Annals of Oncology 27 (Supplement 1): i58 i62, 2016 doi:10.1093/annonc/mdw088 Optimal primary therapy of ovarian cancer M. A. Bookman* Arizona Oncology and US Oncology Research, Tucson, AZ, USA Background: Epithelial ovarian cancer continues to have the highest case fatality ratio of all gynecologic cancers, in spite of ongoing advances in risk-assessment, genomics, tumor biology, cytoreductive surgery, chemotherapy, and molecular-targeted interventions. Primary treatment options for advanced-stage disease not only should reflect current best standards, but also need to be tailored for individual patients, with consideration of local resources. Methods: Formulation of recommendations for optimal primary therapy based on a selective review of data from completed randomized trials, analysis of ongoing trials, and integration with current tumor biology, within the context of individualized clinical care. Recommendations were presented for discussion during an international meeting of experts in ovarian cancer treatment. Results: Key recommendations include full adjuvant therapy for early-stage high-grade serous cancer; tailored utilization of neoadjuvant chemotherapy based on patient comorbidities, extent of disease, and likelihood of achieving optimal surgical cytoreduction; preferred utilization of carboplatin with weekly paclitaxel as primary therapy; consideration of intraperitoneal cisplatin-based therapy in appropriate patients; avoidance of maintenance chemotherapy; lack of necessity for bevacizumab during primary chemotherapy and primary maintenance; acknowledgement of research opportunities and priorities. Conclusions: Integrated multidisciplinary care, including cytoreductive surgery and platinum-based chemotherapy, remain central to the optimal management of women with advanced-stage ovarian cancer. However, even with recent technical advances, the impact on disease-related mortality is limited, and more attention will be focused on the early integration of research, particularly with neoadjuvant chemotherapy and interval Key words: ovarian cancer, neoadjuvant, cytoreductive surgery, intraperitoneal introduction This overview of optimal primary therapy is based on the judgment of the author, incorporating clinical biology and current research in situations where completed randomized trials often have limited scope, and where mature data are not always available. It is intended as a platform for ongoing dialogue within the context of clinical care and research. Advanced-stage epithelial ovarian, peritoneal, or fallopian carcinoma (EOPFC) remains a serious challenge, with the highest case fatality ratio of all gynecologic cancers, regardless of geography, technical expertise, or financial resources. The majority of tumors are classified as high-grade serous cancer (HGSC), which is associated with uniform functional loss of p53, defective apoptotic signaling, genomic instability, defective DNA repair (including associations with germline mutations in multiple genes), early peritoneal/omental dissemination, hypoxiadriven angiogenic signaling, immunologic evasion, and rapid emergence of drug resistance [1]. However, clinical evidence of *Correspondence to: Dr Michael A. Bookman, Arizona Oncology, 603 N Wilmot Rd, #151, Tucson, AZ 85711, USA. Tel: +1-520-5462817; Fax: +1-520-5462895; E-mail: michael.bookman@usoncology.com deep invasion and spread to extra-peritoneal sites is uncommon, except in the setting of carcinosarcoma, which has been attributed to epithelial mesenchymal transition. Current research has focused on the specialized relationship between tumor cells and the peritoneal-omental microenvironment, including the host immune response. Clinical trials are exploring angiogenesis, drug resistance, DNA repair, immunoregulatory pathways, and stem cell biology, in addition to optimized delivery of chemotherapy and integration with cytoreductive surgery. In spite of these dramatic advances, the core elements of primary therapy, including cytoreductive surgery and platinumbased chemotherapy, remain largely unchanged over the last 25 years. Independent of treatment, a trend toward reduced incidence and mortality has been observed, perhaps related to preventive measures, including broader utilization of oral contraceptives, introduction of risk-reducing surgery, and reduction in the use of post-menopausal hormone replacement [2]. Better clinical management (including supportive care) has contributed to improvements in median progression-free survival (PFS) and overall survival (OS). However, the impact on diseaserelated case fatality ratios for women with advanced-stage disease has been minimal [3]. The Author 2016. Published by Oxford University Press on behalf of the European Society for Medical Oncology. All rights reserved. For permissions, please email: journals.permissions@oup.com.
