Ovarian Cancer

Standard Treatment Options for Early-Stage Ovarian Epithelial Cancer, FTC, and PPC
Surgery with or without chemotherapy
If the tumor is well differentiated or moderately well differentiated, surgery alone may be adequate treatment for patients with stage IA or IB disease.
Surgery includes hysterectomy, bilateral salpingo-oophorectomy, and omentectomy.
The undersurface of the diaphragm is visualized and biopsied. Biopsies of the pelvic and abdominal peritoneum and the pelvic and para-aortic lymph nodes are also performed. Peritoneal washings are routinely obtained.
In patients who desire childbearing and have grade I tumors, unilateral salpingo-oophorectomy may be associated with a low risk of recurrence.

In the United States, except for the most favorable subset of patients (those with stage IA well-differentiated disease), evidence based on double-blinded, randomized, controlled trials with total mortality endpoints supports adjuvant treatment with cisplatin, carboplatin, and paclitaxel.

Evidence (surgery with or without chemotherapy):
1. In two large European trials, the European Organization for Research and Treatment of Cancer-Adjuvant ChemoTherapy in Ovarian Neoplasm trial (EORTC-ACTION) and International Collaborative Ovarian Neoplasm trial (MRC-ICON1 [NCT00002477]), patients with stage IA (grade II) and stage IB (grade III), all stage IC and stage II ovarian epithelial, and all stage I and stage IIA clear cell carcinoma were randomly assigned to receive adjuvant chemotherapy or observation.
    a. The EORTC-ACTION trial required at least four cycles of carboplatin or cisplatin-based chemotherapy as treatment. Although surgical staging criteria were monitored, inadequate staging was not an exclusion criterion.
        ■ Recurrence-free survival (RFS) was improved in the adjuvant chemotherapy arm (hazard ratio [HR], 0.63; P = .02), but overall survival (OS) was not affected (HR, 0.69; 95% confidence interval [CI], 0.44–1.08; P = .10).
        ■ OS was improved by chemotherapy in the subset of patients with inadequate surgical staging.
    b. The MRC-ICON1 trial randomly assigned patients to six cycles of single-agent carboplatin or cisplatin or platinum-based chemotherapy (usually cyclophosphamide, doxorubicin, and cisplatin) versus observation and had entry criteria similar to the EORTC-ACTION trial; however, the MRC-ICON1 trial did not monitor whether adequate surgical staging was performed. When the results of the trials were combined, the difference in OS achieved statistical significance.
        ■ Both RFS and OS were significantly improved; 5-year survival figures were 79% with adjuvant chemotherapy versus 70% without adjuvant chemotherapy.
    c. An analysis of pooled data from both studies demonstrated the following:[Level of evidence: 1iA]
        ■ There was significant improvement in RFS with chemotherapy (HR, 0.64; 95% CI, 0.50–0.82; P = .001) and OS (HR, 0.67; 95% CI, 0.50–0.90; P = .008). These data showed an OS at 5 years of 82% with chemotherapy and 74% with observation, with a 95% CI in the difference of 2% to 12%.[Level of evidence: 1iA]
        ■ An accompanying editorial emphasized that the focus of subsequent trials must be to identify patients who do not require additional therapy among the early ovarian cancer subset. Optimal staging is one way to better identify these patients.

