Main

Tumor budding is a histological finding in epithelial cancers that occurs when tumor cells detach from the invasive tumor margin and migrate into the surrounding stroma.1, 2 It likely represents dedifferentiation from epithelial cancer cells to tumor stem cells, which in turn may redifferentiate and invade locally or migrate to distant sites.3, 4 Tumor budding at the invasive tumor margin is an established negative prognostic factor in resected colorectal cancer.5, 6, 7, 8, 9, 10, 11, 12 Widespread use of tumor budding as a prognostic factor has been impeded by a lack of standardized histopathological criteria, highlighted recently by Mitrovic et al13 who state that ‘although performing a tumor bud count is easy in theory, deciding what is or is not a bud can prove surprisingly difficult in practice’. Recent studies have addressed the role of tumor budding in resected colorectal specimens.14, 15, 16 Tumor budding within the entire tumor, rather than at the invasive front, is termed intra-tumoral budding17 and has been described in both colorectal cancer biopsy and resected specimens.18, 19 This study addresses the prognostic and predictive power of tumor budding in biopsy specimens at the time of diagnosis of locally advanced rectal cancer.

Patients with the same stage of rectal cancer at initial diagnosis can have markedly different outcomes.20, 21, 22 Apart from TNM staging, reliable prognostic markers of rectal cancer are few and, apart from pre-operative carcinoembryonic antigen (CEA),23, 24, 25 not often used clinically.23 Tissue diagnosis usually depends on a superficial biopsy obtained at endoscopy from which adverse features such as tumor depth, lymphovascular invasion and perineural invasion cannot be determined. Current treatment options, including local excision, neoadjuvant therapy and radical surgical excision, are guided by relatively crude measurements of local and distant tumor extent and not by markers of tumor biology, which determine tumor response to treatment and ultimately patient survival. Therapeutic algorithms informed by markers of tumor biology are needed to ensure that maximal treatment, including neoadjuvant chemoradiotherapy, is given only when benefit is likely and, conversely, that local or minimally invasive alternatives are available when biologically appropriate.

The aim of this study was to investigate the potential value of intra-tumoral budding in pre-treatment biopsies of rectal cancer in prediction of pathological response to neoadjuvant chemoradiotherapy and cancer-specific survival.

Materials and methods

Patient Selection

Patients with primary rectal adenocarcinoma were identified from the prospectively maintained Centre for Colorectal Disease database. Staging was performed in accordance with the American Joint Committee on Cancer Guidelines.22 Only patients with locally advanced rectal cancer who underwent neoadjuvant chemoradiotherapy and had subsequent rectal resection were included. Patients were excluded where a pre-treatment tumor biopsy was not available for review, if there were distant metastases at the time of diagnosis or if they had a cancer co-existing with ulcerative colitis, familial adenomatous polyposis or Lynch syndrome.

Patient Management

Pre-treatment tumor characteristics were determined with pelvic magnetic resonance imaging for T and N stage and computed tomography of the thorax, abdomen and pelvis for M stage. Radiotherapy was administered for a total dose of 45–50.4 Gray in 25–28 fractions of 1.8 Gray delivered over 6 weeks. 5-Fluorouracil was given concomitantly by protocol.26, 27 Patients underwent surgical resection with curative intent 6–8 weeks post completion of chemoradiotherapy.

Histopathological Staging

Pathological analysis using a standardized reporting template was performed on resected specimens and staged according to the American Joint Committee on Cancer.22 A tumor regression grade was assessed using a three-tier classification method.28, 29 For the purposes of this study, pre-treatment rectal biopsies were retrospectively reviewed for evidence of intra-tumoral budding through a × 4 lens, ( × 40 magnification) with confirmation of positive cases at × 10 ( × 100 magnification) according to the method described by Giger et al,19 adapted from the Nakamura method of identifying peritumoral budding.30 Intra-tumoral budding was defined as a single cancer cell or a group of <5 detached tumor cells found in the stroma of the biopsy specimen. Thus, any budding seen at × 4 and confirmed at × 10 was deemed positive. An example of intra-tumoral budding is shown in Figure 1. Biopsies were reviewed independently by two pathologists (DG and KS) blinded to the clinical and subsequent pathological outcomes, and in cases of discordance (n=16), a consensus on budding status was obtained.