Annals of Oncology While HGSC is the dominant histology associated with advanced-stage tumors, it is important to recognize that the spectrum of EOPFC incorporates endometrioid, clear cell, lowgrade serous carcinoma, transitional, mucinous, and other less common subtypes [4]. Each of these tumor types is associated with distinct clinical and molecular features, which are guiding the development of tailored treatment regimens and exploratory clinical trials. There is no doubt that the incorporation of moleculartargeted treatment interventions, such as inhibition of tumormediated angiogenesis, modulation of intracellular signal transduction, targeted delivery of therapeutics based on tumorassociated antigens, immune response modifiers, and inhibitors of poly(adp-ribose) polymerase (PARP), will improve outcomes for many women, particularly in the setting of recurrent disease. However, with the possible exception of immune response activation, these interventions are unlikely to have curative potential in the setting of advanced-stage disease. Reductions in mortality will likely require more ambitious strategies to restore defective apoptotic signaling (due to p53 loss-of-function mutations), overcome drug resistance, and target tumor stem-like populations. development of recommendations prevention Increased prevalence of genomic screening for deleterious mutations in BRCA1/2 and other genes will facilitate decisions regarding personal risk management, including early interventions (such as oral contraceptives, dietary changes, lifestyle modifications), late interventions (such as risk-reducing prophylactic oophorectomy and/or salpingectomy), and identification of high-risk family members. Many organizations have already recommended the implementation of universal germline screening for all women with EOPFC [5]. In addition, as the level of genomic detail has expanded, interpretation is becoming more precise, and per-patient costs have been reduced. Over time, these measures should contribute to a small but significant decline in incidence and mortality. Pre-menopausal women should consider using oral contraceptives for at least 6 months, especially if they are nulliparous, or have other risk factors for ovarian cancer. All women with a new diagnosis of ovarian cancer should consider germline screening for mutations in BRCA1/2 and other genes associated with increased risk, regardless of family history. adjuvant therapy of early-stage disease The accuracy of initial surgical staging has improved, and could be further enhanced with high-resolution functional imaging and/or intra-operative molecular probes capable of marking microscopic disease and guiding selective surgical interventions. However, the clinical biology of HGSC is associated with early peritoneal dissemination, and HGSC is responsible for the majority of recurrences among women with apparent early-stage disease. In contrast, early-stage clear cell and mucinous tumors are less likely to be associated with occult peritoneal implantation and are generally cured during primary surgical resection. This is perhaps most clearly illustrated in a retrospective analysis of outcomes according to histologic subtypes from a phase III trial in early-stage disease conducted by the Gynecologic Oncology Group [6]. Based on these findings, it is reasonable to recommend a full course of chemotherapy (6 cycles) for women with early-stage HGSC. timing and scope of cytoreductive surgery With increased disease awareness, improved diagnostics, and high-resolution functional imaging, there will be a modest shift toward reduced-volume advanced-stage disease at diagnosis. At the same time, some women will continue to present with highvolume ascites, malnutrition with hypoalbuminemia, inappropriately low serum creatinine, bowel dysfunction, extensive upper abdominal disease, extra-peritoneal disease, advanced age, and/or associated comorbidities. Current clinical practice has witnessed a substantial increase in the proportion of women who undergo primary neoadjuvant chemotherapy (NACT) with consideration of interval cytoreductive surgery (ICS), based on an analysis of perioperative morbidity and clinical outcomes from two phase III randomized trials in patients with high-risk disease, integrated with the complexities of modern medical care. The goal of primary cytoreductive surgery (PCS) is complete resection of macroscopic disease, leaving no more than microscopic residual disease. The challenge is trying to triage patients for PCS or NACT-ICS based on initial presentation and imaging studies, which can both underestimate and overestimate the extent of tumor. This has triggered active debate regarding selection criteria for NACT and clinical scoring systems, as well as increased utilization of invasive laparoscopic assessment at a number of facilities. However, given the known practical limitations of primary chemotherapy and cytoreductive surgery in the setting of advanced disease with peritoneal dissemination, one might question whether more precise patient selection will actually achieve marked improvement in long-term outcomes at the level of an entire