Introduction

Mucinous carcinoma is a specific morphological subtype of rectal cancer. It is characterised by an abundance of extracellular mucin secreted by overactive neoplastic acinar cells. Mucin-producing tumours have been described in other sites such as breast, prostate, ovary and pancreas and share some generic characteristics [1]. The abundance of mucin within the tumour complex gives it a unique appearance both histologically and radiologically. Contrast this with signet cell carcinomas, which are a specific type of mucin-producing tumour in which single neoplastic cells demonstrate large amounts of intracytoplasmic mucin that results in displacement of the nuclei.

Over the years, there has been debate on the precise histological definition of mucinous tumours, which has made interpretation of studies problematic. The current accepted definition, initially proposed by Jass, is based on the presence of a minimum of 50 % of mucin to tumour volume [2]. Previous studies have used varying proportions of mucin to define these tumours [3, 4]. They represent 5–20 % of all colorectal cancers [5–8] and are generally considered to have a worse prognosis than non-mucinous tumours of the rectum. Furthermore, they are also suggested to have a limited response to oncological treatments.

This paper reviews the clincopathological characteristics of mucinous rectal cancer and presents the evidence on treatment response to determine whether these tumours may be considered as a unique entity.

Methods

Identification of studies

An electronic search was carried out using MEDLINE (1965–2013), EMBASE (1980–2013), CINAHL (1982–2013) and the Cochrane library databases. The following medical subject heading (MeSH) terms and keywords were used as follows: The keywords “mucinous”; “mucoid”; “colloid”; “colorectal”; “rectal”; “tumours/tumors”; “cancer” were used in combination. The “related articles” function was used to broaden the search, and all abstracts, studies and citations retrieved were scanned for subject relevance. The latest date of this search was 1 March 2013.

Complete articles of all potentially relevant manuscripts were retrieved and evaluated for inclusion. To be included in this review process, additional references from the collective libraries of the senior authors were identified. Reference lists of all relevant publications were hand-searched for additional studies missed by the search strategy, and this cross-referencing continued until no further relevant publications were identified.

Study inclusion criteria and data extraction

Study methodology was carried out in accordance with the “Preferred Reporting for Systematic Reviews and Meta-Analyses” (PRISMA) recommendations for improving the standard of systematic reviews. Where multiple studies describing the same patient population were identified, the most recent publication was used unless additional information was imparted by earlier work (Fig. 1).

Fig. 1
figure 1

Flowchart showing the systematic methodology used to extract articles for review

Full size image

Results

Epidemiology and clinical characteristics

Mucinous carcinoma accounts for between 5 and 20 % of all colorectal cancers [5–8]. It is relatively rare to find such tumours in the rectum, being most commonly found in the proximal colon. One of the largest studies examining the outcomes of mucinous tumours investigated 16,991 patients and found that just over 18 % of all mucinous tumours were found in the rectum compared with 29 % of non-mucinous tumours [9]. A recent meta-analysis of outcomes in mucinous colon cancer, which included some studies with rectal tumours, reported that mucinous tumours were more frequently found proximal to the splenic flexure compared to non-mucinous adenocarcinoma [10]. Conversely, mucinous tumours are more common than non-mucinous in the proximal colon. A study using the National Cancer Database in the USA reported that 60 % of mucinous tumours were right-sided compared to 42 % of non-mucinous tumours [11]. Consorti et al. [12] conducted a prospective study of mucinous and non-mucinous tumours and reported 31 % of mucinous tumours in the rectum compared to 44.3 % non-mucinous.

There is a higher incidence in younger patients [13–15]. A study by Dozois et al. [16], which investigated the characteristics of colon and rectal cancer in patients below the age of 50, found that patients with mucinous tumours had a mean age at presentation of 42.2 years. Furthermore, rectal cancers were found to be more common below the age of 50 (49.1 vs. 21.9 %). Other studies have shown similar results [17, 18]. Wu and colleagues used a cut-off of 39 years in their comparison of mucinous and non-mucinous tumours, and although this included patients with colonic cancer as well, they found mucinous tumours more commonly in patients <39 years old [19]. A further study of 2,089 patients (of which 144 had mucinous tumours) reported a mean age at presentation of 54.2 ± 16.25 years. This was statistically less than patients with non-mucinous adenocarcinoma who had a mean age at presentation of 58.73 ± 13.62 years [20].