2. The GOG-0157 trial evaluated whether six cycles of chemotherapy were superior to three cycles for patients with early-stage, high-risk epithelial ovarian cancer after primary surgery. Eligible patients were those with stage IA grade 3 or clear cell histology, stage IB grade 3 or clear cell histology, all stage IC, and all stage II. Patients were randomly assigned to receive either three or six cycles of the combination of paclitaxel (175 mg/m2 administered over 3 hours) and carboplatin dosed (area under the curve, 7.5) over 30 minutes and given every 21 days. The primary endpoint was RFS, and the study was powered to detect a 50% decrease in the recurrence rate at 5 years. A total of 427 patients were eligible.[Level of evidence: 1iiDi]◦
No significant difference in cumulative incidence of recurrence was found when three cycles (25.4%) were compared with six cycles (20.1%) (HR, 0.76; 95% CI, 0.5–1.13) or OS for three cycles (81%) versus six cycles (83%) (HR, 1.02; P = .94).[Level of evidence: 1iiDi]
    • As expected, the use of six cycles was associated with increased grade 3 or 4 neurologic toxic effects and increased grade 4 hematologic toxic effects.
    • Although surgical staging was required for study entry, an audit revealed that 29% of the patients had either incomplete documentation of their surgery or insufficient surgical effort.
    • In a post hoc analysis of the patients who underwent complete surgical staging, three additional cycles of chemotherapy decreased the risk of recurrence by only 3%. The cumulative incidence of recurrence within 5 years was 18% for women with stage I disease and 33% for women with stage II disease.

Given the increased risk of recurrence in patients with stage II disease and combined with an earlier trial, the Ovarian Committee of the GOG has opted to include patients with stage II disease in advanced ovarian cancer trials. The interpretation of this study, including findings on subset analyses, has been a source of controversy.

3. Patients with stage II ovarian cancer were enrolled in a Japanese Gynecology Oncology Group study (JGOG-3016 [NCT00226915]) that tested a weekly dosing schedule versus the conventional every-3-week dosing schedule in first-line ovarian cancer.

Clinical trials evaluating the following treatment approaches have been performed:
    • Intraperitoneal phosphorus P 32 or radiation therapy.
    • Platinum-based systemic chemotherapy alone or in combination with alkylating agents.
    • Platinum-based systemic chemotherapy with paclitaxel.

Standard Treatment Options for Advanced-Stage Ovarian Epithelial Cancer, FTC, and PPC
1. Surgery followed by systemic chemotherapy.
2. Surgery followed by intraperitoneal (IP) chemotherapy.
3. Surgery followed by chemotherapy and bevacizumab.
4. Chemotherapy followed by surgery.
5. Chemotherapy for patients who cannot have surgery (although the impact on OS has not been proven).

Surgery followed by systemic chemotherapy
For patients with residual disease larger than 1 cm after surgery, systemic chemotherapy is the standard. Platinum agents, such as cisplatin or its second-generation analog, carboplatin, given either alone or in combination with other drugs, are the foundation of chemotherapy regimens used. Trials by various cooperative groups (1999–2010) addressed issues of optimal dose-intensity for both cisplatin and carboplatin, schedule, and the equivalent results obtained with either of these platinum drugs, usually in combination with cyclophosphamide.

With the introduction of the taxane paclitaxel, two trials confirmed the superiority of cisplatin combined with paclitaxel when compared with the previous standard treatment of cisplatin plus cyclophosphamide. However, two trials that compared single-agent paclitaxel with either cisplatin or carboplatin (ICON2 and GOG-132) failed to confirm such superiority in all outcome parameters (i.e., response, time-to-progression, and survival) (see Table for a list of these studies).

Based on the evidence, the initial standard treatment for patients with ovarian cancer is the combination of cisplatin or carboplatin with paclitaxel (defined as induction chemotherapy).

Evidence (combination of cisplatin or carboplatin with paclitaxel)
1. GOG-132 was widely regarded as showing that sequential treatment with cisplatin and paclitaxel was equivalent to the combination of cisplatin- or carboplatin-plus-paclitaxel; however, many patients crossed over before disease progression. Moreover, the cisplatin-only arm was more toxic because it utilized a 100 mg/m2 dose.

2. The Medical Research Council study (MRC-ICON3), while having fewer early crossovers, could be interpreted similarly in regard to the impact of sequential treatment on survival.

