Mucinous epithelial ovarian cancers (mEOC) are a relatively rare subset of ovarian cancers. Despite a relatively favourable outcome in early disease, the more frequent advanced presentation is associated with poorer response to platinum/taxane chemotherapies, and poorer survival, compared to serous ovarian cancers. We consider some of the fundamental clinico-pathological and molecular features, and existing clinical trial data regarding mEOC. Underlying molecular differences, between mEOC and serous cancers may contribute to the observed clinical differences, including an increased prevalence of K-RAS mutations in mEOC, more in keeping with gastrointestinal tumours. This observation contributes to the rationale for a trial (“mEOC”) investigating the use of “ovarian” versus “gastrointestinal” style chemotherapy. Looking to potential future approaches, we speculate upon the potential impact of emerging technologies on the future investigation and management of mEOC.
- Mucinous ovarian cancer
- serous ovarian cancer
- clinical trials
- targeted therapies
- breast cancer
- ovarian tumour
Statistics from Altmetric.com
- Mucinous ovarian cancer
- serous ovarian cancer
- clinical trials
- targeted therapies
- breast cancer
- ovarian tumour
Ovarian cancer represents the fifth most common cancer to affect women in the UK, with approximately 6700 cases diagnosed annually. The majority of ovarian cancers are serous epithelial ovarian cancers (sEOCs), with mucinous epithelial ovarian cancers (mEOCs) forming around 12% of diagnosed cases, and hence comparatively rare.1 Data from our own institute, St James's Institute of Oncology, shows that only 10 of the approximately 150 primary ovarian cancers seen annually are mucinous in origin.
Evolving evidence has demonstrated distinct histopathological and clinical contrasts between mEOC and sEOC, as discussed further in this issue, leaving oncologists questioning the current blanket approach to their treatment. Loco-regionally advanced sEOC and mEOC (the most prevalent presentation) is treated with a combination of chemotherapy and debulking surgery. However, there is evidence to suggest that mEOCs may be less sensitive to platinum based chemotherapy and the low prevalence of mEOC patients within randomised controlled trials (RCT) leaves the evidence base uncertain.2–4
Here, we outline clinicopathological evidence supporting the view of mEOCs as a separate clinical entity from sEOCs, how these considerations have led to the opening of a trial (the ‘mEOC’ trial) and finally how emerging technologies may impact upon diagnosis and treatment of mEOC.
Histopathological features of mEOCs
The histological diagnosis of mucinous ovarian tumours is often complex, requiring a multidisciplinary approach and correlation with the clinical picture. Undoubtedly, mEOCs share certain histopathological features with gastrointestinal (GI) tract cancers, and difficulties may present when differentiating occult GI tract tumour metastatic to the ovary from a mEOC (or sEOC) primary.
The clinical importance of a correct histopathological diagnosis of an ovarian tumour cannot be underestimated. Patient outcomes can differ greatly depending on the true underlying diagnosis: a patient with a localised mucinous primary will have a 90% 5-year survival,5 whereas a metastasis, for example, from a pancreatic cancer primary, may offer a life expectancy of <6 months.
