Introduction

Colorectal cancer (CRC) is the fourth leading cause of cancer death worldwide [1]. The presence of colorectal liver metastases (CLM) are associated with a poor prognosis, with a median survival for untreated disease ranging between six and twelve months [2–4]. Surgical intervention is the only chance of long-term survival, with the five year survival ranging between 25 and 58 % [4–6]. Unfortunately, forty to fifty per cent of CRC patients develop CLM [6, 7]. They are either synchronous or metachronous in presentation with approximately equal incidence [4, 6, 7]. There is no clear international definition of what constitutes a synchronous presentation, the 7th edition of the AJCC manual states that staging can be undertaken as part of ‘definitive surgery, as part of primary treatment or within 4 months of diagnosis, whichever is longer’ [8]. However, no consensus exists in the literature with varied interpretations being used in clinical studies including: metastases detected prior to or at the time of surgery, metastases detected within three or twelve months of the CRC diagnosis [5, 9]. Patients who present with synchronous colorectal liver metastases have locally advanced primary tumours and tend to present with a greater metastatic burden than patients who develop metachronous colorectal liver metastases [10, 11]. It has been demonstrated that the presence of synchronous disease is an indicator of poor prognosis [10]. There is no consensus as to why colorectal cancer primaries develop either synchronous or metachronous CLM [5]. Clarifying whether synchronous and metachronous CLM represent different subtypes of metastatic CRC is paramount as it could have important clinical implications. The aim of this study was to ascertain whether there was a biological difference between the two subsets of patients. The reader is advised to refer to Table 1 for a brief description of the markers discussed in this review.

Table 1 Description of molecular markers assessed in the systematic review
Full size table

Methods

A systematic review of the literature was performed to assess the differences in biomarker expression: (1) between patients with synchronous colorectal liver metastases (synchronous group) and patients with metachronous colorectal liver metastases (metachronous group) and (2)to assess differences in biomarker expression between colorectal liver metastases and their respective CRC primaries in both the synchronous and metachronous groups (Fig. 1). The methodology undertaken was based on the guidelines from the preferred reporting items for systematic reviews and meta-analyses (PRISMA) statement [12].

Fig. 1
figure 1

Diagram demonstrating the different groups in which molecular markers were compared

Full size image

Search strategy

An electronic database search was performed in October 2012 using the following resources: MEDLINE from PubMed (1950 to present), Embase, Cochrane and the web of knowledge. The following search headings were used: “colorectal cancer liver metastases”, “colorectal cancer hepatic metastases”, “colorectal cancer synchronous liver metastases”, “colorectal cancer synchronous hepatic metastases”, “colorectal cancer metachronous liver metastases”, “colorectal cancer metachronous hepatic metastases” combined with the Boolean operator ‘AND’ and each of the following terms: “biomarkers” and “molecular markers”. The titles were initially scanned and abstracts of interest were reviewed. All articles reviewed and included in the study had their reference lists scanned and studies found were included in the study if they met the inclusion criteria.

Inclusion criteria

To enter the review the study had to:

(1) compare biomarkers between the defined subgroups (Fig. 1); (2) only include studies assessing liver metastases or primary tumours of colorectal adenocarcinoma origin; (3)differentiate between colorectal liver metastases and extra-hepatic metastases during the tissue analysis.

Exclusion criteria

The following criteria were used to exclude studies from the review:

(1) studies that analysed synchronous and metachronous groups together as one entity; (2) studies that analysed CLM with other types of metastases as one entity; (3) studies that did not clearly define whether CLM were synchronous or metachronous in origin; (4)animal studies; (5) articles that were conference abstracts, editorials, commentaries/letters or reviews.

Data extraction

Two authors (AS, PG) independently collected and tabulated the data into an electronic spread-sheet. Any differences in collated data between the two authors were discussed and agreement was reached by consensus. The specific data items collected were the following: first author/institute, year of publication, year of study, study design, groups being compared, study sample size, molecular markers being assessed, manner of molecular marker assessment, and significant/non-significant findings. The study quality was assessed by two authors independently using the Newcastle-Ottawa Scale for assessing the quality of non-randomised trials.

