Abstract
Background
Synchronous colorectal cancer liver metastasis (CRCLM) is closely associated with poor prognosis; however, the precise factors contributing to this adverse outcome remain uncertain. In the tumor microenvironment of CRCLM, tumor budding (TB) serves as a morphologic hallmark of invasive potential, while tertiary lymphoid structures (TLS) are indicative of organized anti-tumor immune contexture. This study aimed to integrate these biomarkers to predict the survival of patients with CRCLM.
Methods
This multicenter retrospective study enrolled 107 patients with synchronous CRCLM who underwent curative resection for colorectal cancer. An external validation cohort including 30 patients with CRCLM were also employed. The density of TLS and the TB count were quantified via hematoxylin and eosin (H&E) staining and immunohistochemical (IHC) staining on primary tumor specimens. Relationships between clinicopathological data, overall survival (OS), and recurrence-free survival (RFS) were analyzed. The TLS/TB index was derived by integrating TLS and TB densities. The clinical role and prognostic value of the TLS/TB index in patients with CRCLM were analyzed, and a nomogram was constructed to predict the 3-year survival of patients.
Results
High TLS density was significantly associated with prolonged overall survival (OS; p < 0.001) and recurrence-free survival (RFS; p< 0.001) in patients with CRCLM. Conversely, higher TB grade predicted a worse OS (p < 0.001) and RFS (p < 0.001). The TLS and TB exhibited a significant inverse correlation (r = -0.581, p < 0.001). The TLS/TB index demonstrated superior prognostic accuracy in the training cohort (AUC = 0.89) compared with histological differentiation (AUC = 0.70), TB grade (AUC = 0.85) and TLS density (AUC = 0.80), with consistent performance in the validation cohort. Calibration curves showed acceptable agreement between predicted and observed survival in both cohorts.
Conclusion
These results demonstrated that tumor budding and tertiary lymphoid structures are inversely correlated in CRCLM. Furthermore, the TLS/TB-based nomogram may effectively enhanced the accuracy of prognosis prediction in patients with CRCLM, potentially facilitating the formulation of individualized treatment strategies.
Keywords: colorectal cancer liver metastasis, nomogram, prediction, tertiary lymphoid structures, tumor budding
1. Introduction
Colorectal cancer (CRC) is a malignancy characterized by high clinical aggressiveness and its development is closely associated with dietary and lifestyle factors. According to recent data, CRC accounts for an estimated 1.9 million new cases and 904,000 deaths globally, representing nearly one-tenth of all cancer cases and deaths. It is the third most common cancer and the second leading cause of cancer-related deaths worldwide. Currently, with the widespread adoption of colonoscopy screening, the 5-year survival rate of patients with early stage CRC is approximately 90% (1). However, for advanced or metastatic CRC, the 5-year survival rate drops to approximately 13% (1). Blood from the gastrointestinal tract enters the liver via the portal vein, facilitating the dissemination of colorectal cancer cells to the liver (2). Consequently, the liver is the most frequent site of metastasis in CRC (3). Approximately 50% of CRC patients develop liver metastases during the course of their disease (4). Despite advancements in surgical techniques and targeted therapies, the prognosis of patients with colorectal cancer liver metastasis (CRCLM) remains poor (5, 6). Understanding the specific factors influencing CRCLM patient outcomes and developing effective targeted therapies based on these factors is critically important for this patient population. Therefore, there is an urgent need for novel models to predict survival and recurrence in patients with liver metastases from colorectal cancer.
Numerous studies indicate that the tumor microenvironment (TME) plays a pivotal role in the progression of CRCLM (7, 8). The TME comprises cancer cells, non-cancerous cells (e.g., fibroblasts and immune cells), and extracellular matrix (ECM) (9, 10). The immune landscape within TME is intrinsically linked to cancer invasion and metastasis (11, 12). As tumors develop and invade surrounding tissues within the TME, they often manifest as single tumor cells or small clusters, a phenomenon known as tumor budding (TB). Tumor budding has long been associated with the epithelial-mesenchymal transition (EMT) and is considered a hallmark of invasive and migratory behavior in colorectal cancer (13, 14). Research demonstrates that TB is an independent adverse prognostic factor in CRC patients (15). Thus, TB serves as a marker for highly aggressive tumor subclones and is associated with poor clinical outcomes (16, 17).
In parallel, the host immune contexture plays a significant role in modulating tumor progression. Tertiary lymphoid structures (TLS) are ectopic lymphoid formations that develop in nonlymphoid organs during chronic inflammation, including cancer, infection, and autoimmune diseases (18). Within tumors, TLS represents the frontline of the anti-tumor immune response. They facilitate the rapid presentation of antigens by dendritic cells to T cells and provide a niche conducive to the activation, proliferation, and differentiation of T and B cells (19). The presence of intratumoral TLS is associated with a reduced risk of recurrence and improved survival in patients with various malignancies, including colorectal cancer (20, 21). Our group’s previous work similarly confirmed that TLS constitutes a significant protective factor against tumor progression (15, 22–25).
In summary, since the immune landscape of the TME correlates with the response to immunotherapy, gaining a deeper understanding of the TME immune landscape in CRCLM holds significant promise for these patients (26). Therefore, we aimed to develop a clinical model capable of reflecting the dynamic interplay within the tumor immune microenvironment to some degree. Currently, we posit that the presence of TB serves as a morphological indicator of a high invasive potential, while TLS reflects an organized host immune architecture within the microenvironment. Building upon our group’s prior research, the objective of this study was to jointly analyze TB and TLS as independent prognostic factors, leveraging their combined potential to predict outcomes in patients with CRCLM. Furthermore, this model aims to provide an integrative morphologic framework to assess the equilibrium between host immune organization and the invasive phenotype of the tumor. While acknowledging the inherent complexity of the tumor microenvironment, we hypothesize that the TLS/TB index may serve as a comprehensive tool for predicting survival and recurrence in CRCLM patients.
2. Materials and methods
2.1. Study design and patient cohort
This study retrospectively analyzed 210 patients who underwent surgery at multiple medical centers between January 2019 and January 2021, all of whom underwent curative resection for colorectal cancer. Patients meeting the following criteria were included: age ≥ 18 years, curative resection for colorectal cancer, histopathologically confirmed colorectal cancer liver metastasis, and compliance with follow-up, while excluding those with synchronous metastases to other organs, a history of autoimmune disease requiring immunosuppressive medication, perioperative mortality, or incomplete pathological/missing follow-up data. Thus, all enrolled patients had synchronous liver metastases at the time of primary colorectal cancer diagnosis. Following stringent screening, 107 CRCLM patients were enrolled, with the study approved by the Ethics Committee of each participating medical center. Written informed consent for surgery and for the use of clinical data for research purposes was obtained from all patients. An independent external validation cohort comprising 30 patients with CRCLM was enrolled from February 2021 to February 2022 at the same institutions, using the same inclusion and exclusion criteria. The validation cohort was temporally separated from the training cohort to ensure independence, with the patient selection process detailed in Figure 1. All procedures performed in our study were in line with the strengthening the reporting of cohort, cross-sectional and case-control studies in surgery (STROCSS) criteria (27).
Figure 1.
Flowchart of the study. A retrospective analysis was conducted on the clinical data of 194 patients with colorectal cancer liver metastases from Nanjing Drum Tower Hospital and 73 patients from Jiangsu Cancer Hospital. After applying strict inclusion and exclusion criteria, a total of 137 patients with colorectal cancer liver metastases were ultimately enrolled. Among them, 107 patients treated between January 2019 and January 2021 were assigned to the training cohort, while 30 patients treated between February 2021 and February 2022 were assigned to the validation cohort.
