Abstract
Objective
Preoperative MRI staging offers a noninvasive approach to assess local tumor invasion and distant metastasis. Attention has been paid to investigating the impact of preoperative abdominal MRI staging on the prognosis of patients with stage II-III colon cancer.
Methods
Patients who underwent radical resection for stage II-III colon cancer at Yunnan Cancer Hospital between 2008 and 2017 were retrospectively analyzed and divided into preoperative MRI and non-MRI groups. Propensity score matching (PSM) was applied to balance patient characteristics. Overall survival (OS) and Disease-free survival (DFS) were compared using the Kaplan–Meier method and log-rank test. Additionally, Cox regression was used to evaluate the correlation between preoperative MRI and prognosis, and subgroup analyses further compared the two groups. For multivariable analysis, two models were constructed and compared.
Results
Among 1086 patients included, 29.28% underwent preoperative abdominal MRI. Patients who received preoperative MRI demonstrated better OS (log-rank p = 0.002), particularly those with stage III colon cancer (p < 0.05). After PSM, the MRI group exhibited significantly better 5-year OS (90.0% vs. 82.0%, P = 0.019) and a lower risk of death (HR, 1.66; 95% CI, 1.07–2.56; P = 0.024). Subgroup analysis revealed that the OS of preoperative MRI was significant in stage III patients. However, no significant difference was observed in DFS between the two groups. Furthermore, compared to the MRI group, the non-MRI group showed a statistically significant increased risk in model 1 (OR = 1.63, 95% CI: 1.08–2.44, P = 0.020) and model 2 (OR = 1.79, 95% CI: 1.17—2.73, P = 0.007) in terms of OS.
Conclusion
Preoperative abdominal MRI significantly enhances OS in patients with stage II-III colon cancer, especially in those with stage III disease. However, the specific role of MRI in treatment planning, as well as the potential impact of different treatment approaches on outcomes, warrants further investigation into the treatment decision-making processes.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12885-025-14652-5.
Keywords: Magnetic resonance imaging, Colon cancer, Overall survival, Propensity score matching, Multicollinearity analysis
Introduction
Colon cancer is a global digestive system neoplasm characterized by nonspecific early symptoms and a poor prognosis [1, 2]. At the time of initial diagnosis, approximately one-third of patients with colon cancer have lymph node metastases but lack typical symptoms [3]. These characteristics limit therapeutic options and contribute to a diminished quality of life. Early and accurate diagnosis and treatment are crucial for improving patients'prognosis [4, 5]. When the endoscopic examination is obstructed or tissue sampling is unsuccessful, imaging studies are instrumental in aiding clinicians in diagnosing tumors and effectively planning subsequent treatments [6]. Preoperative imaging, particularly MRI staging, offers a noninvasive approach to assess local tumor invasion and distant metastasis [6, 7].
As medical technology advances, MRI diagnosis has become increasingly common and has garnered significant attention [8]. MRI demonstrates higher specificity than CT scans in preoperatively assessing colorectal cancer, especially for the detection of rectal cancer [9]. To facilitate surgeons in achieving negative incisal margins, MRI has become the preferred staging method for rectal cancer [10–14]. Furthermore, MRI is the preferred imaging modality for evaluating rectal cancer locally [15]. It has become the preferred staging method for rectal cancer, effectively detecting tumor invasion of the intestinal wall, extramural blood vessels, and serosa invasion [16]. Unlike rectal cancer, colon cancer lacks fixed anatomical landmarks. This distinction impacts MRI’s utility, as tumor mobility and variable serosal relationships complicate staging accuracy in colonic segments. However, for individuals with locally advanced colon cancer (stage II-III), MRI may offer advantages in identifying extramural venous infiltration and determining the T stage [17]. Notably, MRI was performed to assess the presence or absence of invasion of other organs in advanced cancer and to provide detailed anatomical information for early-stage cancers (T1N0) to ensure complete resection [8]. Therefore, MRI should be considered a valuable staging tool for the preoperative assessment of colon cancer and surgical planning. However, the precise impact of MRI on the staging of colon cancer and its prognostic utility remains to be fully elucidated.
In our study, we conducted a single-center retrospective cohort study on patients with resectable stage II-III colon cancer, investigating the survival benefits of staging abdominal MR images.
Patients and methods
The study protocol was approved by the ethics committee of Yunnan Cancer Hospital (KY2019141), and given its retrospective nature, the requirement for informed consent was waived. All the data were anonymized. A total of 1836 patients with stage II-III colon cancer diagnosed at Yunnan Cancer Hospital from January 1, 2008, to December 31, 2017, were included. Demographic information about patients, perioperative clinical results, and pathological findings are acquired from a database and examined in electronic medical records, together with disease status at the most recent follow-up. Patients were subdivided into MRI and non-MRI groups based on whether they received an abdominal MRI before surgery. The patient inclusion and exclusion criteria are provided in Fig. 1. Patients were excluded if they had a history of a malignant neoplasm within the last 5 years (n = 12), received preoperative chemotherapy (n = 434), had missing data (n = 219), or underwent MRI examinations in other hospitals (n = 85). As a result, 1086 participants with R0 resection were included in the final analysis. 318 patients with colon cancer underwent preoperative MRI examination, and 768 patients with colon cancer did not undergo an MRI examination before the operation. The surgeons had up to eight years of experience in performing colon cancer operations.
Fig. 1.
Patient flowchart showing inclusion/exclusion criteria. Total N = 1086; Final analytic cohort N = 1,263 after exclusions
MRI provides greater tissue differentiation between desmoplastic and neoplastic pericolic invasion than CT, contributing to more accurate preoperative staging. To assess the depth of tumor invasion, the involvement of adjacent tissues, and possible metastases, T1-weighted and axial T2-weighted images(T1WI, T2WI) were acquired at a median of 7 days pre-surgery (Interquartile Range, IQR: 5–14 days). All MRIs were performed using 1.5-T or 3.0-T scanners with standardized protocols, including axial T2-weighted imaging (3 mm slice thickness), diffusion-weighted imaging (b = 1000 s/mm2), and contrast-enhanced T1-weighted imaging. MRI was preferred for its superior diagnostic performance. Two experienced radiologists (each with ≥ 10 years of experience) independently interpreted the images in a blinded manner to the pathological results.
Staging abdomen MRI
The patients who underwent preoperative abdominal MRI adhered to our institution’s standard protocol of 1.5-T or 3.0-T MRI scanners. Three-dimensional gradient-echo T1WI and axial T2WI with turbo spin that is contrast-unenhanced and contrast-enhanced (axial, coronal, and sagittal reconstruction)-fluid axial-echo-attenuated inversion-recovery images were acquired individually. MRI protocols included T2-weighted imaging (TR/TE: 4000–6000/80–100 ms) and diffusion-weighted imaging (b-values: 0, 800 s/mm2) on 3.0-T systems.
Surveillance protocol and outcomes
According to the guidelines [18], the standard postoperative monitoring procedures for patients with stage I-III colon cancer involve gathering medical history, performing a physical exam, and checking blood tumor marker levels. The surveillance schedule included serum tumor marker tests every three months for two years, every six months for 5 years, and annually after that. CT of the chest, abdomen, and pelvis was performed every six months for two years and then annually for a total of 5 years. Unless advanced adenomas are discovered, a colonoscopy is typically performed one year after surgery and every three to five years. The existence of new lesions confirmed by CT, MRI, positron emission tomography (PET), and/or biopsy histology formed the basis for a conclusive diagnosis of recurrence. OS was included as the primary outcome variable in this study and was defined as the time between surgery and death from any cause or the last follow-up. The secondary outcome was DFS. DFS was defined as the time from surgery to the first instance of tumor recurrence, metastasis, or death due to any cause.
