Clinical significance of mucinous component in colorectal adenocarcinoma: a propensity score-matched study

Background This study aims to investigate the clinical significance and prognostic value of mucinous component (MC) in colorectal adenocarcinoma (AC). Methods Patients with colorectal AC and AC with MC (ACMC) (1–100%) underwent surgical resection between January 2007 and February 2018 were retrospectively reviewed. Propensity score matching (PSM) was performed according to a 1:1 ratio. Receiver-operating characteristic (ROC) curve was used to identify the optimal cut-off value of MC ratio for prognostic prediction. The clinicopathological features and 3-year overall survival (OS) of AC patients, mucinous adenocarcinoma (MAC) (MC > 50%) patients, and ACMC (1–50%) patients were compared before and after matching. Multivariable analysis was used for analyzing independent risk factors related to prognosis. Results A total of 532 patients were enrolled in this study. Patients with AC, MAC, and ACMC (1–50%) exhibited different clinicopathological features. However, their 3-year OS rates were similar (82.00% vs. 74.11% vs. 81.48%, P = 0.38). After matching, ROC curve determined 70% as the optimal cut-off value. And patients with ACMC > 70% had a much poorer 3-year OS compared with ACMC (1–70%) patients and AC patients (47.37% vs. 86.15% vs. 79.76%, P < 0.001). In addition, ACMC > 70% was revealed as a risk factor for poor survival in univariate analysis (HR = 1.643, 95%CI = 1.025–2.635, P = 0.039), though not an independent risk factor in multivariable analysis (HR = 1.550, 95%CI = 0.958–2.507, P = 0.074). Conclusions MAC is usually diagnosed at an advanced stage. MAC has a similar survival with AC and ACMC (1–50%) patients before and after matching. Patients with ACMC > 70% exhibited a much poorer OS, and should be given more clinical attention.


Introduction
Colorectal cancer (CRC) ranks the world's fourth most deadly cancer with almost 900,000 deaths annually [1]. CRC has several histological types, and mucinous adenocarcinoma (MAC) comprises about 1.6-25.4% of all CRC cases [2]. According to the World Health Organization (WHO) criteria, MAC is defined as "> 50% of the lesion is composed of pools of extracellular mucin that contain malignant epithelium" [3]. However, 50% is more a cutoff Yes 106 13 11 Tumor location < 0.001 value defining MAC pathologically than a clinical index indicating clinical significance and prognosis. An optimal cut-off value of mucinous proportion better defining its clinical significance is needed. Previous studies have discovered that MAC was associated with young age, advanced tumor stage, accumulation in female patients, and distinct molecular patterns, such as microsatellite instability and activating mutations of the BRAF gene [2,4,5]. When analyzing clinical outcomes, the clinicopathological differences between MAC and AC are potential confounding factors. At present, findings regarding the progressive behavior and survival remain controversial in MAC [6]. Due to the lack of substantiated data, MAC specialized treatment strategy remains unclear and patients with MAC are usually treated along the lines of recommendations for adenocarcinoma (AC) of the CRC [7]. Thus, more solid evidence is needed to evaluate the significance of mucinous component (MC) in AC.
The present study aims to further evaluate the clinical significance and prognostic value of MC in AC. Slides of AC with MC (ACMC) (1-100%) were reviewed, and proportion of MC in AC was re-evaluated. Propensity-score matched (PSM) analysis was conducted to minimize bias. The optimal cut-off value of the MC proportion for prognostic prediction was analyzed. The clinicopathological features and survival of enrolled cases were also depicted before and after matching. Meanwhile, the potential risk factors for poor survival were identified.

Study population
Records of CRC patients underwent surgical resection from January 2007 to February 2018 at the First Affiliated

Follow-up method
Patients were followed up postoperatively every 6 months for 2 years, and then annually for 3-5 years at outpatient clinic. Physical examination, serum tumor markers, including CEA, and abdominal/chest/pelvic imaging using a CT scan were used for surveillance. Colonoscopy was performed at the 1st and 2nd year after surgery. Overall survival (OS) was defined as the period from the surgery to death from any cause.

Pathological evaluation
For each case, the number of paraffin block for pathological evaluation was determined based on the tumor size (1 block per cm). Tumor sections from paraffin blocks were stained with hematoxylin-eosin. The ratio of MC area was separately evaluated by two pathologists, and the mean value was adopted. If the difference in estimated values was 10% or greater, the two pathologists reassessed the specimens to determine the consensus. Finally, tumors, with MC proportion ranging from 1 to 100%, were classified into 10 groups evenly with 10% ingredient per group. Classical gland-forming adenocarcinomas with variable size and configuration of the glandular structures were classified as AC. ACMC was defined as tumors with 1-100% of the lesion being composed of mucin, typically characterized by pools of extracellular mucin that contain malignant epithelium as acinar structures, strips of cells, or single cells. And those with more than 50% mucin in tumor were labelled as MAC. Signet ring cell component was defined as AC with signet ring