Data analysis/quality of studies

All studies were assessed for their level of evidence using the Oxford Centre for Evidence-Based Medicine Levels of Evidence table [13]. It was elected to perform a descriptive review of the data as opposed to a meta-analysis due to the heterogeneity of the studies and markers assessed.

Results

Study selection and quality

A total of 3400 citations including references were found. After exclusion of duplicate citations, a review of the titles and abstracts resulted in 213 articles being reviewed of which 31 met the inclusion criteria (Fig. 2). The overall quality of the studies was poor with a median Newcastle-Ottawa score of four (1-6).

Fig. 2
figure 2

Flowchart demonstrating the search strategy

Full size image

Study design and characteristics

All of the studies were retrospective and the publication year of the articles ranged between 1996 and 2012, with the majority (25) being published in or after the year 2001. None of the studies were randomised and were all observational. Fourteen of the studies had been undertaken in Japan.

Comparison of molecular markers between colorectal cancer liver metastases and colorectal cancer primaries in the synchronous and metachronous groups

There were a total of fifteen studies assessing the difference in molecular marker expression between the colorectal liver metastases and their associated CRC primaries (Table 2). In the metachronous group, the expression of TS, ERCC1, DPD was found to be higher in the liver metastases. The expression of p27 using immunohistochemistry (IHC) was significantly reduced in the liver metastases of the metachronous group. No significant differences in p27 mRNA, p53, Ki-67, AREG, EREG were found in the metachronous group. In the synchronous group, the expression of VEGF, CXCR4 and ISR1 were found to be significantly higher in the liver metastases. The expression of Cyclin E, CCR6 and FAK were lower in the liver metastases of the synchronous group. No significant differences in genetic aberration, protein expression profile or expression of ERCC1, TS, p27, p53, Ki-67, AREG, EREG and EGFR were found in the synchronous group.

Table 2 Comparison of molecular marker expression in colorectal liver metastases and associated primaries
Full size table

Comparison of molecular markers in synchronous group compared to metachronous group

A total of 16 studies compared the tumour marker expression between the two groups (Table 3). In the synchronous group, the CRC primaries were found to have a significantly higher methylation level of MINT1 than the primaries of the metachronous group. No difference in the expression of TGF-α, EGFR, Ki-67, p53, VEGF, CEA, sialyl LeA was found between the primaries of the two groups. The liver metastases in the synchronous group were found to have an increased expression of COX-2 mRNA, TGF-α and a higher Angiopoietin-2/Angiopoietin-1 ratio compared to the liver metastases in the metachronous group. The expression of CD83 and EGFR mRNA were found to be higher in the liver metastases of the metachronous group compared to the liver metastases of the synchronous group. No significant difference in the expression of COX-2 or EGFR was found between the liver metastases of the two groups using IHC. No significant differences between the metastases of the two groups were found in the following markers: VEGF, angiopoietins, genetic aberrations, Ki-67, TP, CD31, CD34, c-erb-2 and ZEB2.

Table 3 Expression of molecular markers in synchronous group compared to metachronous group
Full size table

Discussion

Our review highlighted that the majority of the studies were rarely validated by further studies and generally used semi-quantative methods to analyse expression levels (Tables 2, 3). Furthermore, as other authors have found, very few of the included studies defined the time interval employed in their definitions of what constituted either a ‘metachronous’ or ‘synchronous’ presentation and it thus makes interpretation of the comparative findings difficult (Tables 2, 3) [5, 14]. Furthermore, the overall quality of the studies included was poor: nine studies did not use any controls, only 11 studies achieved a Newcastle-Ottawa score of five or more and small sample sizes were frequently used.

