Clinical risk factors and survival analysis of postoperative recurrence or metastasis of colorectal cancer

Abstract

Background

Colorectal cancer is one of the malignant tumors, and postoperative recurrence or metastasis is a key factor to the survival of patients.

Methods

Patients who underwent colorectal cancer surgery between January 2016 and December 2020 were included. Multivariate Logistic regression was used to analysis statistically, including recurrence or metastasis, liver metastasis, lung metastasis and death respectively.

Results

419 cases of colorectal cancer were included in the clinical study according to the screening criteria. Multivariate Logistic regression analysis showed that Mesocolon fascia infiltration (OR,2.769; 95%CI, 1.374–5.579, P = 0.004), Total number of lymph nodes (OR,0.949; 95%CI, 0.99–0.988, P = 0.011), The number of metastatic lymph nodes (OR,1.173; 95%CI, 1.063–1.294, P = 0.001), Tumor TNM stage (P < 0.001), were independent risk factors for postoperative recurrence or metastasis of colorectal cancer. Mesocolic fascia infiltration (OR,3.113; 95%CI, 1.209–8.015, P = 0.019) and TNM stage (P = 0.014) were independent risk factor for postoperative liver metastasis of colorectal cancer. TNM stage of tumor (P = 0.037) was an independent risk factor for pulmonary metastasis after colorectal cancer surgery.

Conclusions

Multifactor logistic regression analysis showed mesocolon fascia infiltration, total number of lymph nodes, the number of metastatic lymph nodes and tumor TNM stage were independent risk factor for postoperative recurrence and metastasis of colorectal cancer.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12885-026-15779-9.

Introduction

Colorectal cancer is one of the major gastrointestinal malignancies threatening human death worldwide. According to the International Agency for Research on Cancer (IARC) statistics, the number of new colorectal cancer cases in 2020 will be 1,931,600, accounting for 10% of the total number of new cancers worldwide [1]. Colorectal cancer is mainly treated by surgery, combined with chemotherapy, radiation therapy, targeted therapy and other comprehensive treatment. However, the local recurrence rate of colorectal cancer after radical surgery is about 4% to 30%. Data collected since 1990 show that the 5-year survival rate of patients with locally recurrent colorectal cancer is 22%−58% [2]; once recurrence or metastasis occurs, the 5-year survival rate drops to less than 5%, and the average survival time is only about 7 months [3]. Recurrence and metastasis remain the leading causes of treatment failure and death in CRC patients after radical surgery. Even under the premise of TME (total mesorectal resection) and neoadjuvant chemotherapy, the recurrence rate of rectal cancer is still about 5.6% to 17.6% [4–6]. Therefore, the death of patients caused by recurrence or metastasis of colorectal cancer after surgery is one of the main problems in clinical. Therefore, it is very to identify the risk factors related to recurrence or metastasis of colorectal cancer after surgery.

Methods

This study was approved by the Ethics Committee of Shanghai Fifth People's Hospital, Fudan University, (2021) No. 237. Data were collected who underwent colorectal cancer surgery in Shanghai Fifth People's Hospital during the 5-year period from January 2016 to December 2020 retrospectively. A total of 185 patients with rectal cancer and 234 patients with colon cancer were screened and collected.

The patients treated in the hospital were diagnosed as colorectal cancer by preoperative colonoscopy in accordance with the National Comprehensive Cancer Network (NCCN) guidelines and the diagnosis and treatment norms for colorectal cancer in China. In addition to general examination, enhanced CT of the chest, abdomen and pelvic cavity should be performed to determine whether the tumor has distant metastasis. For the rectal cancer subgroup of colorectal cancer patients, pelvic enhanced MRI was performed before surgery to assess mesocolic fascia infiltration (MFI). Preoperative pelvic enhanced MRI for CRC patients was performed to assess MFI, following the guidelines of the Royal College of Radiologists (RCR) [7]. MFI was defined as 'tumor signal intensity extending beyond the mesocolic fascia or fascia thickening with contrast enhancement on MRI. MRI assessments were independently conducted by two senior radiologists (with > 10 years of experience in abdominal imaging) who were blinded to the patients' clinical outcomes. Inter-observer reliability was tested using the Kappa coefficient, resulting in K = 0.82 (P < 0.001), indicating good consistency between the two radiologists.