Definition of study covariates
We extracted data such as age, sex, body mass index (BMI), preoperative carcinoembryonic antigen (CEA), postoperative CEA, surgical approach (open resection or laparoscopic resection or conversion to open surgery), primary site (right colon or left colon), tumor differentiation, mucinous type, lymph node yield, adjuvant chemotherapy (yes or no), chemotherapy regimen, chemotherapy cycle, vascular tumor thrombus, nerve membranous invasion, clinical stages were based on the AJCC/UICC (Version 8), and disease status at the last follow-up date from the Yunnan Cancer Hospital.
Statistical analysis
Continuous variables are reported as means with standard deviations (SDs) or medians with interquartile ranges (IQRs), as appropriate. Categorical variables are presented as counts and percentages, and compared using the chi-square or Fisher’s exact test. Propensity score matching (PSM) with 1:1 nearest neighbor matching and a caliper of 0.15 was used to reduce selection bias and balance baseline characteristics, assessed by standardized mean differences (SMDs), with SMD < 0.1 indicating good balance. Survival outcomes were estimated using the Kaplan–Meier method and compared by the log-rank test. Univariable and multivariable Cox proportional hazards regression identified independent prognostic factors. For multivariable analysis, three strategies were used: (1) backward stepwise selection based on the lowest Akaike Information Criterion (AIC) using the stepAIC function (MASS package, R); (2) adjustment for age group, BMI group, and sex (Model 1); and (3) further adjustment for primary tumor site, surgical approach, pathological stage, lymph node yield, tumor differentiation, mucinous histology, vascular invasion, perineural invasion, tumor deposits, adjuvant chemotherapy, pre- and postoperative CEA (Model 2). Multicollinearity was assessed by tolerance (< 0.1) and variance inflation factor (VIF > 10), and a correlation coefficient < 0.7 between covariates was considered acceptable. Analyses were performed using SPSS 24.0 and R 4.4.2. Two-sided p-values < 0.05 were considered statistically significant.
Results
Patient characteristics
A total of 1086 patients were included in the study, with a mean age of 60(51,68) years. Among them, 643 (59.2%) were males. Recurrence was observed in 237 patients (21.82%) before the initial follow-up. 198 patients (18.1%) died during the follow-up, including 187 individuals from the primary cancer and 11 patients from other diseases. The percentage of postoperative chemotherapy was 73.2% (800/1086). The patients were categorized into two groups based on whether they received preoperative abdomen MRI: non-MRI (n = 768) and MRI (n = 318). Significant differences in surgical approach, primary site, and chemotherapy regimen were found between the groups (all p < 0.001) (Table 1).
Table 1.
Patient demographics and baseline characteristics before Propensity Score Matching
| Characteristic | MRI group | P-value2 | ||
|---|---|---|---|---|
| OverallN = 1,0861 | Non-MRIN = 7681 | MRI N = 3181 | ||
| Age (years) | 60 (51, 68) | 60 (50, 68) | 60 (51, 68) | 0.867 |
| BMI (kg/m2) | 22.4 (20.6, 24.8) | 22.4 (20.6, 24.8) | 22.3 (20.7, 24.8) | 0.970 |
| Age group (years) | 0.750 | |||
| ≥ 65 | 558 (51.4%) | 397 (51.7%) | 161 (50.6%) | |
| < 65 | 528 (48.6%) | 371 (48.3%) | 157 (49.4%) | |
| BMI group (kg/m2) | 0.520 | |||
| < 24 | 729 (67.1%) | 511 (66.5%) | 218 (68.6%) | |
| ≥ 24 | 357 (32.9%) | 257 (33.5%) | 100 (31.4%) | |
| Preoperative CEA (ng/l) | 4 (2, 11) | 4 (2, 11) | 4 (2, 11) | 0.119 |
| Postoperative CEA (ng/l) | 2.02 (1.29, 3.32) | 1.99 (1.27, 3.26) | 2.17 (1.33, 3.40) | 0.141 |
| Gender | 0.146 | |||
| Male | 643 (59.2%) | 444 (57.8%) | 199 (62.6%) | |
| Female | 443 (40.8%) | 324 (42.2%) | 119 (37.4%) | |
| Surgical approach | 0.001 | |||
| OR | 773 (71.2%) | 569 (74.1%) | 204 (64.2%) | |
| LR | 313 (28.8%) | 199 (25.9%) | 114 (35.8%) | |
| Primary site | < 0.001 | |||
| Left colon | 532 (49.0%) | 346 (45.1%) | 186 (58.5%) | |
| Right colon | 554 (51.0%) | 422 (54.9%) | 132 (41.5%) | |
| Tumor differentiation | 0.336 | |||
| Poor/undifferentiated | 65 (6.0%) | 49 (6.4%) | 16 (5.0%) | |
| Well | 594 (54.7%) | 408 (53.1%) | 186 (58.5%) | |
| Moderate | 374 (34.4%) | 270 (35.2%) | 104 (32.7%) | |
| Unknown | 53 (4.9%) | 41 (5.3%) | 12 (3.8%) | |
| Mucinous type | 0.212 | |||
| No | 1,007 (92.7%) | 707 (92.1%) | 300 (94.3%) | |
| Yes | 65 (6.0%) | 52 (6.8%) | 13 (4.1%) | |
| Unknown | 14 (1.3%) | 9 (1.2%) | 5 (1.6%) | |
| Pathological T stage | 0.800 | |||
| T3 | 1,011 (93.1%) | 714 (93.0%) | 297 (93.4%) | |
| T4 | 75 (6.9%) | 54 (7.0%) | 21 (6.6%) | |
| Pathological N stage | 0.581 | |||
| N0 | 605 (55.7%) | 430 (56.0%) | 175 (55.0%) | |
| N1 | 276 (25.4%) | 196 (25.5%) | 80 (25.2%) | |
| N2a | 154 (14.2%) | 103 (13.4%) | 51 (16.0%) | |
| N2b | 51 (4.7%) | 39 (5.1%) | 12 (3.8%) | |
| Pathological stage | 0.772 | |||
| II | 605 (55.7%) | 430 (56.0%) | 175 (55.0%) | |
| III | 481 (44.3%) | 338 (44.0%) | 143 (45.0%) | |
| Lymph node yield group | 0.092 | |||
| < 12 | 138 (12.7%) | 106 (13.8%) | 32 (10.1%) | |
| ≥ 12 | 948 (87.3%) | 662 (86.2%) | 286 (89.9%) | |
| Vascular tumor thrombus | 0.167 | |||
| No | 110 (10.1%) | 70 (9.1%) | 40 (12.6%) | |
| Yes | 64 (5.9%) | 43 (5.6%) | 21 (6.6%) | |
| Unknown | 912 (84.0%) | 655 (85.3%) | 257 (80.8%) | |
| Nerve membranous invasion | 0.333 | |||
| No | 131 (12.1%) | 88 (11.5%) | 43 (13.5%) | |
| Yes | 17 (1.6%) | 10 (1.3%) | 7 (2.2%) | |
| Unknown | 938 (86.4%) | 670 (87.2%) | 268 (84.3%) | |
| TD | 0.317 | |||
| No | 988 (91.0%) | 703 (91.5%) | 285 (89.6%) | |
| Yes | 98 (9.0%) | 65 (8.5%) | 33 (10.4%) | |
| Adjuvant chemotherapy | 0.054 | |||
| No | 286 (26.3%) | 215 (28.0%) | 71 (22.3%) | |
| Yes | 800 (73.7%) | 553 (72.0%) | 247 (77.7%) | |
| Chemotherapy regimen | < 0.001 | |||
| FOLFOX | 386 (35.5%) | 246 (32.0%) | 140 (44.0%) | |
| CAPOX/XELOX | 225 (20.7%) | 176 (22.9%) | 49 (15.4%) | |
| Other | 46 (4.2%) | 39 (5.1%) | 7 (2.2%) | |
| 5-FU/Capecitabine | 142 (13.1%) | 91 (11.8%) | 51 (16.0%) | |
| Unknown | 287 (26.4%) | 216 (28.1%) | 71 (22.3%) | |
| Chemotherapy cycle | 0.133 | |||
| < 6 | 439 (40.4%) | 307 (40.0%) | 132 (41.5%) | |
| ≥ 6 | 361 (33.2%) | 246 (32.0%) | 115 (36.2%) | |
| Unknown | 286 (26.3%) | 215 (28.0%) | 71 (22.3%) | |
| Preoperative CEA group (ng/l) | 0.734 | |||
| < 5 | 607 (55.9%) | 435 (56.6%) | 172 (54.1%) | |
| ≥ 5 | 471 (43.4%) | 327 (42.6%) | 144 (45.3%) | |
| Unknown | 8 (0.7%) | 6 (0.8%) | 2 (0.6%) | |
| Postoperative CEA group (ng/l) | 0.082 | |||
| < 5 | 835 (76.9%) | 580 (75.5%) | 255 (80.2%) | |
| ≥ 5 | 118 (10.9%) | 83 (10.8%) | 35 (11.0%) | |
| Unknown | 133 (12.2%) | 105 (13.7%) | 28 (8.8%) | |
1Median (Q1, Q3); n (%)
2Wilcoxon rank sum test; Pearson's Chi-squared test; Fisher's exact test
After PSM,312 patients underwent preoperative abdominal MRI, and 312 patients did not receive an MRI. There was no significant difference in the baseline clinical and pathological characteristics of patients after PSM, which reduced the imbalance in the distribution of patient characteristics (all SMDs < 0.1) (Table 2). Table 2 shows the adequacy of PSM and the deficiency of differences in clinical pathological attributes between the two groups, and among the non-matched variables, no differences were found. In the PSM cohort, 312 patients with MRI data were compared with 312 patients with non-MRI data (Table 2).
