Ovarian Cancer: Advances in First-Line Treatment Strategies with a Particular Focus on Anti-Angiogenic Agents
- First Online:
- Cite this article as:
- Collinson, F.J., Seligmann, J. & Perren, T.J. Curr Oncol Rep (2012) 14: 509. doi:10.1007/s11912-012-0274-4
- 443 Views
Ovarian cancer is an important health concern worldwide. The majority of patients present with advanced disease, and despite initial chemosensitivity, most relapse and die from their disease. Better therapeutic options are urgently required. Maximal surgical debulking in combination with platinum/taxane chemotherapy has been the standard of care in advanced ovarian cancer since the mid-1990s. Trials investigating the addition of a third chemotherapeutic agent have disappointingly failed to demonstrate benefit. Intra-peritoneal therapy demonstrated improvements in outcomes in some trials, but at the cost of increased toxicity and inconvenience. Encouragingly, prospective data has now demonstrated benefits with bevacizumab in both the first-line and relapsed settings; however, interpretation is complex, particularly considering recent data demonstrating non-inferiority of neo-adjuvant chemotherapy with delayed primary surgery, and other data demonstrating a substantial improvement in outcome as a result of first-line paclitaxel dose fractionation. This article reviews the recent advances in ovarian cancer treatment and discusses current management and key areas for future research.
KeywordsOvarian cancerAnti-angiogenicVEGF inhibitorsBevacizumabIntra-peritoneal therapyDose-fractionationDose-densityNeo-adjuvant chemotherapyInterval debulking therapyParp inhibitorsOlaparib
Epithelial ovarian cancer (OC) remains a significant cause of morbidity and mortality worldwide, with an estimated 21,880 new cases and 13,850 attributable deaths in 2010 . Despite significant effort invested in clinical trials there have been only limited improvements in cure rates . Better treatment strategies are urgently needed.
A series of international consensus meetings have aimed to define standard care in OC. Since the inaugural meeting, the recommended treatment paradigm for advanced disease has been debulking surgery followed by platinum-based chemotherapy, initially cisplatin-based and more recently with carboplatin/ paclitaxel [3, 4]. The most recent consensus meeting in 2010 acknowledged that acceptable variations existed, but advised that their use must be supported by at least one trial demonstrating superiority or non-inferiority.
Initially, improvements to standard platinum/taxane chemotherapy were sought through the addition of a third chemotherapeutic agent with demonstrable activity in the relapsed setting. First line trials including over 10,000 women investigated the addition of drugs, including pegylated liposomal doxorubicin (PLD), topotecan, gemcitabine and epirubicin [5–10], but unfortunately demonstrated only increased toxicity without survival benefits.
A more successful strategy seems likely to be the addition of targeted therapies to chemotherapy. Encouraging data are now available from two first-line trials, both of which reported improvements in outcomes with the addition of bevacizumab, a recombinant humanised monoclonal antibody targeting vascular endothelial growth factor (VEGF), to standard chemotherapy. Two further trials of bevacizumab in the relapsed setting have also produced supportive evidence, but have raised further questions. At this time, the optimal integration of bevacizumab into routine clinical practice remains unclear. Encouraging data are also available relating to the use of neoadjuvant chemotherapy (NACT) with interval debulking surgery (IDS), paclitaxel dose-fractionation, intraperitoneal (IP) chemotherapy, and other novel targeted agents; approaches for which mature data with bevacizumab are not yet available.
This paper will review the important recent trials involving first-line treatment strategies in ovarian cancer, with a particular focus on VEGF inhibitors, and key areas of current and future research in this challenging area.
Vascular Endothelial Growth Factor Inhibitors
Angiogenesis is an essential element of tumour growth and development and is critically dependent on VEGF. Inhibition of VEGF restores the balance of pro- and anti-angiogenic factors, normalising blood vessel structure, improving drug delivery and reducing metastatic potential . In OC, angiogenesis is associated with tumour growth, ascites formation and metastasis [12–14].