Current chemotherapy options differ for primary ovarian cancers, compared with tumours metastastic to the ovary. With the increasing use of molecularly targeted therapy, such as anti-epidermal growth factor receptor (EGFR) monoclonal antibodies for metastatic colorectal cancer (mCRC) and tratuzumab for HER2 positive gastric cancer, there are growing therapeutic options for certain metastatic cancers not yet used as a standard in ovarian cancer.6 ,7
Molecular features of mEOC
Molecular differences between sEOC and mEOC support the argument that mEOCs should be considered as a separate clinical entity from other ovarian epithelial tumours. Several studies show downstream signalling in the EGFR pathway differs from mEOC and sEOC, with a higher proportion of activated ras oncogene mutations in mEOC compared with sEOC.8–11 One study demonstrates 46% of mEOC harbouring a KRAS mutation, compared with 17% of other histological types.12 In a large series of colorectal cancers, the prevalence of a KRAS mutation was observed in 51%, similar to the rate suggested for mEOCs.13 Additionally, activating missense mutations in one of the downstream genes, BRAF, has been commonly found in cancers known to exhibit RAS mutations, but within a single tumour these mutations tend to be mutually exclusive. In the above series of colorectal cancers, the prevalence of BRAF mutations was 10%, with no tumours displaying both mutations. However, in contrast, mEOCs without a KRAS mutation did not demonstrate alternative EGFR signalling activation through a BRAF mutation suggesting other downstream signalling pathways may be important.11
KRAS and BRAF mutations in colorectal cancer are used as biomarkers to guide treatments targeting the EGFR pathway, such as the monoclonal antibodies cetuximab and panitumumab, and mutation testing has been used to stratify patients in several recent clinical trials.14 ,15 Given that EGFR messenger RNA is associated with advanced stage, poor prognosis and chemoresistance in some ovarian cancers, it presents an attractive future therapeutic target.16
Clinical trial evidence for chemotherapy in mEOC
Clinical trial data specifically relating to mEOCs are limited. Challenges include rarity compared with other histological subtypes, difficult histological diagnosis and inclusion in trials of both early and advanced stages, with contrasting outcomes.17 The current evidence base for the management of advanced ovarian cancers includes only a small proportion of patients with mucinous histology (3%–7%).2 ,4
ICON3, one of the largest RCTs of first line chemotherapy in ovarian cancer, included 148 (7%) patients with mucinous histology.2 However, a third of patients had stage 1 or 2 disease, and so only 45 patients with mEOC were randomised to treatment with carboplatin and paclitaxel.2 In this subset, no difference in progression free survival (PFS) or overall survival (OS) was seen, compared with other histological subtypes. There is however evidence from meta-analyses and observational studies that mucinous histology confers a worse prognosis in advanced ovarian cancer (table 1). Resistance to standard chemotherapy may contribute to this, and in vitro studies show that chemoresistance in epithelial ovarian cancer varies greatly between histological subtypes, particularly mEOCs22 with mucinous cells being more resistant to platinum.23
Several studies (summarised in table 1) have demonstrated reduced response rates and inferior overall survival, and PFS, of advanced mEOC compared with sEOC treated with chemotherapy in both the first line5 ,18 ,19 ,21 and in the recurrent settings.20
Alexandre et al recently reviewed the outcomes of mEOC patients involved in four first line chemotherapy RCTs run by the French cooperative group (GINECO)3: 54 (4.8%) of 1118 patients with advanced ovarian cancer had mucinous histology; 60% of these patients achieved a complete or partial response to carboplatin and paclitaxel chemotherapies, compared with 81% in the sEOC group, with 36% of the mEOC group progressing on treatment. Respective median PFS was 11.4 vs 17.5 months and OS was 21.6 versus 47.2 months for mEOC patients compared with sEOC patients (table 1).
Another large meta-analysis of advanced stage patients treated through the Gynecologic Cancer InterGroup (GCIG), by Mackay et al, included only 3% of patients with mucinous histology; the prognosis was similarly poor for mEOC, with median OS of ∼14 months compared with over 40 months for sEOC4 (table 1).
One potential weakness of the studies is however apparent, given the recognised difficulties in reaching an accurate histopathological diagnosis, in that only one of the studies, by Pignata et al, included a central review of pathology.20
This is in stark contrast to early stage ovarian cancer where mEOCs appear to have a relatively good prognosis and commonly display favourable prognostic features, including low-grade histology, and high rates of macroscopic complete surgical resection.4 ,21 ,24 Overall, however, advanced stage disease is more common, and in this setting mEOCs have a worse prognosis than sEOCs. Alexandre's meta-analysis demonstrated that extraperitoneal spread was commoner in the mEOC group, and this was associated with shorter OS. A potential explanation for poorer outcomes is that microscopic residual disease following optimal surgical debulking of advanced mEOC may be resistant to standard adjuvant platinum based chemotherapy.23
Given the similarities between mEOC and GI cancers, an alternative therapeutic strategy may be to treat mEOCs with chemotherapy employed in the management of GI mucinous tumours, that is, oxaliplatin, fluoropyrimidine and irinotecan. Phase II studies demonstrate that oxaliplatin and 5-fluorouracil have activity and an acceptable toxicity profile in platinum pretreated relapsed ovarian cancer in general.25 ,26 In vitro studies conducted in mucinous ovarian cancer cell lines have shown activity of oxaliplatin and 5-fluorouracil in cisplatin resistant experimental models.27 A phase II study demonstrated activity of irinotecan in combination with mitomycin-C in 25 patients with platinum refractory mEOC28 with a response rate of 52% and median OS of 15.3 months.
