Predictive value of preoperative routine examination for the prognosis of patients with pT2N0M0 or pT3N0M0 colorectal cancer

Abstract

BACKGROUND

In recent years, the incidence of colorectal cancer (CRC) has been increasing. With the popularization of endoscopic technology, a number of early CRC has been diagnosed. However, despite current treatment methods, some patients with early CRC still experience postoperative recurrence and metastasis.

AIM

To search for indicators associated with early CRC recurrence and metastasis to identify high-risk populations.

METHODS

A total of 513 patients with pT2N0M0 or pT3N0M0 CRC were retrospectively enrolled in this study. Results of blood routine test, liver and kidney function tests and tumor markers were collected before surgery. Patients were followed up through disease-specific database and telephone interviews. Tumor recurrence, metastasis or death were used as the end point of study to find the risk factors and predictive value related to early CRC recurrence and metastasis.

RESULTS

We comprehensively compared the predictive value of preoperative blood routine, blood biochemistry and tumor markers for disease-free survival (DFS) and overall survival (OS) of CRC. Cox multivariate analysis demonstrated that low platelet count was significantly associated with poor DFS [hazard ratio (HR) = 0.995, 95% confidence interval (CI): 0.991-0.999, P = 0.015], while serum carcinoembryonic antigen (CEA) level (HR = 1.008, 95%CI: 1.001-1.016, P = 0.027) and serum total cholesterol level (HR = 1.538, 95%CI: 1.026-2.305, P = 0.037) were independent risk factors for OS. The cutoff value of serum CEA level for predicting OS was 2.74 ng/mL. Although the OS of CRC patients with serum CEA higher than the cutoff value was worse than those with lower CEA level, the difference between the two groups was not statistically significant (P = 0.075).

CONCLUSION

For patients with T2N0M0 or T3N0M0 CRC, preoperative platelet count was a protective factor for DFS, while serum CEA level was an independent risk factor for OS. Given that these measures are easier to detect and more acceptable to patients, they may have broader applications.

Key Words: Colorectal cancer, Platelet count, Serum carcinoembryonic antigen, Total cholesterol level, Overall survival, Disease-free survival

Core Tip: By analyzing the results of preoperative blood tests for pT2N0M0/pT3N0M0, we found that preoperative platelet count was a protective factor for disease-free survival, while serum carcinoembryonic antigen level was an independent risk factor for overall survival. Given that these measures are easier to detect and more widely available, they may have a broader application in the future.

INTRODUCTION

Colorectal cancer (CRC) is one of the most common tumors of the digestive system, and globally it ranks among the top malignant tumors in terms of overall morbidity and mortality rate[1]. Currently, surgical resection, radiotherapy, chemotherapy and immunotherapy have become the main strategies for the treatment of CRC[2]. However, it remains controversial whether adjuvant therapy should be given to pathologically early-stage CRC (especially stage IIA)[3-5]. It is of great clinical significance to identify the patients with high risk of recurrence, not only to avoid excessive chemotherapy in low-risk populations, but also to assign more precise treatment regimens to high-risk patients in order to reduce the recurrence rate. Therefore, the search for an optimal prognostic biomarker associated with early CRC recurrence and metastasis is needed in clinical practice to improve the prognosis of CRC.

Preoperative blood routine examination, including blood routine examination and biochemical examination, is the basic examination to evaluate the general condition of patients before surgery, which can effectively and promptly reflect the inflammation, metabolism, nutritional status and other indicators of patients. In recent years, an increasing number of studies have shown that routine blood and blood biochemistry tests can be used to predict cancer outcomes[6-8]. Herein, we aimed to analyze routine blood examination results in patients with pathologically diagnosed confirmed early CRC (T2N0M0/T3N0N0) with a view to searching for risk factors associated with recurrence and metastasis of early CRC. In the present study, we found platelet count, as a protective factor for CRC, is positively correlated with DFS of CRC.

