Introduction

Colorectal cancer (CRC) is one of the leading causes of cancer-related death worldwide [1]. Different histological subtypes of CRC may exhibit quite different clinical characteristics and survival outcomes [2]. Most CRCs are classical adenocarcinomas (AC), and less frequent subtypes mainly include mucinous adenocarcinomas (MAC) and signet-ring cell carcinomas [3]. MAC, with extracellular mucin > 50% of the tumor, accounts for 10–15% of CRCs [4].

The results of previous studies investigating the clinicopathological and prognostic features are controversial [5]. Molecular and genetic analyses revealed significantly differences between MAC and AC, indicating a prominently different oncogenic development [6, 7]. MAC is more advanced at diagnosis, located mainly at the right side and has a poor prognosis compared with AC [8, 9]. In addition, MAC has a less firm consistency, which may cause symptoms to arise only when the tumor reaches an advanced stage [10]. Adjuvant chemotherapy should be routinely recommended for patients with MAC stage II, and special attention should be paid during their follow-up for chemoradiotherapy resistance [11]. Therefore, the prognosis and clinicopathological characteristics of MAC deserve further investigations. Previous studies quantitatively assessing the differences of clinicopathological features and prognosis in MAC and AC were systematically reviewed in this meta-analysis.

Methods

The reporting of this systematic review follows the Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA) statement [12].

Search strategy and study selection

The literature was searched using PubMed, China National Knowledge Infrastructure (CNKI), WANFANG data, Embase, and Web of Science in December, 2022. The search free-text terms used were (((((mucinous) OR (mucus OR colloid)) AND (cancer OR tumor OR neoplasm OR adenocarcinoma OR malign))) AND (rectal OR colon OR colorectal OR rectum)) AND (prognosis). The control terms used in PubMed were (colorectal neoplasms) OR (colonic neoplasms) OR (rectal neoplasms) and (adenocarcinoma, mucinous). Articles included should meet the following criteria: (1) sufficient information for calculating HR and 95% CI; (2) prognosis features presented in adenocarcinoma and mucinous adenocarcinoma patients; Articles were excluded if they were:(1)narrative reviews, case reports, congress abstracts; (2) studies without comparison of survival data; (3) neither English nor Chinese articles. In total, 56 studies were included in the final meta-analysis [5, 8,9,10,11, 13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63]. A flow diagram of the study selection and exclusion of duplicate articles process is presented in Fig. 1.

Fig. 1
figure 1

A flow chart of this study

Full size image

Data collection

Two authors reviewed all the involved articles and extracted the data independently. Disagreements between the 2 authors were resolved with consensus. We independently extracted the following data from each article: basic research information (name of the first author, publication time, region, number of patients included, survival analysis method); clinicopathological features of patients (gender, age, tumor location, CEA, tumor size, differentiation, metastasis location, lymphatic invasion, venous invasion, peritoneal metastasis, lymph node metastasis, distant metastasis, the mutation of KRAS and BRAF,MSI status, TNM stage and Duke’s stage); prognosis (the HRs of the ratio of MAC to AC for overall survival (OS), disease-free survival (DFS), as well as their 95% CIs and P values). If available, HRs and 95% CIs were preferentially obtained from multivariate results. Otherwise, they were extracted from univariable outcomes or calculated using Engauge Digitizer version 4.1 (free software down-loaded from http://sourceforge.net) to read the Kaplan–Meier survival curves to get the HRs and 95% CIs [64,65,66].

Quality assessment

We independently assessed the quality of all eligible studies using the Newcastle–Ottawa quality assessment scale (NOS). The NOS criteria evaluates the quality of articles from the following three aspects: (1) subject selection, (2) comparability of subject, (3) clinical outcome. We scored the included studies based on these three aspects. Based on the NOS standard score of 7–9, we defined high-quality quality research, 4–6 points for medium quality research, less than 4 points for low quality research.

Statistical analysis

We used Stata statistical software version 15.0 (Stata Corporation, College Station, Texas, USA) and Review Manager version 5 (Revman; The Nordic Cochrane Centre, Copenhagen, Denmark) to perform comprehensive meta-analysis. We used Odds ratios (ORs) and Hazard ratio with 95% confidence intervals (CIs) to evaluate the clinicopathological and prognostic differences between AC and MAC, respectively. The statistical significance of the pooled OR and HR was evaluated with the Z test and P values, and P < 0.05 was considered statistically significant. Subgroup analysis was conducted to analyze the sources of heterogeneity, which was assessed by means of Q and I2 statistic. Publication bias was assessed by the Begg’s rank correlation method and Egger’s weighted regression method and P value less than 0.05 was considered statistically significant. In addition, we used sensitivity analysis to assess the influence of a single study on pooled HR.