Although recent data have shown significant improvements in PFS with the anti-vascular epidermal growth factor monoclonal antibody bevacizumab,29 ,30 in first line chemotherapy for ovarian cancer EGFR targeted therapies are as yet unproven. Monoclonal antibodies against EGFR have a defined role in treating KRAS wild type mCRC, with evidence of benefit as a single agent in third line treatment, and in combination with standard chemotherapy in the first line setting.6 ,15 Phase II trials have been conducted examining the role cetuximab in ovarian cancer as a single agent31 and in combination with carboplatin.32 Results were modest, with no improvement in PFS, unlike in mCRC. These studies were however limited by: small samples sizes; a lack of assessment of KRAS/BRAF status; and inclusion of all histological subtypes. Given that mEOCs appear to demonstrate similar histopathological features to colorectal cancers, KRAS/ BRAF wild type mEOCs may hypothetically be expected to show similar response rates to anti-EGFR therapies as mCRC; yet, this remains to be shown.14 ,15
In summary, data for advanced mEOCs are limited, showing an inferior response rate to standard chemotherapy, uncertainty around the role of targeted agents and worse outcomes than for other histological subtypes. This stresses the need to examine alternative therapeutic options. The recently updated international evidence based consensus for the management for advanced ovarian cancer maintains the importance of surgical debulking, and platinum/taxane based chemotherapy, and acknowledges emerging evidence on route of administration (intravenous vs intraperitoneal), schedule of chemotherapy (weekly vs 3 weekly) and timing of debulking surgery (up-front vs subsequent to neoadjuvant chemotherapy).33 Importantly, the consensus also recommends that in order to account for recognised differences in genetic profiles and clinical behaviour trials be developed for specific histological subtypes (eg, mEOC), where possible.
The ‘mEOC’ trial
Observations of inferior outcomes with standard ovarian chemotherapy for mEOC underpin the development of the ‘mEOC’ trial. This international randomised phase III study, which is being run through the GCIG, opened in January 2010 and is aiming to recruit 332 women with chemotherapy naïve, advanced (or recurrent) mucinous ovarian cancer. Given the rarity of mEOC, the trial incorporates a 2×2 factorial design with 83 patients randomised per arm on a 1:1:1:1 basis for treatment with carboplatin–paclitaxel or oxalaplatin–capecitabine, with either regime given ±bevacizumab (figure 1).
The study aims to answer two questions. First, is the oxalaplatin and capecitabine chemotherapy regimen superior to carboplatin and paclitaxel? Second, does the addition of the angiogenesis (vascular epidermal growth factor) inhibitor bevacizumab offer an additional survival advantage in keeping with activity in colorectal cancers,35 and consistent with emerging evidence of efficacy in advanced ovarian cancers36–38? Standard secondary objectives will be considered in terms of progression free interval, tumour response rates, toxicity and quality of life.
Successful recruitment to the trial will however be challenging. As already noted mEOC is a rare subtype of ovarian cancer, and within the subtype the proportion of patients presenting with advanced disease requiring chemotherapy is relatively small. Increased expertise through the introduction of site-specialised multidisciplinary care together with increasing diagnostic acumen among specialist gynaecological pathologists may reduce the number of patients diagnosed with mEOC still further (we estimate entering one to two patients per year from our centre, which is a large regional referral centre serving a population of 2.75 million). There is therefore an imperative for all eligible patients to be approached for the trial. The involvement of the GCIG is clearly imperative for the success of the trial, and at an institutional level, there is clearly a key role for the pathologist to play in accurately identifying patients who may be appropriate for the trial.