2.2. Observation index
Patient baseline information, including age, sex, tumor size, and tumor location, was collected. Postoperative pathological reports were analyzed to determine the tumor size, histological differentiation (well/moderate/poor), DNA mismatch repair (dMMR/pMMR), T stage (T1/T2/T3/T4), N stage (N0/N1/N2), surgical margin status (R0 vs. R1), number of liver metastases (≤3 vs>3), size of the largest liver metastasis (≤5cm vs>5cm), and preoperative treatment modality (none, Chemotherapy, or Chemoradiotherapy). Given the low 5-year survival rate of patients with CRCLM, all patients underwent postoperative follow-up for 3 years. Overall survival (OS) was defined as the time from the date of primary colorectal cancer surgery to the date of death from any cause or last follow-up. Recurrence-free survival (RFS) was defined as the time from the date of primary colorectal cancer surgery to the date of first documented disease recurrence (local or distant) or death from any cause, whichever occurred first. Patients who were alive and recurrence-free at the last follow-up were censored.
2.3. H&E and immunohistochemistry staining
Tumor specimens from primary colorectal cancer resections were processed into formalin-fixed, paraffin-embedded (FFPE) blocks. The FFPE blocks were sectioned at 4 μm thickness for H&E staining and IHC staining. TLS were defined as organized aggregates of lymphoid cells, typically consisting of a dense cluster of small lymphocytes with or without discernible germinal center formation, located within the tumor stroma or peritumoral region. Tumor-associated TLS were defined as those located within 7 mm of the tumor boundary (including the tumor area) (28). To further characterize the cellular composition of TLS in CRCLM, immunohistochemical (IHC) staining was performed on serial sections using the following antibodies: helper T cells (CD4; Abcam, ab133616, rabbit monoclonal, 1:200), cytotoxic T cells (CD8; Abcam, ab178089, rabbit monoclonal, 1:100), regulatory T cells (FOXP3; Abcam, ab20034, mouse monoclonal, 1:500), memory T cells (CD45RO; Abcam, mouse monoclonal, 1:1,000), B cells (CD20; Abcam, ab78237, rabbit monoclonal, 1:100), dendritic cells (CD11c; Abcam, ab52632, rabbit monoclonal, 1:500), natural killer cells (NCR1; Abcam, ab224703, rabbit monoclonal, 1:1,000), follicular dendritic cells (CD21; Abcam, ab75985, rabbit monoclonal, 1:100), macrophages (CD68; Wuhan Sevicebio Technology Co., Ltd, GB113150, rabbit monoclonal, 1:500), and tumor-associated neutrophils (CD15; Santa Cruz, SC-21702, mouse monoclonal, 1:100). These markers were used to confirm the lymphoid nature and maturity of TLS, with particular attention to CD20+ B-cell aggregates and CD21+ follicular dendritic cell networks, which indicate organized TLS with germinal center formation. The presence and location of TLS were independently assessed by two pathologists, with ICH-stained sections serving as complementary validation to the H&E-based evaluation. And to explore the relationship between TLS maturity and TB, TLS were further classified into three categories based on their maturation status on H&E-stained sections: Agg (aggregates of small lymphocytes without germinal center formation), FL1 (primary follicle-like structures with early germinal center formation), and FL2 (secondary follicle-like structures with well-defined germinal centers). For each patient, the percentage of each TLS category relative to the total TLS count was calculated. A continuous TLS maturity score was then derived using the formula: (percentage of Agg × 0) + (percentage of FL1 × 1) + (percentage of FL2 × 2). Receiver operating characteristic (ROC) curve analysis for overall survival was used to determine the optimal cutoff value for dichotomizing patients into low and high TLS maturity groups (cutoff for TLS maturity was found to be 0.371). This dichotomized maturity score was subsequently included as a covariate in survival analyses. Tumor budding (TB) was assessed using both H&E staining and IHC for cytokeratin 19 (CK19; antibody ab52625, dilution 1:200, Abcam). The assessment followed the International Tumor Budding Consensus Conference (ITBCC) 2016 recommendations. The use of CK19 IHC was justified based on its well-established role as a highly sensitive and specific marker for epithelial cells, including those in colorectal cancer. This approach enhances the detection of isolated single cells or small tumor cell clusters (≤4 cells) at the invasive front, which can sometimes be difficult to distinguish definitively from stromal or inflammatory cells on H&E staining alone, thereby improving the accuracy and reproducibility of TB scoring (29). To integrate the H&E and IHC data for TB assessment, the following procedure was applied: Initially, the region of highest budding intensity at the invasive front was identified under low-power magnification on H&E-stained slides. Subsequently, consecutive sections stained for CK19 were examined in the corresponding histologic area. The TB count was ultimately performed at ×200 magnification, preferentially using the CK19-stained slides to ensure precise identification of budding foci. In cases where IHC was unavailable or equivocal, the H&E-based evaluation was retained. This integrated approach ensured that the final TB quantification leveraged the enhanced detection capability of CK19 IHC while maintaining consistency across all samples (30). TB was categorized into three grades according to ITBCC criteria: low grade (0–4 buds), intermediate grade (5–9 buds), and high grade (≥10 buds). Both TLS and TB were quantitatively analyzed in colorectal cancer tissues. The assessment of both TLS and TB was performed independently by two pathologists. Prior to formal evaluation, they underwent a calibration session using a training set of slides to align their scoring criteria. The interobserver agreement for the continuous variables (TLS density and TB count) was quantitatively assessed using the Intraclass Correlation Coefficient (ICC). The analysis demonstrated excellent agreement, with an ICC of 0.93 (p<0.001, 95% CI: 0.90-0.95) for TLS density and an ICC of 0.91 (p<0.001, 95% CI: 0.87-0.94) for TB count. For the few instances where initial assessments differed, a predefined consensus procedure was followed: the discrepant cases were reviewed by a third senior pathologist, whose judgment served as the final arbiter.
2.4. Statistical analysis
Statistical analyses were performed using R Studio (version 3.6.3) and SPSS (version 27.0). For comparisons of clinicopathological characteristics between groups, categorical variables were analyzed using the χ² test or Fisher’s exact test, as appropriate. Continuous variables were compared using Student’s t-test or ANOVA. Survival outcomes, including overall survival (OS) and recurrence-free survival (RFS), were estimated using the Kaplan-Meier method, and differences between groups were assessed using the log-rank test. To evaluate the correlations among different variables, Pearson correlation analysis was applied for normally distributed data, and Spearman rank correlation analysis was used for non-normally distributed data. Variables showing significant results in univariate analysis were further analyzed using multivariate Cox regression. For model evaluation, receiver operating characteristic (ROC) curves were generated using the pROC package in R to assess the discriminatory ability of the TLS/TB index, TLS density, TB grade, and histological differentiation for 1−year and 3−year overall survival. The area under the curve (AUC) was calculated with 95% confidence intervals. Calibration curves were constructed using the rms package with 1,000 bootstrap resamples to reduce overfitting, and the agreement between predicted and observed survival probabilities was evaluated visually. Decision curve analysis was performed using the rmda package to quantify the net clinical benefit of the nomogram across a range of threshold probabilities. R software was used for nomogram construction, calibration curve analysis, and decision curve analysis. Statistical significance was set at P < 0.05.