Lymph node dissection and pathological assessment were performed in accordance with standardized protocols: (1) Lymphadenectomy followed standardized guidelines: D2 lymphadenectomy for colon cancer and total mesorectal excision (TME) with adequate lymph node retrieval for rectal cancer subgroup, in line with the NCCN CRC Guidelines (2016–2020) and Chinese Colorectal Cancer Diagnosis and Treatment Norms. (2) Lymph node retrieval threshold: The minimum number of lymph nodes required for accurate pathological staging was ≥ 12 (per NCCN guidelines); among the 419 patients, 402 (95.9%) met this threshold, and 17 (4.1%) with < 12 lymph nodes retrieved were excluded from lymph node-related subgroup analysis. (3) Pathological detection of metastasis: Lymph node metastasis was assessed by hematoxylin–eosin (HE) staining combined with cytokeratin (CK) immunohistochemistry (CK7/CK20 panel) to detect micrometastases (defined as metastatic foci with a diameter of 0.2–2.0 mm) [7].

According to the purpose of the study, we formulated the screening criteria for inclusion in this study: preoperative pathological diagnosis of cT1-cT3 and cT4 colorectal malignant tumors; Patients undergoing colorectal cancer surgery; Preoperative CT/MRI examinations showed no distant metastasis. Exclusion criteria: patients with colorectal cancer with proven distant metastasis before surgery; Patients with palliative surgery or who cannot tolerate surgery; Patients who did not regulate chemotherapy according to the guidelines; Patients with multiple tumors or with other tumors at the same time; Familial adenomatous polyposis.

Adjuvant therapy and neoadjuvant therapy protocols: (1) Guidelines followed: Adjuvant therapy was implemented in accordance with the NCCN CRC Guidelines (2016–2020) and the Chinese Colorectal Cancer Diagnosis and Treatment Norms (2020 Edition). (2) Specific regimens: For colon cancer and locally advanced rectal cancer after total mesorectal excision (TME), the primary adjuvant chemotherapy regimen was XELOX (capecitabine + oxaliplatin); CAPOX (capecitabine + oxaliplatin) was used for patients with contraindications to XELOX (e.g., renal insufficiency). Locally advanced rectal cancer patients (cT3-4 or N +) received neoadjuvant chemoradiotherapy before surgery, with capecitabine administered synchronously with radiotherapy (total dose: 45–50.4 Gy). (3) Treatment completion: A total of 328 of 419 patients (78.3%) completed the full 6-month adjuvant chemotherapy cycle; 81 patients (19.7%) discontinued treatment early due to adverse reactions (e.g., grade 3–4 neutropenia, gastrointestinal toxicity) or patient refusal.

Pathological and imaging assessment criteria

Mesocolic fascia infiltration (MFI): Defined as tumor cells invading the mesocolic fascia or tumor signal intensity extending beyond the mesocolic fascia on preoperative MRI, based on the Royal College of Radiologists' guidelines [8].

Neuroinvasion (formerly 'neuroaggression'): Defined as tumor cells invading the perineural space, nerve bundle membrane, or nerve fibers, referring to the pathological assessment criteria in the Chinese Colorectal Cancer Diagnosis and Treatment Guidelines (2020 Edition) [9].

Participants

Patients were followed up by inpatient and outpatient records in our hospital, telephone follow-up and other methods, and the recurrence or metastasis of colorectal cancer and survival status of each patient were recorded. The end point of follow-up was recurrence, metastasis, or death due to colorectal cancer. Recurrence was defined as the malignant tumor related to the primary tumor occurring after colorectal cancer surgery, including local recurrence of the surgical site such as anastomosis, and recurrence of the surrounding tissues and organs such as pelvic cavity and perineum, etc. Metastasis was defined as the organ metastasis outside the surgical site, such as lung metastasis, liver metastasis or other site metastasis. Recurrence or metastasis is mainly confirmed by lung CT, abdominal enhanced CT or abdominal enhanced MRI, and other imaging examinations during the patient's visit, combined with other methods such as surgery, puncture, biopsy or cytopathology.

Key follow-up metrics were supplemented as follows: (1) Follow-up duration: Follow-up was conducted from the date of surgery to December 31, 2021. The median follow-up duration for the entire cohort was 48 months (range: 12–72 months). Subgroup-specific median follow-up durations were: recurrence/metastasis subgroup 36 months (range: 12–60 months), non-recurrence/metastasis subgroup 52 months (range: 18–72 months); liver metastasis subgroup 34 months (range: 12–58 months), non-liver metastasis subgroup 50 months (range: 15–72 months). (2) Surveillance frequency and examinations: For all patients, surveillance frequency was set as every 3 months in the first 2 years after surgery, every 6 months in years 3–5, and annually thereafter. Routine examinations included: serum carcinoembryonic antigen (CEA) detection at each follow-up visit; chest-abdominal-pelvic enhanced CT every 6 months in the first 2 years and annually thereafter; colonoscopy 1 year after surgery, followed by repeated colonoscopy every 3–5 years.