Morphology and pathological characteristics

Histological diagnosis remains the “gold standard”. Mucin produced by the tumour cells aggregates into “pools” or “lakes”, which produce a distinctive pattern on haematoxylin and eosin stain (H&E stain). Although histopathological detection of the mucinous component of tumours is not particularly challenging, sufficient tissue is required to make a diagnosis. Therefore, there may be under-reporting using biopsy specimens due to insufficient tissue [21], and definitive diagnosis is confirmed on analysis of the resection specimen (Fig. 2).

Fig. 2
figure 2

a Histological micrograph showing areas of mucin within tumour complex. The mucinous areas are seen as large vacuoles in relation to the bowel wall and tumour. b Histological micrograph showing a large mucin pool amongst the tumour complex

Full size image

Although most tumours have the propensity to produce some mucin, Parham was one of the first to describe mucinous or colloid carcinoma in detail as a separate entity defined by the amount of mucin produced by the tumour cells [22]. He found that these tumours were associated with a more aggressive behaviour and pattern of spread than non-mucinous adenocarcinoma. A further study by Symonds and Vickery reviewed the pathological characteristics of mucinous tumours and found that the microscopic appearance of these tumours is of particular interest with regard to their behaviour [5]. They found that mucin deposits comprise the main bulk of the tumour complex and are lined by epithelial cells around their perimeter. They proposed that the spread of mucin separated the muscle bundles but upon reaching the perirectal fat showed pooling. This may explain the perceived aggressiveness of these tumours. One theory is that the tumour cells have the ability to imbibe water causing swelling of the tumour complex. As they splay the muscle structure and pass through the layers of the bowel wall, they disseminate tumour cells. This is similar to the pressure theory put forward by Sugarbaker whereby the mucin exerts a pressure effect on the bowel wall structure, thus disseminating cancer cells into the peritoneal cavity that are taken up into lymphatics [23].

Clinical outcomes and survival

Survival outcomes between mucinous and non-mucinous tumours have been much studied. A recent meta-analysis of mucinous cancers of the colon and rectum showed a worse prognosis for mucinous cancers compared with non-mucinous types, even when corrected for stage at presentation [10]. Previously, suggestions had been made that any difference in clinical outcomes between mucinous and non-mucinous carcinomas were attributed to a more advanced stage at presentation. There is a higher incidence of stage III and IV disease associated with mucinous tumours. They are more likely to have lymph nodes involvement in addition to increased tumour penetration [11, 20, 24]. The association with lymph node involvement has led to suggestions of more radical dissections for this tumour subtype [25].

A study by Hyngstrom et al. [11] found that mucinous subtype was only associated with worse survival outcomes when the tumour was located in the rectum. The hazard ratio (HR) of death was 1.22 (95 % CI 1.16–1.29), for rectal cancers but a HR of 1.03 (95 % CI 1.00–1.06), for colonic cancers. This contrasts with signet cell tumours (a special type of mucinous tumour) who had an increased risk of death regardless of tumour location.

In a study of 97 patients with mucinous colorectal cancer who underwent surgery with curative intent, the clinically significant predictive factors on multivariate analysis included stage at diagnosis, mucinous histology, tumour location, gender and age [26]. Mucinous tumours were found to have a worse survival independent of other factors. Curative surgery was defined as obtaining a margin-free resection; however, details on surgical technique and whether surgery was based on principles of total mesorectal excision (TME) were not provided.

One of the largest studies into the outcomes of mucinous tumours that looked at over 160,000 patients did show a survival advantage for non-mucinous tumours [9]. Overall, for all locations of colorectal cancer, mucinous tumours had a significantly worse 5-year survival (58.1 vs. 62.9 %), and this was also apparent for rectal cancers specifically.

