This study aimed to establish whether computed tomography (CT)–determined sarcopenia is a useful imaging biomarker for postoperative outcome in elderly colorectal cancer (CRC) patients, and construct sarcopenia-based nomograms to predict individual outcomes after surgery.
CT imaging data of 298 elderly CRC patients who underwent surgery in 2012-2014 were retrospectively analyzed. Skeletal muscle mass was determined by CT, and sarcopenia was diagnosed based on the optimal cutoff value determined by X-tile program. The correlation between sarcopenia and risk of preoperative nutrition and postoperative complications was evaluated. A Cox proportional hazards model was used to determine survival predictors. Sarcopenia-based nomograms were developed based on multivariate analysis, and calibrated using concordance index and calibration curves.
A total 132 patients (44.3%) had sarcopenia based on the optimum cutoff values (29.9 cm2/m2 for women and 49.5 cm2/m2 for men). Sarcopenia was an independent risk factor for preoperative nutrition (p < 0.001; odds ratio [OR], 3.405; 95% confidence interval [CI], 1.948 to 5.954) and postoperative complications (p=0.008; OR, 2.192; 95% CI, 1.231 to 3.903). Sarcopenia was an independent predictor for poor progression-free survival (p < 0.001; hazard ratio [HR], 2.175; 95% CI, 1.489 to 3.179) and overall survival (p < 0.001; HR, 2.524; 95% CI, 1.721 to 3.703). Based on multivariate analysis, we produced four nomograms that had better predictive performance.
CT-determined sarcopenia is a useful imaging biomarker for predicting preoperative nutritional risk, postoperative complications, and long-term outcomes in elderly CRC patients. The sarcopenia-based nomograms can provide a scientific basis for guiding therapeutic schedule and follow-up strategies.
Colorectal cancer (CRC) is one of the malignancies with the highest morbidity and mortality worldwide. According to Global Cancer Epidemiological Statistics (GLOBOCAN) 2018, there are about 1.8 million new cases and 881,000 cancer-related deaths worldwide, with the third highest incidence and second highest mortality rate worldwide, which accounts for about one tenth of all cancer cases and deaths [
Recently, there has been increasing evidence that the gradual decline in nutrition and function is associated with disease progression and is one of the leading causes of poor treatment outcomes. Sarcopenia, as an objective nutrient assessment index independent of body mass index (BMI), is of increasing concern. Sarcopenia is a progressive and pervasive disease that is characterized by reduced skeletal muscle mass and function throughout the body, reduced physical ability, decreased quality of life, and increased risk of adverse events such as death [
Therefore, this study retrospectively analyzed the medical history of elderly CRC patients to explore the relationship of CT-determined sarcopenia with muscle loss and preoperative nutritional risk, postoperative complications and long-term prognosis in elderly CRC patients.
This retrospective study included data from the hospital records of 670 consecutive patients who underwent surgery for CRC at the First Affiliated Hospital of Guangxi Medical University, Nanning, China, between January 2012 and December 2014. We excluded eight patients were lost during follow-up, 342 patients aged < 60 years when they underwent CRC resection, and 22 patients were not available for measurement of the area of skeletal muscle in the third lumbar spine by preoperative abdominal CT. In total, 298 elderly CRC patients were eligible for the study (
Postoperative follow-up was performed every 3 months within 2 years after surgery and then every 3-6 months thereafter. The final follow-up visit occurred on September 1, 2019. The follow-up mainly included contacting the patient by telephone and investigating the patient in the clinic. The follow-up plan included physical examination, laboratory examination (tumor markers, blood routine, liver function test, kidney function test, etc.), imaging examination (flat, CT and magnetic resonance imaging) and endoscopy. Progression-free survival (PFS) was determined as the duration between surgery and recurrence, death, or last follow-up. OS was determined as the duration between surgery and death or last follow-up.