3. Results
3.1. Comparison of clinicopathological characteristics of CRCLM patients between training cohort and validation cohort
We retrospectively analyzed 107 CRCLM patients and these patients were used as the training cohort. 30 CRCLM patients were used as an external validation set. There was no significant difference between the training and validation cohorts in demographic and clinical characteristics (Table 1).
Table 1.
Comparison between training cohort and validation cohort of clinicopathological characteristics of CRCLM.
| Variables | Training cohort (n=107) | Validation cohort (n=30) | Statistic | P |
|---|---|---|---|---|
| Age, n (%) | χ²=0.00 | 0.992 | ||
| ≥55 | 75 (70.09) | 21 (70.00) | ||
| <55 | 32 (29.91) | 9 (30.00) | ||
| Sex, n (%) | χ²=2.16 | 0.142 | ||
| male | 28 (26.17) | 4 (13.33) | ||
| female | 79 (73.83) | 26 (86.67) | ||
| Tumor size | χ²=0.39 | 0.535 | ||
| ≤2cm | 61 (57.01) | 19 (63.33) | ||
| >2cm | 46 (42.99) | 11 (36.67) | ||
| Tumor site, n (%) | χ²=0.67 | 0.715 | ||
| Left | 41 (38.32) | 12 (40.00) | ||
| Right | 37 (34.58) | 12 (40.00) | ||
| Rectal | 29 (27.10) | 6 (20.00) | ||
| LN Metastasis, n (%) | χ²=0.05 | 0.830 | ||
| Negative | 23 (21.50) | 7 (23.33) | ||
| Positive | 84 (78.50) | 23 (76.67) | ||
| T stage, n (%) | - | 0.742 | ||
| I | 3 (2.80) | 1 (3.33) | ||
| II | 6 (5.61) | 3 (10.00) | ||
| III | 80 (74.77) | 21 (70.00) | ||
| IV | 18 (16.82) | 5 (16.67) | ||
| Ki-67, n (%) | χ²=0.96 | 0.328 | ||
| ≤60% | 36 (33.64) | 13 (43.33) | ||
| >60% | 71 (66.36) | 17 (56.67) | ||
| Differentiation, n (%) | χ²=0.03 | 0.867 | ||
| Well, Moderate | 66 (61.68) | 18 (60.00) | ||
| Poor | 41 (38.32) | 12 (40.00) | ||
| TLS density, n (%) | χ²=0.00 | 0.964 | ||
| Low | 54 (50.47) | 15 (50.00) | ||
| High | 53 (49.53) | 15 (50.00) | ||
| Surgical Margin, n (%) | χ²=0.43 | 0.512 | ||
| Negative | 102 (95.33) | 30 (100.00) | ||
| Positive | 5 (4.67) | 0 (0.00) | ||
| Vasculature Violations, n (%) | χ²=0.93 | 0.336 | ||
| Negative | 50 (46.73) | 17 (56.67) | ||
| Positive | 57 (53.27) | 13 (43.33) | ||
| Perineural Invasion, n (%) | χ²=2.17 | 0.140 | ||
| Negative | 41 (38.32) | 16 (53.33) | ||
| Positive | 66 (61.68) | 14 (46.67) | ||
| MMR, n (%) | χ²=0.00 | 1.000 | ||
| pMMR | 101 (94.39) | 28 (93.33) | ||
| dMMR | 6 (5.61) | 2 (6.67) | ||
| TB, n (%) | χ²=1.76 | 0.414 | ||
| Low | 34 (31.78) | 13 (43.33) | ||
| Moderate | 53 (49.53) | 11 (36.67) | ||
| High | 20 (18.69) | 6 (20.00) | ||
| Neoadjuvant therapy type, n (%) | χ²=0.18 | 0.914 | ||
| None | 52 (48.60) | 14 (46.67) | ||
| Chemotherapy | 35 (32.71) | 11 (36.67) | ||
| Chemoradiotherapy | 20 (18.69) | 5 (16.67) | ||
| Metastasis size, n (%) | χ²=0.00 | 0.964 | ||
| ≤5cm | 53 (49.53) | 15 (50.00) | ||
| >5cm | 54 (50.47) | 15 (50.00) | ||
| Metastasis number, n (%) | χ²=0.70 | 0.404 | ||
| ≤3 | 55 (51.40) | 18 (60.00) | ||
| >3 | 52 (48.60) | 12 (40.00) | ||
| TLS maturity | χ²=1.53 | 0.216 | ||
| Low | 60 (56.07) | 13 (43.33) | ||
| High | 47 (43.93) | 17 (56.67) |
3.2. Correlation of TLSs and TBs with Prognosis and relationships between the TLSs and TBs in CRCLM
Representative H&E-stained images of TLS and TB in primary colorectal cancer tissues from patients with CRCLM are shown in Figures 2A, B, respectively. Based on histopathological quantification through H&E and IHC, TLS was stratified into low-density/high-density groups and low-count/high-count groups using median cutoffs. Using Kaplan-Meier analysis with the log-rank test, prognostic analysis revealed that both TLS high-density and high-count TLS groups exhibited superior overall survival (OS) and recurrence-free survival (RFS) (Figures 3A, B), whereas higher tumor budding grade correlated with worse clinical outcomes (Figure 3C). Based on these findings, we hypothesized potential interactions between TLS and TB. Subsequent correlation analysis demonstrated significant negative correlations between the TLS parameters (maturity/density/count) and TB counts (Figure 3D). We further explored the relationship between TB and TLS maturity subtypes (Agg, FL1, FL2). As shown in Figure 4, the distribution of TLS maturity subtypes differed significantly across TB grades: higher TB grades were associated with a higher proportion of Agg (immature aggregates) and a lower proportion of FL2 (mature TLS with germinal centers). The TLS maturity score, calculated as (percentage of Agg × 0) + (percentage of FL1 × 1) + (percentage of FL2 × 2), also showed a significant negative correlation with TB count (r = -0.401, p < 0.01).
Figure 2.
H&E and IHC sections showing TLS and TB. (A) H&E‐stained images of TLS in primary colorectal cancer tissues from patients with CRCLM (scale bar = 400 μm). (B) IHC‐stained images of TB in primary colorectal cancer tissues from patients with CRCLM (scale bar = 200 μm). TB, tumor budding; TLS, tertiary lymphoid structure; H&E, Hematoxylin and eosin; IHC, Immunohistochemistry; CRCLM, colorectal cancer liver metastases.
Figure 3.
Correlation of TLSs and TBs with prognosis. (A) The Kaplan-Meier curves demonstrate differences in OS and RFS across varying TLS density. (B) The Kaplan-Meier curves demonstrate differences in OS and RFS across varying TLS numbers. (C) Kaplan-Meier curves demonstrate differences in OS and RFS across varying TB grade. (D) Scatter plots indicated the relationship between TLS and TB. OS, overall survival; RFS, recurrence-free survival; TLS, tertiary lymphoid structure; TB, tumor budding.
Figure 4.
H&E sections showing TLS and different immune cells in TLS. (A) Representative figure of H&E slides to show TLS. (B-K) IHC results showed the different immune cells, which formed the TLS, including (B) CD4 + T cells, (C) CD8 + T cells, (D) CD20 + B cells, (E) CD45RO+ memory T cells, (F) CD21 + follicular dendritic cells, (G) CD11c + dendritic cells, (H) CD15 + granulocytes, (I) CD68 + macrophages, (J) FOXP3 + Treg cells, and (K) NCR1 + natural killer cells. (L) Pie chart figure showed the distribution of various immune cells forming the TLS. (M–O) column chart showed different TLS maturity status between different TB grade. (P) Scatter plots indicated the relationship between TLS maturity and TB counts. TB, tumor budding; TLS, tertiary lymphoid structure; H&E, Hematoxylin and eosin; IHC, Immunohistochemistry. *: p < 0.05; **: p < 0.01; ***: p < 0.001.