MATERIALS AND METHODS

Study setting and participants

In this study, we retrospectively collected the medical records of a total of 513 CRC patients with a pathological diagnosis of pT2N0M0 or pT3N0M0 at the First Affiliated Hospital of Naval Medical University from April 2016 to January 2018, including 198 females and 315 males. Inclusion and exclusion criteria were strictly followed. The inclusion criteria were as follows: (1) Hospitalization for radical resection (curative resection) of CRC; (2) Primary colorectal tumor; and (3) Postoperative pathological diagnosis of pT2N0M0 or pT3N0M0. Exclusion criteria: (1) Comorbidity with severe other systemic diseases; (2) Concurrent or heterochronic multiple primary CRC; (3) Comorbidity with other systemic malignancies; (4) Hereditary or familial CRC; (5) Preoperative neoadjuvant therapy; (6) Preoperative receipt of blood transfusion therapy; and (7) Emergency surgery or palliative surgery. The study protocol was approved by the Ethics Committee of the First Affiliated Hospital of Naval Medical University.

Study outcomes

The results of blood routine, blood biochemistry, and tumor markers [carcinoembryonic antigen (CEA) and carbohydrate antigen 19-9] were collected from the date of surgery after admission. The specific items included white blood cell count, lymphocyte count, monocyte count, neutrophil count, hemoglobin count, and platelet count. Total bilirubin, direct bilirubin, total protein, albumin, globulin, alanine aminotransferase, aspartate aminotransferase, total bile acid, prealbumin, blood glucose, creatinine, urea, uric acid, total cholesterol, etc. Blood collection was performed in a fasted state with the patient fasting for at least 8 h overnight. Routine blood test was performed by a Sysmex blood cell analyzer (XN9000). An automated biochemical analyzer (Hitachi 7020) was used to measure blood biochemistry and blood lipid profile. Basic information including gender, age, body mass index (BMI), diagnosis, date of surgery and past medical history were collected through the medical record system.

All patients received standardized management and follow-up after surgery. The specific protocol is as follows: (1) Serum CEA level and physical examination were performed every 3 to 6 months for the first 2 years after surgery, and then repeated every 6 months for a total of 5 years; (2) A computed tomography scan of the chest, abdomen, and pelvis was performed once a year within 3 years after surgery; and (3) Colonoscopy was performed within 1 year after surgery. Postoperative surveillance was usually ceased if there was no relapse 5 years after surgery. In some cases, annual surveillance was continued upon request from the patient. Patients were followed up through the disease-specific database and outpatient follow-up records, and some patients with incomplete information were followed up by telephone. Detailed records were kept on patients’ postoperative survival, especially on events such as tumor recurrence, distant metastasis or death, and the time of their occurrence. The primary endpoint events of the study were postoperative disease-free survival (DFS) and overall survival (OS).

Statistical analysis

The Kolmogorov-Smirnov test was employed to assess the normality of the measurement data distribution initially. Normally distributed data are shown as mean ± SD and non-normally distributed data are expressed as the median (lower quartile, upper quartile). For comparison purposes, an independent samples t-test was used when the data distribution was normal and homogeneity of variance was met, otherwise, the Mann-Whitney U test was employed. Categorical variables (such as gender, combined hypertension, combined diabetes mellitus, smoking history, alcohol intake history and family history of CRC) were described based on their frequency and rate, and they were subjected to analysis using the χ² test. Cox regression analysis was used to evaluate the risk factors for recurrence, metastasis and survival. The predictive ability of independent risk factors was evaluated by plotting receiver operating characteristic (ROC) curves and calculating the area under the curve. The Kaplan-Meier survival curve evaluates the impact of CEA on survival. The optimal critical values were obtained by selecting the largest Youden index. The Youden index, which represents the total ability of the test to detect true patients and nonpatients, is the sum of sensitivity and specificity minus one. After calculating the Youden index, the results were sorted, and the maximum value obtained was the optimal critical value. All statistical analyses were conducted using the SPSS 20.0 statistical software package. Statistical significance was attributed to P values < 0.05.

RESULTS

Characteristics of the study participants

A total of 513 patients with pathological T2N0M0 or T3N0M0 CRC were included in the study and these patients were divided into two groups based on different clinical outcomes: DFS group (n = 451) and recurrence/metastasis or death group (non-DFS, n = 62), the median duration of follow-up was 71.85 (66.83, 78.18) months. As shown in Table 1, whether in the DFS group or the non-DFS group, more than half of the patients were male, and the age of the patients was over 65 years old. The proportion of smoking and alcohol consumption in the DFS group was lower than those in the non-DFS group, but the difference was not statistically significant.

Table 1 Clinical characteristics of the colorectal cancer patients.