Since the adoption of the standard combination of platinum plus taxane nearly worldwide, clinical trials have demonstrated the following:
1. Noninferiority of carboplatin plus paclitaxel versus cisplatin plus paclitaxel.
2. Noninferiority of carboplatin plus paclitaxel versus carboplatin plus docetaxel.
3. No advantage but increased toxic effects of adding epirubicin to the carboplatin plus paclitaxel doublet.
4. Noninferiority of carboplatin plus paclitaxel versus sequential carboplatin-containing doublets with either gemcitabine or topotecan; or, triplets with the addition of gemcitabine or pegylated liposomal doxorubicin to the reference doublet as shown below:
    a. From February 2001 to September 2004, 4,312 women with stage III or stage IV ovarian epithelial cancer, FTC, or PPC participating in the GOG-0182 trial were randomly assigned to four different experimental arms or to a reference treatment consisting of carboplatin (area under the curve [AUC], 6) and paclitaxel (175 mg/m2) every 3 weeks for eight cycles. Stratification factors were residual-disease status and the intention to perform interval debulking surgery.
        ■ None of the experimental regimens was inferior.
        ■ Lethal events attributable to treatment occurred in less than 1% of patients without clustering to any one regimen.
        ■ With a median follow-up of 3.7 years, the adjusted relative risk of death ranged from 0.952 to 1.114, with the control arm achieving a progression-free survival (PFS) of 16.0 months and a median overall survival (OS) of 44.1 months.

In this large study consisting of two arms of patients with Féderation Internationale de Gynécologie et d’Obstétrique stage III disease (84% in one arm and 87% in the other arm), the extent of cytoreduction was an important prognostic factor in OS, as expected.
        ■ PFS in patients with residuum larger than 1 cm was 13 months, and OS was 33 months.
        ■ With residuum 1 cm or smaller, PFS was 16 months, and OS was 40 months.
        ■ With microscopic residuum, PFS was 29 months, and OS was 68 months.

In gynecologic cancer, as opposed to breast cancer, weekly paclitaxel was not explored in phase III trials before 2004. The positive results from the Japanese Gynecologic Oncology Group (JGOG) 3016 study (below) have been widely adopted and also led to new divided-dose paclitaxel studies.

Evidence (dose-dense [weekly] treatment schedule): 1. A JGOG trial accrued 637 patients (JGOG-3016 [NCT00226915]) and randomly assigned them to six to nine cycles of weekly (dose-dense) paclitaxel (80 mg/m2) or to the standard every-21-day schedule of paclitaxel at 180 mg/m2. Both regimens were given with carboplatin (AUC, 6) in every-3-week cycles. With a primary endpoint of PFS, an increase from 16 to 21 months in the PFS of the weekly paclitaxel-based regimen was sought. Although more toxic, the weekly paclitaxel regimen did not adversely affect quality of life when compared with the intermittent schedule.[Level of evidence: 1iiDiii]

Other than ethnicity, this trial population differed from other studies in the following ways:
    ◦ A lower median age (57 years).
    ◦ Twenty percent of patients were stage II.
    ◦ Eleven percent of patients were treated in the neoadjuvant setting.
    ◦ Thirty-three percent of patients had histologies other than high-grade serous or endometrioid cancer.

Study results demonstrated the following:
    ◦ At the 1.5-year follow-up after cessation of treatment, the weekly regimen had a median PFS of 28.0 months (95% confidence interval (CI), 22.3–35.4 months), and the intermittent median PFS was 17.2 months (15.7–21.1; hazard ratio (HR), 0.71), favoring the weekly regimen (P = .0015).
    ◦ The 2013 updated results revealed an increase in median survival for the weekly regimen (median OS, 8.3 years vs. 5.1 years; P = .040); the intermittent regimen results are also noteworthy relative to other clinical trials of weekly dosing schedules.

