Colorectal cancer (CRC) is one of the most common cancers and a leading cause of cancer-related death in men and women [1]. CRC is also a major cause of death in Japan, being the leading cause in women and the third most common cause in men [2]. The efficacy of adjuvant chemotherapy for stage II CRC remains controversial, although the benefit of adjuvant chemotherapy for stage III CRC has been established [3,4,5,6]. Many studies have reported that rectal cancer differs from colon cancer in etiology, genetics, clinical manifestation, anatomy, and biological characteristics [7], but it is unclear whether stage II colon and rectal cancer have the same risk factors for recurrence. The purpose of this study was to identify the risk factors for postoperative recurrence in stage II colorectal cancer.

Patients and methods

We retrospectively analyzed the clinical data of 990 patients with stage II CRC in the Division of Gastroenterological Surgery of Saitama Medical University from 2007 to 2016. The surgery was considered therapeutic when there was no macroscopic or microscopic residual cancer after surgery. There were 579 men and 411 women, comprising 619 patients with colon cancer and 371 patients with rectal cancer. Patients receiving preoperative treatment or presenting with intestinal obstruction or perforation were excluded from the analysis.

Peripheral blood samples were collected before surgery. The serum CEA level was determined by radioimmunoassay. The CEA level was considered high at ≥ 5 ng/ml. The resected specimens were pathologically classified according to the 7th edition of the Union for International Cancer Control TNM classification of malignant tumors.

All patients underwent follow up with regular physical and blood examinations, colonoscopy, and computed tomography. All statistical analyses were performed using the SPSS software package version 22.0 for Macintosh (IBM Japan, Tokyo, Japan). The significance of the correlations between the preoperative CEA level and the pathological features was analyzed using the chi-squared test for independence according to each parameter. In order to control for confounding factors, binary logistic regression was used. Wald test was used to evaluate the significance of the association. Survival curves were plotted with the Kaplan–Meier method and analyzed with a log-rank test. P < 0.05 was considered statistically significant.

Results

Clinical characteristics of CRCs

As shown in Table 1, a total of 990 CRC patients were included, comprising 371 with rectal cancer and 619 with colon cancer. The mean ages of colon cancer and rectal cancer patients were 69.5 years and 66.4 years, respectively. Of these, 508 (82.1%) of the colon cancer patients and 285 (76.8%) of the rectal cancer patients were treated laparoscopically. The median follow-up periods were 42.2 months for colon cancer and 41.8 months for rectal cancer. Forty-four recurrences occurred in both the colon cancer (7.1%) and rectal cancer (11.9%) groups. We observed significant differences between the colon and rectal cancer patients regarding sex, average age, postoperative recurrence rate, gross type, serum CEA levels, and vascular invasion (all P < 0.05). Other features were not significantly different, including open or laparoscopic methods, differentiation, invasion depth (T), perineural invasion, cancer diameter, and infiltration of lymphatic vessels.

Table 1 Clinicopathological parameters in stage II colon and rectal cancer
Full size table

Comparisons of clinicopathological parameters between CRC recurrence and non-recurrence

Rectal cancer recurrence was associated with a body mass index greater than 25 kg/m2 (P = 0.0001), a larger tumor size (6.00 ± 0.34; P < 0.001), advanced T stage (P < 0.005), higher serum CEA levels (P < 0.002), poor differentiation or mucinous histology (P < 0.007), perineural invasion (P < 0.001), vascular invasion (P = 0.028), and infiltration of lymphatic vessels (P < 0.001) (Table 2).

Table 2 Clinicopathological parameters for rectal cancer recurrence and non-recurrence
Full size table

Compared with rectal cancer, colon cancer recurrence was associated with just an advanced T stage (P < 0.0001) and higher serum CEA levels (P = 0.009) (Table 3).

Table 3 Clinicopathological parameters for colon cancer recurrence and non-recurrence
Full size table

Correlations between the preoperative CEA levels and the site of recurrence

Local recurrences were significantly more common for rectal cancers with a higher CEA level than for those with a lower CEA level (P < 0.05). However, there was no significant difference in liver metastasis, lung metastasis, or peritoneal spread between the two groups (Table 4).

Table 4 Correlations between the preoperative CEA levels and the site of recurrence in stage II rectal cancer and colon cancer
Full size table

Although the patients with higher CEA levels were more likely to develop colon cancer recurrence, there was no significant difference in the site of recurrence (Table 4).

Correlations between the depth of infiltration and the site of recurrence

Local recurrences were significantly more common for colorectal cancers with T4 infiltration than for those with T3 infiltration (both P < 0.05) (Table 5). There was a higher rate of lung metastasis recurrence in patients with T4 rectal cancer compared with T3 (P = 0.004). In patients with T4 colon cancer, there was a higher rate of peritoneal metastasis (P = 0.014).

Table 5 Correlations between the depth of infiltration and the site of recurrence in stage II rectal cancer and colon cancer
Full size table

In multivariate analysis, a higher CEA level was associated with a higher chance of recurrence of both rectal cancer (odds ratio [OR] 1.011, 95% confidence interval [95% CI] 1.00–1.021, P = 0.048) and colon cancer (OR 1.010 95% CI 1.003–1.017, P = 0.004). In addition, T4 cancer had a higher chance of recurrence in both rectal cancer (OR 3.867, 95% CI 1.547–9.663, P = 0.004) and colon cancer (OR 3.222, 95% CI 1.238–8.390, P = 0.017) (Table 6). In colon cancer, the postoperative recurrence rate of left colon cancer (descending colon, sigmoid colon) was higher than that of right colon cancer (cecum, ascending colon, transverse colon) (OR 2.191, 95% CI 1.091–4.400, P = 0.027). However, there was no significant difference between low rectal cancer and upper rectal cancer.