Table 2.
Baseline covariates before and after propensity score matching
| Variables | Level | Before matching | After matching | ||||
|---|---|---|---|---|---|---|---|
| Non-MRI | MRI | SMD△ | Non-MRI | MRI | SMD△ | ||
| n | 768 | 318 | 312 | 312 | |||
| Gender (%) | Female | 324 (42.2) | 119 (37.4) | −0.098 | 115 (36.9) | 117 (37.5) | 0.013 |
| Male | 444 (57.8) | 199 (62.6) | 0.098 | 197 (63.1) | 195 (62.5) | −0.013 | |
| Surgical approach (%) | LR | 199 (25.9) | 114 (35.8) | 0.207 | 110 (35.3) | 108 (34.6) | −0.013 |
| OR | 569 (74.1) | 204 (64.2) | −0.207 | 202 (64.7) | 204 (65.4) | 0.013 | |
| Primary site (%) | Left colon | 346 (45.1) | 186 (58.5) | 0.273 | 175 (56.1) | 180 (57.7) | 0.033 |
| Right colon | 422 (54.9) | 132 (41.5) | −0.273 | 137 (43.9) | 132 (42.3) | −0.033 | |
| Tumor differentiation (%) | Poor/undifferentiated | 49 (6.4) | 16 (5.0) | −0.062 | 13 (4.2) | 15 (4.8) | 0.029 |
| Moderate | 270 (35.2) | 104 (32.7) | −0.052 | 102 (32.7) | 102 (32.7) | 0.000 | |
| Well | 408 (53.1) | 186 (58.5) | 0.109 | 185 (59.3) | 183 (58.7) | −0.013 | |
| Unknown | 41 (5.3) | 12 (3.8) | −0.082 | 12 (3.8) | 12 (3.8) | 0.000 | |
| Mucinous type (%) | No | 707 (92.1) | 300 (94.3) | 0.099 | 302 (96.8) | 295 (94.6) | −0.097 |
| Yes | 52 (6.8) | 13 (4.1) | −0.135 | 10 (3.2) | 13 (4.2) | 0.049 | |
| Unknown | 9 (1.2) | 5 (1.6) | 0.032 | 0 (0.0) | 4 (1.3) | 0.103 | |
| Pathological stage (%) | II | 430 (56.0) | 175 (55.0) | −0.019 | 168 (53.8) | 172 (55.1) | 0.026 |
| III | 338 (44.0) | 143 (45.0) | 0.019 | 144 (46.2) | 140 (44.9) | −0.026 | |
| Lymph node yield group (%) | < 12 | 106 (13.8) | 32 (10.1) | −0.124 | 26 (8.3) | 32 (10.3) | 0.064 |
| ≥ 12 | 662 (86.2) | 286 (89.9) | 0.124 | 286 (91.7) | 280 (89.7) | −0.064 | |
| Vascular tumor thrombus (%) | No | 70 (9.1) | 40 (12.6) | 0.104 | 32 (10.3) | 37 (11.9) | 0.048 |
| Yes | 43 (5.6) | 21 (6.6) | 0.040 | 21 (6.7) | 21 (6.7) | 0.000 | |
| Unknown | 655 (85.3) | 257 (80.8) | −0.113 | 259 (83.0) | 254 (81.4) | −0.041 | |
△Standardized Mean Difference
Associations of abdominal MRI with survival
The Kaplan‒Meier survival curve illustrates the relationship between preoperative MRI staging and prognosis in colon cancer patients before PSM. Colon cancer patients who received preoperative abdominal MRI had better OS (Fig. 2A, p = 0.002). In contrast, the two groups had no significant difference in DFS (Fig. 2C, p = 0.40). Kaplan–Meier survival curves illustrated the relationship between the preoperative MRI stage and prognosis in colon cancer patients after PSM. The results remain consistent with those before PSM. Colon cancer patients who received preoperative abdominal MRI staging had a better OS (Fig. 2B, p = 0.016). In contrast, the two groups also had no significant difference in DFS (Fig. 2D, p = 0.61).
Fig. 2.
The Kaplan–Meier survival curve illustrates the relationship between preoperative MRI use and prognosis in colon cancer patients before and after PSM. A MRI group had better OS before PSM, (B) No significant difference in DFS before PSM. C MRI group had better OS after PSM, (D) No significant difference in DFS after PSM
The results of the univariate and multivariate analyses of the OS-associated factors before PSM are shown in Table S1. Univariable analysis revealed that preoperative MRI, age, laparoscopic resection (LR), primary site(right colon), tumor differentiation(moderate, poor/undifferentiated), pathological N stage(N1/N2a/N2b), pathological III stage, vascular tumor thrombus(VTT), nerve membranous invasion, tumor deposition (TD), chemotherapy cycle(≥ 6), and preoperative and postoperative CEA ≥ 5 ng/ml were significantly associated with OS (all p < 0.05). According to the multivariate analysis, preoperative MRI, primary site(right colon), pathological N stage(N1/N2a/N2b), nerve membranous invasion, chemotherapy cycle(≥ 6), and postoperative CEA ≥ 5 ng/ml were significantly associated with OS (all p < 0.05). Similarly, univariate and multivariate analyses of the factors associated with DFS are shown in Table S3. Univariable analysis revealed that tumor differentiation(poor/undifferentiated), pathological N stage(N1/N2a/N2b), pathological III stage, lymph node yield group(≥ 12), TD(Yes), chemotherapy regimen (5-FU/Capecitabine), and postoperative CEA ≥ 5 ng/ml were significantly associated with DFS (all p < 0.05). According to the multivariate analysis, the primary site (right colon), pathological N stage(N1/N2a/N2b), lymph node yield group(≥ 12), chemotherapy regimen (5-FU/Capecitabine), and postoperative CEA ≥ 5 ng/ml were significantly associated with DFS (all p < 0.05). After PSM, univariable Cox regression analysis revealed that OS in the MRI group was favorable (HR = 1.65, 95% CI: 1.09–2.48; p = 0.017) (Table 3). In addition, characteristics including tumor differentiation(moderate/well), pathological N stage(N2a, N2b), pathological stage, nerve membranous invasion(yes), tumor deposition(TD)(yes), preoperative and postoperative CEA ≥ 5 ng/ml group were brought into multivariate Cox regression. According to the multivariate Cox regression analysis, MRI was an independent favorable factor for OS (HR = 1.90, 95% CI: 1.25–2.89, p = 0.002) (Table 3). Moreover, the multivariable analysis identified four variables, including tumor differentiation(moderate/well)(HR = 0.32 95%CI:0.15–0.66, p = 0.002; HR = 0.41, 95% CI:0.20–0.84, p = 0.015), pathological N stage(N2a, N2b)(HR = 3.06 95%CI:1.78–5.23, p < 0.001; HR = 5.18, 95% CI:2.37–11.28, p < 0.001), nerve membranous invasion(yes)(HR = 4.68, 95%CI:1.51–14.49, p = 0.007), and postoperative CEA ≥ 5 ng/ml group (HR = 4.15, 95%CI: 2.57–6.71, p < 0.001), as independent predictors of OS (Table 3). However, according to the univariable Cox regression model, MRI was not an independent favorable factor for DFS (HR = 1.09, 95% CI: 0.78–1.54, p = 0.614) (Table 4). However, according to the multivariate Cox regression analysis, MRI was not an independent favorable factor for DFS (HR = 1.14, 95% CI: 0.80–1.60, p = 0.472) (Table 4). Moreover, multivariable analysis revealed four variables, including BMI group (HR 1.52, 95% CI 1.10–2.09; P = 0.011), tumor deposition (TD) (yes/no) (HR 1.94, 95% CI, 1.26–3.01; P = 0.003), chemotherapy regimen (CAPOX/XELOX) (HR, 2.50; 95%CI, 1.44–4.32; p = 0.001), and postoperative CEA group (≥ 5) (HR 2.40, 95% CI, 1.53–3.77; p ≤ 0.001), as independent predictors of DFS (Table 4).