Variables	DFS	Non-DFS	P value
n	451	62
Male, (%)	61.20	62.90	0.890
Age, yr	66.65 ± 11.08	68.48 ± 10.87	0.221
Diabetes mellitus (%)	10.86	12.90	0.666
Hypertension (%)	31.49	30.65	1.000
Smoking history1 (%)	20.40	30.65	0.072
Alcohol intake history2 (%)	11.87	20.97	0.068
Family history of CRC (%)	4.88	4.84	1.000
NAFLD (%)	21.51	29.03	0.195
BMI, kg/m2	23.33 ± 3.22	23.59 ± 2.55	0.538

¹Smoking history was defined in our study as regular daily inhalation of cigarettes that lasted at least until surgery without cessation.

²Alcohol consumption was defined as an average daily intake of ethanol greater than or equal to 30 grams per day (about two drinks a day) without abstinence.

CRC: Colorectal cancer; NAFLD: Non-alcoholic fatty liver disease; BMI: Body mass index; DFS: Disease-free survival.

Univariate and multivariate analysis of DFS

Tumor recurrence and metastasis were used as the endpoint events, we performed univariate Cox regression analysis and found that smoking was a risk factor for recurrence or metastasis [hazard ratio (HR) = 1.840, 95% confidence interval (CI): 1.063-3.185, P = 0.029], while platelet count was a protective factor (HR = 0.995, 95%CI: 0.991-0.999, P = 0.015) (Table 2). Multivariate Cox regression analysis showed platelet count to be a protective factor in patients with recurrence or metastasis (BE = -0.005, SE = 0.002, Wald = 5.957, HR = 0.995, 95%CI: 0.991-0.999, P = 0.015).

Table 2 Univariate Cox regression analysis of disease-free survival.

Variables	BE	SE	Wald	P value	HR (95%CI)
Male, (%)	0.179	0.276	0.420	0.517	1.196 (0.696-2.055)
Age, yr	-.005	0.012	0.169	0.681	0.995 (0.972-1.018)
BMI, kg/m2	0.026	0.041	0.413	0.520	1.027 (0.947-1.113)
Diabetes mellitus (%)	-0.284	0.468	0.369	0.544	0.753 (0.301-1.883)
Hypertension (%)	-0.133	0.257	0.269	0.604	0.875 (0.529-1.447)
Smoking history (%)	0.610	0.280	4.751	0.029	1.840 (1.063-3.185)
Alcohol intake history (%)	0.590	0.324	3.315	0.069	1.805 (0.956-3.407)
Family history of CRC (%)	0.435	0.518	0.705	0.401	1.545 (0.560-4.266)
NAFLD (%)	0.393	0.288	1.857	0.173	1.482 (0.842-2.608)
WBC, 109/L	-0.027	0.077	0.122	0.727	0.974 (0.837-1.132)
LY, 109/L	0.035	0.210	0.028	0.868	1.036 (0.686-1.564)
MO, 109/L	-0.164	0.841	0.038	0.846	0.849 (0.163-4.413)
NEUT, 109/L	-0.037	0.090	0.167	0.683	0.964 (0.808-1.150)
Hb, g/L	0.005	0.006	0.727	0.394	1.005 (0.993-1.017)
PLT, 109/L	-0.005	0.002	5.957	0.015	0.995 (0.991-0.999)
T-BIL, μmol/L	0.006	0.025	0.066	0.798	1.007 (0.958-1.058)
D-BIL, μmol/L	-0.010	0.086	0.013	0.908	0.990 (0.836-1.172)
TP, g/L	0.002	0.021	0.012	0.913	1.002 (0.961-1.045)
ALB, g/L	-0.014	0.030	0.216	0.642	0.986 (0.931-1.045)
GLB, g/L	0.018	0.029	0.389	0.533	1.018 (0.962-1.078)
ALT, U/L	-0.002	0.008	0.036	0.849	0.998 (0.983-1.015)
AST, U/L	-0.016	0.019	0.637	0.425	0.985 (0.948-1.023)
TBA, μmol/L	0.000	0.023	0.000	0.990	1.000 (0.956-1.047)
PA, mg/L	0.002	0.002	0.857	0.355	1.002 (0.998-1.006)
UREA, mmol/L	-0.001	0.088	0.000	0.990	0.999 (0.841-1.186)
FBG, mmol/L	0.119	0.072	2.742	0.098	1.126 (0.978-1.297)
Cr, μmol/L	0.002	0.008	0.072	0.789	1.002 (0.986-1.019)
UA, mmol/L	-0.392	1.119	0.123	0.726	0.675 (0.075-6.055)
HbA1c, %	0.017	0.156	0.011	0.915	1.017 (0.749-1.380)
CEA, ng/mL	0.004	0.004	1.067	0.302	1.004 (0.996-1.013)
CA19-9, U/mL	0.000	0.001	0.071	0.789	1.000 (0.998-1.001)
TC, mmol/L	0.144	0.179	0.648	0.421	1.155 (0.813-1.639)
TG, mmol/L	0.216	0.232	0.868	0.351	1.241 (0.788-1.955)
HDL-C, mmol/L	-0.054	0.501	0.012	0.914	0.948 (0.355-2.529)
LDL-C, mmol/L	0.142	0.223	0.405	0.525	1.152 (0.744-1.784)
Apo-A1, g/L	-0.788	0.803	0.964	0.326	0.455 (0.094-2.194)
Apo-B, g/L	0.707	1.075	0.432	0.511	2.027 (0.246-16.676)
LP-a, mg/dL	-0.001	0.012	0.005	0.944	0.999 (0.977-1.022)

