Colorectal cancer (CRC) is the third most common form of cancer and the second leading cause of cancer-related death in the Western world, with more than 600,000 deaths worldwide each year.1

Currently, the tumor-node-metastasis (TNM) staging system, developed by the American Joint Committee on Cancer (AJCC) and the International Union Against Cancer (UICC), is the primary method for assessing prognosis for individual patients.2

For patients with lymph node metastases (TNM stage III), adjuvant chemotherapy after resection of the primary tumor is recommended, because of high risk for disease recurrence. Adjuvant 5-fluorouracil (5-FU)-based chemotherapy has been found to increase median 5-year survival in stage III colon cancer patients from 51% to 64%.3 Moreover, combinations of 5-FU-based therapy with oxaliplatin have further improved response rates and disease-free survival.4 However, so far no convincing evidence exists for a beneficial effect of postoperative adjuvant chemotherapy in patients without lymph node metastases (TNM stage I and II), while 20%–30% of stage II patients will still develop recurrent disease.5 , 6 One of the dilemmas here is the number of lymph nodes actually required for making an accurate call of stage II colon cancer. In fact, numbers ranging from as low as 4 to as high as 40 nodes have been suggested for adequate staging.7 – 18 Although most current recommendations stick to a minimum number of 10–12 lymph nodes to be investigated, in routine clinical practice this is certainly not always achieved.19 – 21 Besides its pivotal role in accurate staging, a high lymph node yield has been associated with a better prognosis in both stage II and III colon cancer, an observation for which no clear biological explanation has been found.9 , 12 , 15 , 18 , 22 – 25

Extent of surgical resection and thorough pathological evaluation of the resection specimen play an important role in the total number of lymph nodes identified. Moreover, underlying disease mechanisms of colon cancer may have an important effect on lymph node harvest as well. Possibly, biological behavior of the tumor and interactions between tumor and host affect characteristics of peritumoral lymph nodes resulting in improved node yield. One of the underlying biological factors that has been suggested to influence the number of lymph nodes retrieved in the resection specimens is the microsatellite instability status of colon cancer, which in itself is also associated with clinical outcome.26 The aim of the present study was to investigate the association of lymph node yield, MSI status, and recurrence rate in stage II and III colon cancer patients.

Patients and Methods

From 1996 to 2005, a total of 667 patients underwent surgical resection for colon cancer at the Kennemer Gasthuis hospital in Haarlem, the Netherlands. Of these, 454 were classified as TNM stage II (T3–4, N0, M0) or III (T1–4, N1–2, M0) according to the TNM staging system by the AJCC and UICC.27

Data were collected from clinical reports and included date of birth, date of surgery, location of the primary tumor, that is, right-sided (caecal, ascending, transverse) or left-sided (descending, sigmoid, rectosigmoid), adjuvant chemotherapy, date and site of first disease recurrence, and cause of death, that is, colon cancer related or unrelated cause of death. Disease recurrence was defined as either local tumor recurrence or distant metastases, diagnosed by computed tomography imaging and/or histopathology.

From the histopathology reports, information was retrieved about tumor size, tumor and nodal stage, differentiation grade, ulceration, mucinous differentiation, and angioinvasion of the primary tumor. In line with the Dutch national guidelines, lymph node (LN) yield was defined as “high” when 10 or more LNs were retrieved and “low” in case of fewer than 10 LNs.

Patients with a previous history of colorectal malignancy (n = 12) and those with incomplete resections of the primary tumor (macroscopically or microscopically, n = 9) were excluded from this study. Also patients who were lost for follow-up or died within 3 months after surgery (n = 8 and n = 39, respectively) were excluded. Of the 386 eligible patients, microsatellite instability status could not be determined in 54 cases (see the section “Microsatellite Instability Analysis”), which were also excluded from this study.

The remaining study population consisted of 332 stage II and III colon cancer patients: 181 males and 151 females with a mean age of 70 years and a median follow-up period of more than 57 months (Table 1). Adjuvant 5-FU-based chemotherapy was administered to 15.7% of stage II patients and 55.8% of stage III patients.