Table 3.
Univariate Cox regression analysis and multivariate analysis of factors associated with Overall Survival using backward stepwise selection (AIC criterion) after Propensity Score Matching
| Characteristic | Univariable | Multivariable | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| N | Event N | HR | 95% CI | P-value | N | Event N | HR | 95% CI | P-value | |
| MRI group | ||||||||||
| MRI | 312 | 37 | Ref | — | 312 | 37 | Ref | — | ||
| Non-MRI | 312 | 61 | 1.65 | 1.09, 2.48 | 0.017 | 312 | 61 | 1.90 | 1.25, 2.89 | 0.002 |
| Gender | ||||||||||
| Female | 232 | 38 | Ref | — | ||||||
| Male | 392 | 60 | 0.93 | 0.62, 1.40 | 0.733 | |||||
| Age group, years | ||||||||||
| < 65 | 299 | 46 | Ref | — | ||||||
| ≥ 65 | 325 | 52 | 1.02 | 0.68, 1.51 | 0.935 | |||||
| BMI group, kg/m2 | ||||||||||
| < 24 | 420 | 65 | Ref | — | ||||||
| ≥ 24 | 204 | 33 | 1.08 | 0.71, 1.65 | 0.709 | |||||
| Surgical approach | ||||||||||
| OR | 406 | 74 | Ref | — | 406 | 74 | Ref | — | ||
| LR | 218 | 24 | 0.68 | 0.43, 1.08 | 0.106 | 218 | 24 | 0.71 | 0.44, 1.14 | 0.151 |
| Primary site | ||||||||||
| Left colon | 355 | 52 | Ref | — | ||||||
| Right colon | 269 | 46 | 1.21 | 0.81, 1.80 | 0.350 | |||||
| Tumor differentiation | ||||||||||
| Poor/undifferentiated | 28 | 10 | Ref | — | 28 | 10 | Ref | — | ||
| Moderate | 204 | 33 | 0.40 | 0.20, 0.82 | 0.012 | 204 | 33 | 0.32 | 0.15, 0.66 | 0.002 |
| Well | 368 | 51 | 0.41 | 0.21, 0.80 | 0.009 | 368 | 51 | 0.41 | 0.20, 0.84 | 0.015 |
| Mucinous type | ||||||||||
| No | 597 | 93 | Ref | — | ||||||
| Yes | 23 | 3 | 1.19 | 0.38, 3.79 | 0.764 | |||||
| Pathological T stage | ||||||||||
| T3 | 584 | 90 | Ref | — | ||||||
| T4 | 40 | 8 | 1.40 | 0.68, 2.88 | 0.365 | |||||
| Pathological N stage | ||||||||||
| N0 | 340 | 34 | Ref | — | 340 | 34 | Ref | — | ||
| N1 | 172 | 30 | 1.70 | 1.04, 2.77 | 0.035 | 172 | 30 | 1.50 | 0.90, 2.47 | 0.117 |
| N2a | 91 | 25 | 3.19 | 1.90, 5.35 | < 0.001 | 91 | 25 | 3.06 | 1.78, 5.23 | < 0.001 |
| N2b | 21 | 9 | 4.75 | 2.28, 9.91 | < 0.001 | 21 | 9 | 5.18 | 2.37, 11.28 | < 0.001 |
| Pathological stage | ||||||||||
| II | 340 | 34 | Ref | — | ||||||
| III | 284 | 64 | 2.33 | 1.54, 3.54 | < 0.001 | |||||
| Lymph node yield group | ||||||||||
| < 12 | 58 | 13 | Ref | — | ||||||
| ≥ 12 | 566 | 85 | 0.67 | 0.37, 1.20 | 0.182 | |||||
| Vascular tumor thrombus | ||||||||||
| No | 69 | 9 | Ref | — | ||||||
| Yes | 42 | 9 | 1.72 | 0.68, 4.33 | 0.252 | |||||
| Nerve membranous invasion | ||||||||||
| No | 82 | 10 | Ref | — | 82 | 10 | Ref | — | ||
| Yes | 12 | 5 | 4.78 | 1.63, 14.01 | 0.004 | 12 | 5 | 4.68 | 1.51, 14.49 | 0.007 |
| TD | ||||||||||
| No | 566 | 81 | Ref | — | ||||||
| Yes | 58 | 17 | 2.78 | 1.65, 4.71 | < 0.001 | |||||
| Adjuvant chemotherapy | ||||||||||
| No | 168 | 30 | Ref | — | ||||||
| Yes | 456 | 68 | 0.77 | 0.50, 1.18 | 0.231 | |||||
| Chemotherapy regimen | ||||||||||
| FOLFOX | 233 | 35 | Ref | — | ||||||
| CAPOX/XELOX | 115 | 22 | 1.30 | 0.76, 2.21 | 0.341 | |||||
| 5-FU/Capecitabine | 87 | 8 | 0.59 | 0.27, 1.27 | 0.178 | |||||
| Other | 21 | 3 | 0.85 | 0.26, 2.76 | 0.783 | |||||
| Chemotherapy cycle | ||||||||||
| < 6 | 249 | 40 | Ref | — | ||||||
| ≥ 6 | 207 | 28 | 0.83 | 0.51, 1.34 | 0.444 | |||||
| Preoperative CEA group (ng/l) | ||||||||||
| < 5 | 349 | 44 | Ref | — | ||||||
| ≥ 5 | 270 | 54 | 1.71 | 1.15, 2.55 | 0.008 | |||||
| Postoperative CEA group (ng/l) | ||||||||||
| < 5 | 487 | 60 | Ref | — | 487 | 60 | Ref | — | ||
| ≥ 5 | 65 | 27 | 4.16 | 2.64, 6.55 | < 0.001 | 65 | 27 | 4.15 | 2.57, 6.71 | < 0.001 |
Abbreviations: CI Confidence Interval, HR Hazard Ratio
Table 4.