The data collected included the following: general patient information, including sex, age, BMI (low, < 18.5; normal, ≥ 18.5 and < 24; and high, 24), Nutrition Risk Screening 2002 (NRS-2002), American Society of Anesthesiologists (ASA) grade, Surgical method, and postoperative chemoradiotherapy; preoperative blood tests, including serum carcinoembryonic antigen (CEA) (normal, < 5 ng/mL), lymphocyte counts and serum albumin concentration (< 35 g/L was defined as hypoproteinemia); clinicopathological factors, including tumor location, pathological tumor stage (pT category), pathological node stage (pN category), preoperative metastasis, pathological tumor-node-metastasis stage (pTNM stage), tumor perineural/vascular invasion, tumor differentiation, tumor pathological type, and tumor size; postoperative outcomes, including postoperative complications, Clavien-Dindo classification [
Abdominal CT was taken within 1 month before surgery. We selected a cross-sectional CT image of the third lumbar vertebra (L3) to estimate muscle mass and adjusted the CT threshold range from –29 to 150 HU to separate skeletal muscle from other tissues. Psoas major, erector spine, quadratus psoas, transverse abdominis, internal and external oblique muscles, and rectus abdominis in the L3 region were measured. Then, referring to the BMI, the L3 muscle cross-sectional area calculated for each image was normalized by the square of the patient’s height (m2), and the skeletal muscle index (SMI, cm2/m2) was obtained [
The X-tile program [
This research was approved by the Ethics Committee of The First Affiliated Hospital of Guangxi Medical University of China and strictly complied with the provisions of the Helsinki Declaration (approval number: 2019 (KY-E-022)). All patients in this research have signed informed consent.
Sex-specific cutoff values for SMI associated with OS were 29.9 cm2/m2 for women and 49.5 cm2/m2 for men. Using these cutoff values, 44.3% of patients were found to have sarcopenia. The clinicopathological features of the 298 elderly CRC patients are shown in
The main features of elderly CRC patients with sarcopenia were male sex (p=0.002), older age (p < 0.001), lower serum albumin (p=0.002), lower BMI (p < 0.001), and higher malnutrition risk (p < 0.001). Other factors such as ASA grade, pT category, pN category, preoperative metastasis, tumor location, tumor size, perineural/vascular invasion, tumor differentiation, pathological type, and CEA were not related to the presence of sarcopenia (
Currently, the most commonly used preoperative nutritional assessment method is NRS2002. All patients were assessed for NRS2002 before surgery, and 113 were diagnosed with malnutrition. In univariate logistic regression analysis, serum albumin (p < 0.001), BMI (p < 0.001), sarcopenia (p < 0.001), and tumor size (p=0.038) were considered to be correlated with preoperative malnutrition. In multivariate logistic regression analysis, only serum albumin (p=0.001; odds ratio [OR], 0.418; 95% confidence interval [CI], 0.244 to 0.716), BMI (p=0.037) and sarcopenia (p < 0.001; OR, 3.405; 95% CI, 1.948 to 5.954) were independent risk factors for predicting preoperative nutritional risk in elderly CRC patients (
Sixty-six patients (22.1%) suffered from postoperative complications, including anastomotic leak (4 cases), postoperative bowel obstruction (12 cases), wound problems (25 cases), pulmonary complications (12 cases), and other complications (13 cases). There were 23 patients (23.0%) with grade I complications, 34 (9.6%) with grade II complications, three (1.3%) with grade III complications, and six (2.2%) with grade IV complications by Clavien-Dindo classification. Patients with sarcopenia had a higher incidence of total complications (13.1% vs. 9.1%, p=0.001) (
In univariate Cox proportional hazards analysis, patients with sarcopenia had significantly lower PFS and OS than those with non-sarcopenia, respectively (p=0.008; hazard ratio [HR], 1.607; 95% CI, 1.133 to 2.279 and p=0.001; HR, 1.785; 95% CI, 1.252 to 2.545). After adjustment for clinicopathological factors, multivariate Cox proportional hazards analysis showed that sarcopenia was an independent risk factor for predicting long-term survival outcomes (PFS: p < 0.001; HR, 2.175; 95% CI, 1.489 to 3.179 and OS: p < 0.001; HR, 2.524; 95% CI, 1.721 to 3.703) (
In Kaplan-Meier analysis, patients in the sarcopenia group had poorer clinical outcomes in terms of PFS (48.5% vs. 64.5%, log-rank p=0.007) and OS (47.7% vs. 66.9%, log-rank p=0.001) than those in the non-sarcopenia group had (