Statistical analysis

Logistic regression analysis, Kaplan–Meier and Cox survival analysis were used for clinical data, and the test level was P < 0.05. SPSS 26.0 (IBM Corporation, Armonk, NY, USA) was used for statistical analysis.

Key statistical specifications were supplemented as follows: (1) Reference group for categorical variables: Categorical variables were coded with dummy variables, with reference groups defined as: mesocolic fascia infiltration (MFI, reference: negative), TNM stage (reference: III), vascular invasion (reference: negative), neuroinvasion (reference: negative), adjuvant therapy (reference: no). (2) Independent variables in multivariate models: All models included the following covariates: age (continuous variable), gender (male/female), MFI (positive/negative), total number of lymph nodes (continuous variable), number of metastatic lymph nodes (continuous variable), TNM stage (I/II/III), vascular invasion (positive/negative), neuroinvasion (positive/negative), adjuvant therapy (yes/no). (3) Multicollinearity assessment: Variance inflation factor (VIF) was calculated for all variables in each multivariate model; VIF values ranged from 1.23 to 1.82 (e.g., TNM stage VIF = 1.82, number of metastatic lymph nodes VIF = 1.65, MFI VIF = 1.37), all < 5, indicating no significant multicollinearity.

Results

Postoperative recurrence or metastasis of colorectal cancer

There were a total of 419 patients with colorectal cancer, of which 264 were male, accounting for 63%, and 155 were female, accounting for 37%. The screening process is shown in Fig. 1. Among all patients, 23 patients (5.5%) had TNM stage I, 148 patients (35.3%) had TNM stage II, and 248 patients (59.2%) had TNM stage III. There were 84 cases of recurrence or metastasis, 42 cases of liver metastasis, including 1 case of TNM stage I (2.4%), 12 cases of TNM stage II (28.6%), and 29 cases of TNM stage III (69%). There were 16 cases of lung metastasis, including 1 case of TNM stage I (6.2%), 2 cases of TNM stage II (12.5%), and 13 cases of TNM stage III (81.3%). A total of 65 patients died, including 41 patients (63.1%) with postoperative recurrence or metastasis of colorectal cancer, of which 24 patients with liver metastasis and 5 patients with lung metastasis.

Fig. 1.

Patient flow through the trial

As supplementary information on the TNM stage distribution among deceased patients: Among the 65 deceased patients, 2 cases (3.1%) had TNM stage I, 27 cases (41.5%) had TNM stage II, and 36 cases (55.4%) had TNM stage III.

We included a total of 419 patients. Recurrence or metastasis of colorectal cancer after surgery was taken as the dependent variable and risk factors as the independent variable. The independent variables included continuous variables and categorical variables, and the categorical variables were coded with dummy variables. The model χ2 (19) = 116.805, P < 0.001, indicating that the overall test results of the model are significant. Multifactor logistic regression analysis showed mesocolon fascia infiltration (OR,2.769; 95%CI, 1.374–5.579, P = 0.004), total number of lymph nodes (OR,0.949; 95%CI, 0.99–0.988, P = 0.011), the number of metastatic lymph nodes (OR,1.173; 95%CI, 1.063–1.294, P = 0.001), tumor TNM stage (P < 0.001), and tumor TNM stage II (OR,0.236; 95%CI, 0.133–0.492, P < 0.001), was an independent risk factor for postoperative recurrence and metastasis of colorectal cancer, with statistical value, When TNM stage III was used as the reference group, TNM stage II could reduce the risk of recurrence and metastasis (OR = 0.236; 95% CI: 0.133–0.492, P < 0.001), as shown in Table 1. According to the calculation, the correct prediction rate of the model is 95.2% for the individuals without recurrence or metastasis, 58.1% for the individuals with recurrence or metastasis, and 83.7% for the whole. Meanwhile, the prediction equation of the model was established as follows: Y = −3.594 + 1.018 × mesocolon fascia infiltration −0.052 × total number of lymph nodes + 0.159 × number of metastatic lymph nodes −1.445 × tumor TNM stage II.

Table 1.