3.3. Clinical significance of the TLS/TB index in CRCLM patients
Current research on TLS has advanced rapidly, with prevailing evidence suggesting that TLS reflects host immune defense mechanisms and predicts the prognosis of patients with CRCLM. Conversely, TB, as a manifestation of epithelial-mesenchymal transition (EMT), is widely regarded as an indicator of tumor aggressiveness, and higher TB grades correlate with increased invasiveness and poorer clinical outcomes. While various biomarkers have been explored in CRCLM, limited studies have focused on an integrative metric that specifically combines the quantitative density of TLS and TB. We therefore proposed a novel TLS/TB index to evaluate the potential balance between tumor-cell invasive patterns and organized lymphoid formations. To address this, we propose a novel TLS/TB index designed to quantify the balance between local immune response and tumor invasiveness, with the goal of improving prognostic stratification for CRCLM patients.
Correlation analysis revealed a stronger inverse association between TLS density and TB count (r = -0.581, p < 0.001) than between TLS and TB counts or between TLS maturity and TB count. To standardize units (given a fixed microscopic field area of 0.785 mm² for TB quantification), both TLS and TB metrics were converted to density measurements. Specifically, since TB was assessed within a fixed microscopic field area of 0.785 mm², TB counts were converted to TB density (number/mm²) by dividing the count by the area of the field. The TLS/TB index was then defined as the ratio of TLS density (number/mm²) to TB density (number/mm²). This ratio is a dimensionless quantity. Using the median cutoff value (0.0041), patients were stratified into high and low TLS/TB groups (representative H&E/IHC images in Figure 5A). Significant intergroup differences were observed in tumor location (p = 0.046), Liver metastasis number (p=0.015) and histological differentiation (p < 0.001). Kaplan-Meier analysis (Figure 5B) demonstrated superior survival and reduced recurrence in the high TLS/TB index group. Furthermore, the high TLS/TB index group was more frequently associated with left-sided colon and rectal cancers, exhibited better histological differentiation and less Liver metastasis number (Table 2). Univariate analysis (using a threshold of p < 0.05 for inclusion in multivariate analysis) was performed. Prior to conducting multivariate analysis, we assessed the variance inflation factors (VIFs) to evaluate potential multicollinearity among TLS density, TB grade, and the TLS/TB index (31). The VIF values were 2.257 for TLS density, 2.039 for TB grade, and 3.470 for the TLS/TB index, all well below the common threshold of 5, indicating that multicollinearity was not a critical concern in our model. Subsequently, multivariate analysis identified histological differentiation, TLS density, TB grade, and the TLS/TB index as independent prognostic factors for patients with advanced colorectal cancer liver metastases (Table 3). Interestingly, for RFS, only TB grade was identified as an independent predictor, whereas TLS density did not retain independent significance (Table 4). This divergence may reflect the differential roles of these features in the tumor microenvironment. High TLS density likely contributes to a more robust antitumor immune response, which may predominantly impact long-term survival by controlling disease progression, rather than directly preventing early recurrence events that are more strongly driven by tumor invasiveness. Conversely, higher TB grade, reflecting enhanced invasive and migratory capacity, directly increases the risk of early recurrence and also contributes to poorer overall survival.
Figure 5.
Correlation of the TLS/TB index with prognosis. (A) Co‐localization H&E-stained and IHC-stained images of high and low TLS/TB index in primary colorectal cancer tissues from patients with CRCLM. (scale bar = 400 μm). (B) The Kaplan-Meier curves demonstrate differences in OS and RFS across TLS/TB index. OS, overall survival; RFS, recurrence-free survival; TLS, tertiary lymphoid structure; TB, tumor budding; H&E, Hematoxylin and eosin; IHC, Immunohistochemistry.
Table 2.
Clinicopathological characteristics between TLS/TBhigh and TLS/TBlow group of CRCLM patients.
| Variables | TLS/TB index | P | |
|---|---|---|---|
| Low(n = 57) | High(n = 50) | ||
| Age, n (%) | 0.687 | ||
| <55 | 18 (31.58) | 14 (28.00) | |
| ≥55 | 39 (68.42) | 36 (72.00) | |
| Gender, n (%) | 0.686 | ||
| Male | 43 (75.44) | 36 (72.00) | |
| Female | 14 (24.56) | 14 (28.00) | |
| Tumor Site, n (%) | 0.046 | ||
| Right | 28 (49.12) | 13 (26.00) | |
| Left | 17 (29.82) | 20 (40.00) | |
| Rectum | 12 (21.05) | 17 (34.00) | |
| Tumor Size, n (%) | 0.329 | ||
| <5cm | 30 (52.63) | 31 (62.00) | |
| ≥5cm | 27 (47.34) | 19 (38.00) | |
| Perineural Invasion, n (%) | 0.737 | ||
| Negative | 21 (36.84) | 20 (40.00) | |
| Positive | 36 (63.16) | 30 (60.00) | |
| Vasculature Violations, n (%) | 0.525 | ||
| Negative | 25 (43.86) | 25 (50.00) | |
| Positive | 32 (56.14) | 25 (50.00) | |
| LN Metastasis, n (%) | 0.288 | ||
| Negative | 10 (17.54) | 13 (26.00) | |
| Positive | 47 (82.46) | 37 (74.00) | |
| Differentiation,n (%) | <0.001 | ||
| Well, Moderate | 26 (45.61) | 40 (80.00) | |
| Poor | 31 (54.39) | 10 (20.00) | |
| MMR, n (%) | 1.000 | ||
| dMMR | 54 (94.39) | 47 (94.00) | |
| pMMR | 6 (5.61) | 3 (6.00) | |
| T stage, n (%) | 0.510 | ||
| T1 | 2 (3.51) | 1 (2.00) | |
| T2 | 2 (3.51) | 4 (8.00) | |
| T3 | 41 (71.93) | 39 (78.00) | |
| T4 | 12 (21.05) | 6 (12.00) | |
| Ki-67, n (%) | 0.455 | ||
| ≤60% | 21 (36.84) | 15 (30.00) | |
| >60% | 36 (63.16) | 35 (70.00) | |
| Neoadjuvant therapy type, n (%) | 0.793 | ||
| None | 20(50.88) | 23 (46.00) | |
| Chemotherapy | 17(29.82) | 18(36.00) | |
| Chemoradiotherapy | 11(19.30) | 9(18.00) | |
| Metastasis size, n (%) | 0.210 | ||
| ≤5cm | 25 (43.86) | 28 (56.00) | |
| >5cm | 32 (56.14) | 22 (44.00) | |
| Metastasis number, n (%) | 0.015 | ||
| ≤3 | 23 (40.35) | 32 (64.00) | |
| >3 | 34 (59.65) | 18 (36.00) | |
p-values less than 0.05 are shown in bold.
Table 3.