This preclinical rationale led to a number of promising Phase II studies utilising bevacizumab [15–20]. Unlike other solid tumors, impressive response rates (16–21 %) were seen in heavily pretreated patients with single agent treatment [15, 18]. Such data, along with preclinical evidence of synergy with cytotoxic chemotherapy [11, 21], led to combination studies [16, 18], including one demonstrating the feasibility of combining bevacizumab with carboplatin/paclitaxel in the first-line setting .
Comparison of ICON7 and GOG218 first-line bevacizumab ovarian cancer trials
I: Carboplatin AUC 5/6 + paclitaxel 175 mg/m2
I: Carboplatin + paclitaxel
II: Carboplatin + paclitaxel + bevacizumab (concurrent 5/6 cycles and maintenance up to 12 cycles q 3 weekly)
II: Carboplatin + paclitaxel + bevacizumab (concurrent only)
III: Carboplatin + paclitaxel + bevacizumab (concurrent and maintenance up to 16 cycles q 3 weekly)
Number of patients
PS 0/1/2 (%)
Stage I/IIA/≥IIB (%)
Grade 1/2/3 (%)
PFS (also powered for OS)
PFS (was changed from OS initially)
Intent to start ≤ or > 4 weeks after surgery
High risk early stage disease (I/IIA) or advanced disease (optimally or sub-optimally debulked)
Initially planned for only sub-optimally debulked disease, but eligibility extended to include optimally debulked disease (macroscopic)
Up to 18 cycles
Up to 22 cycles
II: 21.8 (HR 0.81, 95 % CI 0.7–0.94, p = 0.004)
II: 11.2 (HR 0.908, 95 % CI 0.76–1.04, p = 0.08) (not significant)
III: 14.1 (HR 0.717, 95 % CI 0.23–0.82, P < 0.0001) *
GOG218, a 3-arm, placebo-controlled trial, reported, at a median follow-up of 17.4 months, a statistically significant improvement in progression-free survival (PFS) from the addition of bevacizumab concurrently and as maintenance therapy for 15 months (arm 3), when compared to chemotherapy alone (arm 1) (10.3 vs. 14.1 months, HR = 0.72; 95 %CI = 0.63–0.83; p < 0.001). This advantage extended to 6 months (12.0 vs. 18.0 months, HR = 0.645; 95 % CI = 0.551–0.756; p < 0.001) when, as requested by the regulatory authorities, the analysis was censored for asymptomatic CA125 progression. There was no significant PFS advantage for bevacizumab used concurrently only (arm 2) (10.3 vs. 11.2 months (HR = 0.9; 95 % CI = 0.76–1.04; p = 0.16). Despite the PFS advantage, there was no statistically significant difference in OS between arm 1 and arm 3 (39.2 vs. 38.7 months, HR = 0.915; 95 % CI = 0.727–1.152; p = 0.45) [23••].
ICON7 was a 2-arm, open-label study of concurrent and maintenance bevacizumab conducted in early-stage patients requiring chemotherapy, as well as in those with advanced disease. At a median follow-up of 28 months, PFS was 17.3 months in the control arm and 19.8 months in the bevacizumab arm, with clear evidence of non-proportional hazards (p < 0.001). Interestingly the maximum separation of the PFS curves occurred at around 12 months, corresponding with the point at which bevacizumab was discontinued, the curves finally converging at 22 months. In the setting of nonproportional hazards, the restricted mean PFS is a more appropriate value to report (i.e. the difference in area under the PFS curves), giving a difference of 1.7 months (24.1 vs. 22.4 months, HR = 0.87; 95 % CI = 0.77–0.99; p = 0.04) [22••].