Figure 1
figure 1

Biopsy samples of rectal adenocarcinoma stained with hematoxylin and eosin at × 200 magnification. (a) Rectal cancer with no intra-tumoral budding. (b) Small groups and single tumor cells in a desmoplastic stroma typical of intra-tumoral budding.

Full size image

Patient Follow-Up

Follow-up was recorded through the prospectively maintained institutional colorectal cancer database. Primary care physicians were contacted as necessary to complete survival data.

Statistical Analysis

Results were analyzed using the Predictive Analytics Software (PASW 18.0.2, SPSS, Chicago, IL, USA). Comparative analyses of quantitative data were performed using the chi-squared test for categorical variables and Student’s t-test for continuous variables. Spearman’s rank test was used for correlation of budding with non-parametric variables. Tables were generated correlating budding status with tumor regression grade for determination of sensitivity and specificity. Univariate logistic regression was used to assess the effects of demographics and tumor characteristics on cancer-specific death, expressed as hazard ratios with 95% confidence intervals (CIs). Variables with P<0.200 in univariate analysis were included in the multivariate logistic regression analysis. Kaplan–Meier curves were determined for analysis of 5-year survival based on budding status using the log rank test. All tests of significance were two-tailed, with P<0.05 indicating statistical significance.

Results

Patient Demographics, Tumor and Treatment Characteristics

Between 2003 and 2010, 185 patients underwent neoadjuvant chemoradiotherapy and subsequent resection for rectal cancer; 89 patients met the eligibility criteria. The median age was 62 years (range 30–84 years), and 64% were male. Initial tumor staging with MR imaging (where available) identified 18% (15/85) patients with American Joint Committee on Cancer Stage 2 cancer and the remaining had Stage 3 disease. Of the 38 patients who had MR re-imaging post neoadjuvant chemoradiotherapy, 20 (53%) exhibited tumor downstaging. Twenty-five patients underwent abdominoperineal resection, 2 had ultra-low anterior resection with coloanal anastomosis, 19 underwent low anterior resection, 36 had standard anterior resection and the remainder either had pelvic exenteration or a Hartmann’s procedure. Overall, 13% had a complete pathological response (tumor regression grade 1). In all, 50.2% had a partial response (tumor regression grade 2) while the remainder were graded as tumor regression grade 3, denoting no response to chemoradiation. Median follow-up was 49 months (range 7–117 months).

Budding in Pre-Treatment Samples

Intra-tumoral budding was identified in the pre-treatment tumor biopsy in 18 of the 89 patients (20%). There were no differences in patient demographics or pre-treatment tumor characteristics between the two groups (Table 1).

Table 1 Comparison of patient demographics or pre-treatment tumor characteristics between those positive or negative for budding in the pre-treatment rectal cancer biopsy
Full size table

Intra-Tumoral Budding Predicts Poor Pathological Response to Chemotherapy

The presence of intra-tumoral budding in pre-treatment biopsies was significantly associated with negative tumor prognostic factors found in the resected tumor specimen (Table 2). Intra-tumoral budding was strongly indicative of higher ypT stage (0.019), residual lymph node involvement (P=0.018), poor differentiation (P=0.005) and lymphovascular invasion (P=0.004). There was a non-significant association with perineural invasion (P=0.060) and higher tumor regression grade (P=0.061). No patient with intra-tumoral budding in the pre-treatment biopsy had a complete pathological response to neoadjuvant chemoradiotherapy (tumor regression grade 1), giving a 100% specificity and positive predictive value for non-response to neoadjuvant chemoradiotherapy (Table 3).

Table 2 Comparison of post neoadjuvant chemoradiotherapy tumor characteristics between those positive or negative for budding in the pre-treatment biopsy
Full size table
Table 3 Non-response to neoadjuvant chemoradiotherapy in those positive or negative for budding in the pre-treatment biopsy
Full size table

Budding in Pre-Treatment Biopsy Predicts Poor Disease-Free Survival and Cancer-Specific Death

The disease-free survival and cancer-specific death rate at 5 years were 72 and 18%, respectively. In patients who developed recurrent disease and those who died, almost half of them exhibited budding in the pre-treatment rectal biopsy. Intra-tumoral budding was an independent predictor of increased local and distant recurrence, as well as cancer-specific death (Table 4 and Figure 2). Univariate and multivariate analysis of factors predisposing to cancer-specific death identified intra-tumoral budding as a significant predictor of cancer death (hazard ratio 3.51, 95%CI 1.03–11.93, P=0.040). The only tumor characteristic at surgical resection that also independently predicted death in multivariate analysis was tumor stage (Table 5). Tumor budding was not assessed in resected specimens.

