Introduction

Colorectal cancer (CRC) is the third most common cancer in humans1. Although there are many therapeutic options nowadays, radical surgery is still the most important treatment for CRC2,3,4,5. At present, the Tumor Node Metastasis (TNM) system is the most commonly used tool to evaluate the survival results in clinical practice. The number of metastatic lymph nodes plays an important role in the TNM system. Many studies demonstrated that a higher number of lymph nodes retrieved would lead to more accurate staging and significantly improved survival outcomes6,7,8,9. However, the N-stage assessment could be influenced easily by the extent of lymph node dissection, surgeons’ technique, tumor location, and the thoroughness of lymph node dissection10,11,12. It leads to an insufficient number of lymph nodes, which is not enough to accurately predict prognosis.

The survival predictive capacity of lymph node ratio (LNR, the number of metastatic lymph nodes divided by the total number of lymph nodes resected) seems to be less dependent on the number of resected nodes. In addition, previous studies reported that LNR might reduce the stage migration phenomenon and consequently, allow more accurate prediction of prognosis than using the number of metastatic lymph nodes alone13,14. The prognostic value of LNR had been demonstrated in many cancers including breast cancer, gastric cancer, pancreatic cancer and CRC15,16,17,18,19.

There was an increasing number of studies supporting that LNR could be used as a marker for predicting CRC survival, however, most of them were small-scale studies20,21,22,23. In addition, the cut-off value of LNR varied among these studies, meanwhile, no consensus had been reached and no studies discussed the impact of LNR on postoperative complications. Therefore, this study aimed to investigate the effect of LNR on the short-term and long-term outcomes of patients after CRC surgery.

Materials and methods

Patients

This was a retrospective study that included 1434 patients who underwent radical CRC surgery from Jan 2011 to Jan 2020 in a single-center hospital. Ethical approval from the ethics committee of the First Affiliated Hospital of Chongqing Medical University was obtained (2022-K396), and informed consent was acquired for all patients. This study was conducted by the Declaration of Helsinki.

Inclusion and exclusion criteria

A total of 5473 CRC patients were collected. Patients who underwent radical CRC surgery in a single clinical center were included in this study. The exclusion criteria were as follows: 1, stage IV CRC patients (n = 875); 2, non-R0 CRC surgery (n = 25); 3, incomplete clinical data (n = 399); and 4, no metastatic lymph nodes were detected (n = 2740). Finally, a total of 1434 eligible CRC patients were enrolled for final analysis. The flow chart of inclusion and exclusion was shown in Fig. 1.

Figure 1
figure 1

Flow chart of patient selection. LNR lymph nodes ratio.

Full size image

Surgery management and follow-up

According to clinical guidelines, this study enrolled all patients who underwent radical resection, including total mesorectal excision or complete mesocolic excision, which was pathologically confirmed as R0 resection. Patients were followed by telephone review.

Definitions

The TNM stage was diagnosed according to the AJCC 8th Edition24. The complications were defined according to the Clavien-Dindo classification, and ≥ III classification complications were defined as major complications25. The time interval from the date of surgery to the time of last follow-up or death was defined as overall survival (OS); while the time interval from the date of surgery to the time of the last follow-up or pathological and imaging diagnosis of tumor recurrence or metastasis was defined as disease-free survival (DFS).

Data collection

The baseline information comprised age, sex, body mass index (BMI), smoking, drinking, type 2 diabetes mellitus (T2DM), tumor location, tumor size, T stage, metastatic lymph nodes, total lymph nodes and LNR. The surgery-related information included lymph-vascular space invasion, nerve invasion, cancer nodules, operation time, estimated blood loss. The short-term outcomes consisted of hospital stay, overall complications, major complications and various postoperative complications. All information was collected from the medical record system and telephone interviews.

LNR optimal cut-off

The median value of LNR was 18.2%, which was divided into the high group and the low group in different T stage.

Propensity score matching

We used PSM to reduce the selection bias in baseline information between the low LNR group and the high LNR group. Patients in the low LNR group were matched to patients in the high LNR group at a 1:1 proportion, and the caliper width was 0.02 SD. Age, sex, BMI, smoking, drinking, T2DM, tumor size, tumor location and T stage were included for matching.