The overall benefit seen in ICON7 was less marked than in the GOG218 study, due to the inclusion of a large proportion of patients with earlier stage and lower risk disease, illustrated by comparing PFS and OS outcomes of the control-arm populations in the two trials. However, in a predetermined but exploratory analysis of a ‘high-risk’ subset of 465 patients similar (but not identical) to those recruited into GOG218 (stage III disease with >1 cm residuum after debulking, or stage IV disease), the statistical test for interaction suggested that the size of the bevacizumab effect differed between this subgroup and the rest of the population (p = 0.06), implying that in some way, the ‘high risk’ population represented a distinct biological subset. Median PFS was improved from 10.5 vs. 16 months and the restricted mean PFS from 14.5 to 18.1 months at 42 months (HR = 0.73; 95 % CI = 0.60–0.93; p = 0.002), a comparable effect size to that seen in GOG218. Final OS data for the entire trial population is awaited, but in an interim analysis requested by regulatory authorities, median OS was improved from 28.8 to 36.6 months for the ‘high-risk’ population (HR 0.64; 95 % CI = 0.48–0.85; p = 0.002). There was no difference in OS for the non-‘high-risk’ subset. The improvements in outcomes attributable to bevacizumab in this ‘high-risk’ subset within ICON7 have now been widely considered by clinicians who consider them to be not just statistically, but also clinically significant.
In both studies, the addition of bevacizumab led to some increased toxicity, although it was generally well tolerated and consistent with the experience from other clinical trials (hypertension, proteinuria and bleeding). Initial concerns regarding gastrointestinal perforation rates (5.4 % in a meta-analysis of patients with OC24)  were not realised; the higher than expected rates observed in the Phase II studies likely related to the heavily pretreated population, commonly with direct OC bowel involvement.
These encouraging data for bevacizumab in the first-line setting are confirmed and reinforced by more recent results from trials conducted in the relapsed setting. The OCEANS trial is a randomised placebo-controlled trial in patients with platinum-sensitive relapsed disease, comparing carboplatin/gemcitabine plus concurrent/maintenance bevacizumab until progression to carboplatin/ gemcitabine alone. There was a statistically significant benefit in terms of PFS, 8.4 vs. 12.2 months (HR = 0.484; 95 % CI = 0.388–0.605; p < 0.0001), along with benefits in terms of RR (57.4 % vs. 78.5 %, p < 0.0001) and duration of response (7.4 vs. 10.4 months, HR = 0.534; 95 % CI = 0.408–0.698) .
The AURELIA trial demonstrated the benefit of bevacizumab in platinum-resistant and refractory relapsed disease. This randomised trial of 361 patients compared physician’s choice of standard chemotherapy (PLD, paclitaxel or topotecan) with or without concurrent/maintenance bevacizumab until progression. It met its primary endpoint and demonstrated an increase in PFS from 3.4 to 6.7 months (HR = 0.48; 95 % CI = 0.38–0.6; p < 0.001). There was also a statistically significant improvement in both RECIST response and CA 125 response with the addition of bevacizumab (11.3 % vs. 27.2 %, p < 0.001 and 11.6 % vs. 31.8 %, p < 0.001 respectively). Concerns regarding increased rates of gastrointestinal perforation in this more advanced population were not realised with 0 % in the control arm vs. 1.7 % (n = 3) in the bevacizumab arm, although patients perceived to be at risk of gastrointestinal perforation were excluded from participating .
With four positive Phase III trials, bevacizumab has a clear role in the treatment of OC; however, its optimal clinical setting is uncertain. In order to determine how best to use bevacizumab, it is necessary also to consider other data regarding the first line management of the disease.