Univariate and multivariate analysis for OS using Cox regression in CRCLM.
| Characteristics | Univariate analysis | Multivariate analysis | ||
|---|---|---|---|---|
| Hazard ratio (95% CI) | P-value | Hazard ratio (95% CI) | P-value | |
| Age | 0.614 | |||
| ≥55 | Reference | |||
| <55 | 1.15 (0.68 ~ 1.94) | |||
| Sex | 0.546 | |||
| male | 1.00 (Reference) | |||
| female | 1.19 (0.68 ~ 2.09) | |||
| Tumor site | ||||
| Right | 1.00 (Reference) | 1.00 (Reference) | ||
| Left | 0.69 (0.40 ~ 1.19) | 0.180 | 0.90 (0.49 ~ 1.64) | 0.725 |
| Rectal | 0.45 (0.23 ~ 0.86) | 0.017 | 0.62 (0.30 ~ 1.28) | 0.196 |
| Tumor size | 0.357 | |||
| <2cm | 1.00 (Reference) | |||
| ≥2cm | 1.26 (0.77 ~ 2.05) | |||
| Differentiation | <0.001 | <0.001 | ||
| Well, Moderate | 1.00 (Reference) | 1.00 (Reference) | ||
| Poor | 3.45 (2.09 ~ 5.69) | 2.70 (1.52 ~ 4.78) | ||
| Perineural Invasion | 0.160 | |||
| Negative | 1.00 (Reference) | |||
| Positive | 1.45 (0.86 ~ 2.44) | |||
| Vasculature Violations | 0.286 | |||
| Negative | 1.00 (Reference) | |||
| Positive | 1.31 (0.80 ~ 2.14) | |||
| LN Metastasis | 0.057 | |||
| Negative | 1.00 (Reference) | |||
| Positive | 1.98 (0.98 ~ 4.01) | |||
| MMR status | 0.441 | |||
| dMMR | 1.00 (Reference) | |||
| pMMR | 1.49 (0.54 ~ 4.10) | |||
| T stage | 0.768 | |||
| I | 1.00 (Reference) | |||
| II | 0.12 (0.01 ~ 1.20) | 0.072 | ||
| III | 0.62 (0.19 ~ 2.00) | 0.426 | ||
| IV | 0.75 (0.21 ~ 2.65) | 0.660 | ||
| TB | <0.001 | |||
| Low | 1.00 (Reference) | 1.00 (Reference) | ||
| Moderate | 11.92 (4.25 ~ 33.46) | <0.001 | 4.62 (1.46 ~ 14.55) | 0.009 |
| High | 44.12 (14.64 ~132.96) | <0.001 | 7.95 (2.23 ~ 28.36) | 0.001 |
| TLS density | <0.001 | 0.030 | ||
| Low | 1.00 (Reference) | 1.00 (Reference) | ||
| High | 0.15 (0.09 ~ 0.27) | 0.47 (0.24 ~ 0.93) | ||
| TLS/TB index | <0.001 | 0.003 | ||
| Low | 1.00 (Reference) | 1.00 (Reference) | ||
| High | 0.07 (0.03 ~ 0.14) | 0.25 (0.10 ~ 0.62) | ||
| Ki67 | ||||
| ≥60% | 1.00 (Reference) | |||
| <60% | 1.17 (0.69 ~ 1.99) | |||
| Neoadjuvant therapy type, n (%) | 0.733 | |||
| None | 1.00 (Reference) | |||
| Chemotherapy | 0.81 (0.47 ~ 1.41) | 0.453 | ||
| Chemoradiotherapy | 0.85 (0.43 ~ 1.68) | 0.648 | ||
| Metastasis size, n (%) | 0.455 | |||
| ≤5cm | 1.00 (Reference) | |||
| >5cm | 1.19 (0.75 ~ 1.90) | |||
| Metastasis number, n (%) | 0.033 | 0.366 | ||
| ≤3 | 1.00 (Reference) | 1.00 (Reference) | ||
| >3 | 1.71 (1.05 ~ 2.79) | 1.30 (0.73 ~ 2.32) | ||
| TLS maturity | 0.046 | 0.618 | ||
| Low | 1.00 (Reference) | 1.00 (Reference) | ||
| High | 0.60 (0.36 ~ 0.99) | 0.87 (0.51 ~ 1.49) | ||
p-values less than 0.05 are shown in bold.
Table 4.
Univariate and multivariate analysis for RFS using Cox regression in CRCLM.
| Characteristics | Univariate analysis | Multivariate analysis | ||
|---|---|---|---|---|
| Hazard ratio (95% CI) | P-value | Hazard ratio (95% CI) | P-value | |
| Age | 0.614 | |||
| <55 | 1.00 (Reference) | |||
| ≥55 | 1.34 (0.82 ~ 2.20) | |||
| Sex | 0.546 | |||
| male | 1.00 (Reference) | |||
| female | 1.13 (0.66 ~ 1.94) | |||
| Tumor site | ||||
| Right | 1.00 (Reference) | 1.00 (Reference) | ||
| Left | 0.67 (0.39 ~ 1.13) | 0.131 | 0.73 (0.41 ~ 1.30) | 0.279 |
| Rectal | 0.42 (0.22 ~ 0.78) | 0.006 | 0.55 (0.27 ~ 1.12) | 0.098 |
| Tumor size | 0.460 | |||
| <2cm | 1.00 (Reference) | |||
| ≥2cm | 1.19 (0.75 ~ 1.91) | |||
| Differentiation | <0.001 | 0.002 | ||
| Well, Moderate | 1.00 (Reference) | 1.00 (Reference) | ||
| Poor | 3.54 (2.19 ~ 5.73) | 2.50 (1.41 ~ 4.45) | ||
| Perineural Invasion | 0.142 | |||
| Negative | 1.00 (Reference) | |||
| Positive | 1.45 (0.88 ~ 2.39) | |||
| Vasculature Violations | 0.150 | |||
| Negative | 1.00 (Reference) | |||
| Positive | 1.42 (0.88 ~ 2.27) | |||
| LN Metastasis | 0.031 | 0.051 | ||
| Negative | 1.00 (Reference) | 1.00 (Reference) | ||
| Positive | 2.09 (1.07 ~ 4.09) | 2.14 (1.00 ~ 4.58) | ||
| MMR status | 0.606 | |||
| dMMR | 1.00 (Reference) | |||
| pMMR | 1.30 (0.48 ~ 3.58) | |||
| T stage | 0.768 | |||
| I | 1.00 (Reference) | |||
| II | 0.10 (0.01 ~ 1.00) | 0.050 | ||
| III | 0.61 (0.19 ~ 1.97) | 0.409 | ||
| IV | 0.78 (0.22 ~ 2.72) | 0.696 | ||
| TB | <0.001 | |||
| Low | 1.00 (Reference) | 1.00 (Reference) | ||
| Moderate | 12.50 (4.92 ~ 31.77) | <0.001 | 6.46 (2.24 ~ 18.64) | <0.001 |
| High | 38.48 (13.96 ~ 106.05) | <0.001 | 12.86 (3.72 ~ 44.44) | <0.001 |
| TLS density | <0.001 | 0.128 | ||
| Low | 1.00 (Reference) | 1.00 (Reference) | ||
| High | 0.09 (0.05 ~ 0.16) | 0.56 (0.27 ~ 1.18) | ||
| TLS/TB index | <0.001 | 0.022 | ||
| Low | 1.00 (Reference) | 1.00 (Reference) | ||
| High | 0.07 (0.03 ~ 0.14) | 0.36 (0.15 ~ 0.86) | ||
| Ki67 | ||||
| <60% | 1.00 (Reference) | |||
| ≥60% | 1.06 (0.65 ~ 1.75) | |||
| Neoadjuvant therapy type, n (%) | ||||
| None | 1.00 (Reference) | |||
| Chemotherapy | 0.99 (0.59 ~ 1.65) | 0.961 | ||
| Chemoradiotherapy | 0.81 (0.41 ~ 1.59) | 0.541 | ||
| Metastasis size, n (%) | 0.849 | |||
| ≤5cm | 1.00 (Reference) | |||
| >5cm | 1.05 (0.64 ~ 1.71) | |||
| Metastasis number, n (%) | 0.029 | 0.409 | ||
| ≤3 | 1.00 (Reference) | 1.00 (Reference) | ||
| >3 | 1.69 (1.06 ~ 2.70) | 1.26 (0.73 ~ 2.15) | ||
| TLS maturity | 0.046 | 0.901 | ||
| Low | 1.00 (Reference) | 1.00 (Reference) | ||
| High | 0.61 (0.38 ~ 0.99) | 0.97 (0.58 ~ 1.62) | ||
p-values less than 0.05 are shown in bold.