population, or among pre-defined subsets of patients. Hopefully, emerging guidelines will undergo prospective validation, but identification of financial resources to support prospective surgical trials has proven difficult, and it will be important to define strategies to integrate surgical questions within other phase III trials. Clearly, there will always be individual clinical, and resourcedriven, factors that will guide decisions for optimal care of a specific patient. In general, patients who are medically appropriate, and judged to have a high likelihood of achieving optimal cytoreduction, should undergo PCS followed by chemotherapy. Those patients with disease-related comorbidities, or distribution of disease that is not amenable to optimal cytoreduction, should undergo NACT with evaluation for ICS. The role of aggressive cytoreduction also remains controversial, based largely on retrospective studies with inherent selection bias favoring more aggressive surgery. The relationship between preoperative disease burden, extent of surgical effort, and long-term clinical outcomes is important, and it remains unclear if aggressive primary surgery can actually offset the negative predictive value associated with high disease burden at initial presentation. For example, among all patients who achieved microscopic residual disease following PCS, patients Volume 27 Supplement 1 April 2016 doi:10.1093/annonc/mdw088 i59
with high disease burden at diagnosis had long-term clinical outcomes that were similar to patients with suboptimal (macroscopic) cytoreduction [7]. Achieving microscopic residual disease emerged as a significant prognostic factor only among patients with low disease burden prior at diagnosis. At present, the selection of patients for aggressive PCS needs to be balanced by an individualized assessment of perioperative risks, likelihood of resecting all macroscopic disease (including local availability of advanced surgical support), and anticipated clinical benefit, based on the best available data. There are not yet any prospective randomized data to validate the role of aggressive molecular profiling As discussed above, germline mutations in BRCA1/2 and other genes associated with homologous recombination DNA repair should ideally be determined at diagnosis, due to the impact of these mutations on prognosis, prediction of benefit from specific treatment interventions (such as PARP inhibitors), and management of personal and family risk. Genomic analysis of tumor specimens for somatic mutations can potentially double the number of patients who might be candidates for PARP inhibition, and is becoming an important consideration, especially in the setting of recurrent disease. More comprehensive genomic profiling and gene expression analysis has revealed activation of pathways associated with signal transduction, growth regulation, angiogenesis, and invasive behavior, but we do not yet have a broad strategy for molecular targeting of EOPFC in the setting of primary therapy. primary chemotherapy Standard-dose carboplatin and paclitaxel remains a well-tolerated and effective treatment intervention. There is no prospective evidence that higher platinum dose intensity (with or without hematopoietic growth factor support), incorporation of additional cytotoxic agents, or substitution of cytotoxic agents, will improve long-term outcomes for patients with HGSC. As a single agent, weekly administration of paclitaxel appears superior to three-weekly paclitaxel in the management of recurrent disease, and this has been partly attributed to an impact of weekly scheduling on tumor-associated angiogenesis [8]. Of interest, early-phase trials have documented tumor responses with doses as low as 40 mg/m 2 per week, compared with a maximally tolerated dose of 80 100 mg/m 2 per week (as a single agent). Weekly paclitaxel has also been associated with improved outcomes when incorporated in primary platinum-based chemotherapy, in the absence of bevacizumab [9]. The most common dose utilized in randomized trials has been 80 mg/m 2 per week (termed dose-dense ), resulting in a high incidence of dose omissions, dose modifications, clinically significant anemia, and addition of hematopoietic growth factors. Whether clinical benefit (in terms of PFS and OS) is associated more strongly with weekly scheduling, per se, or dose-density (at 240 mg/m 2 per cycle) has not been established. However, the toxicity and efficacy of paclitaxel appears to be more closely linked to variations in scheduling and infusion duration, while the relationship between clinical benefit and dose intensity appears less robust. Annals of Oncology As such, in the absence of randomized data, it seems reasonable to permit flexibility in dose selection for individual patients within a range of 60 80 mg/m 2 per week, depending on comorbidities and overall tolerance. In contrast to paclitaxel, there does not appear to be any clinical benefit from fractionated weekly dosing of carboplatin, and