Table 1 Clinical and pathological characteristics of study population
Full size table

Lymph Node Retrieval and Examination

After resection of the colon, the specimens were placed in formalin and allowed to fix for at least 24 hours. After proper fixation, the mesenteric fat was cut into thin slices and lymph nodes were sampled: small lymph nodes not exceeding 5 mm in diameter were included in toto, somewhat larger lymph nodes (diameter 5–10 mm) were cut in half, and lymph nodes larger than 10 mm were sliced in equal intervals and subsequently placed in marked cassettes. After conventional histological staining with hematoxylin and eosin, the lymph nodes were microscopically examined for the presence of metastases.

DNA Isolation

DNA was isolated from formalin-fixed paraffin-embedded (FFPE) colon cancer tissues samples. For each tumor, areas with at least 70% tumor cells were selected from 4-μm sections. Adjacent serial sections of 10 μm were cut and macrodissected. DNA was isolated as previously described (using QIAamp microkit; Qiagen, Hilden, Germany).28 DNA concentrations were measured with a Nanodrop-100 spectrophotometer (Isogen, De Meern, The Netherlands).

Microsatellite Instability Analysis

Tumor samples were analyzed for microsatellite instability (MSI) using MSI Analysis System, Version 1.2 according to the manufacturer’s instructions (Promega, Madison, WI). This PCR-based assay uses 5 mononucleotide repeat markers to determine MSI status. PCR products were separated by capillary electrophoresis using the ABI 3130 DNA sequencer, and output data were analyzed using the accompanying package GeneScan 3100 (Applied Biosystems, Foster City. CA). Tumors were classified as microsatellite instable (MSI) when instability was observed for 2 or more markers. When instability was observed for none or only 1 marker, tumors were considered to be microsatellite stable (MSS). MSI status could be determined in 332 cases, (i.e., 86% of tumor samples, while attempts to characterize the remaining 14% failed because of insufficient quality of the FFPE-derived DNA material).

Statistical Analysis

Differences between sample means were determined using the t test. Differences in proportions between groups were examined using Pearson’s chi-square test. Survival rates were displayed using the Kaplan-Meier method and compared using the log-rank test. All reported P values are 2-sided, and a significance level of .05 was used. Statistical analysis was performed with SPSS 17.0 for Windows, SPSS Inc., Chicago, IL.

Results

Of the 185 stage II patients, 24.9% developed recurrent disease, while for the 147 stage III patients this was 45.6% (P < .01). An overview of disease recurrence rates is displayed in Table 2. No difference in mean lymph node retrieval was observed over time and between involved surgeons and pathologists (data not shown).

Table 2 Disease recurrence rates in 332 stage II and III colon cancer patients
Full size table

High LN yield was observed in 133 patients (40.1%) and low lymph node yield (< 10 LN) in 199 (59.9%) patients. Overall, patients with high LN yield tended to have fewer recurrences compared with patients with low LN yield (28.6% vs 37.7%, P = .09). Considering only stage II patients, there was a significant difference in recurrence rate, namely, 16.4% for patients with high LN (11 of 67) yield and 29.7% for patients with low LN yield (35 of 118; P = .05). Disease-free survival curves are displayed in Fig. 1.

Fig. 1
figure 1

Disease-free survival curves of colon cancer patients with high (≥ 10) and low (< 10) lymph node (LN) yield for (a) stage II and (b) stage III patients

Full size image

Tumors with high LN yield were significantly larger and were located more frequently right-sided compared with tumors with low LN yield (mean 45.4 mm vs 40.4 mm, respectively, P = .03; right-sided 60.8% vs 47.2%, respectively, P = .02).

Tumors with high LN yield were significantly associated with the MSI phenotype, as 26.3% of these tumors were MSI, compared with 15.1% of tumors with low LN yield (P = .01). The mean LN yield of resected MSI tumors was 10.1 compared with 8.6 for MSS tumors (P = .03; Fig. 2a). This difference though, was mainly observed in stage III patients (Fig. 2b, c). For this subgroup, the mean LN yield for MSI tumors was 11.7, compared with 9.1 for MSS tumors (P < .01).