One study found a difference in 5-year survival in stages II and III but not for stages I and IV. Further, when comparing patients of all stages who had undergone supposed “curative” resection, the 5-year survival was 70 % for non-mucinous tumours and 62 % for mucinous tumours; this was considered statistically significant [20].

However, there are reports that have showed no survival differences between mucinous and non-mucinous adenocarcinomas of the rectum. A prospective, case–control study comparing survival outcomes of Dukes C or stage III tumours showed that there was no significant difference in overall survival at 5 years. For both groups, this was approximately 50–55 % [12]. Although the study included colon and rectal cancers, the survival analysis was predominantly rectal tumours.

A Canadian study also found no difference in survival between mucinous and non-mucinous tumours independent of tumour location [27]. However, a sub-analysis per stage of presentation found that stage III mucinous tumours did have a worse overall survival compared with non-mucinous tumours of the same stage.

Radiological diagnosis and measuring response

Magnetic resonance imaging (MRI) has the ability to differentiate mucinous tumours from non-mucinous tumours. This is advantageous in being able to diagnose mucinous carcinoma early in the treatment pathway rather than relying on histological diagnosis following surgery. The signal characteristics depend on imaging protocol—T1-weighted images show low signal intensity, whereas on T2-weighted images, there is a significant hyper-intense signal. In contrast, non-mucinous tumours demonstrate intermediate signal intensity on T2 imaging. The difference in signal characteristics is due to the abundant mucin composition [28, 29].

The signal and enhancement characteristics on MRI may make it difficult to distinguish mucinous tumours from other fluid-containing pathologies such as cysts, fluid collections and even necrotic tumours [30]. Although it is unusual to confuse these benign conditions with cancer during preliminary imaging of the pelvis, the difficulty arises during follow-up imaging after treatment. Figure 3 shows the signal characteristics of mucinous tumour on MRI.

Fig. 3
figure 3

a, b MRI showing a rectal cancer with focal areas of mucin. There is a difference in signal characteristics between the tumour itself and the areas of mucin which have a higher signal. The heterogeneous nature of the signal is indicative of cellular mucin

Full size image

Allen et al. [31] studied the MRI features of rectal cancers and the changes following chemoradiotherapy. They defined mucinous tumours on T2-weighted images as high signal intensity, similar to that of obturator internus muscle. The same areas showed a uniform hyper-intensity after chemoradiation on T2-weighted images. However, there were no other morphological changes or shrinkage seen radiologically with persistent areas of high signal intensity.

Oberholzer correlated the radiological changes on MRI following CRT with pathology in order to describe the accuracy of MRI in predicting response in mucinous tumours [32]. They found that more advanced mucinous tumours (CRM to tumour distance of <5 mm) showed a worse response to CRT in terms of down-staging than non-mucinous tumours leading to increased rates of R1 resections. In addition, there was also disease progression in mucinous tumours only (Fig. 4).

Fig. 4
figure 4

a, b MRI showing areas of mucin which have responded to chemoradiation. The homogeneous signal characteristics as depicted by the arrow suggest an acellular component to the mucin which is thought to represent tumour response

Full size image

A sub-classification of mucinous tumours

Mucinous tumours may fall into one of three categories. There are those tumours, which are diagnosed as mucinous prior to any treatment and remain mucinous throughout their course. For example, they are identified as mucinous on biopsy and/or MRI as well as on the surgical resection specimen. These tumours appear to have the worst prognosis and increased risk of local recurrence [33, 34]. The documentation of the extent of residual cellular mucin is important since the risk of tumour spillage from such mucin pools will increase the risk of local recurrence.

A second group of tumours exist, which become mucinous during treatment. These tumours may have started out as non-mucinous and subsequently show morphological change as a consequence of either natural disease progression or chemoradiotherapy. This group is of particular interest as it may represent a specific type of tumour, which behaves as non-mucinous adenocarcinoma despite having the phenotype of mucinous carcinoma. Nagtegaal et al. [35] studied a number of mucinous tumours and found that those tumours that had become mucinous following short-course radiotherapy (SCRT) had a better prognosis than those tumours that were mucinous prior to treatment.