CRC: Colorectal cancer; NAFLD: Nonalcoholic fatty liver disease; BMI: Body mass index; WBC: White blood cell; LY: Lymphocyte; MO: Monocyte; NEUT: Neutrophil; Hb: Hemoglobin; PLT: Blood platelet count; T-BIL: Total bilirubin; D-BIL: Direct bilirubin; TP: Total protein; ALB: Albumin; GLB: Globulin; ALT: Alanine aminotransferase; AST: Aspartate aminotransferase; TBA: Total bile acid; PA: Prealbumin; UREA: Urea nitrogen; FBG: Fasting blood glucose; Cr: Creatinine; UA: Uric acid; HbA1c: Glycosylated hemoglobin A1c; CEA: Carcinoembryonic antigen; CA19-9: Carbohydrate antigen 19-9; TC: Total cholesterol; TG: Triglyceride; HDL-C: High-density lipoprotein cholesterol; LDL-C: Low density lipoprotein cholesterol; Apo-A1: Apolipoprotein A1; Apo-B: Apolipoprotein B; LP-a: Lipoprotein A; HR: Hazard ratio; CI: Confidence interval.

Univariate and multivariate analysis of OS

A univariate analysis was conducted based on whether the patient died (Table 3). The analysis revealed that age (HR = 1.031, 95%CI: 1.001-1.062, P = 0.040), hemoglobin A1c (HbA1C) (HR = 1.229, 95%CI: 1.038-1.455, P = 0.016), CEA (HR = 1.008, 95%CI: 1.001-1.016, P = 0.027), and total cholesterol (TC) (HR = 1.538, 95%CI: 1.026-2.305, P = 0.037) were risk factors for death in patients. Serum CEA level (HR = 1.008, 95%CI: 1.001-1.016, P = 0.027) and TC level (HR = 1.538, 95%CI: 1.026-2.305, P = 0.037) were identified as independent risk factors for patient mortality in further multivariate analysis (Table 4).

Table 3 Univariate Cox regression analysis of overall survival.