Univariate Cox regression analysis and multivariate analysis of factors associated with Disease-Free Survivalusing backward stepwise selection (AIC criterion) after Propensity Score Matching
| Characteristic | Univariable | Multivariable | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| N | Event N | HR | 95% CI | P-value | N | Event N | HR | 95% CI | P-value | |
| MRI group | ||||||||||
| MRI | 312 | 63 | Ref | — | 312 | 63 | Ref | — | ||
| Non-MRI | 312 | 68 | 1.09 | 0.78, 1.54 | 0.614 | 312 | 68 | 1.14 | 0.80, 1.60 | 0.472 |
| Gender | ||||||||||
| Female | 232 | 56 | Ref | — | 232 | 56 | Ref | — | ||
| Male | 392 | 75 | 0.76 | 0.54, 1.08 | 0.129 | 392 | 75 | 0.67 | 0.47, 0.96 | 0.027 |
| Age group, years | ||||||||||
| < 65 | 299 | 62 | Ref | — | ||||||
| ≥ 65 | 325 | 69 | 1.05 | 0.75, 1.48 | 0.778 | |||||
| BMI group,kg/m2 | ||||||||||
| < 24 | 420 | 74 | Ref | — | 420 | 74 | Ref | — | ||
| ≥ 24 | 204 | 57 | 1.70 | 1.20, 2.40 | 0.003 | 204 | 57 | 1.72 | 1.22, 2.44 | 0.002 |
| Surgical approach | ||||||||||
| OR | 406 | 85 | Ref | — | ||||||
| LR | 218 | 46 | 1.07 | 0.74, 1.53 | 0.730 | |||||
| Primary site | ||||||||||
| Left colon | 355 | 75 | Ref | — | ||||||
| Right colon | 269 | 56 | 1.00 | 0.71, 1.41 | 0.990 | |||||
| Tumor differentiation | ||||||||||
| Poor/undifferentiated | 28 | 9 | Ref | — | ||||||
| Moderate | 204 | 38 | 0.55 | 0.26, 1.13 | 0.103 | |||||
| Well | 368 | 80 | 0.68 | 0.34, 1.36 | 0.277 | |||||
| Mucinous type | ||||||||||
| No | 597 | 123 | Ref | — | ||||||
| Yes | 23 | 6 | 1.35 | 0.59, 3.07 | 0.473 | |||||
| Pathological T stage | ||||||||||
| T3 | 584 | 123 | Ref | — | ||||||
| T4 | 40 | 8 | 1.00 | 0.49, 2.04 | 0.998 | |||||
| Pathological N stage | ||||||||||
| N0 | 340 | 52 | Ref | — | 340 | 52 | Ref | — | ||
| N1 | 172 | 38 | 1.49 | 0.98, 2.27 | 0.060 | 172 | 38 | 1.36 | 0.89, 2.07 | 0.154 |
| N2a | 91 | 32 | 2.71 | 1.74, 4.21 | < 0.001 | 91 | 32 | 2.36 | 1.50, 3.71 | < 0.001 |
| N2b | 21 | 9 | 3.43 | 1.69, 6.96 | < 0.001 | 21 | 9 | 3.67 | 1.80, 7.50 | < 0.001 |
| Pathological stage | ||||||||||
| II | 340 | 52 | Ref | — | ||||||
| III | 284 | 79 | 1.98 | 1.40, 2.81 | < 0.001 | |||||
| Lymph node yield group | ||||||||||
| < 12 | 58 | 18 | Ref | — | ||||||
| ≥ 12 | 566 | 113 | 0.63 | 0.39, 1.04 | 0.073 | |||||
| Vascular tumor thrombus | ||||||||||
| No | 69 | 14 | Ref | — | ||||||
| Yes | 42 | 8 | 0.97 | 0.41, 2.31 | 0.941 | |||||
| Nerve membranous invasion | ||||||||||
| No | 82 | 16 | Ref | — | ||||||
| Yes | 12 | 2 | 1.13 | 0.26, 4.91 | 0.872 | |||||
| TD | ||||||||||
| No | 566 | 108 | Ref | — | ||||||
| Yes | 58 | 23 | 2.64 | 1.68, 4.15 | < 0.001 | |||||
| Adjuvant chemotherapy | ||||||||||
| No | 168 | 28 | Ref | — | ||||||
| Yes | 456 | 103 | 1.33 | 0.88, 2.02 | 0.178 | |||||
| Chemotherapy regimen | ||||||||||
| FOLFOX | 233 | 61 | Ref | — | ||||||
| CAPOX/XELOX | 115 | 28 | 0.90 | 0.58, 1.41 | 0.656 | |||||
| 5-FU/Capecitabine | 87 | 10 | 0.40 | 0.21, 0.79 | 0.008 | |||||
| Other | 21 | 4 | 0.66 | 0.24, 1.82 | 0.421 | |||||
| Chemotherapy cycle | ||||||||||
| < 6 | 249 | 54 | Ref | — | ||||||
| ≥ 6 | 207 | 49 | 1.07 | 0.73, 1.57 | 0.740 | |||||
| Preoperative CEA group (ng/l) | ||||||||||
| < 5 | 349 | 68 | Ref | — | ||||||
| ≥ 5 | 270 | 61 | 1.20 | 0.85, 1.69 | 0.305 | |||||
| Postoperative CEA group (ng/l) | ||||||||||
| < 5 | 487 | 97 | Ref | — | 487 | 97 | Ref | — | ||
| ≥ 5 | 65 | 25 | 2.44 | 1.57, 3.80 | < 0.001 | 65 | 25 | 2.40 | 1.53, 3.77 | < 0.001 |
Abbreviations: CI Confidence Interval, HR Hazard Ratio
Furthermore, the results of the multicollinearity analysis revealed that the tolerance was > 0.2 and that the variance inflation factor (VIF) was < 2 before and after PSM (Tables S2, S4, S6), indicating that there was no collinearity among the independent variables [19], which was relatively stable.
Subgroup analysis of the two groups
To better understand the potential value of preoperative MRI, we conducted analyses stratified by different stages of colon cancer and performed subgroup analyses for patients with stages II-III disease. Patients were stratified based on sex, age, BMI, primary site, surgical approach, pathological stage, mucinous type, lymph node yield group, and adjuvant chemotherapy. Subsequent multivariable Cox regression analyses were performed within each subgroup to assess the prognostic value of MRI. Before and after PSM, stage III patients who underwent preoperative MRI exhibited substantially higher mortality risk compared to non-MRI patients (HR = 2.43, 95% CI: 1.51–3.92; HR = 2.19, 95% CI: 1.29–3.72) (Figs. 3 and 4). In contrast, Stage II patients showed no significant survival difference associated with MRI utilization in either cohort (before PSM: HR = 1.11, 95% CI: 0.64–1.92; after PSM: HR = 1.00, 95% CI: 0.51–1.96) (Figs. 3 and 4).
Fig. 3.
Subgroup analysis before PSM
Fig. 4.
Subgroup analysis after PSM
In terms of OS, compared to the MRI group, the non-MRI group showed a statistically significant increased risk in model 1 and model 2 before and after PSM (Table 5S and Table 5). However, for DFS, the no-MRI group didn’t show a statistically significant difference compared to the MRI group in either model 1 or model 2 before and after PSM (Table 5S and Table 5). This suggests that being in the non-MRI group was associated with worse OS in two models. So the MRI group wasn’t significantly linked to DFS in the models used.
Table 5.