It is being increasingly recognised that not all patients presenting with ovarian cancer are necessarily best treated by primary surgery. It is well established that OC outcomes are related to the success of initial surgery, in terms of the amount and size of disease remaining . However, it is less clear whether surgery has to be the initial treatment option, and many surgeons are concerned about operating on patients in whom they will leave significant residual disease. The chemosensitivity of OC, and the successful use of NACT in other disease settings have led to trials aiming to reduce tumour volume prior to surgery. EORTC55971 randomised 718 patients to primary surgery followed by chemotherapy, or to NACT followed by IDS and further chemotherapy. There was a higher optimal debulking rate in the IDS arm (83 % vs. 48 %), with associated lower post-operative mortality, reduced operative duration and reduced surgical complications (grade 3 hemorrhage, venous complications and infections). The trial was powered for non-inferiority, and no difference was seen in terms of either OS (HR = 0.98; 90 % CI = 0.84–1.13) or PFS (HR = 1.01; 90 % CI = 0.89–1.15), confirming no detriment to patients with operable disease from delayed surgery. The results of the similar UK CHORUS trial (n = 550), which has recently competed accrual, are awaited and joint analysis with the EORTC55971 trial is planned. Additionally, the JGOG trial (JGOG0602) has completed accrual and has compared debulking surgery followed by eight cycles of chemotherapy, with four cycles of NACT followed by IDS and then a further four cycles of chemotherapy. The optimal number of NACT cycles has not been addressed prospectively, but a meta-analysis of retrospective studies did demonstrate a negative relationship with survival with increasing duration of NACT .
Ideally, the results of the outstanding trials should be awaited prior to routine uptake of this approach, since these trials may help to identify sub-populations most likely to benefit from this strategy. The encouraging data from EORTC55971 has, however, led to the widespread adoption of the NACT approach in many centres around the world. Unpublished data from the authors’ own centre in Leeds UK, a comprehensive cancer centre, suggest that around 47 % of patients newly presenting with OC are now treated with NACT, depending on the extent and distribution of disease at presentation, or concerns regarding comorbidity .
Route and Schedule of Drug Administration
Other research to consider when integrating bevacizumab into routine clinical practice is the route and schedule of cytotoxic drug administration.
There is a well-justified rationale for using IP chemotherapy in optimally debulked OC. It enables higher doses of active drugs to be delivered to the peritoneal cavity, the commonest site of relapse, with relative sparing of normal tissues [30–32]. Eight randomised Phase III trials have been conducted [33–38], and three meta-analyses subsequently performed [39–41]. A pooled estimate of the OS treatment HR of 0.79 (95 % CI 0.69–0.90) was estimated in favour of IP treatment, with a 9 month PFS advantage, although limitations included the inclusion of small trials (< 200 patients) and the fact that only three trials had statistically significant results in favour of IP chemotherapy.
Despite an U.S. National Cancer Institute alert in 2006 supporting the use of IP chemotherapy in patients with optimally debulked disease, IP chemotherapy has not been universally adopted. Clinicians’ reluctance may be due to uncertainty in the validity of the above results (all of the studies have used non-contemporaneous arms and variable paclitaxel doses), the statistically significant increase in toxicity (in GOG172, only 42 % of patients completed the planned six cycles of IP chemotherapy, compared with 90 % in the IV arm ), inferior quality of life (QoL) seen during treatment , and practical and economic implications. A cross-trial comparison of GOG172 and GOG158 (the landmark trial showing that carboplatin/paclitaxel was at least as good as cisplatin/paclitaxel) demonstrated an estimated reduced benefit of IP therapy over standard carboplatin/paclitaxel of 3.1 months PFS and 8.2 months OS (compared to the 15.9 month benefit shown in GOG172), with no difference in 2 and 4 year OS . One key confounding factor is the variability in IP schedules, commonly with a day 8 paclitaxel dose incorporated. Based on the promising data from trials of paclitaxel dose-fractionation , it has been hypothesised that the benefit attributed to IP therapy may in part be due to the altering of drug scheduling or intensity, and may be achieved by more simple means.