Results

Study selection and description of the included studies

A total of 8176 articles were obtained through database retrieval. After removing duplicated studies and irrelevant studies (included not related to CRC and not related to MAC) through screening title and abstract, 240 studies were remained. Then, we read the full texts of the articles carefully, articles without prognosis, incomplete and language restrictions were eliminated and 56 studies met our inclusion criteria were finally included in this meta-analysis. The main characteristics of the included studies are presented in Table 1. These studies were published between 1976 and 2022. A total of 803,157 patients (sample sizes ranged from 70 to 164,628) were included in this review. The relationship of OS between AC and MAC was described in 54 studies [5, 8,9,10,11, 13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43, 45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60, 62, 63] and DFS was compared in 13 studies [9, 11, 44, 46,47,48, 52, 53, 55, 59, 61,62,63]. Some included studies explored the relationship of OS and DFS between MAC and AC in colon and rectum and different TNM stages. The features of the included studies are listed in Table 1. All of the eligible studies scored more than five by NOS, revealing a high methodological quality across all studies.

Table 1 Characteristics of studies included in the present meta-analysis
Full size table

Differences of clinicopathological features between MAC and AC

Differences of clinicopathological features between MAC and AC were presented in all included studies. MAC was significantly correlated with multiple clinicopathological features, including female(OR = 0.87, 95% CI: 0.86–0.88, P < 0.01), tumor located in colon(OR = 1.79, 95% CI: 1.76–1.83, P < 0.01), tumor size ≥ 5 cm(OR = 2.26, 95% CI: 2.17–2.36, P < 0.01), CEA ≥ 5 ng/mL (OR = 1.62, 95% CI:1.38–1.88,P < 0.01), advanced T,N,M stages(OR = 0.88, 95% CI: 0.87–0.90, P < 0.01), advanced Duke’s stage(OR = 0.57, 95% CI: 0.46–0.72, P < 0.01), moderate and well differentiation(OR = 0.84, 95% CI: 0.82–0.85, P < 0.01), non-vascular invasion(OR = 0.87, 95% CI: 0.78–0.98, P < 0.01), non- lymphovascular invasion(OR = 0.84, 95% CI: 0.74–0.95, P < 0.01), BRAF mutation(OR = 2.80, 95% CI: 1.99–3.94, P < 0.01), non-KRAS mutation(OR = 0.68, 95% CI: 0.55–0.85, P < 0.01), MSI-H status(OR = 2.80, 95% CI: 1.99–3.94, P < 0.01), lymph node metastasis (OR = 2.59, 95% CI: 2.17–3.11, P < 0.01), non-liver metastasis (OR = 0.51, 95% CI: 0.41–0.63, P < 0.01), non-lung metastasis (OR = 0.62, 95% CI: 0.40–0.95, P = 0.03), except for lymphatic invasion (OR = 1.07, 95% CI: 0.99–1.15, P = 0.09), and perineural invasion(OR = 0.92, 95% CI: 0.79–1.06, P = 0.25) (Table 2).

Table 2 Meta-analysis of the differences of clinicopathological features between MAC and AC in CRC
Full size table

Prognostic value of MAC in CRCs

A total of 54 studies was enrolled to detect the prognostic value of MAC in OS. A random-effect model was used to calculate the pooled HR and 95% CI because excessive heterogeneity existed among studies (P < 0.01, I2 = 86%) (Fig. 2a). Overall, MAC predicted poor OS compared with AC (HR = 1.04, 95% CI: 1.03–1.04, P < 0.01), but the DFS was comparable between MAC and AC (HR = 1.01,95% CI: 0.88- 1.17, P = 0.85) (Fig. 2b). To detect potential heterogeneity, subgroup analyses were performed based on recruitment time, region, tumor location, TNM stage, sample size and data type used for calculating survival. As shown in Table 3, MAC was obviously correlated with OS in cohorts with different recruitment time (recruited before 2012 HR = 1.01, 95% CI: 1–1.03, P = 0.01; recruited after 2012 HR = 1.04, 95% CI: 1.04–1.05, P < 0.01), with tumor located in rectum (HR = 1.09, 95% CI: 1.07–1.11, P < 0.01), from different regions (from eastern Asia HR = 1.06, 95% CI: 1.05–1.06, P < 0.01; from other regions HR = 1.02, 95% CI: 1.02–1.03, P < 0.01), with different sample size (≥ 500 HR = 1.04, 95% CI: 1.03–1.04, P < 0.01; < 500 HR = 1.13, 95% CI: 1.08–1.19, P < 0.01), with different data types (analysed with multivariate method HR = 1.02, 95% CI: 1.02–1.03, P < 0.01; analysed with Univariate or K-M survival curves HR = 1.11, 95% CI: 1.1–1.13, P < 0.01), and with TNM stage II (HR = 1.39, 95% CI: 1.07–1.8, P < 0.01). However, the OS were not statistically different in MAC and AC patients with tumor located in colon (HR = 1.01, 95% CI: 1–1.02, P = 0.21), TNM stage III (HR = 1.02, 95% CI: 0.95–1.09, P = 0.64) and IV (HR = 1.13, 95% CI: 0.91–1.41, P = 0.27) (Table 3). As shown in Table 4, no significant difference was detected in DFS between MAC and AC.