Variables	BE	SE	Wald	P value	HR (95%CI)
Male, (%)	-0.329	0.313	1.102	0.294	0.720 (0.390-1.330)
Age, yr	0.031	0.015	4.211	0.040	1.031 (1.001-1.062)
BMI, kg/m2	0.020	0.049	0.157	0.692	1.020 (0.926-1.123)
Diabetes mellitus (%)	0.098	0.477	0.042	0.838	1.102 (0.433-2.809)
Hypertension (%)	0.014	0.222	0.004	0.949	1.014 (0.656-1.568)
Smoking history (%)	0.418	0.343	1.482	0.223	1.519 (0.775-2.976)
Alcohol intake history (%)	0.492	0.394	1.558	0.212	1.635 (0.755-3.541)
Family history of CRC (%)	0.476	0.600	0.630	0.427	1.610 (0.497-5.215)
NAFLD (%)	0.614	0.329	3.471	0.062	1.847 (0.969-3.523)
WBC, 109/L	-0.054	0.094	0.327	0.568	0.948 (0.789-1.139)
LY, 109/L	0.074	0.249	0.089	0.765	1.077 (0.662-1.754)
MO, 109/L	-1.432	1.110	1.663	0.197	0.239 (0.027-2.105)
NEUT, 109/L	-0.074	0.112	0.434	0.510	0.929 (0.746-1.156)
Hb, g/L	-0.003	0.007	0.149	0.700	0.997 (0.984-1.011)
PLT, 109/L	-0.002	0.002	1.072	0.301	0.998 (0.993-1.002)
T-BIL, μmol/L	-0.070	0.038	3.484	0.062	0.932 (0.866-1.004)
D-BIL, μmol/L	-0.243	0.128	3.611	0.057	0.784 (0.610-1.008)
TP, g/L	-0.005	0.025	0.040	0.841	0.995 (0.946-1.046)
ALB, g/L	-0.056	0.034	2.643	0.104	0.946 (0.885-1.011)
GLB, g/L	0.045	0.033	1.788	0.181	1.046 (0.979-1.116)
ALT, U/L	-0.001	0.009	0.009	0.925	0.999 (0.981-1.017)
AST, U/L	-0.004	0.016	0.052	0.819	0.996 (0.966-1.028)
TBA, μmol/L	-0.009	0.031	0.089	0.765	0.991 (0.932-1.053)
PA, mg/L	0.000	0.003	0.001	0.978	1.000 (0.995-1.005)
UREA, mmol/L	0.022	0.104	0.045	0.832	1.022 (0.834-1.252)
FBG, mmol/L	0.122	0.082	2.242	0.134	1.130 (0.963-1.326)
Cr, μmol/L	-0.011	0.010	1.260	0.262	0.989 (0.969-1.009)
UA, mmol/L	-2.209	1.874	1.390	0.238	0.110 (0.003-4.321)
HbA1c, %	0.206	0.086	5.754	0.016	1.229 (1.038-1.455)
CEA, ng/mL	0.008	0.004	4.896	0.027	1.008 (1.001-1.016)
CA19-9, U/mL	-0.005	0.007	0.476	0.490	0.995 (0.981-1.009)
TC, mmol/L	0.431	0.206	4.350	0.037	1.538 (1.026-2.305)
TG, mmol/L	0.211	0.272	0.601	0.438	1.235 (0.724-2.107)
HDL-C, mmol/L	0.264	0.552	0.229	0.632	1.303 (0.441-3.846)
LDL-C, mmol/L	0.434	0.258	2.820	0.093	1.543 (0.930-2.559)
Apo-A1, g/L	0.848	0.832	1.038	0.308	2.335 (0.457-11.931)
Apo-B, g/L	1.200	1.216	0.974	0.324	3.320 (0.306-35.966)
LP-a, mg/dL	0.006	0.011	0.253	0.615	1.006 (0.983-1.029)

CRC: Colorectal cancer; NAFLD: Nonalcoholic fatty liver disease; BMI: Body mass index; WBC: White blood cell; LY: Lymphocyte; MO: Monocyte; NEUT: Neutrophil; Hb: Hemoglobin; PLT: Blood platelet count; T-BIL: Total bilirubin; D-BIL: Direct bilirubin; TP: Total protein; ALB: Albumin; GLB: Globulin; ALT: Alanine aminotransferase; AST: Aspartate aminotransferase; TBA: Total bile acid; PA: Prealbumin; UREA: Urea nitrogen; FBG: Fasting blood glucose; Cr: Creatinine; UA: Uric acid; HbA1c: Glycosylated hemoglobin A1c; CEA: Carcinoembryonic antigen; CA19-9: Carbohydrate antigen 19-9; TC: Total cholesterol; TG: Triglyceride; HDL-C: High-density lipoprotein cholesterol; LDL-C: Low density lipoprotein cholesterol; Apo-A1: Apolipoprotein A1; Apo-B: Apolipoprotein B; LP-a: Lipoprotein A; HR: Hazard ratio; CI: Confidence interval.

Table 4 Multivariate Cox regression analysis of overall survival.

Variables	BE	SE	Wald	P value	HR (95%CI)
CEA, ng/mL	0.019	0.005	16.075	< 0.001	1.019 (1.010-1.029)
TC, mmol/L	0.006	0.011	0.253	0.615	1.006 (0.983-1.029)

CEA: Carcinoembryonic antigen; TC: Total cholesterol; TG: Triglyceride; HR: Hazard ratio; CI: Confidence interval.