As yet, an agreed optimal IP regime with acceptable toxicity and evidence of superiority over standard carboplatin/paclitaxel has not been defined. Well-designed Phase III trials are required prior to more widespread acceptance, particularly in assessing its role alongside novel biological agents and more intensive IV regimes. Most of the published trials have utilised IP cisplatin; however, the recently completed GOG252 trial has used IP carboplatin, and included bevacizumab; results are awaited. The role of IP chemotherapy after NACT/IDS is also currently being addressed in a number of prospective trials. The PETROC/OV21 Phase II/III study will to compare IP/IV chemotherapy with current standard practice. It is randomising patients with surgically optimally debulked disease after NACT to IP/IV chemotherapy vs. IV carboplatin/paclitaxel. In the initial phase, a comparison between IP cisplatin and carboplatin will select the most deliverable IP arm for the Phase III study.
Weekly Paclitaxel Chemotherapy
Recent data supporting the use of alternative paclitaxel scheduling are interesting, and demonstrate similar benefits to those previously shown in breast cancer . Preclinical studies have demonstrated that frequent low doses of paclitaxel have anti-angiogenic and pro-apoptotic effects in xenograft models . Significant activity (RR 25–40 %) has been demonstrated in relapsed OC, including in disease resistant to three-weekly paclitaxel [47, 48]. A number of different weekly dosages and schedules have been utilised, most reporting good activity, but often with significant haematological toxicities [48–50].
Katsumata et al. performed a Phase III trial in 631 chemo-naïve patients with stage II–IV OC, randomised to standard three-weekly carboplatin/paclitaxel (AUC6/180 mg/m2) or weekly paclitaxel (80 mg/m2) plus three-weekly carboplatin AUC6. After a median follow-up of 29 months, an unprecedented 10.8 month improvement in PFS was reported in the weekly chemotherapy arm (17.2 vs. 28 months, HR = 0.71; 95 % CI = 0.58–0.88; p = 0.002). Haematological toxicities were more frequent in the weekly chemotherapy arm, leading to a higher proportion of patients discontinuing treatment compared with the control arm (61.5 % vs. 72.7 %); however, a higher mean delivered paclitaxel dose was still achieved in the experimental arm . Updated results (median follow-up 6.4 years) confirm the previous benefits shown (PFS HR = 0.76; 95 % CI = 0.62–0.91; p = 0.004 and OS HR = 0.79; 95 % CI = 0.63–0.99; p = 0.039) .
One criticism of this study was that less than half of women included were successfully surgically optimally debulked (compared to 74 % in ICON7 [22••]). Also, even in the control arm, fewer patients than expected completed six cycles of chemotherapy. These factors may have adversely affected the outcomes in the control group. More patients in the control group received over six cycles of treatment (73 % vs. 62 %), potentially due to the additional toxicity in the experimental arm.
The plethora of recent data has led to the design of a number of new trials and studies, designed to investigate how to maximise the potential of all of these encouraging potential treatment approaches. The data concerning paclitaxel dose fractionation are being validated contemporaneously in a number of trials (ICON8-ENGOT/OV-13, MITO7 and GOG262), some of which are also investigating platinum dose-fractionation, based on the rationale of improved tolerability .
GOG262, which closed earlier this year, used arms similar to the JGOG study, but importantly also allowed bevacizumab, but as a per patient choice rather than as a randomised arm, so conclusions may be limited. MITO7 is comparing weekly carboplatin and paclitaxel (AUC2/60 mg/m2 rather than 80 mg/m2 weekly) and has also recently completed accrual.
ICON8-ENGOT OV-13 is a 3-arm trial that is comparing standard treatment with either three-weekly carboplatin and weekly paclitaxel or weekly carboplatin and paclitaxel; this trial also allows the use of NACT/IDS and is currently being redesigned to allow the inclusion of bevacizumab as a randomised option, based on the premise that weekly paclitaxel may also work via an anti-angiogenic effect that may or may not be synergistic with bevacizumab. The results of the trials and future meta-analysis will help further define the use of weekly paclitaxel in this setting.
Future Bevacizumab-Related Research Questions
Further questions remain with respect to bevacizumab dose and duration. It is unclear whether all patients should be treated with bevacizumab, or only those at high risk of disease progression, or only those who have relapsed.