To correct confounding factors, we performed a subgroup multivariate analysis of each factor. 28 of the 35 subgroups indicated that sarcopenia was an independent risk factor for prognosis in PFS (
Factors with p < 0.05 in univariate analysis were included in multivariate analysis. Based on multivariate analysis, we built four nomograms (A, nutritional risk model; B, complication risk model; C, PFS model; D, overall survival model) (
The latest research suggests that sarcopenia in cancer patients is the result of local muscle inflammation. During tumor progression, tumor cells or surrounding cells are stimulated to produce tumor necrosis factor and interleukin-6 inflammatory cytokines [
Gastrointestinal malignancies are often associated with the occurrence of dyscrasia. Studies have shown that about 80% of gastrointestinal cancer patients experience varying degrees of weight loss and muscle loss [
In the present study, about 22.1% of the elderly patients experienced varying degrees of postoperative complications. We determined that sarcopenia is an independent risk factor for postoperative complications, which is consistent with previous studies [
Sarcopenia has been reported to have good prognostic value in many tumors, such as gastric cancer, hepatocellular carcinoma, and pancreatic cancer. However, different studies on the relationship between sarcopenia and long-term prognosis of CRC have yielded inconsistent results [
In survival analysis, we confirmed that sarcopenia is an independent risk factor for long-term outcomes, whether in PFS or OS. Besides, it can be seen from the survival nomograms that sarcopenia has a good impact efficiency, but comparing these risk-related factors, the effects of pN category and metastasis were greater than sarcopenia. It is well known that advanced tumor stages are associated with poor long-term prognosis of malignant tumors, but the prognosis of patients with the same stage is often different. We performed a stratified analysis of patients based on TNM staging and compared survival rates in patients with sarcopenia and non-sarcopenia. The results showed that OS and PFS of the sarcopenia group were significantly lower than those of the non-sarcopenia group in TNM stage I and II, but not in stage IIII and IV. Although in TNM III stage, Patients with sarcopenia still tend to have a poor prognosis, but it is not the main factor affecting prognosis. This may be because, in the early stage of the tumor, sarcopenia has a large impact on the prognosis of the patient. As the tumor progresses, the tumor invasion and metastasis gradually occupy the main position.
Nomograms are considered to be a direct tool for individualized risk assessment for each patient. We constructed four novel and effective nomograms based on sarcopenia for individualized assessment of preoperative nutritional risk, postoperative complications, PFS, and OS in elderly CRC patients. It is worth noting that the results of the C-index and the calibration chart confirm that the nomograms have good prediction accuracy. These four sarcopenia-based nomograms, combining clinically easy-to-use clinicopathological factors, are viable and reliable risk prediction tools for individualized prediction of elderly CRC patients, which may contribute to individualized postoperative follow-up and treatment options. Using sarcopenia-based nomograms to help screen high-risk patients can facilitate early treatment interventions, which is beneficial to improve the prognosis of elderly CRC patients. Resistance training [
CT is considered clinically as an accurate method for assessing skeletal muscle mass. It provides important quantitative information about muscle composition and distribution through high-quality images, spatial accuracy, and location features, as well as the ability to measure muscle mass from a cross-section of an abdomen. Before CRC surgery, an abdominal CT scan or enhanced examination is usually performed to evaluate tumor staging. If the L3 skeletal muscle mass is measured at the same time, it will not increase any cost, which can reduce the burden on the patient. Other detection methods for skeletal muscle reduction include dual-energy X-ray absorption, nuclear magnetic resonance imaging, and bioelectrical impedance. Due to the complexity of the other methods, the testing process requires additional patient inspection costs. The prospects for clinical application are far from good than CT examinations. However, the measurement of skeletal muscle mass is still not a routine measurement item in CT examination. More studies are still needed to elucidate the importance of preoperative skeletal muscle mass measurement.
