Introductıon

Pathologic staging systems such as Dukes’, Astler-Coller’s, and the UICC/TNM classifications [1,2,3], which consider only parameters regarding depth of tumor penetration and lymph node involvement, are most widely used to predict long-term survival after a potentially curative resection in colorectal cancer (CRC). However, patients with the same stage of disease do not always have similar oncologic outcomes even if complete radical surgery was performed. This prognostic difference is thought to be due mainly to variations in the biological aggressiveness of primary tumors of the same stage [4].

CRC is not a homogeneous disease; rather it contains different molecular and pathological entities expressing a wide range of clinical behavior. Therefore, traditional pathological staging systems are insufficient to estimate the biological behaviour of rectal cancer. Concerning the prognosis of rectal cancer, it is thought that the qualitative characteristic of the true biological activity of the tumor, e.g. tumor budding (TB), can have important value t. TB, defined as undifferentiated cancer cells in the form of small aggregates existing on the invasive side of the lesion, is a characteristic microscopic feature of tumor dedifferentiation, which is the first sign of tumor invasion [5, 6]. In recent years, the meaning of TB in CRC has begun to be researched in terms of tumor biology, invasion, and metastasis. The prognostic importance of TB has been investigated in series of patients with CRC in various stages [7]. TB in CRC patients is related to both their biological status and clinical condition. Recent studies have demonstrated that the intensity of TB at the invasive site of CRC has a correlation with local recurrence, lymph node metastasis, and 5-year-survival. However, most of the studies have inherent limitations because of the small number of patients, short follow-up time, undeclared full oncological outcomes, and the lack of standard pathological reporting of TB [8,9,10,11,12].

For rectal cancer the clinical and prognostic importance of TB has not been investigated in resected specimens from both patients who had received neoadjuvant chemoradiotherapy (CRT) and those who had not. It has been a well-known assumption that unexpected and unpredictable local recurrence and distant metastasis in patients with prognostically a “grey zone” stage are due to the undetermined biological aggressiveness of the primary tumor in case of no additional treatment. Hence, the pathological evaluation of TB may be very important in determining which patients are going to receive neoadjuvant CRT.

The aim of this study was to investigate the following:

  • the prognostic importance of TB in patients who received and neoadjuvant CRT and those who did not,

  • the association between TB and oncological outcomes,

  • whether a relationship between TB and the other well-investigated pathological prognostic factors could be demonstrated or not, in rectal cancer patients.

Materıals and methods

This study was approved by Dokuz Eylul University, Faculty of Medicine Research Ethics Committee (approval no: 381-GOA).

Patients

The clinicopathological data and oncological outcomes of 437 patients who had surgery for rectal cancer at Dokuz Eylul University Hospital between January 2000 and June 2010 were evaluated. Exclusion criteria were: tumors other than adenocarcinomas, synchronous or metachronous cancer, cancer complicating familial adenomatous polyposis or inflammatory bowel disease and ypT0N0 tumors. The preoperative work-up included general clinical examination, digital rectal examination, a complete blood test, carcinoembryonic antigen (CEA), colonoscopy and directed tumor biopsy, computed tomography (CT) scan of the chest and abdomen, and pelvic phase-array magnetic resonance imaging (MRI). Tumor location was determined by rigid proctosigmoidoscopy. The treatment regimen was planned according to the patient’s age, tumor location, clinical stage, and World Health Organization performance status in a multidisciplinary team approach. All patients with locally advanced rectal cancer (cT3–T4N0 or cTanyN+) received neoadjuvant chemoradiotherapy (1.8 Gy/day, 5 days/week to a total of 25 fractions over a period of 5 weeks for a total of 4500 + 5-fluorouracil (FU) 225 mg/m2/day infusion for 5 days/week during a period of 5 weeks). All included patients had radical surgery strictly according to the principles of total mesorectal excision (TME) [13]. In the CRT group patients had surgery 8 weeks after the completion of CRT. All patients were followed up with physical examination and serial assay of the serum concentrations of CEA every 3 months for 2 years and every 6 months thereafter. Colonoscopy, abdominal ultrasound/CT scan were additionally performed every 6 months for 2 years and yearly thereafter. Disease progression was defined as local recurrence and/or development of distant metastasis.

Histopathology

Histopathology slides of radically resected specimens were prospectively evaluated by two experienced pathologists who were blinded to the clinical data and patient outcomes. Formalin-fixed and paraffin-embedded tissue was sectioned and stained with hematoxylen and eosin for microscopic examination. TB along the invasive margin was examined in addition to routine pathological findings. The presence of budding was determined according to the criteria proposed by Hase at al. [6], whereby budding is defined as an isolated single cancer cell or a cluster composed of ≤ 5 undifferentiated cancer cells appearing to bud from a large cancer gland on the invasive side. Tumor slides were initially scanned at 20 × magnification for areas with the highest density of tumor buds. The selected area was then examinated under 400 × magnification and number of the tumor buds was determined with the same light microscope for each case. The resulting number of tumor buds was considered to be the degree of TB and the term “budding intensity” was used for this number. Budding intensity was scored as follows: none (0), mild (1–5 buds), moderate (6–10 buds), and severe (> 10 buds) (Fig. 1). Two groups were created, TB-1 (none, mild) and TB-2 (moderate, severe), for statistical analysis [14]. The pathologic stage was evaluated according to the 7th UICC/TNM staging system.