Results of the ROC curve analysis of CEA for CRC

The ROC curve analysis was performed to evaluate the predictive ability of serum CEA level and TC level for CRC. When the optimal critical value of CEA was set at 2.740 ng/mL, the sensitivity and specificity were 0.750 and 0.415, respectively. When the optimal critical value of TC was set at 4.495 mmol/L, the sensitivity and specificity were 0.875 and 0.407, respectively (Figure 1A).

Figure 1 The receiver operating characteristic curve and Kaplan-Meier survival analysis of carcinoembryonic antigen for colorectal cancer patients. A: The receiver operating characteristic curve analysis; B: The Kaplan-Meier survival analysis. ROC: Receiver operating characteristic curve; CEA: Carcinoembryonic antigen; TC: Total cholesterol.

Results of the Kaplan-Meier survival analysis of CEA for CRC

Survival curves of patients were plotted by dividing them into two categories based on the cutoff value of CEA (Figure 1B), which was defined as “1” for CEA ≥ 2.74 ng/mL and “0” for CEA < 2.74 ng/mL. The result analysis revealed that the survival rate of patients with CEA values higher than the cutoff was lower than that of patients with low CEA values, but the difference was not statistically significant (P = 0.075).

DISCUSSION

In the multivariate analysis of this study, our results identified platelet count as a protective factor against tumor recurrence and metastasis in patients with early-stage CRC. As an important component of the blood, platelets play an important role in the normal process of thrombosis and hemostasis[9,10]. Over the past few years, an increasing number of studies have confirmed that blood platelets also play an important role in the growth and migration of tumor cells[11-14]. Josa et al[15] analyzed the postoperative platelet count levels of 357 patients with primary CRC and found that the postoperative increase in platelet count was associated with poor prognosis of the tumor. Meanwhile, many studies have found that platelet-related indicators, such as the platelet-monocyte ratio and the platelet-albumin ratio, are also closely associated with poor clinical outcomes[16,17]. Wan et al[18] retrospectively analyzed 1513 patients undergoing CRC surgery and found that patients with platelet count greater than 400 × 10⁹/L within 1 month before surgery had significantly higher postoperative mortality and distant metastasis rates than those with normal platelet count. However, in univariate and multivariate analysis, we found that platelet count was a protective factor for recurrence and metastasis in patients with early CRC (pT2N0M0/pT3N0N0) (P = 0.015), which means that patients with higher platelet count had a lower risk of postoperative local recurrence and metastasis. This was different from the findings of Wan et al[18]. We considered the possible reason for the low level of platelet count in the patients included in this study, with only 16.4% (84 patients) having platelet count higher than 300 × 10⁹/L and the majority of patients having platelet count below normal. In addition, we divided the patients into a high platelet count group (platelet count > 300 × 10⁹/L) and a low platelet count group (platelet count ≤ 300 × 10⁹/L) and found no difference in the risk of postoperative recurrence and metastasis between the two groups. Therefore, we hypothesized that an elevated platelet count within the normal range is a protective factor against recurrence and metastasis in pT2N0M0/pT3N0N0 CRC. However, a high-level of clinical evidence is needed to support this speculation. In spite of this, as an important part of blood routine examination, platelet count is easier to obtain and promote. Further researches, which confirm the correlation between platelets and colorectal cancer prognosis, will have a profound impact on primary prevention and postoperative surveillance of colorectal cancer. Next, we plan to expand the sample size to further clarify the real significance of platelets in predicting the prognosis of colorectal cancer.