Statistical analysis

Continuous variables were expressed as the mean ± standard deviation (SD), and categorical variables were expressed as n (%). Chi-square tests and independent-sample t-tests were used to compare the differences between the high LNR group and the low LNR group. Univariate and multivariate logistic regression analyses were performed to identify independent predictive factors for OS and DFS. Kaplan–Meier method was used to estimate the OS and DFS between the two groups in different T stages. Data were analyzed using the SPSS (version 22.0) statistical software. And a value of p < 0.05 was considered statistically significant.

Ethics approval and informed consent

The study was approved by the ethics committee of the First Affiliated Hospital of Chongqing Medical University was obtained (2022-K396), and informed consent was acquired from all patients. This study was conducted by the Declaration of Helsinki.

Results

Patients

A total of 1434 eligible CRC patients were enrolled in the study according to the inclusion and exclusion criteria. The low LNR group had 730 (50.9%) patients and the high LNR group had 704 (49.1%) patients. After PSM analysis, 618 patients were included in each group. (Fig. 1) The baseline information, surgery-related information and short-term outcomes of all patients is shown in Table 1.

Table 1 Clinical characteristics.
Full size table

Comparison of baseline information before and after PSM

After comparing the baseline information, the high LNR group had more older patients (p = 0.021), higher BMI (p = 0.016), higher proportions of rectal tumors (p < 0.01) and large tumor size (≥ 4 cm) (p = 0.020) than the low LNR group. Sex, smoking, drinking, T2DM, T stage had no significant difference (p > 0.05). There was no significant difference in all baseline information between the two groups after the 1:1 proportion PSM (p > 0.05) (Table 2).

Table 2 Comparison baseline characteristics between the high LNR group and the low LNR group before and after PSM.
Full size table

Comparison of surgery-related information and short-term outcomes before and after PSM

Before PSM, the high LNR group had a higher proportion of lymph-vascular space invasion (p < 0.01), more intraoperative estimated blood loss (p < 0.01) and 30-day death (p = 0.035). After PSM, the high LNR group had a longer hospital stay (p < 0.01). There was no significant difference in other information before and after PSM, including nerve invasion, cancer nodules, operation time, overall complications, major complications, anastomotic leakage, incision Infection, lymphatic fistula, intestinal obstruction, hypoproteinemia, cardiovascular complications, pulmonary complications, urinary complications, disturbance of water and electrolyte balance, vein thrombosis, and abdominal infection (p > 0.05) (Table 3).

Table 3 Summary of outcomes between the high LNR group and the low LNR group before and after PSM.
Full size table

Univariate and multivariate logistic regression analyses of the OS

Through using univariate and multivariate logistic regression analyses, we found that age (p < 0.01, HR = 1.026, 95% CI 1.016–1.035), tumor location (p = 0.024, HR = 1.289, 95% CI 1.034–1.607), lymph-vascular space invasion (p < 0.01, HR = 1.942, 95% CI 1.402–2.689), cancer nodules (p < 0.01, HR = 1.660, 95% CI 1.289–2.139), tumor size (p < 0.01, HR = 1.357, 95% CI 1.085–1.696), LNR (p < 0.01, HR = 2.035, 95% CI 1.609–2.573), and overall complications (p < 0.01, HR = 1.704, 95% CI 1.358–2.139) were independent risk factors for OS. The other factors including sex, BMI, T2DM, T stage, nerve invasion, smoking and drinking had no predictive value for OS (Table 4).

Table 4 Univariate and multivariate analysis of overall survival.
Full size table

Univariate and multivariate logistic regression analyses of the DFS

Univariate and multivariate logistic regression analyses were used to find out independent predictors for DFS. We found that age (p < 0.01, HR = 1.018, 95% CI 1.010–1.027), tumor location (p = 0.031, HR = 1.250, 95% CI 1.021–1.531), T stage (p = 0.014, HR = 1.212, 95% CI 1.040–1.412), lymph-vascular space invasion (p < 0.01, HR = 1.666, 95% CI 1.204–2.304), cancer nodules (p < 0.01, HR = 1.575, 95% CI 1.247–2.990), LNR (p < 0.01, HR = 1.996, 95% CI 1.613–2.471), and overall complications (p < 0.01, HR = 1.478, 95% CI 1.192–1.832) were identified as independent risk factors for DFS. The other factors including sex, BMI, T2DM, nerve invasion, smoking drinking and tumor size had no predictive value for DFS (Table 5).