There were some limitations to our study. First, as a single-center and retrospective study, there were problems such as insufficient sample size and selection bias. Second, the definition of sarcopenia was based on muscle mass on CT, and we did not investigate muscle strength or function. Based on the overall survival and overall survival time, this study used the X-tile program to determine the optimal cutoff value, which may have certain subjectivity. However, there are no clear diagnostic criteria for CT-determined sarcopenia in elderly CRC patients. In our opinion, this method can provide some reference value for subsequent research. Finally, sarcopenia-based nomograms are designed based on a limited population. Although the calibration curve suggested that they have good effectiveness, it is not verified by other independent teams. Therefore, the validity and practicability of the nomograms should be verified in the multicenter and large-sample population in the future.
This study is believed to be the first to explore the relationship between sarcopenia and preoperative nutrition, postoperative complications, and long-term outcomes of elderly CRC patients. The results suggest that CT-determined sarcopenia is a useful imaging biomarker for predicting preoperative nutritional risk, postoperative complications, and long-term outcomes in elderly CRC patients. The sarcopenia-based nomograms can provide a scientific basis for guiding therapeutic schedule and follow-up strategies.
Conflicts of interest relevant to this article was not reported.
This research was supported by the 2019 Innovation Project of Guangxi Graduate Education (JGY2019052).
The process of patients’ inclusion and exclusion in elderly colorectal cancer patients.
Kaplan-Meier survival curves of sarcopenia and non-sarcopenia groups of all elderly colorectal cancer patients. (A) Kaplan-Meier progression-free survival (PFS) curves of all patients. (B) Kaplan-Meier overall survival (OS) curves of all patients.
Kaplan-Meier survival curves of sarcopenia and non-sarcopenia groups of elderly colorectal cancer patients based on each TNM stage. (A) Kaplan-Meier progression-free survival (PFS) curves of TNM I stage patients. (B) Kaplan-Meier PFS curves of TNM II stage patients. (C) Kaplan-Meier PFS curves of TNM III stage patients. (D) Kaplan-Meier overall survival (OS) curves of TNM I stage patients. (E) Kaplan-Meier OS curves of TNM II stage patients. (F) Kaplan-Meier OS curves of TNM III stage patients.
Subgroup multivariate survival analysis of sarcopenia in elderly colorectal cancer patients. (A) Subgroup multivariate progression-free survival (PFS) analysis of sarcopenia. (B) Subgroup multivariate overall survival (OS) analysis of sarcopenia. HR, hazard ratio; CI, confidence interval; ALB, albumin; BMI, body mass index; ASA, American Society of Anesthesiologists; CEA, carcinoembryonic antigen.
Construction of sarcopenia-based nomograms in elderly colorectal cancer patients. (A) Sarcopenia-based nomograms of nutritional risk. (B) Sarcopenia-based nomograms of complication risk. (C) Sarcopenia-based nomograms of progression-free survival (PFS) (D) Sarcopenia-based nomograms of overall survival (OS). BMI, body mass index.
The calibration curves for predicting nutritional risk (A), complication risk (B), progression-free survival (PFS) (C), and overall survival (OS) (D) in elderly colorectal cancer patients. The X axis presents the predicted probability and the Y axis shows the actual probability. The calibration lines fit along with the 45 reference.
