Clinicopathological characteristics
A total of 3,029 patients with primary colorectal cancer (splenic flexure to rectal cancer) were referred to Kurume University Hospital from November 1974 to March 2017. Among these patients, 1,948 met the inclusion criteria. In total, 360 cases were excluded from the investigated cohort because R0 resection could not be performed. The final study population consisted of 1,614 patients (Fig. 2). Among these patients, 23 cases (1.4%) had No.253 LNM.
Patients’ characteristics are shown in Table 1. Patients’ median age was 64 years (range, 16–92 years), and 1,014 patients (63%) were men and 600 (37%) were women. No significant differences were identified between No.253 LNM-positive and No.253 LNM-negative patients in terms of sex, BMI, CEA levels, year of surgery, tumor site, tumor diameter, number of retrieved nodes, adjuvant therapy, or perineural invasion. A univariate analysis revealed that age, T category, number of metastatic LNs, TNM stage, neuropathy, tumor grade, lymphatic invasion, and venous invasion were associated with No.253 LNM (Table 1). Multivariate analysis demonstrated that three or more metastatic LNs was an independent predictor of No.253 LNM positivity (odds ratio [OR] 26.816, 95% confidence interval [CI] 5.839–123.151, p<0.0001). No other clinicopathological factors were independent predictors for No.253 LNM. To determine independent characteristics associated with No.253 LNM, we identified three or more metastatic LNs (odds ratio (OR) 26.816, 95% confidence interval (CI) 5.839–123.151, p<0.0001) as an independent predictor of No.253 LNM positivity (Table 1). Previous reports identified T4 as a risk factor for No.253 LNM [4, 18], but in this study, 20 (%) of the No.253 LNM-positive group had pT4, but this was not an independent risk factor (OR 2.615, 95% CI 0.726–9.420, p=0.109).
Surgical method
No.253 LN dissection was performed in all cases; these details are summarized in Table 2. Left-hemicolectomy (LHC) was performed in all splenic flexure cases and 61% of descending colon cancer cases. Sigmoidectomy was performed in 95% of cases with sigmoid colon cancer, and anterior resection (AR) was performed in 97% of cases with rectosigmoid cancer and in 96% of the cases with upper rectal cancer. For the cases with lower rectal cancer, 39% of cases were treated with abdominoperineal resection (APR), 35% of cases with AR, and 24% of cases with intersphincteric resection (ISR).
Distribution of metastatic lymph nodes in left colorectal cancer
The distribution of metastatic LNs in left-sided colorectal cancer in this study is summarized in Table 3. No No.253 LNM was observed in cases with tumors located in the splenic flexure or descending colon. In cases with sigmoid colon cancer, LNM was observed in 273 of 498 cases, of which 60.0% cases had No.241 LNM and 2.9% cases had No.253 LNM. In rectosigmoid cancer, LNM was observed in 159 of 289 cases, of which 67.9% cases had No.251 LNM and 1.9% cases No.253 LNM. In cases with upper rectal cancer, LNM was observed in 135 of 255 cases. Of these, 70.4% had No.251 LNM, 3.0% had No.253 LNM, and 5.9% had lateral LNM (LLNM). In cases with lower rectal cancer, LNM was observed in 287 of 505 cases. Of these, 61.3% had No.251 LNM, 2.8% had No.253 LNM, and 17.8% had LLNM.
Univariate and multivariate analyses of risk factors for overall survival
According to the univariate analysis, no significant associations were identified between tumor site, tumor diameter, perineural invasion, or postoperative complications (Table 4). Multivariate analysis revealed that year of surgery (1995–2017 [193 months] vs 1974–1994 [180 months]; HR 16.491, 95% CI 12.922–21.046, p<0.0001), age (<56 years [337 months] vs ≥56 [170 months]; HR 1.207, 95% CI 1.025–1.422, p=0.024), CEA level (≤5 ng/ml [210 months] vs <5 ng/ml [170 months]; HR 1.176, 95% CI 1.003–1.380, p=0.046), surgery length (<278 mins [213 months] vs ≥278 mins [179 months]; HR 1.669, 95% CI 1.410–1.976, p<0.0001), intraoperative bleeding volume (<280 ml [229 months] vs ≥280 [175 months]; HR 0.52, 95% CI 0.435–0.622, p<0.0001), No.253 LNM (No.253 LNM negativity [200 months] vs No.253 LNM positivity [57 months]; HR 0.431, 95% CI 0.187–0.994, p=0.048), lymphatic vessel invasion (no [205 months] vs yes [149 months]; HR 1.255, 95% CI 1.023–1.539, p=0.029), and postoperative adjuvant chemotherapy (yes [212 months] vs no [174 months]; HR 0.84, 95% CI 0.723–0.975, p=0.022) were associated with OS (Table 4).
Recurrence and survival
To identify the clinical importance of the No.253 LNM in left-side colorectal cancer, OS and DFS were compared between the No.253 LNM-negative group and the No.253 LNM-positive group. As shown in Figure 3, DFS and OS of the No.253 LNM-positive group was significantly worse than that of the No.253 LNM-negative group. Five-year OS rates were 49.1% in the No.253 LNM-positive group and 78.4% in the No.253 LNM-negative group, with a median follow-up time of 60 months (range, 1–341 months) (p=0.002; Fig. 3).
During the follow-up period after radical surgery, tumor recurrence was observed in 329 of 1,614 cases (20%) in this study. Of the 329 cases, 11 cases (47%) were in the No.253 LNM-positive group and 318 cases (20%) were in the No.253 LNM-negative group. n=11 (47%) vs No.253 LNM-negative group n=318 (20%). Especially, lung metastasis was significantly greater in the No.253 LNM-positive group than in the No.253 LNM-negative group (OR 3.765, p=0.013; Table 5). The No.253 LNM-positive group was more likely to have lung metastasis recurrence than the No.253 LNM-negative group (OR 3.765, p=0.013; Table 5).