The need for concurrent bevacizumab during chemotherapy has been questioned, although the improved response rates with bevacizumab in ICON7 (complete/partial response 67 % vs. 48 %, p < 0.001), and the early separation of the OS curves for the ‘high-risk’ subgroup whilst patients were still receiving chemotherapy, suggest synergy with chemotherapy and a proportion of patients potentially salvaged by early exposure to bevacizumab.
The clearest evidence of bevacizumab benefit has been shown in the first-line setting for patients at highest risk of progression (GOG218 and the ‘high-risk’ subgroup of ICON7); based on available evidence to date, this seems the most appropriate group to treat. However, currently in many centres this sub-population is often treated by NACT/IDS, and there is as of yet no trial data informing the optimal scheduling of bevacizumab and surgery. There are particular concerns about the use of bevacizumab in close relationship to surgery because of the potential negative effect of anti-angiogenic drugs on wound healing. It is therefore unknown whether bevacizumab should be incorporated into NACT, utilised only post-IDS, or at relapse. The adaptation of ICON8 to include bevacizumab may answer this question. In the meantime, many centres are adopting a pragmatic extrapolation of the data, reserving bevacizumab for patients treated with neo-adjuvant chemotherapy who have significant residual disease after IDS.
Both ICON7 and GOG218 excluded those patients with a poor performance status, who are often the subset with significant ascites and oedema and unfit for surgery. It is these patients who are hypothesised to have the most VEGF-driven cancers, and it is intriguing to postulate whether this group might therefore derive significant benefit from cautiously administered bevacizumab. There are, however, currently no data to support its safe or effective use in this setting, and due to the fragility and numbers of these patients, prospective studies may be difficult to robustly perform.
In both ICON7 and GOG218, the maximal treatment effect occurred when maintenance bevacizumab was stopped. It is theorised that additional benefit may have been gained if bevacizumab been continued until disease progression, and that the reason no clear benefit was seen in better prognosis patients was due to premature bevacizumab cessation prior to patients reaching their median PFS. The first-line setting the ENGOT-OV15/AGO-OVAR 17 trial, which is comparing 15 with 30 months of bevacizumab, will inform the optimal duration question. The preclinical mechanistic data and results from the OCEANS and AURELIA trials support the use of bevacizumab until progression, but we do not have direct evidence to support this in the first-line setting. Continuation to progression is, however, at the expense of a slight reduction in QoL, significantly increased expense and patient inconvenience, all requiring careful consideration. Patients not receiving bevacizumab first-line treatment clearly still stand to gain substantial advantage from its use at relapse.
It is interesting to speculate why no OS advantage was seen in GOG218, despite it being conducted in patients at high risk of recurrence. One possible explanation is the unblinding at the time of relapse, and at the time of analysis of the primary endpoint, when there were still a significant proportion of patients receiving blinded therapy, leading to significant crossover (47 %). The survival outcome for patients in the control arm and research arm of GOG218 are very similar to those in the research arm of the high risk subset of ICON7, whereas those in the control arm of ICON7, where there was very little cross-over, were substantially inferior. The significant cross-over further supports the hypothesis that relapsing patients who do not receive bevacizumab first-line can to some extent be salvaged by its subsequent use.
Further questions will arise, for example following first-line bevacizumab, should bevacizumab be restarted at relapse, or even be continued past progression with alteration in chemotherapy, as suggested in other solid tumour sites . The results of the ENGOT-OV17/MITO16/MaNGO2 trial, which will compare bevacizumab plus chemotherapy to chemotherapy alone at progression, after first-line chemotherapy plus bevacizumab, will help to answer this.