The relationships between the sarcopenic and clinicopathological factors of elderly colorectal cancer patients
Feature | Case (n=298) | Sarcopenic (n=132) | Nonsarcopenic (n=166) | p-value | |
---|---|---|---|---|---|
Sex (male/female) | 197 (66.1)/101 (33.9) | 119 (90.2)/13 (9.8) | 78 (47.0)/88 (53.0) | 61.143 | < 0.001 |
Age (yr) | 68.46±6.67 | 70.14±7.434 | 67.13±5.67 | 3.846 | < 0.001 |
ALB (g/L) | 35.81±4.12 | 34.81±4.26 | 36.60±3.84 | –3.819 | < 0.001 |
BMI (kg/m2) | 22.10±6.67 | 20.43±3.23 | 23.42±3.01 | –8.220 | <0.001 |
NRS2002 (non-malnutrition/malnutrition) | 185 (62.1)/113 (37.9) | 57 (43.2)/75 (56.8) | 128 (77.1)/38 (22.9) | 35.953 | < 0.001 |
ASA grade (I-II/III-IV) | 141 (47.3)/157 (52.7) | 62 (47.0)/70 (53.0) | 79 (47.6)/87 (52.4) | 0.011 | 0.915 |
pT category (T1-2/T3-4) | 88 (29.5)/210 (70.5) | 42 (31.8)/90 (68.2) | 46 (27.7)/120 (72.3) | 0.596 | 0.440 |
pN category (N0/N1/N2) | 178 (59.3)/79 (26.5)/41 (14.2) | 84 (63.6)/33 (25.0)/15 (11.4) | 94 (56.6)/46 (27.7)/26 (15.7) | 1.796 | 0.407 |
Preoperative metastasis (no/yes) | 273 (91.6)/25 (8.4) | 122 (92.4)/10 (7.6) | 151 (91.0)/15 (9.0) | 0.204 | 0.651 |
Tumor location (colon/rectal) | 146 (49.0)/152 (51.0) | 69 (52.3)/63 (47.7) | 77 (46.4)/89 (53.6) | 1.020 | 0.313 |
Tumor size (cm) | 4.91±2.04 | 5.00±2.19 | 4.84±1.92 | 0.683 | 0.495 |
Perineural/vascular invasion (negative/positive) | 245 (82.2)/53 (17.8) | 112 (84.8)/20 (15.2) | 133 (80.1)/33 (19.9) | 1.124 | 0.289 |
Pathological type (protrude/infiltrating/ulcerative) | 67 (22.5)/36 (12.1)/195 (65.4) | 36 (27.3)/13 (9.8)/83 (62.9) | 31 (18.7)/23 (13.9)/112 (67.4) | 3.632 | 0.163 |
Differentiation (poor/medium and high) | 21 (7.0)/277 (93.0) | 6 (4.5)/126 (95.5) | 15 (9.0)/151 (91.0) | 2.264 | 0.132 |
CEA (normal/high) | 165 (55.4)/133 (44.6) | 75 (56.8)/57 (43.2) | 90 (54.2)/76 (45.8) | 0.201 | 0.654 |
Hospital stays (day) | 14.69±7.45 | 15.25±7.74 | 14.25±7.20 | 1.148 | 0.252 |
Total complications | 66 (22.1) | 39 (13.1) | 27 (9.1) | 11.400 | 0.001 |
Clavien-Dindo classification grade I | 23 (7.7) | 13 (4.4) | 10 (3.4) | 1.510 | 0.219 |
Clavien-Dindo classification grade II | 34 (11.4) | 20 (6.7) | 14 (4.7) | 3.283 | 0.070 |
Clavien-Dindo classification grade III | 3 (1.0) | 2 (0.7) | 1 (0.3) | 0.615 | 0.433 |
Clavien-Dindo classification grade IV | 6 (2.0) | 4 (1.3) | 2 (0.7) | 1.242 | 0.265 |
Values are presented as number (%) or mean±standard deviation. ALB, albumin; BMI, body mass index; ASA, American Society of Anesthesiologists; CEA, carcinoembryonic antigen.