Multivariate analysis of MRI group and overall survival or disease-free survival after propensity score matching
| Variable | Overall survival | Disease-free survival | ||||||
|---|---|---|---|---|---|---|---|---|
| Model 11 | Model 22 | Model 11 | Model 22 | |||||
| OR (95%CI) | P value | OR (95%CI) | P value | OR (95%CI) | P value | OR (95%CI) | P value | |
| MRI group | ||||||||
| Yes | Ref | Ref | Ref | Ref | ||||
| No | 1.63 (1.08–2.44) | 0.020 | 1.79 (1.17–2.73) | 0.007 | 1.02 (0.72–1.44) | 0.912 | 1.05 (0.73–1.50) | 0.788 |
1Model 1: adjusted by age group, BMI group, gender
2Model 2: adjusted by Model1 plus primary site, surgery approach, pathological stage, lymph nodes yield, tumor differentiation, mucinous type, Vascular tumor thrombus,Nerve membranous invasion, TD adjuvant chemotherapy, preoperative CEA, postoperative CEA
Discussion
The results of this study indicate that, both before and after matching, preoperative abdomen MRI staging is associated with better prognosis and OS for patients with stage II-III colon cancer, particularly those with stage III disease. Specifically, patients with colon cancer who underwent preoperative MRI had an 8% higher 5-year OS rate than those who did not have the MRI. This significant association suggests that preoperative abdominal MRI provides detailed anatomical information about the local tumor stage, including the depth of invasion, the involvement of adjacent tissues, and possible metastases [20, 21]. This detailed information is essential for accurate staging and planning individualized treatment strategies, which can improve R0 resection rates and surgical technique sensitivity [22, 23].
Despite the growing interest in using MRI for colon cancer staging, current guidelines from major oncology societies such as the European Society of Medical Oncology (ESMO), the National Comprehensive Cancer Network (NCCN), and the Chinese Society of Clinical Oncology (CSCO) do not universally recommend preoperative abdomen MRI for colon cancer patients [18, 22, 23]. Our study supports the need to reconsider these recommendations, as MRI demonstrated significant benefits in predicting OS, especially in stage III patients. This aligns with previous studies highlighting MRI's superior tissue differentiation capabilities compared to CT, particularly in identifying extramural venous invasion and other high-risk features [8, 18, 24]. For instance, Kim TH et al. [8] reported that MRI has moderate sensitivity and specificity in identifying colon cancer extramural vascular invasion, which can be critical for accurate staging and treatment planning. In addition, MRI has the advantage of being radiation-free compared to CT.
This study aimed to pinpoint the criteria for judiciously applying preoperative MRI in patients with resectable colon cancer, guided by the American Joint Committee on Cancer's 8th edition of clinical staging. Such comprehensive information is crucial for developing an individualized treatment plan. With precise staging, clinicians can better assess the tumor severity and select the most appropriate treatment strategy, improving patient prognosis. To evaluate the clinical utility of prognostic assessments in colon cancer patients, we conducted a Cox regression analysis, which confirmed that MRI is a significant independent predictor of 5-year OS in patients with stage II-III colon cancer. Therefore, MRI seems to be of important clinical value as a prognostic marker of colon cancer. By providing detailed imaging information of the tumor, preoperative MRI can help doctors better understand the characteristics of the cancer and develop more accurate treatment plans. However, MRI had a significant impact on OS, but it did not significantly affect DFS. MRI optimizes surgical planning and lymphadenectomy, resulting in greater R0 resection rates, but it doesn’t address micrometastatic disease, which preoperative therapy handles better. Uncertainty exists regarding the utility of MRI in patients with clinically early-stage colon cancer who have a low probability of occult metastases. Our analysis revealed that the absence of a significant difference in DFS between the MRI and non-MRI groups after PSM is intriguing. DFS is a critical measure of treatment efficacy, reflecting the duration of patient survival without disease recurrence. This discrepancy may be attributed to the fact that MRI's detailed anatomical information helps achieve R0 resection but does not always reduce the risk of microscopic residual disease or distant micrometastases, which are key factors influencing DFS [25, 26]. While MRI may not directly lower the incidence of distant micrometastases, its ability to refine surgical resection and potentially guide preoperative therapy may indirectly reduce systemic recurrence and enhance DFS.
The colon is characterized by high mobility, particularly in the sigmoid and transverse colon segments. This mobility arises from peritoneal attachments and intraperitoneal positioning, leading to variable tumor location and potential motion artifacts on MRI. No equivalent structure to the MRF exists in the colon. T staging relies on serosal involvement or adjacent organ invasion, which is less dependent on fixed anatomical boundaries. Given that, it has been a dilemma for MRI to evaluate the depth of colon cancer invasion. Nevertheless, Bompou et al. [27] highlighted that MRI outperformed CT in assessing T1/T2 versus T3/T4 staging, lymph node positivity, and extramural venous invasion. Additionally, the biological behavior of the tumor, the efficacy of preoperative therapy, and patient-specific factors may play a more significant role in determining DFS. It is also possible that the sample size or the study duration was insufficient to detect a difference in DFS, which may have emerged with a longer follow-up. Other variables that could impact the prognosis, like whether the tumor has spread to blood vessels outside the intestinal wall or whether lymph nodes have metastasized, can also be found with MRI. In the subgroup analysis, preoperative abdomen MRI was related to the OS of stage III patients but not stage II patients. This suggests that the utility of preoperative abdominal MRI may be stage-dependent. Thus, accurate staging with MRI could indirectly influence the appropriateness and timing of adjuvant chemotherapy, which may have an impact on prognosis. For early-stage cancer, preoperative abdominal MRI may prevent underestimation of tumor extent, reducing the risk of inadequate surgical resection and potential recurrence [28]. As we know, pathological tumor staging is the reference used to assess MRI accuracy. For example, early-stage tumors may appear more clearly on MRI, while advanced-stage tumors may cause errors in staging due to tumor extension or metastasis [29]. In addition, different pathological stages of colon cancer may also require various chemotherapies after the operation. Howerve, MRI's primary contribution is in surgical decision-making and achieving optimal resection margins, rather than directly determining chemotherapy regimens. For the detection of T3-4 colon cancer, MRI demonstrates superior performance compared to CT [30]. While MRI’s ability to assess tumor invasion depth (T stage) is limited by colonic mobility and anatomical variability, its utility in stage III colon cancer likely stems from enhanced lymph node staging accuracy. Unlike rectal cancer, where mesorectal fascia involvement dictates treatment decisions, colon cancer staging relies exclusively on lymph node status for stage III classification. However, the lack of improvement in DFS suggests that MRI may not directly reduce the risk of distant micrometastases. Its role in optimizing surgical resection and potentially guiding preoperative therapy could have an indirect effect on reducing systemic recurrence and improving DFS. Future studies should explore MRI’s role in identifying high-risk N1 subgroups who may benefit from escalated treatment.
Several limitations of our study must be acknowledged. First, our retrospective single-center study introduces potential biases inherent in observational research. Unmeasured confounders (e.g., surgeon experience, non-fixed segments such as the sigmoid and transverse colon, institutional MRI availability) may persist. Future studies should incorporate those points by using multicenter data. While this retrospective study suggests a prognostic role for preoperative MRI in colon cancer, prospective multicenter trials are needed to confirm these findings and address potential selection biases. Second, the economic burden associated with MRI may limit its widespread use, potentially skewing our patient cohort toward those with better access to healthcare resources. While MRI incurs higher costs, its ability to reduce unnecessary surgeries in stage III patients may offset expenses in high-volume centers. Third, the heterogeneity in the results produced by different MRI scans may be attributed to tumor characteristics, genetic mutations (RAS, BRAF, MSI, etc.) status, patient-specific factors, regional differences, and the quality of the MRI equipment [22, 31]. Fourth, since our study is retrospective, its characteristics determine that the results will be biased. This study did not collect detailed data on the specific diagnostic results from MRI that influenced treatment planning, and cannot clearly explain the mechanism behind this effect. Future research should address these limitations by conducting multi-center, prospective studies with larger populations to evaluate the cost-effectiveness of preoperative MRI, especially in resource-limited areas [22, 31]. Additionally, exploring the potential integration of preoperative MRI with emerging diagnostic technologies and biomarkers could enhance its predictive accuracy and clinical utility [21, 30]. Further research is also needed to elucidate the role of MRI in improving clinical outcomes in patients with early-stage colon cancer, where the benefits may be less pronounced. Furthermore, we should create a prospective study and undertake greater in-depth research.