Post-therapeutic change leading to mucinous differentiation has previously been well documented; however, the clinical significance of this is now better understood. For example, on baseline imaging, such as MRI, this corresponds to a tumour that is of entirely intermediate signal intensity with no areas of high signal intensity mucin. After neo-adjuvant chemoradiation or radiotherapy alone, necrosis of the tumour can result in mucinous degeneration. In such cases, degeneration of the tumour results in high signal intensity pools within the previously documented intermediate signal intensity tumour stroma, and can therefore be interpreted as evidence of treatment response. In a further study by Allen et al. [31] in which the MRI features of rectal carcinoma were studied before and after chemoradiation, an additional 4 tumours were added to the original 7 tumours, which were identified as mucinous by MRI. Tumours were classified as mucinous on signal intensity of the supposed mucinous area having greater signal intensity than the obturator internus muscle. Interestingly, although chemoradiation may increase the degree of mucinous differentiation, these tumours do show a response to treatment. They show tumour regression and down-staging, meaning that induced mucinous change does not appear to be a poor prognostic factor. The European Organisation for Research and Treatment of Cancer (EORTC) trial investigated the effect of adding chemotherapy to pre-operative radiotherapy and the role of adjuvant chemotherapy in stages II and IV rectal cancer. As part of the analysis, they found that in addition to there being an increased proportion of mucinous tumours found following neo-adjuvant treatment, this was further enhanced in the group with chemo- and radiotherapy than just pre-operative radiotherapy alone [36]. This phenomenon is not fully understood; however, one explanation offered by the College of American Pathologists is that mucinous change is a feature of response to pre-operative treatment [33].

The final group of mucinous tumours are those that demonstrate acellular mucin on histological analysis. The relevance of acellular mucin remains controversial. A recent analysis of 108 prospectively collected post-treatment specimens showed acellular mucin pools in 16 cases. The presence of acellular mucin pools had no impact on recurrence-free survival [37]. Acellular mucin may be regarded as a type of treatment response and not as residual tumour although this is still contentious. Cellular mucin on T2 imaging is hyper-intense but contains areas of more intermediate signal corresponding to the histologically demonstrated malignant cells, cords and vessels. After treatment, the necrosis of these morphological features results in the formation of acellular mucin—namely pools of featureless high signal intensity fluid-like signal on the T2-weighted images which when compared with pre-treatment scans contain no or minimal intermediate signal intensity. The clinical outcomes of these tumours that show colloid or mucin pools with little or no tumour activity have been studied by Rullier et al. [38]. They differ in behaviour and morphology from pre-existing mucinous tumours and commonly have a much greater component of mucin in the tumour complex, sometimes as much as 80 %. The pools of mucin are also less basophilic than pre-existing mucin. The debate as to whether this variant of mucin should be considered as complete response—ypT0N0; or consider these mucin pools to contain residual tumour continues. Rullier showed some similarities with tumours that have shown down-staging and some similarities with tumours that have shown no response at all. It may be prudent not to consider mucin response as complete response.

The effect of chemo/radiotherapy

Neo-adjuvant chemo/radiotherapy (CRT) is an integral part of the management of rectal cancer. CRT may be given for a combination of reasons based on the evidence of poor prognostic factors. Several studies have shown that patients with mucinous tumours respond more poorly to radiotherapy than non-mucinous tumours. One such study of 136 patients compared the outcome of 25 mucinous tumours with 111 non-mucinous tumours. Although the results showed decreased down-staging in the mucinous group, this was not seen to affect overall survival or disease-free survival [39]. Interestingly, there was a benefit for the mucinous tumours that received neo-adjuvant CRT. None of the mucinous tumour achieved complete response, but almost half were down-staged by one stage. Sengul et al. [40] reported a statistically significant difference in tumour regression grade (TRG) between the two subtypes and a much smaller decrease in T and N staging. However, only 16 patients were studied and the small numbers may account for the results. The underlying biology of these tumours may explain their different responses to oncological therapy.