2. In a phase III trial (MITO-7 [NCT00660842]), the outcomes of 406 patients assigned to weekly paclitaxel (60 mg/m2) administered with weekly carboplatin (AUC, 2) were compared with those of 404 patients receiving the conventional every-3-week regimen of paclitaxel and carboplatin.[Level of evidence: 1iiA]
    ◦ The results failed to confirm the superiority of this particular weekly schedule (18.3 months PFS for the weekly arm vs. 17.3 months PFS for the standard arm [HR, 0.96; 95% CI, 0.80–1.16]).
    ◦ The treatments did not differ in toxic effects. A decrease in quality of life (assessed by the Functional Assessment of Cancer Therapy Ovarian Trial Outcome Index questionnaire) was not seen in the weekly arm compared with the every-3-week arm.
3. GOG-0262 [NCT01167712] is a phase III study that compared weekly paclitaxel (80 mg/m2) to every-3-week dosing (175 mg/m2), both with the conventional every-3-week carboplatin (AUC 6) regimen.[Level of evidence: 1iiDiii] An option to give bevacizumab every 3 weeks beginning with cycle two and continuing until cycle six and followed by bevacizumab alone for 1 year, as in GOG-0218, was included for both arms. This option was applied in about 84% of all patients.
    ◦ Overall, the weekly paclitaxel regimen failed to prolong PFS compared with the every-3-week regimen (14.7 months vs. 14.0 months), with an HR for progression or death of 0.89 (95% CI, 0.74–1.06).
    ◦ However, among patients not receiving bevacizumab, the weekly paclitaxel arm had significantly prolonged PFS (14.2 months vs. 10.3 months), with an HR of 0.62 (95% CI, 0.40–0.95; P = .03)
    ◦ The weekly paclitaxel regimen had a higher rate of grade 3 or 4 anemia (36% vs. 16%) and grade 2 to 4 sensory neuropathy (26% vs. 18%).

Surgery followed by intraperitoneal (IP) chemotherapy The pharmacologic basis for the delivery of anticancer drugs by the IP route was established in the late 1970s and early 1980s. When several drugs were studied, mostly in the setting of minimal residual disease at reassessment after patients had received their initial chemotherapy, cisplatin alone and in combination received the most attention. Favorable outcomes from IP cisplatin were most often seen when tumors had shown responsiveness to platinum therapy and with small-volume tumors (usually defined as tumors <1 cm).

In the 1990s, randomized trials were conducted to evaluate whether the IP route would prove superior to the intravenous (IV) route. IP cisplatin was the common denominator of these randomized trials.

Hyperthermic peritoneal chemotherapy (HIPEC) is another pharmacologically-based modality to enhance the antitumor effects via direct drug delivery to peritoneal surfaces. It was initially tested against mucinous tumors of gastrointestinal origin. Increasingly, HIPEC is being applied to ovarian cancers, with considerable variation in patient selection, drugs administered, and time at target temperatures (most often 30 minutes at 42°C). While exploratory trials are ongoing in the setting of recurrent ovarian cancer, such modalities should not be used as a substitute for intraperitoneal cisplatin-based regimens following initial therapy. The role of HIPEC remains experimental in the treatment of patients with high-grade serous ovarian cancers.

