis of 34 studies with a total of 7821 CRC patients at various stages, and found that TB was significantly correlated with LNM (odds ratio [OR]: 4.94, 95%CI: 3.96-6.17; P < 0.00001), 5-year PFS (OR: 5.50, 95%CI: 3.64-8.29; P < 0.00001) and 5-year OS (OR: 4.51, 95%CI: 2.55-7.99; P < 0.0001) . It was concluded that TB was a risk factor for predicting LNM at 5 years, PFS at 5 years, and OS at 5 years in CRC patients, and its inclusion in CRC staging facilitated further risk stratification; however, there was a lack of studies on stage III-IV CRC in a meta-analysis by Rogers et al[16] and the definition and evaluation criteria of TB varied greatly between different studies. Similarly, Graham et al[20] evaluated 553 CRC patients at various stages and found that high-grade TB was an independent predictor of poor disease-specific Survival. Therefore, TB could also be a pathological marker that is clinically practical to further refine risk classification and guide decision-making in stage III-IV CRC.

Interestingly, our study also found that the Bd 1 group had a higher proportion of TILs at the tumor invasive margin, which may be associated with better prognosis. This is consistent with the findings of Nearchou et al[21], who found the deletion of TILs in high-grade TB by exploring the effects of different lymphocyte distribution patterns and TB quantity and density on prognosis in stage II CRC through automatic imaging analysis and machine learning workflow.

During the development of cancer, EMT promotes cancer development by empowering mesenchymal phenotypes associated with highly aggressive tumor cells[22]. TB demonstrates the dynamic process of EMT; that is, tumor cells in the aggressive front acquire an interstitial phenotype. This increases the malignant potential of vascular invasion and distant metastasis, and consequently, metastases to lymph nodes or distant parenchymal organs through the vasculature[23-25].

E-cadherin is an intercellular adhesion protein and pivotal regulatory factor in this process[26,27]. Numerous studies have claimed that in different solid tumors, the expression of E-cadherin is reduced or absent on the cell surface of the tumor invasion front, especially in TB[28-31]. This loss is usually accompanied by a decrease in β-catenin expression, indicating that the WNT pathway in cells may be activated. Hence, budding tumor cells can degrade the extracellular matrix (ECM) and invade and migrate through the surrounding stroma. In addition, it has been found that the expression of laminin 5γ2 is usually increased in response to tumor invasion[32]. In the gene ontology study, RNA sequencing data from the TB region and the corresponding central tumor region were compared and analyzed. The expression of genes involved in integrin-mediated cell adhesion, migration, cytoskeleton rearrangement and ECM degradation were found to be significantly different between the two regions[33]. Koelzer et al[34] found that there is an immune escape mechanism in EMT, in which tumor cells lose the expression of their major histocompatibility complex, allowing them to evade recognition and attack by the immune system.

Finally, our study included a substantial number of 547 CRC patients. TB was evaluated independently by two pathologists and re-evaluated by a third pathologist when the results were inconsistent, ensuring a high level of reliability in the TB assessment. Furthermore, the ITBCC 2016 recommendations were followed to evaluate TB specifically in patients with stage III-IV CRC, thereby investigating its impact on patient prognosis.

However, this study had some limitations. First, this study was a retrospective analysis conducted at a single center, which may introduce some degree of bias as the study population may not represent all CRC patients in China. Second, using a single index as a risk factor still has some drawbacks, which can be solved by constructing a prognosis model with a multi-index multidimensional association algorithm. Furthermore, the accuracy of the assessment results based on TILs on H&E slides must be improved to provide better medical strategies and achieve individualized precision treatment for patients with CRC. Future efforts should focus on utilizing artificial intelligence techniques, such as digital pathology methods, to enhance the accuracy and reproducibility of TB and TIL assessments. Additionally, integrating TB with the immunoscore could refine the risk stratification of stage III-IV CRC.

TB has an independent predictive prognostic value for PFS and OS in patients with stage III-IV CRC. It is recommended to complete the TB report of stage III-IV CRC cases in the standardized pathological report to further refine risk stratification.

We thank all the patients and their family members involved in the study without which the advancements described herein would have not been made possible.