there is concern that weekly dosing may achieve inadequate peak serum levels, and low area under the curve (AUC) for concentration over time, limiting tumor penetration and direct cytotoxicity. The MITO-7 trial compared a standard threeweekly regimen of carboplatin and paclitaxel with weekly carboplatin (AUC = 2) and weekly paclitaxel (60 mg/m 2 ), achieving comparable PFS with improved tolerability on the weekly regimen [10]. The lack of superiority associated with the weekly regimen is most likely related to fractionation of carboplatin, rather than the lower dose of weekly paclitaxel, but this has not been firmly established. The ICON-8 trial is comparing three-weekly carboplatin and paclitaxel (Arm 1), with carboplatin and weekly paclitaxel (Arm 2), or weekly carboplatin combined with weekly paclitaxel (Arm 3), and will hopefully resolve some of this uncertainty. At present, the utilization of carboplatin (AUC = 5 or 6 on day 1) with weekly paclitaxel (60 80 mg/m 2 days 1, 8, and 15) on either a 21-day (continuous) or 28-day (1 week without treatment) schedule is reasonable, and can be tailored to meet the needs of individual patients, based on the timing of surgery, extent of residual disease, vital organ function, and comorbidities (such as malnutrition). intraperitoneal (IP) chemotherapy Randomized trials, with some caveats, have documented improved long-term clinical outcomes associated with IP cisplatin (with or without IP paclitaxel), at the expense of increased toxicity [11]. Multiple practical modifications have been adopted over the years to improve patient tolerability, including a reduction in the dose of cisplatin and substitution of carboplatin, but only two of the modified regimens are undergoing validation within prospective randomized trials, NRG-GOG0252 (NCT00951496) and JGOG3019 ipocc (NCT01506856). While regional pharmacokinetic advantage, or dose-intensity, is frequently cited as the primary rationale for IP therapy, the actual mechanism remains unknown. The relationship between tumor implants, the omentum, and the peritoneal cavity is highly specialized, and it is likely that IP drug administration has an impact on the tumor microenvironment, independent of any direct impact on the tumor. In fact, the development of loculated adhesions, fibrosis, high interstitial pressures, and leaky capillaries can prevent cytotoxic drugs from directly penetrating many tumor nodules. In addition, both cisplatin and carboplatin are rapidly absorbed from the peritoneal cavity, achieving blood AUCs (and associated systemic toxicity) that closely match intravenous (IV) drug administration. As a result, a large proportion of activated platinum that is localized to tumor implants is actually delivered through the circulation, regardless of IP or IV administration. It is frequently stated that IP therapy is only effective in smallvolume disease, but that is actually a bias and misinterpretation of available data, as IP therapy has only been prospectively i60 Bookman Volume 27 Supplement 1 April 2016
Annals of Oncology evaluated in patients with small-volume residual disease, and patients with suboptimal residual disease were excluded from prior studies. In fact, accrual on GOG0104 was extended to enroll a higher proportion of patients with disease 0.5 cm, based on the prevailing bias regarding efficacy in small-volume disease [12]. However, the primary analysis of OS in this population did not demonstrate a statistically significant benefit, while an exploratory analysis of the entire population (which included patients with larger residual disease) achieved a hazard ratio (HR) for OS of 0.77 with P 0.02. Fortunately, two ongoing trials (GOG0252 and JGOG ipocc) have enrolled a substantial number of patients with suboptimal residual disease to address this point. Although IP chemotherapy continues to rely largely on cisplatin, primary IV chemotherapy has transitioned to carboplatin, with additional modifications, including the use of weekly paclitaxel, raising questions about the optimal control arm for comparison with IP therapy. Both ongoing IP trials include IV arms with carboplatin and weekly paclitaxel, and new IP regimens with carboplatin. In addition, GOG0252 includes a modified IP regimen with cisplatin at 75 mg/m 2 and paclitaxel administered over 3 h, based on non-randomized feasibility reports from several centers. We await data to determine how current IV regimens compare with modified IP regimens with either cisplatin or carboplatin. Finally, with increased utilization of NACT and ICS, questions have been raised regarding the potential role of IP therapy following interval cytoreduction. While technically feasible, there are no prospective randomized trials that have validated this approach. Enrollment has been completed on NCIC-OV21 (NCT00993655), comparing IP versus IV platinum-based chemotherapy following NACT and ICS, and data from the primary analysis of PFS are pending. At present, while awaiting data from ongoing