Shin et al. [41] reported on the effectiveness of pre-operative CRT on 23 patients with mucinous rectal cancer. Seven patients showed minimal down-staging of one stage, and one patient showed disease progression. There were no cases of pathological complete response unlike the non-mucinous group where this was the case in 16 % of patients. Recurrence was not statistically significant, but there was a difference in 5-year survival (64.8 vs. 79.8 %). This was more apparent in more advanced stages of disease. They compared the response to CRT of 23 mucinous tumours with 345 non-mucinous tumours. T-stage down-staging was much worse in the mucinous group although there was no difference in nodal down-staging. However, the numbers of mucinous tumours in this study were again small. A study by Qiu found that mucinous type along with T4 stage on multivariate analysis negatively affected tumour response [42]. These findings may be explained by underlying gene expression which results in chemo-resistivity. Enzymes that interfere with pyrimidine and platinum-based compounds are overexpressed in mucinous tumours [43].

Mucin or colloid production by a tumour may represent a measure of response to neo-adjuvant treatment. The production of mucin without the presence of tumour cells is a common phenomenon after radiotherapy in particular [32, 37, 44–46]. Rullier et al. [38] found that 39 patients out of 200 demonstrated a colloid response to pre-operative radiotherapy, of which the vast majority (n = 34) showed no regression in tumour grade. But this is a separate phenomenon to radiation causing mucinous tumours. A significant proportion of tumours that arise following pelvic irradiation for other cancers are seen to be mucinous—approximately 25 % [47].

Yet the role of adjuvant therapy in the treatment for mucinous tumours may be changing. Previous studies have suggested that mucinous tumours do not response well to 5-FU-based chemotherapy after surgery [48, 49]. Negri et al. [49] studied the effect of 5-FU-based adjuvant chemotherapy on survival in colorectal cancers. Median overall survival was 11.8 vs. 17.9 months, which showed statistical significance on multivariate analysis. The view had been that mucinous tumours show little response to adjuvant therapy, but this may have been due to inadequate oncological surgery. The acceptance of TME in rectal cancer has changed outcomes for all stages of disease. More recent studies have accounted for this factor as a reason for local, and in the case of nodal disease, distant recurrence. A recent report by Catalano studying the outcomes of mucinous colon cancer in stage II and stage III disease has shown a survival benefit for patients with mucinous tumours that are offered adjuvant chemotherapy. This was shown on multivariate analysis with a hazard ratio of 2.73 [50].

The biology of mucinous rectal cancer

Cancer of the colon and cancer of the rectum demonstrate many biological and molecular differences. This raises the possibility of different pathways governing tumorigenesis in these malignancies. The differences in clinicopathological such as location of tumour may be a reflection of the underlying molecular profiles or different embryological origin of the different parts of the gut.

Mucin is produced as a protein by a group of specific mucin-producing genes, which are prefixed by the name—“MUC”. There are several genes within this family, and they are able to produce two distinct types of mucin—a transmembranous type and a secretory type. These genes are not part of a gene family in the truest sense of the term as their gene products vary dramatically in their structure and function. However, it is the characterisation of specific parts of the gene products, which have led to them sharing the name “MUC”. The two types of mucins are classified according to whether they are secreted mucins or transmembranous mucins. MUC1 produces a transmembranous mucin, which is associated with a poor prognosis. MUC1 expression is low in mucinous tumours but in those tumours that show overexpression, there is an association with worse prognosis and chemoresistivity to certain chemotherapy agents [51]. Conversely, the MUC5 gene has been shown to confer a survival benefit [52] and unusually is overexpressed in some mucinous tumours. However, MUC5 also shows a greater association with proximal location of the colorectum, which may explain the poor prognosis of mucinous rectal cancer; that this may be due to relative underexpression of MUC5. MUC2 on the other hand is almost exclusively associated with mucinous tumours. It may play a role in p53 mutation. There are also reports of MUC2 and liver metastases [53]. Kim et al. [54] studied the prevalence and expression of specific mucin core proteins in a variety of colorectal polyps in addition to mucinous and non-mucinous tumours. MUC2 and MUC5 were most frequently found in mucinous tumours and villous adenomas.