Fig. 2
figure 2

Box plot analysis comparing number of lymph nodes retrieved between colon cancer patients with MSS tumors and patients with MSI tumors for the total patient population (MSS: n = 267, MSI: n = 65, [a]) and separately for stage II (MSS: n = 147, MSI: n = 38, [b]) and stage III (MSS: n = 120, MSI: n = 27, [c]) patients

Full size image

MSI vs MSS Tumors

Of all tumors, 19.6% appeared to be MSI and 80.4% MSS. Disease recurrence occurred in 24.6% of the patients with MSI tumors, compared with 36.3% of the patients with MSS tumors (P = .07). This difference was mainly attributable to stage II patients, as in this group the recurrence rate was 13.2% for patients with MSI tumors and 27.9% in MSS cases (P = .06). For stage III patients, recurrence rates were 40.7% for MSI tumors and 46.7% in case of MSS tumors (P = .6). Disease-free survival curves are depicted in Fig. 3.

Fig. 3
figure 3

Disease-free survival curves of colon cancer patients with MSS tumors and patients with MSI tumors for (a) stage II and (b) stage III patients

Full size image

MSI tumors revealed a poor histological differentiation in 30.8% compared with only 10.5% in cases of MSS tumors (P < .01). MSI tumors were located more often right-sided, compared with MSS tumors (62.9% vs 24.6%, respectively, P < .01) and were significantly larger (mean 54.2 mm vs 39.6 mm, P < .01).

Within the patient population with MSI tumors, a trend toward better disease-free survival was seen for those with high LN yield compared with patients with low LN yield, as shown in Fig. 4a. For patients with MSS tumors, a similar trend was observed (Fig. 4b). When these analyses were performed stratified for disease stage, no significant differences in survival between patients with high LN yield and those with low LN yield were seen (data not shown).

Fig. 4
figure 4

Disease-free survival curves of microsatellite instable (MSI) (a) and microsatellite stable (MSS) (b) colon cancer patients with high (≥10) and low (<10) lymph node (LN) yield

Full size image

Multivariate Analysis

Multivariate analysis included sex, age, disease stage, MSI status, tumor location, tumor diameter, differentiation grade, presence of mucinous differentiation, ulceration, and angioinvasion. MSI phenotype was the strongest independent factor associated with high lymph node yield (odds ratio 2.3, 95% confidence interval 1.2–4.4).

Discussion

In the present study, high LN yield in the resection specimen was associated with improved disease-free survival. This was observed mainly in stage II patients, while for stage III patients only a trend toward better survival was seen in cases of high LN yield. Regarding stage II patients, we identified high-risk patients based on NCCN criteria.29 We analyzed T-stage, histologic grade, vascular invasion, and number of lymph nodes sampled of all tumor specimens. These characteristics are included in Table 1 and Fig. 2. For these characteristics, no significant difference was observed between patients with low lymph node count (<10) and those with high lymph node count (≥10), as shown in Table 1.

Several studies strongly confirm the association between high LN yield and improved survival in colorectal cancer patients, both in stage II and III disease.10 , 15 , 18 , 25 , 30 , 31

The causal factors for this association are subject of debate. Stage migration has been postulated as a factor. Patients with lower numbers of nodes analyzed could be falsely designated as stage I or stage II, when the nodes examined contain no metastases, while additional node samples would have revealed tumor and thus stage III would have been assigned.

Moreover, resection of lymph nodes itself may have a therapeutic effect. For patients with positive lymph nodes, a higher number of recovered nodes has been associated with better survival.15 , 25 , 31 – 33 This was not strongly confirmed by the present study as stage III patients with high LN yield showed only a trend toward better disease-free survival.