Table 5 Univariate and multivariate analysis of disease-free survival.
Full size table

Comparison between the high LNR group and the low LNR group in patients with rectal or colon cancer

After comparing the baseline characteristics between the high LNR group and the low LNR group in rectal cancer patients, the high LNR group had a higher BMI (p = 0.033), more intraoperative estimated blood loss (p < 0.01). In colon cancer patients, the high LNR group had more older patients (p = 0.039), more intraoperative estimated blood loss (p = 0.030) and a higher proportion of overall complications (p = 0.020). The other baseline characteristics had no significant differences (p > 0.05). (Table S1, S2).

Kaplan–Meier curves in different T stages

Kaplan–Meier method was used to estimate the OS and DFS between the two groups in different T stages. We found that the high LNR group had a worse OS (Fig. 2) in T3 (p < 0.01) and T4 (p < 0.01) as well as a worse DFS (Fig. 3) in T3 (p < 0.01) and T4 (p < 0.01).

Figure 2
figure 2

Kaplan–Meier survival curve for the impact of high LNR group and low LNR group on the OS of patients in T stage I–IV. (a) OS of T1 stage CRC patients (b) OS of T2 stage CRC patients (c) OS of T3 stage CRC patients (d) OS of T4 stage CRC patients. LNR lymph nodes ratio; OS overall survival.

Full size image
Figure 3
figure 3

Kaplan–Meier survival curve for the impact of high LNR group and low LNR on the DFS survival of patients in T stage I-IV. (a) DFS of T1 stage CRC patients (b) DFS of T2 stage CRC patients (c) DFS of T3 stage CRC patients d. DFS of T4 stage CRC patients. LNR lymph nodes ratio; DFS disease-free survival.

Full size image

Discussion

A total of 1434 CRC patients who underwent radical CRC surgery were included in this study. And there were 618 patients in the low LNR group and the high LNR group after PSM. In multivariate logistic regression analyses, we found that age, tumor location, lymph-vascular space invasion, cancer nodules, tumor size, LNR, and overall complications were independent risk factors for OS. Age, tumor location, T stage, lymph-vascular space invasion, cancer nodules, LNR, and overall complications were identified as independent risk factors for DFS. Furthermore, it was worth noting that the high LNR group had a worse OS and DFS in T3 and T4 stages.

Many previous studies confirmed the role of LNR in predicting the survival of CRC. Jiang K et al. divided LNR into four groups based on quartiles and found that the value of LNR at 0.167–0.562 were significant prognostic factors for OS and DFS26. Shumacher et al. reported that an LNR of 0.18 was an important prognostic factor for DFS, but not for OS27. In terms of the best cut-off value, Schumacher's results were consistent with the current study. However, there was no agreement on the cut-off value of LNR and no studies discussed the impact of LNR on postoperative complications. Therefore, this study aimed to investigate the effect of LNR on the short-term and long-term outcomes of patients after CRC surgery.

In our research, we found that LNR was a very powerful prognostic factor for OS and DFS in CRC, which was consistent with previous studies. Furthermore, we expect to explore the effect of LNR on prognosis in different T stages. Thus, we divided LNR into the high group and low group according to the median, and we found that high LNR had a worse OS and DFS under T3 and T4 stages. This meant that LNR was more prognostically significant under T3 and T4 stages for CRC patients.

Lymph node metastasis was the most important way of metastasis in CRC28,29. More lymph node metastasis would require lymph node dissection, which not only increases the difficulty of surgery, but also increases the length of hospital stay. In our research, we divided LNR values into high and low groups according to the median and found that high LNR was associated with a longer hospital stay. Previous studies had shown that lymph node metastasis increases the risk of postoperative complications30. However, we just found a longer hospital stay in the high LNR group, the other postoperative complications had no statistical significance. The specific reasons might need further research to explain.

To our knowledge, this current study had the largest sample size in western China, and for the first time revealed the relationship between LNR and short-term complications in CRC patients. However, some limitations existed in this current study as well. First, the number of patients in the T1 and T2 stages were small, which might lead to data bias. Second, there was a lack of chemotherapy information in our study, which might affect the reliability of the survival analysis. Third, we suggeated further prognostic analysis more focused on the collinearity between LNR and other classic prognostic indicators. Finally, this study was a retrospective study conducted in a single center in western China, thus, the current results applied to restricted regions. Therefore, multicenter prospective randomized controlled trial studies should be conducted further to verify the applicability of our findings.

In conclusion, we found that the relationship between LNR and postoperative complications was not statistically significant, however, LNR was identified as an independent predictor for OS and DFS. Furthermore, high LNR had a worse OS and DFS under T3 and T4 stages. Therefore, LNR was more prognostically significant under T3 and T4 stages for CRC patients.