Survival-related univariate analysis revealed that glycosylated HbA1C, serum TC and CEA were risk factors associated with postoperative survival. HbA1C is the combination of blood glucose and hemoglobin in human blood, which can be an indication of changes in blood glucose levels over the last 2-3 months[19]. Epidemiological data show that diabetes mellitus, especially type 2 diabetes mellitus, is associated with the development and poor prognosis of various types of tumors[20-22]. HbA1C was associated with the incidence of lung, colorectal and breast cancer in a prospective cohort study by Srour et al[23]. The results of a Mendelian randomization study by Murphy et al[24] showed that high levels of glycated hemoglobin increased the risk of CRC. The hyperglycemia caused by diabetes increases the levels of advanced glycation end products (AGEs), which have been shown to promote the growth of colon cancer in vitro[25,26]. The main mechanism of this process is the binding of AGEs to its receptor RAGE, which leads to oxidative stress in the organism and also promotes the inflammatory process[27,28]. Free radicals produced by oxidative stress can directly or indirectly damage DNA, protein and other cellular components, and then induce gene mutations, leading to the occurrence of cancer[29]. At the same time, activation of various inflammatory factors such as interleukin (IL)-6 and IL-18 can increase the survival rate of tumor cells and accelerate the process of cancer[30]. Similar to diabetes, hyperlipidemia has been shown in numerous studies to be involved in the development of cancer, including CRC, prostate cancer and breast cancer[31,32]. In a cohort study, hypercholesterolemia was associated with a significant increase in the risk of CRC[33]. However, we found that HbA1C and hypercholesterolemia were not independent risk factors for survival-related CRC when we performed multivariate analysis. We suspect that the possible reason is that these metabolic markers are closely related, and patients with multiple metabolic abnormalities are very common[34]. Chen et al[35] investigated the relationship between metabolic syndrome and metabolic comorbid conditions and early-onset CRC and found that patients with one, two and three metabolic comorbidities had an increased risk of early-onset CRC by 9%, 12% and 31%, respectively, compared with patients without metabolic comorbidities. Therefore, we suspect that HbA1C and hypercholesterolemia are not independent, but interact and act synergistically in CRC progression. In addition, multivariate analysis of survival-related indicators showed that serum CEA, also known as C stage, was an independent risk factor for early CRC (T2N0M0/T3N0M0), consistent with some previous findings. A retrospective cohort study involving 1367 patients with CRC found that age and preoperative serum albumin and CEA level significantly affected the prognosis of patients with CRC undergoing surgery, and preoperative CEA level was an independent risk factor for OS in patients with stage II and III CRC[36]. Furthermore, Thirunavukarasu et al[37] followed 9083 patients with American Joint Committee on Cancer (AJCC) stage I-IV colon cancer for 27 months (median follow-up time). Multivariate analysis showed that compared with patients with normal CEA (C0 stage) at the same tumor-node-metastasis (TNM) stage, elevated CEA (C1 stage) before treatment can increase the overall mortality by 60% and significantly reduce the OS rate after surgery. Elevated CEA before treatment is an independent risk factor for the prognosis of patients with colon cancer. Interestingly, the study found that under the same TNM staging conditions, patients with C1 stage had a significantly worse prognosis than patients with C0 stage, which is similar to the lymph node staging in TNM staging. In other words, the prognosis of early-stage colon cancer patients with negative preoperative C1 stage lymph nodes may be similar to that of patients with positive lymph nodes. The inclusion of lymph node negative patients in our study, which excludes the effect of N-staging, also confirms the above findings. However, in the current treatment mode, such patients with node-negative C1 stage early CRC may not receive postoperative adjuvant chemotherapy[38], which may be the cause of postoperative tumor metastasis or recurrence in some patients. Based on the results of these studies, the AJCC recommended that CEA level be included in the TNM staging system of colon cancer for risk stratification patients[39]. Taking this into account, we calculated the cutoff value of CEA as 2.74 ng/mL according to the prognosis of the patients, and divided the patients into high and low groups based on this value. The subsequent analysis showed that the survival rate of patients with CEA > 2.74 ng/mL was lower than that of patients with low CEA level, but the difference in survival between the two groups was not statistically significant (P = 0.075). Limited by the small sample size and short follow-up period, the effect of CEA on patient survival needs to be further investigated.

Our study has some limitations. First of all, in the studies on platelets, the proportion of patients with platelet count higher than the normal value is small, which may be the reason why the results of this part of the study are different from those of some studies, so we need to expand the sample size for further analysis. At the same time, we only included patients with pathologically negative lymph nodes, so it was not possible to cross-analyze CEA level and N stage.

CONCLUSION

In conclusion, our study found that preoperative platelet count and serum CEA level were independent protective factor for DFS and risk factor for OS in patients with T2N0M0 or T3N0M0 CRC, respectively. Due to the simplicity of blood tests, the measurement of these indicators may also be more acceptable to patients in clinical application than traditional histology-based biomarkers and may have a broader application prospect. However, these findings still require further confirmation from studies with a higher level of evidence.