Clinicopathological characteristics and prognosis of patients with mucinous adenocarcinoma originating from the left colon, right colon, or rectum: a nationwide retrospective study in China.

Introduction
Colorectal cancer (CRC) is the third most commonly diagnosed cancer worldwide (1). Mucinous adenocarcinoma (MAC) accounts for approximately 10–20% of all CRC cases (2). The World Health Organization defines MAC as a lesion in which more than 50% is composed of pools of extracellular mucin containing malignant epithelium (3). In terms of prognosis, our team’s previous studies demonstrated that a percentage of MAC component >70% was a significant risk factor for poor survival [hazard ratio (HR) =1.643; 95% confidence interval (CI): 1.025–2.635, P=0.03] (4). Furthermore, a meta-analysis conducted by our group, which included 56 studies with a total of 803,157 patients, revealed that MAC is an adverse prognostic factor for overall survival (OS) in CRC patients (5). Recent studies have shown that MACs have more distinct genetic backgrounds and clinicopathological and prognostic features than those of non-mucinous CRCs (6), and an increasing number of studies have suggested that MACs should be treated as single tumours.
The prognostic significance of primary tumour location in CRC has garnered considerable interest. Different tumour locations are associated with distinct genetic backgrounds, clinicopathological features, and prognoses. CRCs originating from the right-sided colon (RS) tend to have worse OS and disease-free survival (DFS) than those originating from the left-sided colon (LS) (7). Nonetheless, the impact of tumour location on the clinicopathological features and prognosis of MAC remains unclear. In addition, Chinese patients often present with more distal cancers and at younger ages compared to Western populations (8). And Asians exhibit a higher KRAS mutation rate than Caucasians, while Caucasians show a significantly higher BRAF mutation rate (9).
This longitudinal international cohort study aimed to investigate the influence of tumour location on clinicopathological features, recurrence, and survival in patients with MAC. The enrolled MAC patients were divided into different groups on the basis of tumour site, and clinicopathological and prognostic characteristics were compared among the different groups. We present this article in accordance with the STROBE reporting checklist (available at https://jgo.amegroups.com/article/view/10.21037/jgo-2025-627/rc).

Methods

Study population
The clinicopathological data of CRC patients who underwent curative surgery and were pathologically diagnosed with MAC across 23 hospitals in China from 2016 to 2021 were collected. Data from 2016–2021 MAC patients in the Surveillance, Epidemiology, and End Results (SEER) database were also collected. The URL is: https://seer.cancer.gov/data-software/. In this study, all MAC patients were included, whereas those with multiple primary tumours, multiple CRCs, stage IV MAC, and insufficient clinical and/or pathological information were excluded (Figure 1). In the Chinese cohort, data, including sex, age, tumour location [RS, LS, and rectum (RC)], tumour size, perineural invasion (PNI), lymphovascular invasion (LVI), T stage, N stage, and tumour-node-metastasis (TNM) stage, were collected from clinical records. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Institutional Review Board of The First Affiliated Hospital of Shandong First Medical University & Shandong Provincial Qianfoshan Hospital [approval No. YXLL-KY-2024(116)]. Given that this study is a retrospective study, it is objectively impossible to obtain the informed consent of the participants themselves. For the above reasons, informed consent was waived for all patients in this study. The other hospitals were also informed and agreed to the study.

Follow-up method
Postoperative follow-up was conducted every 6 months for 2 years, followed by annual assessments for 3 to 5 years at the outpatient clinic. Follow-up included physical examinations, tumour marker assessments, computed tomography (CT) scans, and colonoscopies to detect postoperative recurrence. CT scans were typically performed every 6 months for 5 years, whereas colonoscopies were performed in the first 3 years.

Data selection
Patients were categorized into three groups based on tumour location: the RS group, the LS group, and the RC group. The RS includes the caecum, appendix, ascending colon, hepatic flexure and the right side of the transverse colon, whereas the LS encompasses the left side of the transverse colon, splenic flexure, descending colon and sigmoid. Data on age, sex, tumour size, LVI, and PNI at presentation were recorded for each patient according to the Union for International Cancer Control (8th edition) TNM classification. Follow-up data were obtained through patient interactions in the hospital or via telephone communication.

Study endpoint
The endpoint of this study was survival time. Deaths from any cause, recurrence or metastasis were considered events. Patients still alive at the last follow-up, with or without disease, were censored. Survival time was measured from the date of surgery to the date of death or the last known follow-up. OS was defined as the period from surgery to death from any cause, and DFS was defined as the period from surgery to recurrence or any death from any cause.