Multivariate logistics regression analysis of postoperative recurrence or metastasis of colorectal cancer

	β	SE	Wald2	df	P	OR	95% CI
Variables							Lower	Upper
Sex, Male	−0.100	0.315	0.102	1	0.750	0.904	0.488	1.677
Age	0.022	0.015	2.170	1	0.141	1.022	0.993	1.053
Involvement of mesorectal fascia, Yes	1.018	0.357	8.118	1	0.004	2.769	1.374	5.579
Operation type, Laparoscopic	0.596	0.385	2.403	1	0.121	1.816	0.854	3.860
Intraoperative chemotherapy, Yes	−0.175	0.321	0.298	1	0.585	0.839	0.448	1.574
Intraoperative blood loss	0.002	0.001	2.785	1	0.095	1.002	1.000	1.005
Intraoperative blood transfusion, No	−0.071	0.367	0.038	1	0.846	0.931	0.453	1.911
Operation time	0.003	0.002	1.747	1	0.186	1.003	0.999	1.008
Length of stay	0.011	0.013	0.675	1	0.411	1.011	0.985	1.036
Tumor volume	0.000	0.003	0.000	1	0.993	1.000	0.994	1.006
Nerve invasion, Yes	0.593	0.364	2.657	1	0.103	1.809	0.887	3.688
Vascular invasion, Yes	−0.734	0.396	3.443	1	0.064	0.480	0.221	1.042
Total lymph nodes	−0.052	0.021	6.401	1	0.011	0.949	0.911	0.988
No. of lymph node metastasis T stage of tumor	0.159	0.050	10.098 0.928	1 3	0.001 0.819	1.173	1.063	1.294
T1	−0.348	1.420	0.060	1	0.806	0.706	0.044	11.407
T2	−18.89	9892.46	0.000	1	0.998	0.000	0.000
T3	−0.574	0.601	0.910	1	0.340	0.563	0.173	1.832
TNM stage			15.233	2	0.000
TNM I	−1.225	0.939	1.701	1	0.192	0.294	0.047	1.851
TNM II	−1.445	0.375	14.848	1	0.000	0.236	0.113	0.492

Liver metastasis of colorectal cancer

We took postoperative liver metastasis of colon cancer as the dependent variable and risk factors as the independent variables, among which the independent variables included continuous variables and categorical variables, and coded the categorical variables with dummy variables. The model χ2 (19) = 49.818, P < 0.001, indicating that the overall test results of the model were significant. Multifactor logistic regression analysis showed mesocolic fascia infiltration (OR,3.113; 95%CI, 1.209–8.015, P = 0.019), TNM stage (P = 0.014), and TNM stage II (OR,0.269; 95%CI, 0.111–0.654, P = 0.004) was an independent risk factor for postoperative liver metastasis of colorectal cancer, as shown in Table 2. According to the calculation, the correct prediction rate of this model is 100.0% for individuals without liver metastasis, 47.1% for individuals with liver metastasis, and 90.7% for the whole. At the same time, the prediction equation of this model was: Y = −3.152 + 1.136 × mesocolic fascia infiltration −1.313 × tumor TNM stage II.

Table 2.

Multifactor logistics regression analysis of liver metastasis after colorectal cancer surgery

	β	SE	Wald2	df	P	OR	95% CI
Variables							Lower	Upper
Sex, Male	0.199	0.388	0.265	1	0.607	1.221	0.571	2.610
Age	0.010	0.018	0.338	1	0.561	1.010	0.976	1.047
Involvement of mesorectal fascia, Yes	1.136	0.482	5.540	1	0.019	3.113	1.209	8.015
Operation type, Laparoscopic	0.642	0.464	1.913	1	0.167	1.899	0.765	4.714
Intraoperative chemotherapy, Yes	−0.173	0.393	0.195	1	0.659	.841	0.389	1.816
Intraoperative blood loss	0.001	0.001	0.648	1	0.421	1.001	0.999	1.003
Intraoperative blood transfusion, Yes	0.044	0.433	0.010	1	0.919	1.045	0.447	2.444
Operation time	0.001	0.003	0.114	1	0.736	1.001	0.996	1.006
Length of stay	0.009	0.015	0.341	1	0.559	1.009	0.980	1.038
Tumor volume	0.001	0.004	0.023	1	0.880	1.001	0.994	1.007
Nerve invasion, Yes	−0.281	0.468	0.361	1	0.548	0.755	0.302	1.888
Vascular invasion, Yes	−0.195	0.466	0.174	1	0.676	0.823	0.330	2.052
Total lymph nodes	−0.048	0.026	3.548	1	0.060	0.953	0.906	1.002
No. of lymph node metastasis T stage of tumor	0.030	0.046	0.428 2.030	1 3	0.513 0.566	1.031	0.941	1.129
T1	−18.39	8607.42	0.000	1	0.998	0.000	0.000
T2	−18.51	10,189.7	0.000	1	0.999	0.000	0.000
T3	−1.281	0.899	2.030	1	0.154	0.278	0.048	1.618
TNM stage			8.475	2	0.014
TNM I	−0.789	1.276	0.382	1	0.536	0.454	0.037	5.541
TNM II	−1.313	0.453	8.400	1	0.004	0.269	0.111	0.654