Univariate and multivariate logistic regression analysis of factors associated with malnutrition in elderly CRC
Feature | Univariate analysis |
Multivariate analysis |
||
---|---|---|---|---|
OR (95% CI) | p-value | OR (95% CI) | p-value | |
0.920 (0.560-1.512) | 0.743 | - | - | |
1.537 (0.951-2.483) | 0.079 | - | - | |
0.319 (0.195-0.520) | < 0.001 | 0.418 (0.244-0.716) | 0.001 | |
< 0.001 | 0.037 | |||
Low | 1.000 | 1.000 | ||
Normal | 0.233 (0.115-0.472) | < 0.001 | 0.381 (0.179-0.813) | 0.013 |
High | 0.164 (0.075-0.359) | < 0.001 | 0.390 (0.162-0.938) | 0.036 |
4.142 (2.515-6.823) | < 0.001 | 3.405 (1.948-5.954) | < 0.001 | |
0.688 (0.430-1.101) | 0.119 | - | - | |
0.895 (0.537-1.491) | 0.670 | - | - | |
0.673 | ||||
N0 | 1.000 | - | - | |
N1 | 0.783 (0.450-1.361) | 0.385 | - | - |
N2 | 0.869 (0.431-1.756) | 0.696 | - | - |
2.237 (0.978-5.116) | 0.056 | - | - | |
1.462 (0.913-2.339) | 0.114 | - | - | |
1.694 (1.031-2.784) | 0.038 | 1.505 (0.856-2.645) | 0.155 | |
0.733 (0.390-1.378) | 0.335 | - | - | |
0.590 | ||||
Protrude | 1.000 | - | ||
Infiltrating | 0.741 (0.322-1.706) | 0.481 | - | - |
Ulcerative | 0.750 (0.427-1.319) | 0.318 | - | - |
0.636 (0.239-1.688) | 0.363 | - | - | |
1.095 (0.684-1.752) | 0.707 | - | - |
CRC, colorectal cancer; OR, odds ratio; CI, confidence interval; ALB, albumin; BMI, body mass index; ASA, American Society of Anesthesiologists; CEA, carcinoembryonic antigen.
Univariate and multivariate logistic regression analysis of factors associated with complications in elderly CRC
Feature | Univariate analysis |
Multivariate analysis |
||
---|---|---|---|---|
OR (95% CI) | p-value | OR (95% CI) | p-value | |
0.739 (0.407-1.343) | 0.322 | - | - | |
1.083 (0.618-1.898) | 0.780 | - | - | |
0.557 (0.321-0.968) | 0.038 | - | - | |
0.142 | - | |||
Low | 1.000 | - | ||
Normal | 1.095 (0.511-2.349) | 0.815 | - | - |
High | 0.552 (0.225-1.353) | 0.194 | - | - |
2.159 (1.237-3.766) | 0.007 | 2.192 (1.231-3.903) | 0.008 | |
1.395 (0.802-2.429) | 0.239 | - | - | |
2.484 (1.230-5.015) | 0.011 | 2.337 (1.127-4.847) | 0.023 | |
0.890 | - | |||
N0 | 1.000 | - | ||
N1 | 1.167 (0.622-2.187) | 0.631 | - | - |
N2 | 1.036 (0.456-2.357) | 0.932 | - | - |
2.583 (1.102-6.058) | 0.029 | 1.988 (0.811-4.874) | 0.133 | |
1.053 (0.609-1.820) | 0.853 | - | - | |
0.906 (0.517-1.588) | 0.731 | - | - | |
1.332 (0.673-2.638) | 0.410 | - | - | |
0.139 | - | |||
Protrude | 1.000 | - | ||
Infiltrating | 1.140 (0.378-3.440) | 0.816 | - | - |
Ulcerative | 1.966 (0.933-4.140) | 0.075 | - | - |
0.541 (0.209-1.402) | 0.206 | - | - | |
1.129 (0.652-1.954) | 0.665 | - | - | |
2.035 (1.141-3.629) | 0.016 | 1.693 (0.929-3.084) | 0.086 |
CRC, colorectal cancer; OR, odds ratio; CI, confidence interval; ALB, albumin; BMI, body mass index; ASA, American Society of Anesthesiologists; CEA, carcinoembryonic antigen.