3.4. Construction and validation of the nomogram
Based on the results of the univariate and multivariate analyses in the training cohort, a nomogram clinical prediction model was constructed by assigning individual scores to each independent risk factor. The total score, obtained by summing these individual scores, corresponds to the model-predicted probability of overall survival (OS). Histological differentiation, TLS density, TB grade, and TLS/TB index were identified as independent risk factors for OS in patients with CRCLM. These factors were integrated into a nomogram for colorectal cancer liver metastases (Figure 6A). The calibration curve demonstrated excellent consistency between the nomogram-predicted outcomes and actual observations (Figure 6B). The nomogram was subsequently applied to the independent external validation cohort (n=30). The ROC curves for the validation cohort showed good consistency with those of the training cohort for both 1−year and 3−year survival (Figures 7A, B), and the calibration curves similarly demonstrated good agreement between predicted and observed survival (Figures 7C, D), suggesting the model’s potential generalizability and clinical utility.
Figure 6.
Construction and validation of the nomogram for CRCLM. (A) The nomogram to predict 1-year, 2-year and 3-year survival probability for CRCLM. (B) The calibration curves of 1-year and 3-year survival probability in CRCLM. (C) The ROC curve of 1-year survival probability in CRCLM, comparing the risk-score model with Differentiation, TLS index, TB and TLS/TB index alone. (D) The ROC curve of 3-year survival probability in CRCLM, comparing the risk-score model with Differentiation, TLS index, TB and TLS/TB index alone. (E) Decision curve analysis based on the predictive model of the TLS‐TB profile and the Differentiation of patients with CRCLM. CRCLM, colorectal cancer liver metastases; TLS, tertiary lymphoid structure; TB, tumor budding; ROC, Receiver Operating Characteristic.
Figure 7.
Model performance in training and validation cohorts. (A) The ROC curve of 1-year survival probability in CRCLM, comparing the training cohort with validation cohort. (B) The ROC curve of 3-year survival probability in CRCLM, comparing the training cohort with validation cohort. (C) Calibration curves for 1-year survival probability of CRCLM, comparing the training and validation cohorts. (D) Calibration curves for 3-year survival probability of CRCLM, comparing the training and validation cohorts. CRCLM, colorectal cancer liver metastases; ROC, Receiver Operating Characteristic.
3.5. Comparison of the accuracy and clinical utility between the two models
Although histological differentiation and the TLS/TB index demonstrate independent prognostic significance in patients with advanced colorectal cancer, it remains unclear which of these two models better predicts outcomes in patients with CRCLM. ROC curve in the training cohort analysis revealed that the TLS/TB index (AUC = 0.89) outperformed the risk stratification models based solely on histological differentiation (AUC = 0.70), TLS density (AUC = 0.80) and TB grade (AUC = 0.84) (Figures 6C, D), indicating superior predictive precision. Furthermore, decision curve analysis confirmed that our index provided a greater net clinical benefit compared to histological differentiation (Figure 6E).
4. Discussion
Colorectal cancer liver metastasis (CRCLM) is a highly aggressive malignancy, and current prognostic assessments rely on the integration of multidimensional biological characteristics (32). In recent years, tertiary lymphoid structures (TLS) have garnered significant attention as biomarkers of antitumor immune responses (28). This study further confirms that high TLS density correlates with prolonged overall survival (OS) (33, 34), consistent with prior research indicating that TLS enhances antitumor immunity by recruiting cytotoxic T cells and promoting TLS maturation (35, 36). We also evaluated TLS maturity using a quantitative scoring system based on germinal center formation. Although higher maturity scores were associated with better survival in univariate analysis, this effect was not independent after adjusting for TLS density, suggesting that the overall density of TLS—rather than their maturity status per se—may be the dominant prognostic factor in our synchronous CRCLM cohort. However, the tumor microenvironment is not governed by a single factor. Tumor budding (TB), a morphological manifestation of epithelial-mesenchymal transition (EMT), enhances the capacity of tumor cells to detach from the primary site and initiate metastasis. Our findings demonstrated that higher TB counts are associated with poorer patient prognosis (37). Furthermore, correlation analysis revealed an inverse relationship between TLS density and TB density. The observed inverse correlation between TLS and TB may stem from their opposing roles within the tumor microenvironment. Currently, there is a lack of research on the biological mechanisms underlying their negative correlation. We hypothesize that, on one hand, a robust TLS-mediated immune response—characterized by the production of cytokines and the activity of cytotoxic lymphocytes—may directly suppress the outgrowth and dissemination of tumor cell clusters at the invasive front, thereby inhibiting TB formation. On the other hand, an aggressive TB phenotype, driven by EMT and associated with an immunosuppressive microenvironment, may actively inhibit the recruitment and function of lymphoid cells, impairing the initiation and maturation of TLS. This dynamic seesaw between local immune surveillance and tumor invasiveness is captured by the TLS/TB index. This hypothesis requires further experimental validation.
The integration of TLS and TB into a unified index provides a hypothesis−driven integrative morphologic metric reflecting the balance between host immune organization and the invasive phenotype. Given that direct mechanistic evidence linking these morphologic features remains limited, this index serves as a surrogate tool to capture the local microenvironmental equilibrium in clinical practice. The results demonstrate that the TLS/TB index effectively stratifies patients into distinct prognostic subgroups: those with a high ratio may exhibit an “immune-active” status, while those with a low ratio tend toward an “invasion-dominant” phenotype. This model should be interpreted as a complementary prognostic tool rather than a definitive measure of underlying biological mechanisms. From a clinical translation perspective, the stratification capability of the TLS/TB index may inform personalized postoperative treatment strategies. As a hypothesis-generating observation, patients with a high TLS/TB ratio, indicative of a pre-existing, active anti-tumor immune contexture, might be hypothesized to derive greater benefit from immunotherapy. Conversely, patients with a low TLS/TB ratio, signaling dominant invasive and metastatic potential, may require intensified or novel adjuvant chemotherapy regimens targeting EMT and metastasis-initiating cells, or be considered for more frequent surveillance. It is important to note that the current cohort did not include patients treated with immunotherapy, and no response data were available. Therefore, these speculations should be interpreted with caution and require prospective validation in future studies. This model provides a rationale for future studies to test these hypotheses in clinical trials. Building upon our group’s prior findings identifying TLS as a protective factor and TB as an independent risk factor for colorectal cancer, this study introduces an integrative approach through the TLS/TB index. By quantifying this dynamic equilibrium, we developed a comprehensive model for predicting survival in patients with CRCLM. Our findings reconfirm the tumor-promoting role of TB and tumor-suppressive function of TLS. In this study, the TLS/TB index was identified as an independent prognostic factor, showing promising predictive performance not only over individual TLS density or TB grade, but also over the conventional histological differentiation parameter. These results underscore that the interplay between immune defense and invasion-metastasis mechanisms within the tumor microenvironment may play a pivotal role in determining recurrence and survival outcomes in patients with advanced cancer patients.