Fig. 2
figure 2

Forest plots: Summary hazard ratios (HRs) and 95% confidence intervals (CIs) of colorectal cancer OS (a) and DFS (b) for Mucinous subtype

Full size image
Table 3 Subgroup analyses for overall survival
Full size table
Table 4 Subgroup analyses for disease-free survival
Full size table

Publication bias and sensitivity analysis

The funnel plots of publication bias of 54 studies for OS and 13 studies for DFS showed basic symmetry (Fig. 3A and B). Evaluation of publication bias using Begg’s and Egger’s tests also showed that no publication bias existed (P value of Begg’s test, 0.893 and 0.760 for OS and DFS, respectively; P value of Egger’s test, 0.065 and 0.373 for OS and DFS, respectively). Furthermore, to evaluate the results of meta-analysis, sensitivity analysis was conducted. No significant change was found in the results when any one study was excluded, confirming the robustness and reliability of meta-analysis results on both OS and DFS (Table 5).

Fig. 3
figure 3

Begg’s funnel plots of the natural logarithm of the hazard ratios (HRs) and the SE of the natural logarithm of the HRs for the included studies reported with OS (a) and DFS (b)

Full size image
Table 5 The influence of individual study on the pooled estimate for outcomes
Full size table

Discussion

This systemic review and meta-analysis explore the differences of clinicopathological features and prognosis between MAC and AC, it covers a wide range of time from 1976—2022, 56 articles and 803,157 patients. Previous studies have confirmed that MAC often presents with advanced stages [36, 41, 52, 67]. This meta-analysis also showed that MAC was significantly associated with advanced TNM stage and lymph node metastasis. Besides, rate of lymphovascular invasion was also higher in MAC than in AC. These results indicate the essential role of mucin in local development and metastasis in MAC.

However, lymphatic invasion (OR = 1.07, 95%CI: 0.99–1.15, P = 0.09) and perineural invasion (OR = 0.92, 95%CI: 0.79–1.06, P = 0.25) were not correlated with MAC as revealed in this study. Compared with AC, MAC was reported to be associated with different molecular features, such as MSI and mutations in BRAF, KRAS [68]. This meta-analysis suggests that MAC has higher KRAS mutation and lower BRAF mutation tendency. MSI-H is a well-established prognostic biomarker for better survival [49]. Consistent with previous studies [9, 28, 29, 33, 69], MAC patients exhibited a higher rate of MSI-H, which might provide clues for using immunotherapy in MAC.

The prognosis of MAC patients is a pivotal topic but has always been in controverse, this meta-analysis suggests that MAC has a poor OS but comparable DFS compared with AC. Previous studies have suggested that the OS of MAC changes with tumor location and stage [9, 29, 41]. Subgroup analysis in this study found that MAC was correlated with OS in cohorts with TNM stage II and tumor located in rectum, but not with III or IV stage and tumor located in colon. So the prognosis of patients with MAC also varies and should be evaluated according to the localization of the tumor and TNM staging. We included 13 articles to explore the relationship of DFS between AC and MAC, however, no difference was found.

This meta-analysis has several limitations. First, heterogeneity in the meta-analysis is significant, although sensitivity analysis revealed the prognostic value of MAC by removing each study individually. Second, the number of patients recruited in some included studies was relatively small. Third, the included studies comprise a long time period and cover a diverse geographical origin, differences in treatment may affect survival analysis for MAC. Fortunately, publication bias was not detected for 54 studies for OS and 13 studies for DFS, and sensitivity analysis revealed that no significant change was found in the results when any 1 study was excluded. Finally, although we use Begg’s and Egger’s tests to assess the publication bias in our meta-analysis, some missing data were inevitable.

Conclusions

MAC has greatly different clinicopathological features compared with AC. And MAC indicated a poor OS relative to AC, though the DFS was comparable. This evidence suggests that MAC should be regarded as an unique cancer to treat, and further studies are needed to better define the mechanism of MAC initiation and development.