Other Anti-Angiogenic Agents
We are currently at an early stage in our understanding of the molecular biology of OC. The genomic revolution has provided a plethora of new drugs and agents to test. We currently understand that it is early tumour development that is most reliant on VEGF pathways, and after VEGF inhibition other pathways may become activated. This provides a rationale for potential increased efficacy with multi-targeted, anti-angiogenic drugs [54–56]. Thus far, these agents have not looked promising as single agents [57, 58], and combinations of targeted therapies including bevacizumab have proved intolerable .
Nintedanib (BIBF1120), a triple angiokinase inhibitor targeting three receptor classes, VEGF, platelet-derived growth factor receptor (PDGF), and fibroblast growth factor receptor (FGFR), appears promising. A randomised Phase II trial in 84 patients who had responded to initial chemotherapy for platinum-sensitive relapsed disease, demonstrated prolonged PFS in patients treated with maintenance nintedanib compared to placebo; rates of maintained PFS at 36 weeks were 16.3 % vs. 5.0 % in the nintedanib and placebo groups (HR = 0.65; 95 % CI = 0.42–1.02; p = 0.06). Predominant toxicities were gastrointestinal . The ongoing LUME-Ovar-1 Phase III trial investigates the role of concurrent and maintenance nintedanib in combination with first-line chemotherapy.
Cedirinib, another multi targeted anti-angiogenic agent is currently under investigation in the ICON6 Phase II/III clinical trial, which investigates its role as concurrent and maintenance therapy in platinum-sensitive relapsed OC. Other anti-angiogenic drugs under investigation in OC include further tyrosine kinase inhibitors (dorafenib, vandetanib and j1-101), an anti-angiopoetin antibody (AMG386) and VEGF-trap (aflibercept).
Other Promising Drugs
There are further promising novel drugs in drugs in development other than anti- angiogenic agents. The testing of Poly (ADP-ribose) polymerase (PARP) inhibitors represents an important step towards personalised therapy. To date, their use is confined to cancers associated with germline BRCA1/2 mutations, affecting proteins essential for maintenance of genomic integrity . ICEBERG3, a multicentre Phase II trial, evaluated olaparib compared to PLD in platinum-sensitive relapsed disease . There was no significant difference in PFS of the combined olaparib arms and PLD arm (HR = 0.88; 95 % CI = 0.51–1.56; p = 0.66); however, RRs with olaparib were similar to another Phase II study (41 %) . The authors noted a higher than anticipated benefit in the PLD arm (7.1 months; 95 % CI = 3.7–10.7 months) compared to 4 months in a previous study , perhaps suggesting that BRCA1/2 mutated OC may have a more favourable response to PLD than other patients. Additionally, it has been suggested that imbalances in patient characteristics between the trial groups may have underestimated the benefit of olaparib .
Another Phase II trial has evaluated the role of maintenance olaparib in patients with relapsed platinum-sensitive BRCA1/2-mutated OC . An improvement in PFS was seen with olaparib compared to placebo (4.8 months vs. 8.4, HR = 0.35; 95 % CI = 0.25–0.49), and whilst mature OS data is awaited, no initial OS benefit was seen. In both trials, olaparib has been well tolerated. Further trials, combining PARP inhibitors and chemotherapy based on preclinical synergism, are in progress [67–69]; however, problems have been encountered due to dose-limiting myelosuppression.
There is increasing recognition of a BRCA1/2 deficient OC phenotype, which may represent a separate clinical entity to sporadic disease; patients are more likely to be younger, have serous histology, platinum-sensitive disease, visceral metastases and improved survival [70, 71], but may lack a strong family history. This represents the potential for a molecular subclassification in OC, which, as in other cancers, could lead to tailored therapeutic strategies. Additionally, as family history detects only a fraction of patients with a BRCA1/2 mutation , recognising this phenotype and confirming a BRCA1/2 mutation may increase the number of patients suitable for trials with olaparib, as well as providing important information for patients and their families. In the future, further sub-classification based upon additional phenotypic and molecular information may also be possible.