This study has several limitations. First, the retrospective multi-center design, while valuable, may introduce selection bias, necessitating further validation of the model’s generalizability through prospective multicenter cohorts. Second, the current quantification of TLS and TB relies on semi-quantitative pathological assessment, which could be enhanced by integrating digital pathology and artificial intelligence algorithms to improve standardization. Third, TLS were quantified based on density without subclassification by maturity (e.g., presence of germinal centers) or spatial localization (intratumoral vs. peritumoral). Given that TLS maturity has been associated with differential immune functions and prognostic implications, this represents a limitation that may affect biological interpretation. Forth, the TLS/TB index cannot fully represent the complex interactions among myriad components within the tumor microenvironment, with the roles of other elements such as fibroblasts, endothelial cells, and additional immune cells requiring further investigation. Fifth, while TB was categorized according to standardized ITBCC criteria, TLS density and the TLS/TB index were dichotomized using median cutoffs in the absence of established clinical thresholds, a limitation that future studies with larger cohorts may address by identifying optimal cutoffs. Additionally, while we developed a quantitative TLS maturity score, its lack of independent prognostic significance may be attributable to sample size limitations. Future studies with larger cohorts are needed to clarify the relative contributions of TLS density versus maturity in CRCLM.
In summary, the TLS/TB index provides a novel perspective for prognostic prediction of CRCLM by integrating bidirectional regulatory information along the “immunity-invasion” axis. Future research should focus on exploring the associations between this index and genomic features (e.g., microsatellite instability), as well as treatment responses to facilitate its translation into clinical practice.
5. Conclusion
These results demonstrate that the TLS/TB index serves as an integrative prognostic metric that reflects the balance between immune organization and invasive phenotype in CRCLM. This model may facilitate more granular risk stratification when used in conjunction with conventional clinicopathological parameters. After adjustment for established clinical covariates, the index remained independently associated with patient outcomes. External validation provided preliminary support for its predictive accuracy and generalizability, and its performance compared favorably with that of conventional prognostic parameters in this cohort. The integration of this index into clinical models may enhance risk stratification and facilitate personalized therapeutic strategies for patients with synchronous CRCLM.
Funding Statement
The author(s) declared that financial support was received for this work and/or its publication. This work was supported by the Nanjing Drum Tower Hospital Clinical Research Special Fund (2022-LCYJ-MS-26, Q Li).
Footnotes
Edited by: Xiao Liang, Sichuan University, China
Reviewed by: Junfeng Du, Seventh Medical Center of PLA General Hospital, China
Yinghao Lyu, Sichuan University, China
Data availability statement
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding authors.
Ethics statement
The studies involving humans were approved by The Committee on Medical Ethics of Nanjing Drum Tower Hospital. The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation was not required from the participants or the participants’ legal guardians/next of kin in accordance with the national legislation and institutional requirements. Written informed consent was obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.
Author contributions
JZ: Conceptualization, Data curation, Methodology, Validation, Formal analysis, Software, Writing – original draft. HZ: Data curation, Methodology, Software, Writing – original draft. LQ: Methodology, Software, Writing – original draft, Formal analysis. ML: Formal analysis, Software, Writing – original draft. JM: Investigation, Project administration, Supervision, Validation, Writing – review & editing. HTZ: Supervision, Validation, Writing – review & editing, Conceptualization, Data curation, Resources, Software. XS: Data curation, Validation, Writing – review & editing, Formal analysis. YM: Data curation, Project administration, Supervision, Validation, Writing – review & editing, Software. QL: Conceptualization, Data curation, Methodology, Project administration, Resources, Supervision, Validation, Writing – review & editing.
Conflict of interest
The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Generative AI statement
The author(s) declared that generative AI was not used in the creation of this manuscript.
Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
References
- 1. Bray F, Laversanne M, Sung H, Ferlay J, Siegel RL, Soerjomataram I, et al. Global cancer statistics 2022: Globocan estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA: A Cancer J For Clin. (2024) 74:229–63. doi: 10.3322/caac.21834. PMID: [DOI] [PubMed] [Google Scholar]
- 2. Shin AE, Giancotti FG, Rustgi AK. Metastatic colorectal cancer: Mechanisms and emerging therapeutics. Trends Pharmacol Sci. (2023) 44:222–36. doi: 10.1016/j.tips.2023.01.003. PMID: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Leiphrakpam PD, Newton R, Anaya DA, Are C. Evolution and current trends in the management of colorectal cancer liver metastasis. Minerva Surg. (2024) 79:455–69. doi: 10.23736/s2724-5691.24.10363-2. PMID: [DOI] [PubMed] [Google Scholar]
- 4. Capdevila J, Saura C, Macarulla T, Casado E, Ramos FJ, Tabernero J. Monoclonal antibodies in the treatment of advanced colorectal cancer. Eur J Surg Oncol (EJSO). (2007) 33:S24–34. doi: 10.1016/j.ejso.2007.09.025. PMID: [DOI] [PubMed] [Google Scholar]
- 5. Jin Z, Li Y, Yi H, Wang M, Wang C, Du S, et al. Pathogenetic development, diagnosis and clinical therapeutic approaches for liver metastasis from colorectal cancer (review). Int J Oncol. (2025) 66. doi: 10.3892/ijo.2025.5728. PMID: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Calderon Novoa F, Ardiles V, de Santibañes E, Pekolj J, Goransky J, Mazza O, et al. Pushing the limits of surgical resection in colorectal liver metastasis: How far can we go? Cancers. (2023) 15. doi: 10.3390/cancers15072113. PMID: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Chen Y, Liang Z, Lai M. Targeting the devil: Strategies against cancer-associated fibroblasts in colorectal cancer. Trans Res. (2024) 270:81–93. doi: 10.1016/j.trsl.2024.04.003. PMID: [DOI] [PubMed] [Google Scholar]
- 8. Wu K, Zhang G, Shen C, Zhu L, Yu C, Sartorius K, et al. Role of T cells in liver metastasis. Cell Death Dis. (2024) 15:341. doi: 10.1038/s41419-024-06726-2. PMID: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Xiao Y, Yu D. Tumor microenvironment as a therapeutic target in cancer. Pharmacol Ther. (2021) 221:107753. doi: 10.1016/j.pharmthera.2020.107753. PMID: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Morabito M, Thibodot P, Gigandet A, Compagnon P, Toso C, Berishvili E, et al. Liver extracellular matrix in colorectal liver metastasis. Cancers. (2025) 17. doi: 10.3390/cancers17060953. PMID: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Zhou S-N, Pan W-T, Pan M-X, Luo Q-Y, Zhang L, Lin J-Z, et al. Comparison of immune microenvironment between colon and liver metastatic tissue in colon cancer patients with liver metastasis. Digestive Dis Sci. (2020) 66:474–82. doi: 10.1007/s10620-020-06203-8. PMID: [DOI] [PubMed] [Google Scholar]