trials, it is appropriate to consider IP therapy with a cisplatin-based regimen in patients with limited residual disease following primary maintenance chemotherapy There is no convincing evidence that extended therapy with cytotoxic agents will improve long-term outcomes in high-risk early-stage disease or advanced-stage disease [13, 14]. However, there is one ongoing phase III trial (GOG0212, NCT00108745) comparing observation with maintenance paclitaxel or polyglutamated paclitaxel in patients in clinical complete remission. Accrual was completed in 2014, and primary analysis for OS is pending. At present, maintenance therapy with cytotoxic agents should not be considered as standard-of-care for women in remission following primary therapy. incorporation of anti-angiogenic interventions There is no question that hypoxia-driven angiogenic pathways are activated in HGSC, with frequent over-production of vascular endothelial growth factor (VEGF), which is also associated with the development of ascites. In addition, a number of phase III trials have validated the ability of bevacizumab to target VEGF in different clinical settings, with a consistent improvement in the HR for PFS. Improvements in OS have been more difficult to demonstrate, related to treatment crossover at progression, and access to other related and unrelated treatments for management of recurrent disease. Unfortunately, it has also been nearly impossible to identify clinically relevant predictive biomarkers that could identify patient populations most likely to benefit. Analysis of patients who elected to receive bevacizumab on GOG0262 (NCT01167712) demonstrated no difference in PFS between carboplatin with three-weekly paclitaxel compared with carboplatin with dose-dense weekly paclitaxel, suggesting that bevacizumab might not have additive anti-angiogenic effects when administered with weekly paclitaxel [15]. This hypothesis was further supported in an exploratory subset analysis of patients who elected to receive treatment without bevacizumab. In this analysis, the weekly paclitaxel regimen had improved PFS compared with three-weekly paclitaxel without bevacizumab. Table 1. Summary of author s recommendations: optimal primary therapy, ovarian cancer, 2015 All women with a new diagnosis of ovarian cancer should consider germline screening for mutations in BRCA1/2 and other genes associated with increased risk, regardless of family history. Women with apparent early-stage HGSC a should receive a full course of adjuvant chemotherapy (6 cycles), due to the risk of occult intraperitoneal spread. Patients who are medically appropriate, and judged to have a high likelihood of achieving optimal cytoreduction, should undergo PCS followed by chemotherapy. Those patients with disease-related comorbidities, or distribution of disease that is not amenable to optimal cytoreduction, should undergo NACT with evaluation for ICS. Selection of patients for aggressive PCS needs to be balanced by an individualized assessment of perioperative risks, likelihood of resecting all macroscopic disease (including local availability of advanced surgical support), and anticipated clinical benefit, based on the best available data. Utilization of carboplatin (AUC = 5 or 6 on day 1) with weekly paclitaxel (60 80 mg/m 2 days 1, 8, and 15) on either a 21-day (continuous) or 28-day (one week without treatment) schedule is reasonable, and can be tailored to meet the needs of individual patients, based on the timing of surgery, extent of residual disease, vital organ function, and comorbidities. It is appropriate to consider IP therapy with a cisplatin-based regimen in patients with limited residual disease following primary Maintenance therapy with cytotoxic agents should not be considered as standard-of-care for women in remission following primary therapy. Addition of bevacizumab with primary chemotherapy (and as extended primary maintenance) is not necessary, particularly when using a regimen that incorporates weekly paclitaxel. a HGSC, high-grade serous carcinoma; PCS, primary cytoreductive surgery; NACT, neoadjuvant chemotherapy; ICS, interval cytoreductive surgery; AUC, area under the curve for concentration and time; IP, intraperitoneal. Volume 27 Supplement 1 April 2016 doi:10.1093/annonc/mdw088 i61