Evidence (surgery followed by IP chemotherapy):
1. The use of IP cisplatin as part of the initial approach in patients with stage III optimally debulked ovarian cancer is supported principally by the results of three randomized clinical trials (SWOG-8501, GOG-0114, and GOG-0172 [NCT00003322]). These studies tested the role of IP drugs (IP cisplatin in all three studies and IP paclitaxel in the last study) against the standard IV regimen.
    ◦ In the three studies, superior PFS and OS favoring the IP arm were documented.
Specifically, the most recent study, GOG-0172, demonstrated the following:[Level of evidence:1iiA]◦
A median survival of 66 months for patients on the IP arm versus 50 months for patients who received IV administration of cisplatin and paclitaxel (P = .03).[7][Level of evidence:1iiA]
    ◦ Toxic effects were greater in the IP arm because of the cisplatin dose per cycle (100 mg/m2); sensory neuropathy resulted from the additional IP chemotherapy and from the systemic administration of paclitaxel.
    ◦ The rate of completion of six cycles of treatment was also less frequent in the IP arm (42% vs. 83%) because of the toxic effects and catheter-related problems.[Level of evidence: 1iiA]
An updated combined analysis of GOG-0114 and GOG-0172 included 876 patients with a median follow-up of 10.7 years and reported the following results.
    ◦ Median survival with IP therapy was 61.8 months (95% CI, 55.5–69.5) compared with 51.4 months (95% CI, 46.0–58.2) for IV therapy.
    ◦ IP therapy was associated with a 23% decreased risk of death (adjusted hazard ratio [AHR], 0.77; 95% CI, 0.65–0.90; P = .002).
    ◦ IP therapy improved the survival of patients with gross residual (≤1 cm) disease (AHR, 0.75; 95% CI, 0.62–0.92; P = .006).
    ◦ Risk of death decreased by 12% for each cycle of IP chemotherapy completed (AHR, 0.88; 95% CI, 0.83–0.94; P < .001).
    ◦ Factors associated with poorer survival included clear and mucinous versus serous histology (AHR, 2.79; 95% CI, 1.83–4.24; P < .001), gross residual versus no visible disease (AHR, 1.89; 95% CI, 1.48–2.43; P < .001), and fewer versus more cycles of IP chemotherapy (AHR, 0.88; 95% CI, 0.83–0.94; P < .001).
    ◦ Younger patients were more likely to complete the IP regimen, with a 5% decrease in probability of completion with each year of age (odds ratio, 0.95; 95% CI, 0.93–0.96; P < .001).

Accordingly, efforts are under way by the GOG to examine some modifications of the IP regimen used in GOG-0172 to improve its tolerability (e.g., to reduce by ≥25% the total 3-hour amount of cisplatin given; and, to shift from the less practical 24-hour IV administration of paclitaxel to a 3-hour IV administration.)

2. A Cochrane-sponsored meta-analysis of all randomized IP-versus-IV trials showed an HR of 0.79 for disease-free survival and 0.79 for OS, favoring the IP arms.

3. In another meta-analysis of seven randomized trials assessing IP versus systemic chemotherapy conducted by Cancer Care of Ontario, the relative ratio (RR) of disease progression at 5 years based on the three trials that reported this endpoint was 0.91 (95% CI, 0.85–0.98), and the RR of death at 5 years based on six trials was 0.88 (95% CI, 0.81–0.95) for the IP route.

Surgery followed by chemotherapy and bevacizumab
Two phase III trials (GOG-0218 [NCT00262847] and ICON7 [NCT00483782]) have evaluated the role of bevacizumab in first-line therapy for ovarian epithelial cancer, FTC, and PPC after surgical cytoreduction. Both trials showed a modest improvement in PFS when bevacizumab was added to initial chemotherapy and continued every 3 weeks for 16 and 12 additional cycles, as a maintenance phase.

Evidence (surgery followed by chemotherapy and bevacizumab):
1. GOG-0218 was a double-blinded, randomized, controlled trial that included 1,873 women with stage III or IV disease, all of whom received chemotherapy—carboplatin (AUC 6) and paclitaxel (175 mg/m2 for six cycles). Forty percent of the women had suboptimally resected stage III disease, and 26% had stage IV disease. The primary endpoint of the study was PFS.[Level of evidence:1iDiii] Participants were randomly assigned to receive the following:
    ◦ Chemotherapy plus placebo (cycles 2–22) (the control group).
    ◦ Chemotherapy plus bevacizumab (15 mg/kg cycles 2–6), followed by placebo (cycles 7–22) (the bevacizumab-initiation group).
    ◦ Chemotherapy plus bevacizumab (15 mg/kg cycles 2–22) (the bevacizumab-throughout group).