- 12. Khan SU, Fatima K, Malik F, Kalkavan H, Wani A. Cancer metastasis: Molecular mechanisms and clinical perspectives. Pharmacol Ther. (2023) 250:108522. doi: 10.1016/j.pharmthera.2023.108522. PMID: [DOI] [PubMed] [Google Scholar]
- 13. Huang L-B, Yang T-H, Chen H-N. Predictive role of tumor budding in T1 colorectal cancer lymph node metastasis. Gastroenterology. (2021) 161:732–3. doi: 10.1053/j.gastro.2020.12.053. PMID: [DOI] [PubMed] [Google Scholar]
- 14. Lugli A, Zlobec I, Berger MD, Kirsch R, Nagtegaal ID. Tumour budding in solid cancers. Nat Rev Clin Oncol. (2020) 18:101–15. doi: 10.1038/s41571-020-0422-y. PMID: [DOI] [PubMed] [Google Scholar]
- 15. Mao Y, Wang X, Xi L, Dong M, Song P, Miao J, et al. Prediction values of tertiary lymphoid structures in the prognosis of patients with left- and right-sided colon cancer: A multicenter propensity score-matched study. Int J Surg. (2023) 109:2344–58. doi: 10.1097/js9.0000000000000483. PMID: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Hatthakarnkul P, Quinn Jean A, Ammar A, Lynch G, Van Wyk H, McMillan Donald C, et al. Molecular mechanisms of tumour budding and its association with microenvironment in colorectal cancer. Clin Sci. (2022) 136:521–35. doi: 10.1042/cs20210886. PMID: [DOI] [PubMed] [Google Scholar]
- 17. Qu Q, Wu D, Li Z, Yin H. Tumor budding and the prognosis of patients with metastatic colorectal cancer: A meta-analysis. Int J Colorectal Dis. (2023) 38:141. doi: 10.1007/s00384-023-04423-8. PMID: [DOI] [PubMed] [Google Scholar]
- 18. Yang F, Yang J, Wu M, Chen C, Chu X. Tertiary lymphoid structures: New immunotherapy biomarker. Front Immunol. (2024) 15:1394505. doi: 10.3389/fimmu.2024.1394505. PMID: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Zhao L, Jin S, Wang S, Zhang Z, Wang X, Chen Z, et al. Tertiary lymphoid structures in diseases: Immune mechanisms and therapeutic advances. Signal Transduct Target Ther. (2024) 9:225. doi: 10.1038/s41392-024-01947-5. PMID: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Zhang C, Wang X-Y, Zuo J-L, Wang X-F, Feng X-W, Zhang B, et al. Localization and density of tertiary lymphoid structures associate with molecular subtype and clinical outcome in colorectal cancer liver metastases. J Immunother Cancer. (2023) 11. doi: 10.1136/jitc-2022-006425. PMID: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Varghese A, Hess SM, Chilakapati S, Conejo-Garcia JR, McGray AJR, Zsiros E. Tertiary lymphoid structures: Exploring opportunities to improve immunotherapy in ovarian cancer. Front Immunol. (2025) 16:1473969. doi: 10.3389/fimmu.2025.1473969. PMID: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Langouo Fontsa M, Padonou F, Willard-Gallo K. Tumor-associated tertiary lymphoid structures in cancer: Implications for immunotherapy. Expert Rev Clin Immunol. (2024) 20:839–47. doi: 10.1080/1744666x.2024.2380892. PMID: [DOI] [PubMed] [Google Scholar]
- 23. Yu A, Cao M, Zhang K, Yang Y, Ma L, Zhang X, et al. The prognostic value of the tertiary lymphoid structure in gastrointestinal cancers. Front Immunol. (2023) 14:1256355. doi: 10.3389/fimmu.2023.1256355. PMID: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Cui X, Gu X, Li D, Wu P, Sun N, Zhang C, et al. Tertiary lymphoid structures as a biomarker in immunotherapy and beyond: Advancing towards clinical application. Cancer Lett. (2025) 613:217491. doi: 10.1016/j.canlet.2025.217491. PMID: [DOI] [PubMed] [Google Scholar]
- 25. Peyraud F, Guegan J-P, Vanhersecke L, Brunet M, Teyssonneau D, Palmieri L-J, et al. Tertiary lymphoid structures and cancer immunotherapy: From bench to bedside. Med. (2025) 6:100546. doi: 10.1016/j.medj.2024.10.023. PMID: [DOI] [PubMed] [Google Scholar]
- 26. Wang Y, Zhong X, He X, Hu Z, Huang H, Chen J, et al. Liver metastasis from colorectal cancer: Pathogenetic development, immune landscape of the tumour microenvironment and therapeutic approaches. J Exp Clin Cancer Res. (2023) 42:177. doi: 10.1186/s13046-023-02729-7. PMID: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Riaz AA, Ginimol M, Rasha R, Ahmed K, Ahmed A, Catrin S, et al. Revised strengthening the reporting of cohort, cross-sectional and case-control studies in surgery (Strocss) guideline: An update for the age of artificial intelligence. Premier J Sci. (2025) 2. doi: 10.70389/PJS.100081 [DOI] [Google Scholar]
- 28. Schumacher TN, Thommen DS. Tertiary lymphoid structures in cancer. Science. (2022) 375:eabf9419. doi: 10.1126/science.abf9419. PMID: [DOI] [PubMed] [Google Scholar]
- 29. Satoh K, Nimura S, Aoki M, Hamasaki M, Koga K, Iwasaki H, et al. Tumor budding in colorectal carcinoma assessed by cytokeratin immunostaining and budding areas: Possible involvement of C-Met. Cancer Sci. (2014) 105:1487–95. doi: 10.1111/cas.12530. PMID: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Lugli A, Kirsch R, Ajioka Y, Bosman F, Cathomas G, Dawson H, et al. Recommendations for reporting tumor budding in colorectal cancer based on the International Tumor Budding Consensus Conference (Itbcc) 2016. Mod Pathol. (2017) 30:1299–311. doi: 10.1038/modpathol.2017.46. PMID: [DOI] [PubMed] [Google Scholar]
- 31. Marcoulides KM, Raykov T. Evaluation of variance inflation factors in regression models using latent variable modeling methods. Educ Psychol Meas. (2019) 79:874–82. doi: 10.1177/0013164418817803. PMID: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Al Bandar MH, Kim NK. Current status and future perspectives on treatment of liver metastasis in colorectal cancer. Oncol Rep. (2017) 37:2553–64. doi: 10.3892/or.2017.5531. PMID: [DOI] [PubMed] [Google Scholar]
- 33. Tokunaga R, Nakagawa S, Sakamoto Y, Nakamura K, Naseem M, Izumi D, et al. 12‐Chemokine signature, a predictor of tumor recurrence in colorectal cancer. Int J Cancer. (2020) 147:532–41. doi: 10.1002/ijc.32982. PMID: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Chen X, Wu P, Liu Z, Li T, Wu J, Zeng Z, et al. Tertiary lymphoid structures and their therapeutic implications in cancer. Cell Oncol. (2024) 47:1579–92. doi: 10.1007/s13402-024-00975-1. PMID: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Fridman WH, Sibéril S, Pupier G, Soussan S, Sautès-Fridman C. Activation of B cells in tertiary lymphoid structures in cancer: Anti-tumor or anti-self? Semin Immunol. (2023) 65:101703. doi: 10.1016/j.smim.2022.101703. PMID: [DOI] [PubMed] [Google Scholar]
- 36. Vion R, Roulleaux-Dugage M, Flippot R, Ouali K, Rouanne M, Clatot F, et al. Induction of tertiary lymphoid structures in tumor microenvironment to improve anti-tumoral immune checkpoint blockade efficacy. Eur J Cancer. (2025) 225:115572. doi: 10.1016/j.ejca.2025.115572. PMID: [DOI] [PubMed] [Google Scholar]
- 37. Grigore A, Jolly M, Jia D, Farach-Carson M, Levine H. Tumor budding: The name is Emt. Partial Emt. J Clin Med. (2016) 5. doi: 10.3390/jcm5050051. PMID: [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding authors.