Of increasing interest and importance is the identification and validation of predictive biomarkers identifying groups of OC patients most likely to benefit from specific treatment strategies. Predictive markers are routinely utilised in other cancer sites; for example, epidermal growth factor receptor mutations as a predictive marker for gefitinib in non-small cell lung cancer , and KRAS mutations as a negative predictive marker for cetuximab and panitumumab in advanced colorectal cancer . Molecular selection of patients leads to improvements in efficacy and cost-effectiveness, but also avoidance of toxicity and the ability to utilise alternative therapeutic strategies in those unlikely to respond.
Increasing understanding of the molecular mechanisms involved in tumour growth and development has identified candidate biomarkers for anti-angiogenic therapies, but as yet none have been robustly validated. Potential biomarkers include VEGF levels in primary tumour tissue , VEGF polymorphisms , serum VEGF levels , and VEGF receptor expression levels . However, none of these biomarkers have shown consistent predictive values across studies, and prospective validation is necessary to identify which, if any, are appropriate for clinical use.
The benefit seen with first-line bevacizumab in the high risk setting, or in second-line therapy in patients at inevitable risk of further relapse, suggests that these clinical phenomena are surrogates for aspects of ovarian cancer biology still to be determined; hopefully, future translational studies will identify these patients more accurately.
Both ICON7 and GOG218 prospectively collected tissue and blood samples. It is hoped that ongoing translational work will provide the opportunity not only to validate or refute candidate prognostic and predictive biomarkers, but also provide the opportunity for further biomarker discovery, ultimately leading to the ability to determine patients who have the greatest chance of benefit from this expensive new approach.
The use of NACT/IDS allows post-NACT tissue to be compared to initial biopsy tissue for translational research questions, particularly relating to drug resistance and sensitivity. Trials such as ICON8, which includes patients treated in the neoadjuvant setting, will take advantage of this opportunity.
Addressing OC Heterogeneity
OC covers a heterogeneous range of diseases, and it is becoming apparent that specific subtypes are likely to need specific tailored therapies. The survival benefits with weekly paclitaxel chemotherapy were not seen in the subset of clear cell or mucinous OC51. Alternative strategies are necessary, and are currently being tested, for example, in the international mEOC trial in mucinous disease. Over recent years, the differing molecular profiles of high and low grade serous OC have been defined [79, 80], and this information is now being used to assist in the design of clinical trials; for example, low grade tumours are usually not associated with p53 mutations, but often have activating mutations of the MAPK/MEK pathway, and there are ongoing studies in this population using MEK1/2 inhibitors, e.g. AZD6244.
Despite improved understanding of the pathogenesis of OC in recent years, and prior to the results of the recent positive prospective trials with bevacizumab, IP therapy and alternate paclitaxel scheduling, there had been few significant advances in this disease. The recent positive results with bevacizumab have identified the first new effective drug in 15 years of research. These results are extremely exciting but have raised a multiple questions regarding how best to incorporate these strategies into routine practice and also research.
Ongoing prospective randomised trials will help identify which patients should receive bevacizumab in first-line rather than in the relapsed setting, the optimal duration of bevacizumab therapy, and how should it be combined with alternative treatment strategies including weekly paclitaxel, NACT/IDS and IP treatment. In the meantime, patients at high risk of relapse at first presentation or at first relapse should be strongly considered for bevacizumab therapy.
Translational research alongside future clinical trials will be essential to assist in defining patient cohorts most suited to the different therapies available. The increasing use of NACT and IDS facilitates the availability of tissue after systemic therapy exposure, and this should be exploited. Ongoing laboratory research continues to discover more promising and relevant targets for drugs, including the phosphoinositide 3-kinases and Ras/Raf pathways and the Notch/hedgehog signaling pathways, and trials now are beginning with dugs targeting these pathways.
It seems that there is a broad seam of questions to be exploited by future well-designed prospective clinical trials ideally incorporating biological questions as part of their design.
No potential conflicts of interest relevant to this article were reported.