Results from the trial demonstrated the following:
    ◦ There was no difference in PFS between the control group and the bevacizumab-initiation group.
    ◦ There was a statistically significant increase in PFS in the bevacizumab-throughout group when compared with the control group (14.1 months vs. 10.3 months), with an HR of disease progression or death of 0.717 in the bevacizumab-throughout group (95% CI, 0.625–0.824; P < .001).
    ◦ Median OS was 39.3 months for the control group, 38.7 months for the bevacizumab-initiation group, and 39.7 months for the bevacizumab-throughout group.
    ◦ Quality of life was not different between the three groups. Hypertension grade 2 or higher was more common with bevacizumab than with placebo.
    ◦ There were more treatment-related deaths in the bevacizumab-throughout arm (10 of 607, 2.3%) than in the control arm (6 of 601, 1.0%).

2. ICON7 randomly assigned 1,528 women after initial surgery to chemotherapy—carboplatin (AUC, 5 or 6) plus paclitaxel (175 mg/m2 for six cycles)—or to chemotherapy plus bevacizumab (7.5 mg/kg for six cycles), followed by bevacizumab alone for an additional 12 cycles. Nine percent of patients had early-stage, high-grade tumors; 70% had stage IIIC or IV disease; and 26% had more than 1 cm of residual tumor before initiating chemotherapy. PFS was the main outcome measure.[Level of evidence: 1iiDiii]a.
    a. Median PFS was 17.3 months in the control group and 19 months in the bevacizumab group. HR for disease progression or death in the bevacizumab group was 0.81 (95% CI, 0.70–0.94; P = .004).
    b. Grade 3 or higher adverse events were more common in the bevacizumab group, with an increase in bleeding, hypertension (grade 2 or higher), thromboembolic events (grade 3 or higher), and gastrointestinal perforations.
    c. Quality of life was not different between the two groups.
    d. In 2015, the ICON7 authors reported an updated survival analysis.
        ■ There was no significant difference with 44.6 months (95% CI 43.2–45.9) in patients on standard chemotherapy versus 45.5 months (44.2–46.7) in patients receiving bevacizumab with the chemotherapy induction, and then completing 1 year of bevacizumab maintenance (log-rank P = 0.85).

In summary, the evidence does not support the use of bevacizumab as front-line therapy because the gain in PFS comes with increased toxicity, without improvement in OS or quality of life.

Chemotherapy followed by surgery
Two phase III studies compared the outcome of standard primary cytoreductive surgery (PCS) with that of neoadjuvant chemotherapy (NACT) followed by interval cytoreductive surgery; both studies (described below) demonstrated that PFS and OS were noninferior with use of PCS.

Evidence (chemotherapy followed by surgery):
1. Between 1998 and 2006, a study led by the European Organization for the Research and Treatment of Cancer (EORTC) Gynecological Cancer Group, together with the National Cancer Institute of Canada Clinical Trials Group (EORTC-55971 [NCT00003636]), included 670 women with stages IIIC and IV ovarian epithelial cancer, FTC, and PPC.[Level of evidence: 1iiA] The women were randomly assigned to undergo primary debulking surgery followed by at least six courses of platinum-based chemotherapy or to receive three courses of neoadjuvant platinum-based chemotherapy followed by interval debulking surgery, and at least three more courses of platinum-based chemotherapy.