Several explanations have been postulated to explain why some resection specimens of colon cancer have high lymph node yields and others do not. It has been suggested that a low number of nodes recovered in a specimen is a reflection of inadequate surgical resection or pathological examination.34 – 36 Guidelines recommending a certain minimum number of investigated lymph nodes are also based on the assumption that increased effort by the surgeon and pathologist will lead to higher lymph node counts. However, large studies that show inferior survival in colon cancer patients with fewer lymph nodes demonstrated that this relationship could not be explained by factors such as extent of surgical resection and pathologic processing. 30 , 37 Moreover, the fact that most societies nowadays recommend a minimum number of 10 to 12 lymph nodes to be investigated, and this is only achieved in about one-third to one-half of the patients, also indicates that other factors besides surgical and pathologic skills may play a role in lymph node yield.19 – 21 Alternatively, a low number of nodes may not necessarily represent the quality of care a patient has received and instead may be influenced by underlying biological characteristics of the tumors.

One of the underlying biological factors that have been suggested to be of influence is the microsatellite instability status of colon cancer. Microsatellite instability is 1 of 2 major distinct colorectal oncogenic pathways, the other being chromosomal instability.38 – 40 Microsatellite instability is observed in about 15% of sporadic colorectal cancers and is caused by a defect in the DNA mismatch repair system. Colon cancers with the MSI phenotype have been associated with a better prognosis compared with MSS tumors.38 , 40 , 41 One of the factors involved here could be the antitumor immune response, which differs between MSI and MSS colorectal cancers. In general, the immune system recognizes neoplasia poorly, but in MSI colon cancer with infiltrating lymphocytes, it has been shown that mechanisms of T-cell cytotoxicity are activated.42

Truncated peptides produced by frameshift mutations that are common in MSI cancers because of failing DNA mismatch repair may be immunogenic and contribute to the host immune response resulting in the migration of activated T-cells into the malignant epithelium of the tumor.43 – 50 Moreover, marked lymphocytic, so-called “Crohn’s-like” infiltrates are a hallmark of MSI colorectal cancers.51 Microscopically, hyperplastic changes are seen in lymph nodes draining colorectal tumors exhibiting prominent antitumor immune reactions, and these tumors were found to have larger and more detectable lymph nodes.52 , 53 These findings suggest a relation between MSI status of the primary tumor and lymph node yield in the resection specimen. This is supported by the present study, as MSI tumors were significantly associated with a high number of lymph nodes harvested (Fig. 2a).

The association between MSI phenotype and high lymph node yield as was observed in the present study has been reported before in 2 smaller and more heterogeneous study populations compared with the present one.54 , 55 One study consisted of 82 stage I (n = 27) and II (n = 55) both colon (n = 52) and rectal (n = 30) cancers, the other was based on 121 stage I–III colon cancers (12 stage I, 71 stage II, and 38 stage III patients).54 , 55 In the present study, rectal cancers were excluded, because in general these tumors are treated with preoperative (chemo)radiation therapy, which is known to influence lymph node retrieval of the resection specimen.56 – 58 In both earlier studies, heterogeneous patient populations were investigated and analysis did not include, in contrast to the present study, stratification for disease stage, possibly because the relatively small sample sizes would not allow for meaningful subgroup analysis.

The present study consisted of 185 stage II and 147 stage III colon cancer patients. This is the first study reporting a significant association between MSI tumors and high LN yield in stage III colon cancer. Interestingly, the effect of MSI on lymph node yield was highest in stage III cancer, that is, in those tumors in which metastatic spread to regional lymph has already occurred. This could be caused by an additional boost of the immune response due to more intimate exposure of lymphoid tissue to tumor cells than in stage II tumors.

In conclusion, the present study showed a strong association between MSI phenotype and high lymph node yield in colon cancer, mainly so in stage III disease. The biology of MSI colon cancers can provide explanations for this, while at the same time contributing to a better prognosis. Conversely, a less favorable outcome in patients with lower lymph node yields would then not only be attributable to understaging due to missed positive lymph nodes, but also to tumor intrinsic factors.