Statistical analysis
Chi-squared analysis, Fisher’s exact test, or Kruskal-Wallis analysis of variance (ANOVA) was used to assess differences in categorical variables where appropriate. Kaplan-Meier curves were generated to assess survival, which was compared via log-rank tests. Univariable and multivariable survival analyses with Cox regression models were conducted to identify prognostic factors. HR is reported with 95% CI. The adjusted variables included sex, age, tumour location, LVI, PNI, tumour size, and TNM stage. A two-sided P value ≤0.05 was considered statistically significant. All of the statistical analyses were conducted via SPSS software (version 25.0 for Windows, IBM SPSS Statistics, IBM Corporation, Armonk, NY, USA).

Results

Study cohort characteristics and clinicopathological features
A total of 2,646 patients in the SEER cohort and 2,439 patients in the Chinese cohort met the inclusion criteria. The SEER cohort comprised 1,636 (61.8%) patients whose disease originated from the RS, 697 (26.4%) from the LS, and 313 (11.8%) from the RC. In the Chinese cohort, 993 (40.7%) patients had their disease originating from the RS, 526 (21.6%) from the LS, and 920 (37.7%) from the RC (Table 1). There were significant differences in sex, age, tumour size, and pT category among the three groups in the SEER cohort. Compared with those in the LS and RC groups, patients in the RS group were more likely to be female (P<0.001 and P=0.004), be older (P<0.001 and P<0.001), and belong to a higher pathological T category (P<0.001 and P=0.005), (Table 2). Additionally, patients in the RC group were more likely to exhibit a higher pathological N category (P<0.001 and P=0.02) and TNM category (P<0.001 and P=0.005) than those in the RS and LS group (Table 2). In the Chinese cohort, similar trends were noted, with RS patients also being more likely to be female (P<0.001 and P=0.03), have larger tumours (P<0.001 and P<0.001), and present with a lower pathological N category (P<0.001 and P<0.001) and TNM category (P<0.001 and P<0.001) than LS and RC patients (Table 3). Furthermore, patients in the LS group had a greater pathological T category (P<0.001) than did those in the RC group (Table 3). These findings indicate that MAC tumours from different locations have different clinicopathological characteristics.

Prognosis of MAC patients based on tumour location
To further elucidate the differences among different tumour locations, the prognoses of the included patients were depicted. The 3-year OS rates for MAC patients were 88.3% in the Chinese cohort and 76.5% in the SEER cohort. For patients in the Chinese cohort, the 3-year OS rate for RC patients was 85.3%, which was lower than that for RS (90.7%) and LS (88.7%) patients, but the difference was not statistically significant (P=0.06) (Figure 2A). The trends in the DFS rates were similar and were 87.5%, 84.2%, and 81.5% for the RS, LS and RC groups, respectively (P=0.06) (Figure 2B).
When colon cancers were evaluated as a unified category (including RS and LS), the 3-year OS rate for RCs in the Chinese cohort was 85.3%, whereas it was 90.0% for colon cancer patients (P=0.02) (Figure 2C). For DFS, significant differences were also noted in the Chinese cohort, with the 3-year DFS rate at 86.4% for colon cancer patients and 81.5% of RC patients (P=0.03) (Figure 2D). In the SEER cohort, the 3-year OS rates for the RC, RS and LS groups were 69.0%, 77.8%, and 76.8%, respectively (P<0.001) (Figure 3A). When evaluating colon cancers as a unified category (including RS and LS), the OS rates for RCs and colon cancers in the SEER cohort were 69.0% and 77.5% (P<0.001) (Figure 3B). In total, significant prognostic differences existed between the Chinese cohort and the SEER cohort, and RC patients exhibited relatively poorer survival than did RS patients and LS patients.