Comparison of molecular marker expression between colorectal liver metastases and their respective colorectal cancer primaries

The process of the development of colorectal cancer liver metastases is still not fully understood and is complex [15, 16]. It is believed that a ‘subpopulation’ of cancer cells within the primary tumour evolve and develop the ability to metastasise [4]. It is recognised that CLM can exhibit biological differences to their matched primaries due to the micro-environment of the liver and the necessary genetic alterations required for the CRC tumour cells to survive the different steps of metastatic development [4, 17].

p27 has been demonstrated to correlate with advanced stages of CRC. [18, 19] In the metachronous group, its expression was reduced in the liver metastases suggesting that there was a ‘post-translational’ degradation of the protein in the liver metastases. Currently little is known about the role of Cyclin E in the development of CLM [20]. The expression of cyclin E was found to be significantly reduced in the liver metastases of the synchronous group [20]. It is postulated that this finding is as a result of the liver microenvironment resulting in a reduced rate of proliferation of CLM compared to the primaries [20].

Thymidylate synthase (TS) is the target of 5-fluorouracil (5-FU) and excision repair cross-complementing factor 1 (ERCC1) confers the ability to repair platinum related DNA damage [21, 22]. Kobayashi H et al. [21], demonstrated that there was no quantitative difference in expression of TS and ERCC1 expression between the CLM and their respective primaries in the synchronous group. Conversely, the expression levels of both TS and ERCC1 were significantly higher in the metastases of the metachronous group compared to their matched CRC primaries [21]. This finding would suggest that there is a difference in biology between the CLM of the synchronous and metachronous groups [21].

The expression of chemokine receptors in CRC cells has led to the belief that they play an important role in the development of CRC metastases [16, 23, 24]. An interesting finding of the review, was that the expression level of chemokine receptor CCR6 was decreased in the CLM of the synchronous group compared to their matched primaries [23]. The ligand (CCL20) for CCR6 is found predominantly in the periportal area of the liver, and it is postulated that this may be one of the mechanisms by which the CRC cells metastasize to the liver [23, 25]. The reduced expression of CCR6 found in the synchronous CLM was probably as a result of ligand binding and the subsequent degradation of the chemokine receptors [23]. CXCR4 is the most commonly expressed chemokine in CRC whilst its ligand is highly expressed in normal liver parenchyma and has been shown to have an important role in the growth of CRC liver metastases [16, 24, 26]. Indeed, this seems substantiated by the fact the expression of CXCR4 was significantly elevated in the liver metastases of the synchronous group compared to their respective CRC primaries [24]. The liver has a naturally high expression of the CXCR4 receptor ligand (CXCL12) and it is thought CRC cells with an increased expression of CXCR4 may have an increased ability to metastasise to the liver by a ‘homing’ mechanism [24].

Vascular endothelial growth factor (VEGF) has been shown to correlate with advanced CRC, lymphatic invasion and metastases [26]. Kim et al. [27], demonstrated, using immunohistochemistry, that the expression of VEGF was significantly increased in the CLM of the synchronous group compared to their CRC primaries. The increased expression of VEGF maybe as a result of a type two error or it may suggest that VEGF plays an important role in the progression of synchronous CLM and perhaps relates to the aggressive nature of synchronous CLM compared to metachronous CLM. Insulin Receptor Substrate 1 (Irs1), is thought to be involved in the β-catenin signalling pathway and is considered to have a role in CRC progression [28, 29]. The CLM in the synchronous group had a significantly higher expression of Irs1 suggesting a possible role in the development of metastases [28]. Focal adhesion kinase (FAK) is thought to play a role in metastatic adenocarcinomas [16, 30]. The expression of FAK was lower in the liver metastases of the synchronous group compared to their CRC primaries [30]. Although a small sample size of ten patients was assessed, it may denote that a reduced motility of the metastatic cells confers an advantage once the metastatic CRC cells are established in the liver.

Several studies did not demonstrate any difference in expression between the CLM and their respective primaries. In a recent study, the expression of EGFR ligands in the primary tumours correlated with their respective liver metastases in both the metachronous and synchronous groups (Table 1) [31]. In addition, studies assessing genetic aberrations and protein expression profile in the synchronous group did not demonstrate any significant differences between the primaries and their respective CLM (Table 2). This would suggest that some of changes required for metastatic progression occur at a primary level and are maintained at a metastatic level. This seems to confirm the hypothesis that the metastatic genetic profile arises in the primary tumour and ‘is maintained in the distant metastases’ [32].