Fig. 1
figure 1

Tumor budding indicated with arrows, showing small clusters of poorly differentiated or undifferentiated cells, ahead of the invasive front of rectal cancer; a no budding (H&E × 200); b mild (1–5 buds) (H&E × 400); c moderate (6–10 buds) (H&E × 400); d severe (> 10 buds) (H&E × 400)

Full size image

Statistical analysis

An independent variables t test, the Mann–Whitney U test, and the Chi square test were used during data evaluation for single variable analysis. A model was composed using the significant relation detected in the univariate analysis and then the multivariate logistic regression analysis was performed. The Kaplan–Meier method was used for the survival analysis and the log rank test was applied for the comparison of the groups. The significantly different variables were evaluated by the Cox regression model. The results were evaluated as 95% confidence interval and significance was set at p < 0.05.

Results

A total of 230 patients out of 432 patients were included in this study; 117 patients with cT3–T4N0 or cTanyN(+) locally advanced rectal cancer who received neoadjuvant CRT, and 113 patients with cT2–T3N0 rectal cancer who did not receive CRT preoperatively. There were 96 (41.7%) women, and 134 (58.3%) men and the mean age was 62.1 ± 12 years. Excluding 15 patients without follow-up data, median follow-up time was 40.1 ± 27.5 months. Local recurrence had occurred in 13 patients during the follow-up period. For all patients the 5-year overall survival (OS) and disease-free survival (DFS) rates were 66% and 62%, respectively.

The association between the intensity of TB and clinicopathological variables in all patients is shown in Table 1. In multivariate analysis, the presence of lymphatic invasion (HR 3.99; 95% CI 1.84–8.69), venous invasion (HR 2.48; 95% CI 1.04–5.92), local recurrence (HR 3.87; 95% CI 1.05–14.26), and the number of metastatic lymph nodes (HR 1.16; 95% CI 1.02–1.32) were shown to be independent prognostic factors for TB.

Table 1 The association between clinicopathological variables and tumor budding
Full size table

Multivariate Cox regression analysis revealed that radial margin status (HR 2.16; 95% CI 1.18–3.96) and intensity of TB (HR 2.64; 95% CI 1.46–4.77) had an independent effect on overall survival (Table 2).

Table 2 The association between clinicopathological variables and overall survival
Full size table

The association between clinicopathological variables and overall survival in patients who received neoadjuvant CRT and those who did not is shown in Table 3. In multivariate analysis, TB (HR 4.28; 95% CI 1.60–11.49), radial margin status (HR 2.53; 95% CI 1.09–5.90), distant metastasis (HR 2.33; 95% CI 1.18–4.60), and venous invasion (HR 4.48; 95% CI 2.14–9.39) were strong prognostic factors of overall survival for the patients who underwent CRT while TB (HR 4.87; 95% CI 2.10–11.28) and distant metastasis (HR 4.30; 95% CI 1.81–10.22) were independent predictors in patients not undergoing CRT.

Table 3 The association between clinicopathological variables and overall survival in the CRT (+ve) and CRT (−ve) groups
Full size table

In the Cox proportional hazards model for all patients, worse disease free survival was correlated with involved lymph nodes (HR 2.78; 95% CI 1.60–4.87), venous invasion (HR 1.76; 95% CI 1.00–3.09), and with radial margin status (HR 2.31; 95% CI 1.27–4.22) as shown in Table 4.

Table 4 Determination of prognostic factors for disease-free survival using Cox regression analysis
Full size table

In the subgroup of patients who had CRT decreased disease free survival was associated with lymph node involvement (HR 4.39; 95% CI 1.70–11.33) and radial margin status (HR 2.56; 95% CI 1.12–5.90) while only lymph node involvement (HR 2.33; 95% CI 1.16–4.66) was significant predictor of DFS in patients who had not received CRT (Table 5).

Table 5 Multivariate analysis of the prognostic factors of DFS in the CRT (+ve) and CRT (−ve) groups
Full size table

A separate analysis of pT2–T3NO rectal tumors was performed. In multivariate analysis, the presence of venous invasion (HR 6.20; 95% CI 1.99–19.36) was shown to be an independent predictor for TB in pT2–T3N0 rectal cancer. Multivariate Cox regression analysis showed that venous invasion (HR 26.22; 95% CI 2.63–261.76) was the only independent risk factor for overall survival. In multivariate analysis, no factor was independently predictive of survival in the CRT-treated group, while venous invasion (HR 15.60; 95% CI 1.34–182.09) was a strong prognostic factor in patients with pT2–T3N0 rectal cancer who did not receive CRT. In the Cox proportional hazards model, TB (HR 4.40; 95% CI 1.10–17.74) was significantly associated with decreased DFS in CRT-treated patients with T2–T3N0 tumors.