Conclusion
In conclusion, preoperative abdominal MRI significantly improves the OS of patients with stage II-III colon cancer, especially in those with stage III disease. This highlights the potential value of preoperative abdominal MRI as a prognostic indicator. It provides detailed imaging information of cancer, helping doctors better understand its characteristics and develop more accurate treatment plans. However, the specific role of MRI in treatment planning, as well as the potential impact of different treatment approaches on outcomes, requires further investigation into the treatment decision-making processes. Further researches are warranted to fully elucidate the role of MRI in improving the clinical outcomes of patients with colon cancer.
Supplementary Information
Supplementary Material 1: Table S1. Univariate Cox regression analysis and multivariate analysis of factors associated with Overall Survival using backward stepwise selection (AIC criterion). Table S2. Variance Inflation Factor and Tolerance for variables in the final Cox regression model of Overall Survival derived from backward stepwise selection. Table S3. Univariate Cox regression analysis and multivariate analysis of factors associated with Disease Free Survival using backward stepwise selection (AIC criterion). Table S4. Variance Inflation Factor and Tolerance for variables in the final Cox regression model of Disease Free Survival derived from backward stepwise selection. Table S5. Multivariate analysis of the MRI group and Overall Survival or Disease-Free Survival Before Propensity Score Matching. Table S6. Variance Inflation Factor and Tolerance for variables in the final Cox regression model of Overall Survival derived from backward stepwise selection after Propensity Score Matching. Table S7. Variance Inflation Factor and Tolerance for variables in the final Cox regression model of Disease Free Survival derived from backward stepwise selection after Propensity Score Matching
Acknowledgements
We would like to acknowledge the hard and dedicated work of all the staff who implemented the intervention and evaluation components of the study.
Abbreviations
- MRI
Magnetic resonance imaging
- CI
Confidence interval
- CEA
Carcinoembryonic antigen
- HR
Hazard ratios
- IQR
Interquartile range
- OS
Overall Survival
- DFS
Progression-free survival
- TD
Tumor deposition
- PSM
Propensity score-matched
- T1WI
T1-weighted images
- T2WI
T2-weighted images
- VTT
Vascular tumor thrombus
- LR
Laparoscopic resection
- OR
Open resection
Authors’ contributions
Z.H.L. and H.J.P made substantial contribution to conception and design, data acquisition and drafted the article; M.M L, L.Z.L, H.J.P, and Z.H. L collected and analyzed data, M.S. B, X.T.T,and X.L edited and reviewed the article; K.Y.and Z.H.L supervised the work. All authors read and approved the final manuscript.
Funding
This study was supported by the National Natural Scientific Foundation of China [82360345]; National Natural Scientific Foundation of China [82001986]; Yunnan Provincial Science and Technology Department Science and Technology Programme Projects[202201AY070001159]; Dazhou Science and Technology Program [2022YYJC0014]; Chengdu Medical College Research Foundation Project [CYZYB24-17].
Data availability
No datasets were generated or analysed during the current study.
Declarations
Ethics approval and consent to participate
This study was approved by the Ethics Committee of Yunnan Cancer Hospital (KY2019141).. All procedures were in accordance with the ethical standards of the Clinical Research Committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.The need for informed consent was waived due to the retrospective nature of the study and use of deidentified data, as deemed unnecessary by our institution, with no ethics committee involved.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Maoshu Bai, Lizhu Liu, Mengmei Liu, and Xin Liu contributed equally to this work.
Contributor Information
Kun Yu, Email: kinfish@163.com.
Hongjiang Pu, Email: puhongjiang@qq.com.
Zhenhui Li, Email: lizhenhui621@qq.com.
References
- 1.Zhang J, Bao Y. Value of MSCT plus MRI in the detection of colon cancer. Evid Based Complement Alternat Med. 2022;2022:6507865. 10.1155/2022/6507865. [DOI] [PMC free article] [PubMed] [Google Scholar] [Retracted]
- 2.Siegel RL, Wagle NS, Cercek A, Smith RA, Jemal A. Colorectal cancer statistics, 2023. CA Cancer J Clin. 2023;73(3):233–54. 10.3322/caac.21772. [DOI] [PubMed] [Google Scholar]
- 3.Zhang W, Hou Z, Zhang L, et al. A log odds of positive lymph nodes-based predictive model effectively forecasts prognosis and guides postoperative adjuvant chemotherapy duration in stage III colon cancer: a multi-center retrospective cohort study. BMC Cancer. 2024;24(1): 1088. 10.1186/s12885-024-12875-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Park JW, Jung KH, Lee JH, Moon SH, Cho YS, Lee KH. 89Zr anti-CD44 immuno-PET monitors CD44 expression on splenic myeloid cells and HT29 colon cancer cells. Sci Rep. 2021;11(1):3876. 10.1038/s41598-021-83496-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Rajedadram A, Pin KY, Ling SK, Yan SW, Looi ML. Hydroxychavicol, a polyphenol from Piper betle leaf extract, induces cell cycle arrest and apoptosis in TP53-resistant HT-29 colon cancer cells. J Zhejiang Univ Sci B. 2021;22(2):112–22. 10.1631/jzus.B2000446. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Li S, Dai Y, Chen J, Yan F, Yang Y. MRI-based habitat imaging in cancer treatment: current technology, applications, and challenges. Cancer Imaging. 2024;24(1): 107. 10.1186/s40644-024-00758-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Qiu H, Wang M, Wang S, Li X, Wang D, Qin Y, Xu Y, Yin X, Hacker M, Han S, Li X. Integrating MRI-based radiomics and clinicopathological features for preoperative prognostication of early-stage cervical adenocarcinoma patients: in comparison to deep learning approach. Cancer Imaging. 2024;24(1):101. 10.1186/s40644-024-00747-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Kim TH, Woo S, Han S, Suh CH, Vargas HA. The diagnostic performance of MRI for detection of extramural venous invasion in colorectal cancer: a systematic review and meta-analysis of the literature. AJR Am J Roentgenol. 2019;213(3):575–85. 10.2214/AJR.19.21112. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Beets-Tan RGH, Lambregts DMJ, Maas M, Bipat S, Barbaro B, Curvo-Semedo L, Fenlon HM, Gollub MJ, Gourtsoyianni S, Halligan S, et al. Magnetic resonance imaging for clinical management of rectal cancer: Updated recommendations from the 2016 European Society of Gastrointestinal and Abdominal Radiology (ESGAR) consensus meeting. Eur Radiol. 2017;28:1465–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Jhaveri KS, Hosseini-Nik H. MRI of rectal cancer: an overview and update on recent advances. AJR Am J Roentgenol. 2015;205(1):W42-55. 10.2214/AJR.14.14201. [DOI] [PubMed] [Google Scholar]
- 11.Bipat S, Glas AS, Slors FJ, Zwinderman AH, Bossuyt PM, Stoker J. Rectal cancer: local staging and assessment of lymph node involvement with endoluminal US, CT, and MR imaging–a meta-analysis. Radiology. 2004;232(3):773–83. 10.1148/radiol.2323031368. [DOI] [PubMed] [Google Scholar]