Molecular marker expression in the synchronous group compared to the metachronous group

We hypothesised that the expression of molecular markers in the primaries of the two groups would be different in view of the known clinico-pathological differences that exist between synchronous and metachronous colorectal liver metastases. However, the majority of studies demonstrated no differences in expression. In view, of the limited number of studies comparing the primaries in the two groups as well as the limited number of molecular markers assessed, it is not possible at this stage to draw any firm conclusions.

Cyclooxygenase-2 (COX-2) has been demonstrated to be up-regulated in colorectal adenocarcinomas and correlates with CRC progression and the presence of CLM [33, 34] Nakamato et al. [35] using immunohistochemistry demonstrated no difference in the expression of COX-2 between the two groups. However, Pantaleo et al. [36] using Reverse transcription polymerase chain reaction (RT-PCR) demonstrated that the expression of the COX-2 gene was elevated in the CLM of the synchronous group. Epidermal growth factor receptor plays an important role in the progression and metastatic potential of advanced colorectal cancers [37, 38]. Interestingly, Pantaleo et al. [36] demonstrated using RT-PCR and enzyme-linked immunosorbent assay that EGFR was significantly overexpressed in metachronous group . However, another study using immunohistochemistry found no difference in the expression of EGFR between the two groups [39]. A direct comparison of the findings between the studies assessing COX-2 and EGFR using IHC and RT-PCR cannot be made due to the different laboratory techniques used. Pantaleo et al. [36] findings are interesting and would imply that the synchronous group represent a different biological entity to the metachronous group . However, they cannot be interpreted as significant without further validation. The expression of TGF-α was found to be higher in the liver metastases of the synchronous group compared to the liver metastases of the metachronous group, although this difference was not significant it may denote a biological difference [39].

An important aspect of CLM development is the ability of metastatic tumour cells to evade immunological responses during migration and invasion of the liver. Mature dendritic cells are known to increase in number in response to CLM and increased numbers are associated with a reduced rate of growth of the metastases [40]. One study demonstrated that the metastases in the metachronous group were found to have a significantly higher number of mature dendritic cells [41]. This would suggest that both groups of metastases elicit a different immunological response and could explain the difference in tumour aggression between the two groups [41].

A recent study, demonstrated that a primary tumour in situ in the synchronous group resulted in a higher Ang-1/Ang-2 ratio in the liver metastases compared to either a synchronous group with their primary tumour resected or the metachronous group [42]. It was also demonstrated that the adjacent liver parenchyma in the synchronous group with their primary in situ had significantly higher expression levels of angiogenic factors including VEGF [42]. These findings suggest that the primary tumour has an important role in the progression of colorectal liver metastases by creating a ‘permissive soil’ for the metastases to proliferate. These results could lead one to hypothesise that any differences demonstrated between the metastases of the synchronous and the metachronous groups could be related to the presence or absence of the primary tumour.

Conclusion

The review of the literature confirms that both synchronous and metachronous colorectal liver metastases ‘evolve’ and exhibit different biological characteristics to their respective colorectal cancer primaries. Although there are conflicting results, the systematic review suggests that biological differences between the liver metastases of the synchronous and metachronous groups may exist and are consistent with the clinically more aggressive nature of synchronous colorectal liver metastases. Whether these differences are as a result of the host immunological response or denote that synchronous or metachronous colorectal liver metastases represent different tumour subtypes remains underdetermined. Determining whether these two groups of patients have biologically distinct metastases is crucial as it could improve and ‘tailor’ current oncological management according to the timing of the liver metastases presentation. One of the most interesting questions arising from this review is whether any differences at a metastatic level are present at a primary tumour level. Indeed, if differences are detectable within the primary tumour this could have important clinical implications at a pre-operative biopsy stage as well as in routine post-operative surveillance. It is clear that important changes occur within the primary tumour and are maintained throughout the metastatic cascade. In addition, recent evidence seems to suggest that the presence of the primary tumour may have an influence on the biological characteristics of the liver metastases. The review has served to highlight that the main focus of recent research has been to determine whether biological differences exist at a metastatic level and it is has thus not been possible to determine whether differences between the two groups occur at primary tumour stage. Future research should include comparison of the primaries between the two groups.