In addition to the pattern of mucin gene expression, mucinous tumours also express other distinct genetic features. For example, there is a lower frequency of p53 mutations and conversely a higher frequency of K-Ras mutations [55]. p53 mutations are interesting as they are associated with rectal cancers but not common in mucinous subtype [56, 57]. There may be heterogeneity in p53 mutations that accounts for different phenotypic effects on individual tumours.

Discussion

The results of this study have highlighted specific characteristics of mucinous rectal cancer, which distinguishes it from non-mucinous disease. These tumours are more commonly found in the proximal colon; however, when present in the rectum, they appear to behave more aggressively. The location of tumours may reflect the underlying molecular profile although this has not been definitively proven. It is also unclear as to whether fair comparisons can be made between the clinical outcomes of colon and rectal cancer; the survival outcomes, response to chemotherapy and ability of the tumour to metastasise are not immediately comparable between non-mucinous tumours. The more aggressive behaviour of mucinous rectal, rather than colonic tumours, may simply be due to anatomical location and that the mechanism of spread described above has a more immediate effect in the narrow confines of the pelvis.

The results of the demographic profile of mucinous tumours consistently showed that they are more common in younger patients. Different studies used varying cut-offs; however, the trend was that these tumours manifest earlier in younger patients. Whilst the explanation to this observation will most likely lie in the genetic make-up of these tumours, this may simply reflect their aggressive nature. Whilst non-mucinous tumours may take longer to progress through carcinogenesis, mucinous tumours may accelerate through this pathway and thus present earlier. Another explanation may be related to the stage at presentation. Early tumours that are confined to the bowel wall may not present with symptoms compared to more locally advanced tumours. Mucinous tumours are more commonly seen at an advanced stage, which may be due to aggressive spread through the bowel wall causing symptoms requiring investigation.

The actual response of these cancers to oncological treatment is becoming clearer. The challenge over the years in both mucinous and non-mucinous tumours with regard to survival outcomes following varying types of oncological treatment has been interpreting the results in the light of pathological scrutiny of the surgical technique. Several of the historical studies did not include operative details or grade the quality of the specimens. The importance of this has been highlighted by Quirke [58]. Therefore, attempting to apply the results of these studies, which do not describe surgical technique and histopathological audit, may not be relevant to today’s practice. However, more recent studies, which include TME surgery and note the importance of a positive CRM, have challenged the traditional view of chemoresistance. And whilst the survival benefit of adjuvant chemotherapy has not been specifically documented in any high-quality randomised trial for this type of rectal cancer, it is becoming clear that adjuvant chemotherapy may indeed have a role to play in mucinous cancer. Equally, whether pre-operative radiotherapy leads to down-staging is another issue, which must be resolved with greater accuracy. If these issues do become clearer, it will allow clinicians to decide whether and when radio- or chemotherapy should be given and what the survival benefit will be.

The acceptance of TME as the optimal surgical technique for rectal cancer has helped standardise practise. A discussion on the modifications of this technique is beyond the scope of this review; however, the association of an increased nodal burden in mucinous disease should not impact on technique. If the principles of TME are followed and the mesorectal nodes are removed en-bloc with the specimen, there should be prognostic relevance to local recurrence although this may be a stronger indication for adjuvant chemotherapy.

Conclusion

Mucinous tumours of the rectum remain difficult to treat and are still associated with worse survival outcomes than non-mucinous tumours. However, with a universally accepted definition, better pre-operative detection using MRI and high-quality oncological surgery, we may see improvement in outcomes in the future. This will aid further study of these subtypes and allow clinicians to risk stratify patients in terms of oncological treatment.