Utilization of bevacizumab remains an area of controversy, with international differences in regulatory approval and patterns of care. However, the weight of evidence favors treatment in the setting of recurrent disease, with greater likelihood of achieving clinical benefit for an individual patient through control of disease-related symptoms. In summary, it is not necessary to incorporate bevacizumab with primary chemotherapy, particularly when using a regimen with weekly paclitaxel. clinical research integrated with primary therapy Efficient screening of new targeted agents remains a high-priority, and it would be advantageous to enroll patients receiving NACT in randomized phase II trials to obtain pre- and posttherapy biospecimens. Several proposals are moving forward within national groups, and in collaboration with the pharmaceutical industry, with a focus on tumor stem cell populations, incorporation of metformin, restoring defective apoptotic signaling, and immunoregulatory check-point inhibition. Promising regimens could be expanded as phase III trials, using an adaptive design framework. In addition, 15% of patients will demonstrate clinical evidence of platinum refractory disease during primary therapy. In these patients, the incorporation of early functional imaging (such as perfusion CT, diffusion MRI, or PET) could identify patients with resistant disease within the first 3 weeks of treatment and provide an opportunity to explore alternative regimens during primary chemotherapy. discussion The accelerated pace of molecular diagnostics and drug development has not yet translated into wide-scale success for patients with advanced-stage EOC, although a number of promising interventions are under development. Standard-of-care recommendations for primary therapy in 2015 reflect our current knowledge and interpretation of existing clinical data, as well as the limitations of those data, and the fiscal burdens of healthcare (Table 1). Perhaps most importantly, these recommendations also reflect the high lethality associated with advanced-stage disease, regardless of treatment selection, and the limited success of even the most advanced surgical and chemotherapeutic interventions, highlighting the importance of innovative research. disclosure The author has declared no conflicts of interest. references Annals of Oncology 1. Bowtell DD, Böhm S, Ahmed AA et al. Rethinking ovarian cancer II: reducing mortality from high-grade serous ovarian cancer. Nat Rev Cancer 2015; 15: 668 679. 2. Sopik V, Iqbal J, Rosen B, Narod SA. Why have ovarian cancer mortality rates declined? Part I. Incidence. Gynecol Oncol 2015; 138: 741 749. 3. Sopik V, Iqbal J, Rosen B, Narod SA. Why have ovarian cancer mortality rates declined? Part II. Case-fatality. Gynecol Oncol 2015; 138: 750 756. 4. Bookman MA, Gilks CB, Kohn EC et al. Better therapeutic trials in ovarian cancer. J Natl Cancer Inst 2014; 106: dju029. doi: 10.1093/jnci/dju029. 5. Lancaster JM, Powell CB, Chen LM, Richardson DL. SGO Clinical Practice Committee. Society of Gynecologic Oncology statement on risk assessment for inherited gynecologic cancer predispositions. Gynecol Oncol 2015; 136: 3 7. 6. Chan JK, Tian C, Fleming GF et al. The potential benefit of 6 vs. 3 cycles of chemotherapy in subsets of women with early-stage high-risk epithelial ovarian cancer: an exploratory analysis of a Gynecologic Oncology Group study. Gynecol Oncol 2010; 116: 301 306. 7. Horowitz NS, Miller A, Rungruang B et al. Does aggressive surgery improve outcomes? Interaction between preoperative disease burden and complex surgery in patients with advanced-stage ovarian cancer: an analysis of GOG 182. J Clin Oncol 2015; 33: 937 943. 8. Markman M, Blessing J, Rubin SC et al. Phase II trial of weekly paclitaxel (80 mg/m 2 ) in platinum and paclitaxel-resistant ovarian and primary peritoneal cancers: a Gynecologic Oncology Group study. Gynecol Oncol 2006; 101: 436 440. 9. Katsumata N, Yasuda M, Isonishi S et al. Long-term results of dose-dense paclitaxel and carboplatin versus conventional paclitaxel and carboplatin for treatment of advanced epithelial ovarian, fallopian tube, or primary peritoneal cancer (JGOG 3016): a randomised, controlled, open-label trial. Lancet Oncol 2013; 14: 1020 1026. 10. Pignata S, Scambia G, Katsaros D et al. Carboplatin plus paclitaxel once a week versus every 3 weeks in patients with advanced ovarian cancer (MITO-7): a randomised, multicentre, open-label, phase 3 trial. Lancet Oncol 2014; 15: 396 405. 11. Tewari D, Java JJ, Salani R et al. Long-term survival advantage and prognostic factors associated with intraperitoneal chemotherapy treatment in advanced ovarian cancer: a Gynecologic Oncology Group study. J Clin Oncol 2015; 33: 1460 1466. 12. Alberts DS, Liu PY, Hannigan EV et al. Intraperitoneal cisplatin plus intravenous cyclophosphamide versus intravenous cisplatin plus intravenous cyclophosphamide for stage III ovarian cancer. N Engl J Med 1996; 335: 1950 1955. 13. Mannel RS, Brady MF, Kohn EC et al. A randomized phase III trial of IV carboplatin and paclitaxel 3 courses followed by observation versus weekly maintenance low-dose paclitaxel in patients with early-stage ovarian carcinoma: a Gynecologic Oncology Group Study. Gynecol Oncol 2011; 122: 89 94. 14. Mei L, Chen H, Wei DM et al. Maintenance chemotherapy for ovarian cancer. Cochrane Database Syst Rev 2013; 6: CD007414. 15. Chan J, Brady M, Penson. et al. Phase III trial of every-3-weeks paclitaxel vs. Dose dense weekly paclitaxel with carboplatin +/ bevacizumab in epithelial ovarian, peritoneal, fallopian tube cancer: GOG 262 (NCT01167712). Int J Gynecol Cancer 2013; 23(Supp 1): 9. i62 Bookman Volume 27 Supplement 1 April 2016