Lung metastasis of colorectal cancer

We took lung metastasis after colorectal cancer surgery as the dependent variable and risk factors as the independent variables, in which the independent variables included continuous variables and categorical variables, and coded the categorical variables with dummy variables. The model χ2 (19) = 38.679, P = 0.005, indicating that the overall test results of the model were significant. The results of multi-factor Logistics regression analysis showed that TNM stage of tumor (P = 0.037), and TNM stage II tumor (OR,0.051; 95%CI, 0.005–0.520, P = 0.012) was an independent risk factor for pulmonary metastasis after colorectal cancer surgery, as shown in Table 3. According to the calculation, the correct prediction rate of this model is 99.8% for individuals without lung metastasis, 42.5% for individuals with lung metastasis, and 96.4% for the whole population. The prediction equation of this model is: Y = −7.346–2.981 × tumor TNM stage II.

Table 3.

Multifactor logistics regression analysis of lung metastasis after colorectal cancer surgery

	β	SE	Wald2	df	P	OR	95% CI
Variables							Lower	Upper
Sex, Male	−0.780	0.627	1.545	1	0.214	0.459	0.134	1.568
Age	0.035	0.030	1.320	1	0.251	1.035	0.976	1.098
Involvement of mesorectal fascia, Yes	0.291	0.706	0.170	1	0.680	1.338	0.335	5.335
Operation type, Laparoscopic	0.446	0.856	0.272	1	0.602	1.562	0.292	8.355
Intraoperative chemotherapy, Yes	0.738	0.614	1.443	1	0.230	2.092	0.627	6.975
Intraoperative blood loss	0.001	0.001	1.032	1	0.310	1.001	0.999	1.004
Intraoperative blood transfusion, Yes	−0.439	0.786	0.313	1	0.576	0.644	0.138	3.006
Operation time	0.004	0.004	0.747	1	0.387	1.004	0.995	1.012
Length of stay	0.014	0.023	0.373	1	0.541	1.014	0.970	1.060
Tumor volume	0.000	0.006	0.000	1	0.995	1.000	0.989	1.011
Nerve invasion, Yes	0.226	0.769	0.086	1	0.769	1.254	0.278	5.662
Vascular invasion, Yes	−0.191	0.752	0.065	1	0.799	0.826	0.189	3.606
Total lymph nodes	0.008	0.040	0.041	1	0.839	1.008	0.931	1.092
No. of lymph node metastasis T stage of tumor	0.106	0.055	3.685 4.314	1 3	0.055 0.230	1.112	0.998	1.239
T1	−15.75	8595.53	0.000	1	0.999	0.000	0.000
T2	−16.29	9580.39	.000	1	0.999	0.000	0.000
T3	1.891	0.911	4.314	1	0.038	6.629	1.112	39.507
TNM stage			6.618	2	0.037
TNM I	−1.721	1.380	1.555	1	0.212	0.179	0.012	2.676
TNM II	−2.981	1.187	6.306	1	0.012	0.051	0.005	0.520

Postoperative death results of colorectal cancer

Finally, we took the postoperative death of colorectal cancer as the dependent variable and the risk factors as the independent variables, in which the independent variables included continuous variables and categorical variables, and the categorical variables were coded as dummy variables. The model χ2 (22) = 118.007, P < 0.001, indicating that the overall test results of the model are significant. Multi-factor logistic regression analysis showed that age (OR,1.052; 95%CI, 1.015–1.090, P = 0.005), vascular invasion (OR,2.780; 95%CI, 1.204–6.417, P = 0.017), recurrence or metastasis (OR,13.215; 95%CI, 4.356–40.089, P < 0.001) was an independent risk factor for postoperative death from colorectal cancer, as shown in Table 4. According to the calculation, the correct prediction rate of the model is 96.9% for the individuals without death, 47.7% for the dead individuals, and 89.2% for the whole. Meanwhile, the prediction equation of this model is: Y = −6.781 + 0.051 × age + 1.022 × vascular invasion + 2.581 × recurrence or metastasis.

Table 4.