Discussion

To date, the prognostic and clinical importance of tumour budding have not been investigated in the irradiated macroscopic specimens of rectal cancer from patients who received neoadjuvant chemotherapy. The literature indicates that TB in CRC cases was investigated before 2000 when neoadjuvant CRT was uncommon [7, 11, 15,16,17]. Our study is important, because it is, to the best of our knowledge, the first study to evaluate the potential prognostic relationship of TB in a relatively large series of CRT-treated patients with locally advanced rectal cancer. Our results showed that as the intensity of TB increased, the OS rates sharply decreased. When patients who received CRT were grouped with patients who did not receive CRT, there was a significant relationship between TB and worse OS but such a robust association was lacking between TB and DFS. In a study by Kinoshita et al. [18], TB was reduced after preoperative CRT. In our study we observed that in patients who had received neoadjuvant CRT TB also affected OS. TB was not identified as an independent prognostic factor for DFS in both patients who underwent CRT and patients who did not. However, TB was identified as a a prognostic factor for DFS in the subset of patients with T2–T3N0 tumors who underwent CRT.

It is remarkable that TB had prognostic value in early stage rectal cancer given the fact that adjuvant chemotherapy and radiotherapy is not administered in early stage cancer TB may be used as an indication for such treatment. TB in CRC patients with T3N0 was also reported to be an independent prognostic factor in a study by Wang et al. [19]. TB must be evaluated and reported in pathology reports and should be considered in the multidisciplinary tumor board discussions to assist in decision making about additional treatment.

The oncologic literature indicates that as the depth of tumor invasion increases the OS rate of rectal cancer patients decreases [20]. Our study also confirmed that the OS rates were higher in patients with early stage rectal cancer.

Findings in our study are consistent with those in studies supporting the notion that the presence of lymphatic invasion and venous invasion reduces OS in patients with CRC [21]. Lymphatic invasion and venous invasion are the established prognostic factors defined as Category I in the classification made by College of American Pathologists [22].

TB was more frequent in patients who developed a local recurrence. In a study by Hase et al [6], patients with moderate and severe TB had significantly higher local recurrence rates and significantly lower 5-year and 10-year survival rates than patients with no TB or mild TB. Also according to this study, 5-year survival of Dukes B cases with moderate or severe tumor budding was reported to be lower than that of Dukes C cases with no TB or mild TB. In our study, the presence of distant metastasis was found to negatively affect OS whether patients had received neoadjuvant CRT or not. Distant metastasis in patients with CRC was proved to be an indicator of poor prognosis for survival by several studies [23].

The presence of lymph node involvement reduces OS significantly in patients with CRC and the higher the numbers of involved lymph nodes the worse the effect on survival rates. Lymph node involvement is defined as the second most powerful “postoperative outcome indicator”, the first one being the occurence of distant metastasis [3]. As well as the number of involved lymph nodes, the number of total lymph nodes in the surgical specimen affects the prognosis of both stage II (lymph node negative) and stage III (lymph node positive) disease directly [24]. In our study, lymph node involvement was found to be a factor that reduced OS and DFS rates through all groups in univariate analysis. In multivariate analysis, lymph node involvement was found to be an independent prognostic factor for DFS whether patients received neoadjuvant CRT or not.

The intactness of the mesorectum and the radial margin status are among the most important prognostic factors for local and distant metastasis and for survival [25, 26]. In a study by Birbeck et al. [27], the effect of radial margin involvement in rectal cancer on the survival rate has been investigated. A higher rate of local recurrence was reported in cases of radial margin involvement. It has been demonstrated that radial margin status could be used as the prognosticator of survival after rectal cancer surgery. In our study, radial margin involvement reduced OS significantly in all patients whether they had received neoadjuvant CRT or not. Moreover, when the radial margin was involved, DFS rates decreased in patients treated with CRT.

Although not used for staging, pre-treatment CEA levels, radial margin status, the presence/absence of perineural invasion, the microsatellite instability status and tumor regression grade (for CRT-treated patients) are the recommended prognostic factors according to the TNM guidelines [3]. Despite the fact that many studies in the literature report the prognostic value of TB, it is still not used sufficiently in clinical practice.

The drawbacks of this study are mainly the non-measured interobserver agreement and the relatively small number of patients in the series, particularly in the subgroup analysis.

Conclusions

TB has a prognostic value at least as strong as lymph node involvement. We recommend that TB be included in the routine pathological examination of rectal carcinoma. Because of its independent prognostic value the intensity of TB should be used to determine whether adjuvant treatment is indicated after chemotradiation and may be used as a biological criterion for patients who are unresponsive to treatment and for the use of more intensive systemic chemotherapy protocols. More studies are needed to confirm the prognostic significance of TB in rectal cancer patients, especially in the subset treated with CRT. If our results are confirmed TB should be included among the negative prognostic factors listed in the TNM classification.