Univariate and multivariate survival analyses of clinicopathological characteristics in elderly colorectal cancer patients
Feature | Progression-free survival |
Overall survival |
||||||
---|---|---|---|---|---|---|---|---|
Univariate |
Multivariate |
Univariate |
Multivariate |
|||||
HR (95% CI) | p-value | HR (95% CI) | p-value | HR (95% CI) | p-value | HR (95% CI) | p-value | |
0.914 (0.629-1.329) | 0.639 | - | - | 0.916 (0.627-1.339) | 0.652 | - | - | |
1.180 (0.829-1.679) | 0.359 | - | - | 1.209 (0.846-1.727) | 0.297 | - | - | |
0.649 (0.458-0.920) | 0.015 | 0.813 (0.563-1.175) | 0.271 | 0.651 (0.458-0.927) | 0.017 | 0.797 (0.550-1.156) | 0.232 | |
0.262 | - | 0.393 | - | |||||
Low | 1.000 | - | 1.000 | - | ||||
Normal | 1.305 (0.781-2.181) | 0.309 | - | - | 1.160 (0.700-1.920) | 0.565 | - | - |
High | 0.955 (0.536-1.702) | 0.876 | - | - | 0.870 (0.491-1.540) | 0.633 | - | - |
1.607 (1.133-2.279) | 0.008 | 2.175 (1.489-3.179) | < 0.001 | 1.785 (1.252-2.545) | 0.001 | 2.524 (1.721-3.703) | < 0.001 | |
1.705 (1.192-2.438) | 0.003 | 0.919 (0.595-1.419) | 0.702 | 1.753 (1.219-2.521) | 0.002 | 0.944 (0.606-1.469) | 0.797 | |
2.317 (1.474-3.640) | < 0.001 | 1.330 (0.811-2.181) | 0.259 | 2.217 (1.409-3.488) | 0.001 | 1.238 (0.750-2.044) | 0.404 | |
< 0.001 | < 0.001 | < 0.001 | < 0.001 | |||||
N0 | 1.000 | 1.000 | 1.000 | 1.000 | ||||
N1 | 2.067 (1.378-3.099) | < 0.001 | 1.824 (1.141-2.916) | 0.012 | 2.061 (1.366-3.110) | 0.001 | 1.822 (1.129-2.941) | 0.014 |
N2 | 4.247 (2.727-6.615) | < 0.001 | 3.382 (1.955-5.851) | < 0.001 | 4.289 (2.744-6.703) | < 0.001 | 3.442 (1.995-5.936) | < 0.001 |
6.559 (4.153-10.361) | < 0.001 | 5.028 (2.999-8.431) | < 0.001 | 6.498 (4.109-10.275) | < 0.001 | 5.387 (3.208-9.046) | < 0.001 | |
0.923 (0.651-1.307) | 0.651 | - | - | - | 0.325 | - | - | |
1.117 (0.775-1.609) | 0.554 | - | - | 1.069 (0.741-1.541) | 0.722 | - | - | |
1.612 (1.070-2.428) | 0.022 | 0.949 (0.603-1.494) | 0.822 | 1.458 (0.952-2.234) | 0.083 | - | - | |
0.199 | - | 0.270 | ||||||
Protrude | 1.000 | - | 1.000 | - | - | |||
Infiltrating | 1.416 (0.743-2.696) | 0.290 | - | - | 1.405 (0.738-2.677) | 0.301 | - | - |
Ulcerative | 1.534 (0.962-2.446) | 0.072 | - | - | 1.470 (0.920-2.349) | 0.107 | - | - |
0.453 (0.260-0.790) | 0.005 | 0.433 (0.239-0.782) | 0.006 | 0.461 (0.259-0.820) | 0.008 | 0.442 (0.240-0.814) | 0.009 | |
2.140 (1.503-3.049) | < 0.001 | 1.414 (0.962-2.080) | 0.078 | 2.039 (1.427-2.913) | < 0.001 | 1.397 (0.951-2.053) | 0.089 | |
1.294 (0.907-1.848) | 0.155 | - | - | 1.267 (0.885-1.814) | 0.197 | - | - | |
1.117 (0.785-1.590) | 0.539 | - | - | 1.142 (0.800-1.631) | 0.463 | - | - |
HR, hazard ratio; CI, confidence interval; ALB, albumin; BMI, body mass index; ASA, American Society of Anesthesiologists; CEA, carcinoembryonic antigen.