Statistical analysis
The data are presented as the mean and SD or as the median and range. For differences in categorical variables, chi-square analysis, Fisher exact test or Kruskal-Wallis ANOVA test was performed where appropriate. Survival was depicted with Kaplan-Meier curves and compared using log-rank tests. Univariable and multivariable survival analyses using Cox regression models were performed to identify prognostic factors. Hazard ratios (HRs) were presented with 95% confidence intervals (95%CI). Propensity-score matched (PSM) analysis was conducted to minimize bias. The 1:1 matching process was performed by using the nearest neighbor matching method, with a maximum caliper width of 0.03 times the standard deviation of the logit (propensity score). Variables adjusted included gender, age, history of smoking and alcoholism, ASA class, comorbidities, preoperative CEA, CA19-9, albumin, and HGB levels, occult blood status, operative factors and tumor factors. Receiveroperating characteristic (ROC) curve was used to identify the optimal cut-off value of MC ratio for prognostic prediction. At each ratio, the sensitivity and specificity for survival were determined and plotted, thereby generating a ROC curve. According to the (0, 1) criterion, the point on of the curve with the shortest distance to the coordinate (0, 1) was chosen as the cut-off value. Twosided P ≤ 0.050 was considered statistically significant. All statistical analyses were performed using the SPSS software program (version 22.0 for Windows, IBM SPSS Statistics, IBM Corporation, Armonk, NY).

Discussion
MAC has different clinicopathological features compared with AC [2,8]. Consistently with previous reports [2,8,9], our data revealed that MAC was associated with higher rate of right-sided location, poor differentiation, advanced T stage before matching, which indicated that MAC is more advanced at diagnosis. In addition, our results showed that MAC patients have a higher rate of albumin< 40 g/dl, HGB < 110 g/L and presence of occult blood, these parameters were seldom analyzed in previous studies. Our data suggested that MAC patients need more nutritional support and improvement in general conditions prior to surgery.
The survival of MAC patients or ACMC (1-50%) patients has always been controversial in previous studies  [2,4,[8][9][10][11][12][13][14][15][16][17][18][19][20][21][22][23][24]. The retrospective nature of these studies may be an essential factor leading to the difference. Two PSM studies, minimizing confounding factors statistically, discovered that MAC was a prognostic factor in TNM stage II patients [6,25]. This study found that the survival of patients in the AC group, the ACMC (1-50%) group, and the MAC group were similar both before matching and after matching. However, MAC exhibited a relatively low 3-year OS compared to ACMC (1-50%) and AC in TNM stage II patients after matching, though no statistical significance was detected (P = 0.051). However, the detailed mechanisms of MAC patients with TNM stage II exhibited poorer survival need further investigation. It has been recommended that adjuvant chemotherapy should be routinely performed for patients with stage II MAC, and special attention should be paid during follow-up because of the risk of peritoneal or local recurrence [25].
To further define the clinical significance of MC in CRC. Our study re-evaluated the MC proportion more accurately and 70% was found to be a cut-off value for predicting prognosis, which is rarely reported in previous studies. Patients with MC > 70% displayed a much poorer 3-year OS compared with patients with ACMC 1-70% and AC patients both in all patients and stage-matched (TNM stage II and stage III) patients. In addition, MC > 70% was demonstrated to be a risk factor of poor OS in univariable analysis, though not an independent risk factor to multivariable analysis. Thus, the effect of mucin on survival may be associated with its proportion in the lesion, and MC > 70% may serve as a biomarker for poor prognosis.
To better understand the cause of diverse clinical behaviors, numerous studies have focused on discovering the gene expression profiling in MAC [26][27][28][29].  [26]. Besides, low frequency of mutations in the p53 gene or overexpression of p53 protein and loss of heterozygosity in the DCC gene have been reported [30,31]. Genome-wide analysis found that MAC displayed 182 upregulated and 135 downregulated genes compared with AC [29]. The most upregulated genes included those involved in cellular differentiation and mucin metabolism, and altered biologic pathways included those associated with mucin substrate metabolism, amino acid metabolism, and the mitogen-activated protein kinase cascade [29]. Consistently, MUC2, which is one of the glycosylated proteins, was reported to be overexpressed in MAC [32,33]. In addition, MAC overexpresses both TYMS and GSTP1, biomarkers indicating resistance to 5-FU and oxaliplatin [34]. These findings may partially illustrate the different phenotypes of MAC.
In conclusion, this study detected that MAC is usually diagnosed as an advanced stage. MAC patients have a similar survival with AC patients and ACMC (1-50%) patients before and after matching. Mucin accounting for more than 70% in the lesion is a more valuable cut-off score of predicting poor survival. Patients with MC > 70% should be given more clinical attention. However, data was retrospectively reviewed in this study, although PSM was conducted to adjust for known confounding factors, some degree of selection bias cannot be ruled out.