Table 4 Relationship of intra-tumoral budding with long-term outcomes
Full size table
Figure 2
figure 2

Kaplan–Meier survival curve relating to budding status in the pre-treatment biopsy. Patients with tumor budding in initial biopsy have a significantly lower 5-year disease-free survival rate (33.33 vs 77.46%, P≤0.001) and lower cancer-specific survival rate (61.11 vs 87.32%, P=0.021).

Full size image
Table 5 Univariate and multivariate logistic regression analysis of factors predisposing to cancer-specific death
Full size table

Discussion

This study has found that the identification of budding in a pre-treatment biopsy of rectal cancer (intra-tumoral budding) is strongly predictive of a poor response to neoadjuvant chemoradiotherapy and is associated with poor long-term prognosis. This observation must now be considered in the context that at the time of diagnosis of locally advanced rectal cancer, clinical characteristics and tumor stage poorly predict tumor response to neoadjuvant chemoradiotherapy and thus overall prognosis.31, 32 Other studies indicate that serological markers such as CEA are markers for tumor response to neoadjuvant chemoradiotherapy.23, 24, 25 Identification of patient subsets who are not likely to exhibit a complete pathological tumor response to neoadjuvant chemoradiotherapy must inform treatment decisions and help tailor individualized management. As there are no additional staining protocols required, intra-tumoral budding assessment can be performed at the time of pathological diagnosis at little or no extra cost.

This study was not designed to address the issue of reproducibility of budding in biopsy specimens. Apart from agreeing to use the criteria published by Giger et al,19 no training or validation sets were used. In this study, tumor budding status was assigned based on a consensus of two pathologist’s interpretation. The study does not validate this method of identifying intra-tumoral budding compared with other methods but does identify two groups of patients with markedly different outcomes. Budding was denoted as either ‘present’ or ‘absent’ in the biopsy specimens—this is a practical solution, as biopsy samples often give a low yield of tumor material and often do not allow for assessment of multiple high-power fields. Interestingly, our positivity rate (18/89, 20%) was very similar to the 17% intra-tumoral budding rate reported by Giger et al.19 Although their study addressed reproducibility, only a small number of biopsy cases (n=14) were subjected to intra-observer analysis.

Previous studies have correlated tumor budding in resected specimens with higher tumor regression grade.33, 34 This study demonstrates that budding in the initial biopsy predicts residual adverse features post neoadjuvant chemoradiotherapy (higher ypT stage, positive lymph nodes, poorly differentiated tumors and lymphovascular invasion). Intra-tumoral budding was associated with tumor regression grade 2 or 3 in 100% of cases. This suggests that tumors with budding at initial diagnosis are unlikely to exhibit a complete response to neoadjuvant chemoradiotherapy. However, only 25% of non-responsive cases (tumor regression grade 2 and 3) showed intra-tumoral budding; thus additional markers of non-response are required.

The present study found no differences in baseline clinical or tumor parameters between those biopsies with or without intra-tumoral budding. The poor prognostic outcomes in the former group suggest an aggressive phenotype, which may now be identified at initial biopsy. This group may be a target for escalation of therapy; however, it remains to be determined what course of action is most suited to these patients. Certainly they are a group in whom further study is required to improve outcomes.

The results are limited by the retrospective nature of database studies. Type II errors (such as the apparent lack of correlation of tumor budding with perineural invasion) may have occurred due to the sample size. Use of intra-tumoral budding as a prognostic marker will be limited by its diagnostic reproducibility. A recent study clarified some issues surrounding this;14 reproducibility of this currently described method is a significant issue and will need to be demonstrated in future studies.

Intra-tumoral budding is a marker of poor prognosis and poor response to neoadjuvant chemoradiotherapy. If confirmed in other series, intra-tumoral budding may be an indication to avoid neoadjuvant chemoradiotherapy in its present form.