Factors affecting OS and DFS in MAC patients
Univariate and multivariate analyses were conducted to assess the factors associated with the prognosis of MAC patients. In the Chinese cohort, age ≥60 years (HR =1.835, 95% CI: 1.252–2.69, P=0.002), LVI (HR =2.492, 95% CI: 1.699–3.653, P<0.001), PNI (HR =2.003, 95% CI: 1.349–2.973, P=0.001), and advanced TNM stage (HR =4.436, 95% CI: 2.91–6.762, P<0.001) were identified as risk factors for poor OS (Table 4). Multivariate analysis further revealed that age ≥60 years (HR =1.838, 95% CI: 1.253–2.696, P=0.002) and advanced TNM stage (HR =4.009, 95% CI: 2.603–6.175, P<0.001) were independent risk factors for poor OS (Table 4). In terms of DFS, in the Chinese cohort, a tumour size ≥5 cm (HR =1.327, 95% CI: 1.078–1.508, P=0.01) and advanced TNM stage (HR =1.595, 95% CI: 1.152–2.209, P=0.005) were significant risk factors (Table 5). In the SEER cohort, age ≥60 years (HR =1.891, 95% CI: 1.561–2.292, P<0.001), tumour size ≥5 cm (HR =1.561, 95% CI: 1.306–1.876, P<0.001), and advanced TNM stage (HR =2.567, 95% CI: 2.173–3.031, P<0.001) were identified as independent risk factors for poor OS (Table 4). And compared with the RC, the RS was identified as an independent predictor of improved 3-year OS (HR 0.688, 95% CI 0.548–0.865, P=0.001).

Factors affecting OS in patients with different tumour locations
To verify the factors affecting the prognosis of patients with different tumour locations, further statistical analysis was conducted. In the Chinese cohort, multivariable analysis revealed that age ≥60 years (HR =2.357, 95% CI: 1.155–4.808, P=0.02) and TNM stage (HR =5.158, 95% CI: 2.631–10.11, P<0.001) were independent risk factors affecting OS in patients in the RS group. Similarly, in the SEER cohort, multivariate analysis revealed that age ≥60 years (HR =1.883, 95% CI: 1.416–2.504, P<0.001), tumour size ≥5 cm (HR =1.353, 95% CI: 1.075–1.702, P=0.01), and TNM stage (HR =2.61, 95% CI: 2.108–3.232, P<0.001) were independent risk factors for poor OS (Table 6). Among patients with LS in the Chinese cohort, advanced TNM stage (HR =12.13, 95% CI: 2.798–52.6; P=0.001) was identified as an independent risk factor for poor OS. In the SEER cohort, advanced age ≥60 years (HR =1.839, 95% CI: 1.321–2.561, P<0.001), tumour size ≥5 cm (HR =1.73, 95% CI: 1.226–2.441, P=0.002), and TNM stage (HR =3.351, 95% CI: 2.383–4.711, P<0.001) were identified as risk factors for poor OS (Table 7). For patients with RC in the Chinese cohort, univariable analysis indicated that age ≥60 years (HR =1.72, 95% CI: 1.004–2.947, P=0.048), LVI (HR =2.813, 95% CI: 1.609–4.919, P<0.001), PNI (HR =3.661, 95% CI: 2.124–6.31, P<0.001), and advanced TNM stage (HR =2.501, 95% CI: 1.368–4.572, P=0.003) were associated with poor cause-specific survival. Multivariate analysis confirmed that age ≥60 years (HR =1.783, 95% CI: 1.041–3.056, P=0.03), PNI (HR =2.79, 95% CI: 1.456–5.306, P=0.002), and advanced TNM stage (HR =2.12, 95% CI: 1.142–3.935, P=0.02) were independent prognostic factors. In the SEER cohort, poor cause-specific survival was associated with age ≥60 years (HR =1.879, 95% CI: 1.24–2.849, P=0.003) and tumour size ≥5 cm (HR =2.799, 95% CI: 1.692–4.632, P<0.001) in the multivariate analysis (Table 8). For MAC patients with disease originating from the colon in the Chinese cohort, multivariate analysis confirmed that age ≥60 years (HR =1.986, 95% CI: 1.143–3.452, P=0.02) and advanced TNM stage (HR =5.972, 95% CI: 3.261–10.936, P<0.001) were independent prognostic factors. In the SEER cohort, poor cause-specific survival was associated with age ≥60 years (HR =1.801, 95% CI: 1.475–2.225, P<0.001), tumour size ≥5 cm (HR =1.451, 95% CI: 1.198–1.756, P<0.001) and advanced TNM stage (HR =2.804, 95% CI: 2.341–3.358, P<0.001) according to multivariable analysis (Table 9). In total, the risk factors for poor prognosis differ across different subgroups of patients, and age, tumour size and TNM stage should be the focus of follow-up regimens.