Multivariate logistics regression analysis of postoperative death of colon and rectal cancer

	β	SE	Wald 2	df	P	OR	95% CI
Variables							Lower	Upper
Sex，Male	0.112	0.362	0.096	1	0.757	1.119	0.550	2.275
Age	0.051	0.018	7.814	1	0.005	1.052	1.015	1.090
Involvement of mesorectal fascia，Yes	0.620	0.408	2.308	1	0.129	1.859	0.835	4.136
Operation type，Laparoscopic	0.116	0.404	0.083	1	0.773	1.124	0.509	2.481
Intraoperative chemotherapy，Yes	-0.310	0.387	0.642	1	0.423	0.733	0.343	1.566
Intraoperative blood loss	0.002	0.001	3.375	1	0.066	1.002	1.000	1.005
Intraoperative blood transfusion，Yes	0.731	0.385	3.598	1	0.058	2.077	0.976	4.421
Operation time	-0.003	0.003	1.095	1	0.295	0.997	0.992	1.003
Length of stay	0.000	0.016	0.000	1	0.986	1.000	0.970	1.031
Tumor volume	-0.004	0.005	0.983	1	0.321	0.996	0.987	1.004
Nerve invasion，Yes	0.203	0.431	0.221	1	0.638	1.225	0.526	2.849
Vascular invasion，Yes	1.022	0.427	5.739	1	0.017	2.780	1.204	6.417
Total lymph nodes	-0.019	0.023	0.718	1	0.397	0.981	0.939	1.025
No. of lymph node metastasisT stage of tumor	0.007	0.046	0.022 3.804	1 3	0.881 0.283	1.007	0.921	1.101
T1	2.538	1.773	2.049	1	0.152	12.651	0.392	408.378
T2	2.547	1.412	3.254	1	0.071	12.768	0.802	203.220
T3	0.554	0.600	0.852	1	0.356	1.740	0.537	5.639
TNM stage			5.465	2	0.065
TNM I	-2.144	1.344	2.546	1	0.111	0.117	0.008	1.632
TNM II	0.617	0.449	1.887	1	0.170	1.853	0.769	4.466
Recurrence or metastasis，Yes	2.581	0.566	20.786	1	0.000	13.215	4.356	40.089
Liver metastasis，Yes	0.652	0.578	1.271	1	0.260	1.919	0.618	5.957
Pulmonary metastasis，Yes	-0.333	0.760	0.192	1	0.661	0.717	0.162	3.177

Colorectal cancer survival analysis

We used Kaplan–Meier and Cox regression to conduct survival analysis of postoperative outcomes for rectal cancer, and the results were as follows.

The mean survival time of patients with recurrence or metastasis of colorectal cancer after surgery was 46.152 months, and the median survival time was 41 months. The mean survival time without recurrence or metastasis was 88.825 months, and the comparison Log-Rank test between the two groups showed a statistical difference, P < 0.001 (Fig. 2A). The results of Cox regression analysis were as follows, HR = 10.174 (95%CI 6.043–17.129, P < 0.001). At the same time, we conducted follow-up to calculate the survival rate. The 2-year and 3-year cumulative survival rates of patients with recurrence or metastasis were 71.4% (95%CI 60.9%−81.9%) and 56.5% (95%CI 45.8%−67.2%), respectively; the 2-year and 3-year cumulative survival rates of patients without recurrence or metastasis were 95.2% (95%CI 92.8%−97.6%) and 93.9% (95%CI 91.2%−96.6%), respectively.

Fig. 2.

Kaplan–Meier survival curves for colorectal cancer patients (A) Survival comparison between recurrence/metastasis (R/M) and non-recurrence/metastasis (Non-R/M) subgroups (Log-Rank P < 0.001); B Survival comparison between liver metastasis and non-liver metastasis subgroups (Log-Rank P < 0.001); C Survival comparison between lung metastasis and non-lung metastasis subgroups (Log-Rank P = 0.018); D Survival comparison among TNM stage I, II, and III subgroups (Log-Rank P < 0.001).Note: All survival curves include 95% confidence interval bands; follow-up duration is shown on the x-axis, and overall survival probability is shown on the y-axis

The mean survival time of liver metastases after colorectal cancer surgery was 40.785 months, and the median survival time was 37 months. The mean survival time of non-liver metastases was 85.974 months, and the comparison Log-Rank test between the two groups showed a statistical difference (Fig. 2B), P < 0.001. The results of Cox regression analysis were as follows, HR = 7.719 (95%CI 4.606–12.937, P < 0.001). The 2-year and 3-year cumulative survival rates were 65.6% and 50.8% for patients with liver metastasis, and 93.3% and 90.7% for patients without liver metastasis.

The mean survival time of lung metastasis after colorectal cancer surgery was 45.883 months, and the median survival time was 56 months. The mean survival time of non-pulmonary metastasis was 82.935 months, and P = 0.018 in comparison Log-Rank test between the two groups showed a statistical difference (Fig. 2C). The results of Cox regression analysis were as follows, HR = 2.874 (95%CI 1.146–7.207, P = 0.024). At the same time, we conducted follow-up to calculate the cumulative survival rate. The 2-year and 3-year cumulative survival rates of patients with lung metastasis were 78.1% and 70.3%, respectively, and the 2-year and 3-year cumulative survival rates of patients without lung metastasis were 91.0% and 87.4%, respectively.