Table 6 Multivariate logistic regression analysis evaluating possible risk factors associated with recurrence
Full size table

Correlations of the preoperative CEA levels and depth of infiltration with the survival rate

The overall survival rate was significantly lower in both colon and rectal cancer patients with high levels of CEA and in T4 patients (P = 0.005, 0.006, 0.0044, and 0.006, respectively) (Figs. 1a, b and 2a, b). High levels of serum CEA and T4 reduced the disease-free survival (P = 0.005, 0.001, 0.000, and 0.000) (Figs. 3a, b and 4a, b).

Fig. 1
figure 1

Survival outcomes for colorectal cancer patients with CEA ≧ 5 vs. CEA < 5

Full size image
Fig. 2
figure 2

Survival outcomes for colorectal cancer patients with T3 vs. T4

Full size image
Fig. 3
figure 3

Disease-free survival for colorectal cancer patients with CEA ≧ 5 vs. CEA < 5

Full size image
Fig. 4
figure 4

Disease-free survival for colorectal cancer patients with T3 vs. T4

Full size image

Prognostic differences between the left colon and the right colon

To compare the prognosis of histological parameters determined in univariate analysis, Cox’s proportional hazard regression model was applied. For DFS, left colon, vascular invasion, infiltrating pattern, and T4 were all shown to be independent risk factors. For OS, signet ring cell carcinoma, mucinous carcinoma, poorly differentiated carcinoma, vascular invasion, infiltrating pattern etc. were shown to be independent risk factors (Table 7). For the DFS Kaplan-Meier survival curve, the left colon also has a relatively poor prognosis. However, in the OS curve, there is no difference between the left colon and the right colon (Fig. 5a, b). There is no statistical difference in the ratio of T4:T3 (P = 0.337) and CEA ≥ 5 ng/ml (P = 0.32) in left colon cancer and right colon cancer (detailed data not shown).

Table 7 Cox proportional hazard regression model comparing the effects of different parameters on the prognosis of patients with stage II Colon cancer
Full size table
Fig 5
figure 5

Prognosis of left colon cancer vs. right colon cancer

Full size image

Discussion

The question of whether colon and rectal cancer should be treated as a single entity or as two separate entities remains controversial [8]. The treatment of stage II CRCs has also been extensively debated [7, 9]. Our findings indicate that these two groups show considerable differences in many clinic-pathological characteristics. Significant differences were seen between the two groups in sex, average age, postoperative recurrence, gross type, surgical procedure (open/laparoscopic), serum CEA level, and vascular invasion. Therefore, our results showed that, in stage II CRC, the characteristics of colon cancer and rectal cancer were different, which is why they should be considered two separate entities.

Out of 4,244 primary CRC patients, 990 cases (23.3%) of stage II CRC were found in our hospital in the past 10 years. In view of the different characteristics of rectal cancer and colon cancer, we independently analyzed their relapse characteristics. The depth of infiltration and level of the preoperative tumor marker CEA were directly related to recurrence in the colon cancer recurrence group compared with the non-recurrence group. This is consistent with the results of another study [10]. However, rectal cancer recurrence is much more complicated. It was not only related to the serum CEA level and depth of infiltration, but also obesity, tumor size, tumor type, and postoperative pathological lymphatic infiltration and perineural invasion. Therefore, CEA level and depth of infiltration are the common factors for the recurrence of the stage II CRC. These results are largely consistent with those of another study that also found that CEA and CA19-9 were factors associated with recurrence [11, 12].

We then looked at whether CEA was associated with the site of tumor recurrence. Although CEA is a complex glycoprotein that is the most commonly used tumor marker for CRC, it is highly nonspecific. In colon cancer, we did not find any relationship between CEA and the location of tumor recurrence but it was significantly associated with the local recurrence of rectal cancer. Local recurrence of rectal cancer involves lymph nodes near the sacrum, lymph nodes around the great arteries, and lateral lymph nodes. In this study, the preoperative CEA level and depth of infiltration (T4) were risk factors and prognostic factors for recurrence of stage II colon cancer. The different biological characteristics between rectal cancer and colon cancer lead to diverse recurrence factors and prognostic factors.

We found that the incidence of postoperative recurrence for left colon cancer was higher than that of right, and left colon is one of the independent risk factors of poorer DFS, which was consistent with some reports [13, 14]. The majority of the literature differs, reporting that there is no difference [15] or that the right colon has a relatively poor prognosis [16,17,18,19]. The inclusion of the rectum in the left colon is the biggest difference between these reports. This study also shows that many biological characteristics of the colon and rectum are different. The distinction between the descending colon and the sigmoid colon from the rectum may be more conducive to future treatment.

But more surprisingly, there was no difference in OS between the left colon and the right colon. This may be because the left colon, when compared with the right colon, even though RAS/BRAF are wild-type, has more advantages in the choice of chemotherapy drugs cetuximab or panitumumab [20]. It is important to recognize the limitations of this study. The chemotherapy effect of CRC is closely related to KRAS and BRAF gene mutations or microsatellite instability (MSI) [21]. In future analysis, research on genetic components needs to be strengthened. Second, this study is a single-center retrospective study with a relatively small sample size, and we look forward to future multi-center clinical studies.

Conclusion

Our results indicate that the preoperative CEA level and depth of infiltration (T4) are high-risk factors associated with recurrence and are prognostic factors in stage II colorectal cancer. Left colon is also a risk factor for the postoperative recurrence of stage II colon cancer, and special attention should be paid during follow-up.