There is also a potential weakness to the trial rationale in that the data supporting the use of oxaliplatin and capecitabine are based upon trial evidence derived from all histological types of mCRC cancers, not just the mucinous subgroup. Several, albeit small, retrospective analyses of colorectal cancer clinical trials suggest lower response rates in mucinous, compared with adenocarcinomas, with fluorouracil,39 oxaliplatin and irinotecan chemotherapies.40 Therefore actual response rates to these drugs for mEOC may also be lower than predicted.
Future approaches to the treatment of mEOC
Taking a broader perspective, the management of mEOC represents part of a wider debate regarding the approach to the treatment of all cancers. We are increasingly stratifying treatment based on pathological or molecular features. We have, for many years, stratified hormonal manipulation for breast cancer based on oestrogen receptor expression41 and more lately, used HER 2-neu expression to direct the use of trastuzumab.42–44 A more refined understanding in other cancers illustrates how advances in understanding molecular biology may drive progress. Key examples of targeted therapies include the tyrosine kinase inhibitors imatinib for GI stromal tumours with c-KIT receptor mutations45 and gefitinib in patients with specific activating EGFR mutations in non-small cell lung cancer.46 Notable improvements in response rates and survival in these cancers may help to overcome the scepticism concerning the use of targeted therapies in solid tumours.
In recognition of these advances, the translational arm of the ‘mEOC’ trial offers an opportunity to explore the progress of ‘personalised medicine’.
Increasingly, recognising specific genetic alterations that drive tumour growth may allow us to identify subsets of patients who will benefit from specific therapies. This paradigm may hold greater importance for rare tumours such as mEOC, where small numbers create a challenge to gather randomised trial data to support treatment. The rapid progress in DNA sequencing technology and rapidly falling cost may offer the potential for routine rapid assessment of the cancer genome, within a few decades.47 Microarray technology is already being employed to identify tissues of origin in cases of unknown primary cancers48 ,49 and may be applicable to differentiate mEOC from sEOC and also from tumours metastatic to the ovary in the future. Existing data from gene expression profiling of mEOC indicate activated genes affecting drug resistance, signal transduction, proliferation and apoptosis.50 Further refining knowledge of these pathways may lead to the identification of key mutations driving the oncogenic phenotype, and therefore provide potential targets for intervention. Improvements in functional (protein) based assays may also have a significant role to play.
The intelligent application and interpretation of genomic data in the context of clinicopathological details can now be allied to the proliferation of drugs targeting various signalling pathways. This approach promises a future of individualised treatment options and would have significant implications for clinical and pathological practice. An emerging demand for a tumour sequencing service and concomitant bioinformatic/interpretative service would lead to a number of challenging questions: Should services be provided based on histological or molecular phenotype? How would this work practically? Who would determine the standards for genomic testing? How would physicians be educated regarding the interpretation of molecular data?
It is likely that there would need to be a centralised provision of services, perhaps on a regional basis initially, to offer the necessary sequencing technologies and bioinformatics expertise to achieve standardised diagnostic outcomes. Even if samples are to be processed centrally, collection, storage and transportation remain a significant issue. Fresh frozen tissue for gene expression analyses in particular can be challenging.49 Therefore, keen attention to sample handling requirements (eg, tissue procurement, time to freezing and sample storage) and provision of standardised (possibly web based) protocols will be vital for effective use of tissue samples.
While personalised medicine may form a greater part of our future practice, the present day requires support for trials such as ‘mEOC’ to inform the evidence base for a rare ovarian cancer subtype and translational research to stimulate discovery of future treatment options.
JDN and JS contributed equally to writing of this article.
Competing interests None.
Provenance and peer review Commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.