We analyzed the TNM stage of colorectal cancer by stratification. The mean survival time of TNM stage I was 92.425 months, TNM stage II was 84.084 months, and TNM stage III was 76.124 months. The 2-year and 3-year cumulative survival rates were 98.4% for TNM I, 91.9% and 88.7% for TNM II, and 86.6% and 79.9% for TNM III, respectively. For the total cohort, the 2-year and 3-year cumulative survival rates by TNM stage were: stage I 98.4% (95%CI 95.8%−100.0%) and 98.4% (95%CI 95.8%−100.0%), stage II 91.9% (95%CI 87.6%−96.2%) and 88.7% (95%CI 83.5%−93.9%), stage III 86.6% (95%CI 82.1%−91.1%) and 79.9% (95%CI 73.2%−86.6%), respectively. Log-Rank test showed significant differences between stages (P < 0.001). At the same time, we compared the tumor TNM stage between groups, TNM I vs TNM II, P = 0.031; TNM I vs TNM III, P = 0.001; TNM II vs TNM III, P = 0.011. This indicated that TNM staging was statistically different among colorectal tumors (Fig. 2D).

In addition, we also conducted a survival analysis of selected clinical factors with statistical significance.

Mesocolic fascia infiltration showed that the mean survival time of positive patients was 76.307 months, while that of negative patients was 87.835 months. The comparison Log-Rank test between the two groups showed a statistical difference (P < 0.001). The results of Cox regression analysis were as follows, HR = 2.887 (95%CI 1.654–5.041, P < 0.001). The 2-year and 3-year cumulative survival rates of patients with lung metastasis were 86.8% and 80.5%, respectively, and the 2-year and 3-year cumulative survival rates of patients without lung metastasis were 93.9% and 92.1%.

The mean survival time of patients with neuroinvasion was 75.735 months, and that of patients without neuroinvasion was 82.965 months. Comparison Log-Rank test between the two groups showed a statistical difference (P = 0.028). The results of Cox regression analysis were as follows, HR = 1.757 (95%CI 1.055–2.925, P < 0.030). The 2-year and 3-year cumulative survival rates were 87.4% and 79.6% for patients with lung metastasis, and 90.5% and 87.2% for patients without lung metastasis.

Tumor sites were divided into right colon, transverse colon, left colon, sigmoid colon and rectum, and stratification analysis was performed. Kaplan–Meier survival analysis results were as follows: The average survival time of patients with right colon cancer was 81.884 months, transverse colon cancer 75.5 months, left colon cancer 82.073 months, sigmoid cancer 78.465 months, and rectal cancer 68.484 months. The 2-year and 3-year cumulative survival rates of right colon cancer were 89.2% and 84.7%. Transverse colon cancer was 92.9%, 78.6%, left half colon cancer 93.8%, sigmoid colon cancer 90, 3%, 86.0%, respectively. Rectal cancer was 91% and 88.9%, respectively. At the same time, we conducted a one-to-group comparison of tumor sites, and no statistical difference was found.

We also conducted stratified analysis of T stages of colorectal cancer, and the results were as follows: the mean survival time of patients with T1 stage was 92.125 months, T2 stage was 78.267 months, T3 stage was 84.00May, and T4 stage was 79.453 months. The 2-year and 3-year cumulative survival rates were 100% in T1, 93.3% in T2 and 90.5% and 88.8% in T3, respectively. In T4, 88.9% and 83.7% respectively. At the same time, no statistical difference was found in the comparison of tumor T stage between groups.

Discussion

In our study, the difference in influencing factors of statistical results is related to the size of the sample number of patients included in the study, as well as other confounding factors such as the lack of patient follow-up, the difference in the operation level of different surgeons, and the difference in clinical treatment plans. We established strict inclusion criteria when screening patients to maximize the impact of confounding bias on statistical results. In our study, risk factors associated with postoperative recurrence or metastasis of colorectal cancer were mesocenteric fascia invasion, number of metastatic lymph nodes, total number of lymph nodes and tumor TNM stage of the tumor. The OR value of mesocenteric fascia infiltration was 2.769. In hepatic metastasis of colorectal cancer, mesocolic fascia infiltration, TNM stage were related. These influencing factors are consistent with our clinical treatment of colorectal cancer.