- 12.Beets-Tan RG, Lambregts DM, Maas M, Bipat S, Barbaro B, Caseiro-Alves F, Curvo-Semedo L, Fenlon HM, Gollub MJ, Gourtsoyianni S, Halligan S, Hoeffel C, Kim SH, Laghi A, Maier A, Rafaelsen SR, Stoker J, Taylor SA, Torkzad MR, Blomqvist L. Magnetic resonance imaging for the clinical management of rectal cancer patients: recommendations from the 2012 European Society of Gastrointestinal and Abdominal Radiology (ESGAR) consensus meeting. Eur Radiol. 2013(9). 10.1007/s00330-013-2864-4. [DOI] [PubMed]
- 13.Taylor FG, Quirke P, Heald RJ, Moran BJ, Blomqvist L, Swift IR, Sebag-Montefiore D, Tekkis P, Brown G, Magnetic Resonance Imaging in Rectal Cancer European Equivalence Study Study Group. Preoperative magnetic resonance imaging assessment of circumferential resection margin predicts disease-free survival and local recurrence: 5-year follow-up results of the MERCURY study. J Clin Oncol. 2014;32(1):34–43. 10.1200/JCO.2012.45.3258. [DOI] [PubMed] [Google Scholar]
- 14.Xie H, Zhou X, Zhuo Z, Che S, Xie L, Fu W. Effectiveness of MRI for the assessment of mesorectal fascia involvement in patients with rectal cancer: a systematic review and meta-analysis. Dig Surg. 2014;31(2):123–34. 10.1159/000363075. Epub 2014 Jun 17. [DOI] [PubMed] [Google Scholar]
- 15.Beets-Tan RG, Lambregts DM, Maas M, et al. Magnetic resonance imaging for the clinical management of rectal cancer patients: recommendations from the 2012 European Society of Gastrointestinal and Abdominal Radiology (ESGAR) consensus meeting. Eur Radiol. 2013;23:2522–31. [DOI] [PubMed] [Google Scholar]
- 16.Nerad E, Lambregts DM, Kersten EL, Maas M, Bakers FC, van den Bosch HC, Grabsch HI, Beets-Tan RG, Lahaye MJ. MRI for local staging of colon cancer: can MRI become the optimal staging modality for patients with colon cancer? Dis Colon Rectum. 2017;60(4):385–92. 10.1097/DCR.0000000000000794. [DOI] [PubMed] [Google Scholar]
- 17.Rollvén E, Holm T, Glimelius B, Lörinc E, Blomqvist L. Potentials of high resolution magnetic resonance imaging versus computed tomography for preoperative local staging of colon cancer. Acta Radiol. 2013;54(7):722–30. 10.1177/0284185113484018. [DOI] [PubMed] [Google Scholar]
- 18.Argiles G, Tabernero J, Labianca R, Hochhauser D, Salazar R, Iveson T, Laurent-Puig P, Quirke P, Yoshino T, Taieb J, et al. Localised colon cancer: esmo clinical practice guidelines for diagnosis, treatment and follow-up. Ann Oncol. 2020;31(10):1291–305. 10.1016/j.annonc.2020.06.022. [DOI] [PubMed] [Google Scholar]
- 19.Marcoulides KM, Raykov T. Evaluation of variance inflation factors in regression models using latent variable modeling methods. Educ Psychol Meas. 2019;79(5):874–82. 10.1177/0013164418817803. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Hunter C, Blake H, Jeyadevan N, Abulafi M, Swift I, Toomey P, Brown G. Local staging and assessment of colon cancer with 1.5-T magnetic resonance imaging. Br J Radiol. 2016;89(1064). 10.1259/bjr.20160257. Epub 2016 May 26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Rafaelsen SR, Dam C, Vagn-Hansen C, Møller J, Rahr HB, Sjöström M, Lindebjerg J, Hansen TF, Pedersen MRV. CT and 3 tesla MRI in the TN staging of colon cancer: a prospective, blind study. Curr Oncol. 2022;29(2):1069–79. 10.3390/curroncol29020091. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Benson AB, Venook AP, Al-Hawary MM, Arain MA, Chen YJ, Ciombor KK, Cohen S, Cooper HS, Deming D, Garrido-Laguna I, et al. NCCN guidelines insights: rectal cancer, version 6.2020. J Natl Compr Canc Netw. 2020;18:806–15. 10.6004/jnccn.2020.0032. [DOI] [PubMed] [Google Scholar]
- 23.Chen L, Hu H, Yuan Y, Weng S. CSCO guidelines for colorectal cancer version 2024: updates and discussions. Chin J Cancer Res. 2024;36:233–9. 10.21147/j.issn.1000-9604.2024.03.01. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Dam C, Lindebjerg J, Jakobsen A, Jensen LH, Rahr H, Rafaelsen SR. Local staging of sigmoid colon cancer using MRI. Acta Radiol Open. 2017;6(7):2058460117720957. 10.1177/2058460117720957. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Westberg K, Othman B, Suzuki C, Blomqvist L, Martling A, Iversen H. Magnetic resonance imaging as a predictor of surgical outcome in patients with local pelvic recurrence of colorectal cancer. Eur J Surg Oncol. 2021;47(8):2119–24. 10.1016/j.ejso.2021.04.021. [DOI] [PubMed] [Google Scholar]
- 26.Lord AC, Corr A, Chandramohan A, Hodges N, Pring E, Airo-Farulla C, Moran B, Jenkins JT, Di Fabio F, Brown G. Assessment of the 2020 NICE criteria for preoperative radiotherapy in patients with rectal cancer treated by surgery alone in comparison with proven MRI prognostic factors: a retrospective cohort study. Lancet Oncol. 2022;23(6):793–801. 10.1016/S1470-2045(22)00214-5. [DOI] [PubMed] [Google Scholar]
- 27.Bompou E, Vassiou A, Baloyiannis I, Perivoliotis K, Fezoulidis I, Tzovaras G. Comparative evaluation of CT and MRI in the preoperative staging of colon cancer. Sci Rep. 2024;14(1). 10.1038/s41598-024-68147-7. PMID: 39060367. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Park YJ, Park KJ, Park JG, et al. Prognostic factors in 2230 Korean colorectal cancer patients: analysis of consecutively operated cases. World J Surg. 1999. 10.1007/pl00012376. [DOI] [PubMed] [Google Scholar]
- 29.Mazaki J, Tsukamoto S, Miyake M, Moritani K, Sakamoto R, Shida D, Kanemitsu Y. Circumferential resection margin status as a predictive factor for recurrence in preoperative MRI for advanced lower rectal cancer without preoperative therapy. Dis Colon Rectum. 2021;64(1):71–80. 10.1097/DCR.0000000000001769. [DOI] [PubMed] [Google Scholar]
- 30.Liu LH, Lv H, Wang ZC, Rao SX, Zeng MS. Performance comparison between MRI and CT for local staging of sigmoid and descending colon cancer. Eur J Radiol. 2019;121: 108741. 10.1016/j.ejrad.2019.108741. [DOI] [PubMed] [Google Scholar]
- 31.Plodeck V, Platzek I, Streitzig J, Nebelung H, Blum S, Kühn JP, Hoffmann RT, Laniado M, Michler E, Hoberück S, Zöphel K, Kotzerke J, Fritzmann J, Weitz J, Radosa CG. Diagnostic performance of 18F-fluorodeoxyglucose-PET/MRI versus MRI alone in the diagnosis of pelvic recurrence of rectal cancer. Abdom Radiol (NY). 2021;46(11):5086–94. 10.1007/s00261-021-03224-3. Epub 2021 Aug 17. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Material 1: Table S1. Univariate Cox regression analysis and multivariate analysis of factors associated with Overall Survival using backward stepwise selection (AIC criterion). Table S2. Variance Inflation Factor and Tolerance for variables in the final Cox regression model of Overall Survival derived from backward stepwise selection. Table S3. Univariate Cox regression analysis and multivariate analysis of factors associated with Disease Free Survival using backward stepwise selection (AIC criterion). Table S4. Variance Inflation Factor and Tolerance for variables in the final Cox regression model of Disease Free Survival derived from backward stepwise selection. Table S5. Multivariate analysis of the MRI group and Overall Survival or Disease-Free Survival Before Propensity Score Matching. Table S6. Variance Inflation Factor and Tolerance for variables in the final Cox regression model of Overall Survival derived from backward stepwise selection after Propensity Score Matching. Table S7. Variance Inflation Factor and Tolerance for variables in the final Cox regression model of Disease Free Survival derived from backward stepwise selection after Propensity Score Matching
Data Availability Statement
No datasets were generated or analysed during the current study.