Discussion
This study first investigated the impact of tumour location on MAC prognosis. In the SEER cohort, RC was associated with worse OS than RS was, while RS and LS had comparable OS. A similar conclusion was reached in the Chinese cohort. Compared with RS, RC was associated with worse DFS outcomes, whereas RC and LS resulted in similar DFS rates. Thus, tumour location has distinct prognostic implications for recurrence after curative resection and overall mortality in MAC patients, and primary tumour location may function as a prognostic biomarker. Our findings emphasize the clinical implications of tumour location for MAC prognosis.
Several factors may contribute to the differences among patients with distinct tumour locations. The embryological origins may be the primary reason. The RS develops from the embryonic midgut, whereas the LS and RC derive from the hindgut. This embryological divergence is correlated with variations in genetic carcinogenic pathways (10). Since the vascular supply also differs (with the RS being supplied by the superior mesenteric artery and the LS and rectum being supplied by the inferior mesenteric artery), the patterns of lymph node metastasis differ by tumour location. Specifically, lateral pelvic lymph node metastases are pertinent only to rectal tumours. These varying patterns may contribute to diverse clinical, pathological, and biological characteristics. Additionally, we noted differential recurrence sites between these tumour locations. Furthermore, unique clinicopathological characteristics are associated with specific tumour locations. In both the Chinese and SEER cohorts, the pathological T category was more advanced in patients with RS than in those with LS or RC. This discrepancy may stem from clinical challenges in diagnosing RS, which is often asymptomatic due to its more distal location from the anus and the passage of softer bowel content through lesions (11). Conversely, RC patients frequently exhibit a higher pathological N category, likely driven by a greater incidence of metastases to lateral lymph nodes (12). Given that RC patients with lateral lymph node metastases are classified as locally advanced, it is unsurprising that a significant number of patients with RC presented with elevated pathological N categories (13). Multivariate analyses in the SEER cohort confirmed that RC is associated with worse OS than RS is. Although neoadjuvant chemoradiotherapy (nCRT) is the standard of care for locally advanced RC, MAC exhibit low responsiveness to radiation (14). This may be the reason why the OS of RC group is lower than that of colon cancer, and also poses challenges for the treatment of rectal MAC.
With respect to genetic characteristics based on tumour location, mutations in BRAF or KRAS codon 12 are prevalent in RS, whereas non-mutated BRAF/KRAS is found more frequently in LS (15). Furthermore, incidence reports suggest a gradual decrease in BRAF mutation and microsatellite deficient mismatch repair (dMMR) as the tumour location shifts from the RS towards the RC (16,17). Mutant KRAS tumours have been linked to poorer OS, whereas dMMR tumours often exhibit extended survival post-recurrence (18). Therefore, the disparities in prognosis associated with tumour location can be elucidated by these molecular differences pertaining to MAC based on its anatomical site.
Notably, multivariate analysis of the Chinese cohort did not reveal a significant relationship between tumour size and prognosis (P=0.86), whereas in the SEER cohort, tumour size significantly impacted prognosis (P<0.001) (Table 4). Patients with RS exhibited the highest OS and DFS, whereas those with RC exhibited the lowest outcomes, which was consistent across both cohorts. Moreover, the Chinese cohort exhibited higher OS and DFS rates than did the SEER cohort for comparable tumour locations, potentially attributable to differences in access to care. Additionally, racial differences could also account for variations in OS and DFS outcomes (19,20).
There are several limitations in this study. First, prognostic differences based on tumour location might be partially attributed to genetic disparities, such as the high frequency of microsatellite instability and BRAF mutations in the RS (16); however, this does not account for the poorer prognosis observed among RC patients in our study. Consequently, further investigations regarding the relationship between mucus secretion and tumour location are warranted. Second, while the occurrence of tumour perforation at the time of surgery was minimal, those cases were not recorded in our database. This omission is significant, as tumour perforation can influence survival outcomes because its association with lower DFS and higher operative mortality is correlated with proximal obstruction (21). Finally, despite adjustments for known confounders, the potential for residual unknown confounding factors remains.

Conclusions
This nationwide multicentre retrospective study demonstrated that three tumour locations (RS, LS, and RC) associated with MAC exhibit divergent prognostic consequences regarding recurrence following curative resection and overall mortality. Compared with LS and RS, RC is associated with a worse prognosis. Future randomized trials exploring MAC should consider tumour location as a stratification parameter.

Supplementary
The article’s supplementary files as