Mesocolic fascial infiltration, which is diagnosed by preoperative MRI imaging, has been shown in many studies to be one of the risk factors for poor prognosis of rectal cancer. If the mesenteric fascia is infiltrated by tumor cells, what is the prognosis of rectal cancer? Statistical results of a retrospective study on 665 cases of stage II and III colorectal cancer showed that if the mesocolic fascia was involved by tumor, the pelvic recurrence rate of rectal cancer would be significantly increased and the survival time would be shortened, indicating that mesocolic fascia infiltration is one of the important adverse factors for the prognosis of rectal cancer [10], which is consistent with our research conclusion. In our study, mesocenteric fascia infiltration was an independent risk factor for postoperative recurrence or metastasis of colorectal cancer. If MRF is involved, whether it suggests that we should conduct preoperative chemoradiotherapy for this patient to reduce the recurrence or metastasis rate after surgery has not been reached yet. However, for postoperative patients, it can prompt us to develop a more rigorous treatment and follow-up plan.

Postoperative pathological staging of colorectal cancer has long been regarded as the most powerful prognostic indicator of colorectal cancer, and usually determines the selectivity of systemic adjuvant therapy. In pT3 tumors, with or without regional lymph node metastasis, it has been reported that the degree of extramural invasion is an important prognostic feature. In most studies, extramural infiltration of more than 5 mm may have a serious adverse effect on prognosis. The pathological assessment of serous tumor infiltration is relevant only to the upper rectum, which is anterior to the visceral peritoneal covering, which is extremely important. Numerous studies have evaluated serous membrane penetration as a separate pathological variable and demonstrated that it has independent adverse prognostic significance through multivariate analysis. For pT4 tumors that penetrate the visceral peritoneum, the median survival time after surgical resection is significantly shorter than for pT4 tumors that do not involve the serous membrane.

The effectiveness of different surveillance regimens for recurrence or metastasis of colorectal cancer after radical surgery has not been well established. Studies using different surveillance protocols have also found that the proportion of relapses detected at regular follow-up is, on average, earlier than those detected at additional follow-up [11]. Therefore, for those patients with colorectal cancer with high risk factors, it is necessary to establish an individualized follow-up program, which can better carry out early intervention and early treatment, and effectively improve the prognosis of patients.

Limitacions

The limitations of this study are as follows: it is a single-center retrospective study with a small sample size; the follow-up time is limited, and long-term survival outcomes cannot be observed; some baseline data are missing, which may affect the accuracy of the results. Additionally, two critical limitations should be noted: First, this study excluded patients who did not receive guideline-based chemotherapy, which may have introduced selective bias—our cohort may overrepresent patients with better baseline health and treatment compliance, limiting the external validity to 'real-world' colorectal cancer (CRC) populations. Second, we did not include molecular markers known to influence CRC recurrence and survival, such as microsatellite instability-high (MSI-H) and RAS/BRAF mutations. These molecular factors may modify the effect of clinical risk factors (e.g., mesocolic fascia infiltration [MFI] may have a stronger association with recurrence in RAS-mutant tumors), and their exclusion is a limitation of this study. Third, the single-center retrospective design may limit the generalizability of the results, as patient characteristics and treatment practices may vary across institutions. In future studies, we will carry out multi-center prospective studies with larger samples to further verify the conclusions of this study, and include molecular markers (e.g., MSI-H, RAS/BRAF) in the analysis to explore their potential modifying effects.

Conclusion

Multifactor logistic regression analysis showed mesocolon fascia infiltration, total number of lymph nodes, the number of metastatic lymph nodes and tumor TNM stage were an independent risk factor for postoperative recurrence and metastasis of colorectal cancer.

Abbreviations

IARC

International Agency for Research on Cancer

TNM

Tumor, Node, Metastasis staging system

CT

Computed tomography

MRI

Magnetic Resonance Imaging

OR

Odds Ratio

CI

Confidence Interval

P

Probability Value

Non-R/M

Non-recurrence or metastasis

R/M

Recurrence or Metastasis

Authors’ contributions

Haoran Zhu and Yuankun Cai contributed to data collection and statistical analysis. Chongwei Ke assisted with clinical data interpretation. Huipeng Wang designed the study, supervised the research, and drafted the manuscript. All authors reviewed and approved the final version.

Funding

This study was supported by the Shanghai Minhang District High-Level Specialist Key Physician Training Project (No. 2024MZYS17).

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was approved by the Ethics Committee of Shanghai Fifth People's Hospital, Fudan University (approval number: 2021 No. 237). Informed consent to participate was obtained from all participants. The study was conducted in accordance with the Declaration of Helsinki.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.