Methods included efforts to ensure accuracy of diagnosis (e.g., rule out peritoneal carcinomatosis of gastrointestinal origin) and stratification by largest preoperative tumor size (excluding ovaries) (<5 cm, >5 cm–10 cm, >10 cm–20 cm, or >20 cm). Other stratification factors included institution, method of biopsy (i.e., image-guided, laparoscopy, laparotomy, or fine-needle aspiration), and tumor stage (i.e., stage IIIC or IV). The primary endpoint of the study was OS, with primary debulking surgery considered the standard.[Level of evidence: 1iiA]
    ◦ Median OS for primary debulking surgery was 29 months, compared with 30 months for patients assigned to neoadjuvant chemotherapy.
    ◦ The HRdeath in the group assigned to neoadjuvant chemotherapy followed by interval debulking, as compared with the group assigned to primary debulking surgery followed by chemotherapy, was 0.98 (90% CI, 0.84–1.13; P = .01 for noninferiority).[Level of evidence: 1iiA]
    ◦ Perioperative and postoperative morbidity and mortality were higher in the primary debulking surgery group (7.4% severe hemorrhage and 2.5% deaths, compared with 4.1% severe hemorrhage and 0.7% deaths in the neoadjuvant group).
    ◦ The strongest independent predictor of prolonged survival was the absence of residual tumor after surgery.
    ◦ The subset of patients achieving optimal cytoreduction (≤1 cm residuum), whether after primary debulking surgery or after neoadjuvant chemotherapy followed by interval debulking surgery, had the best median OS.

2. Between 2004 and 2010, a group of 87 hospitals in the United Kingdom and New Zealand enrolled 550 women with stage III or IV ovarian epithelial cancer and randomly assigned them to undergo PCS followed by six cycles of chemotherapy or primary (neoadjuvant) chemotherapy for three cycles, followed by surgery and three additional cycles of chemotherapy. In contrast to the EORTC study, the chemotherapy consisted of conventional carboplatin (AUC, 5 or AUC, 6) and paclitaxel (175 mg/m2, in 76% of patients), or carboplatin alone (23% of patients), or nonpaclitaxel chemotherapy (1% of patients).[Level of evidence: 1iiA]
    ◦ A minimization method was used to randomly assign patients in a 1:1 ratio. Participants were stratified by randomizing center, largest radiologic tumor, and prespecified chemotherapy regimen. The primary endpoint was to establish noninferiority, with the upper bound of a one-sided 90% CI for the HRdeath at less than 1.18. As of May 2014, 451 deaths had occurred, and the HRdeath favored NACT, with the upper bound of the one-sided 90% CI 0.98 (95% CI, 0.72‒1.05).
    ◦ The most common grade 3 or 4 postoperative adverse event was hemorrhage in both groups, with 8 women (3%) having this problem with PCS versus 14 (6%) in the NACT group. Grade 3 and 4 toxic events from chemotherapy occurred in 110 (49%) of 225 women randomly assigned to PCS, and in 102 (40%) of the 253 women receiving NACT, with one fatal event, neutropenic sepsis, occurring in the primary chemotherapy group.

These studies and additional observational and partially published phase III studies have led to the publication of a Clinical Practice Guideline on behalf of the Society of Gynecologic Oncology and the American Society of Clinical Oncology.

Consolidation and/or maintenance therapy
Phase III trials of consolidation and/or maintenance therapy have been carried out with cytotoxic drugs that contribute to the treatment of recurrent ovarian cancer, vaccines, and radioimmunoconjugates listed below with mostly negative results, and with some biologicals (such as bevacizumab, discussed in a separate section above). These treatments have included the following:
• IP cisplatin (four cycles).
• Yttrium Y 90-labeled radioimmunoconjugate plus IP chemotherapy.
• IV topotecan (four cycles).
• Oregovomab vaccination.
• High-dose chemotherapy with hematopoietic support.
• Monthly paclitaxel (12 cycles).
• Erlotinib.

Trials ongoing with anti-angiogenic drugs (other than bevacizumab) and PARP inhibitors are described in the Treatment Options under Clinical Evaluation section that follows.

Treatment Options Under Clinical Evaluation
Trials are ongoing with anti-angiogenic drugs (other than bevacizumab) and with PARP inhibitors.
PARP is a family of enzymes involved in base-excision repair of DNA single-strand breaks. In patients with homologous recombination deficiency (HRD), including patients with germline BRCA1 or BRCA2 (gBRCA) mutations or with nongermline HRD–positive tumors, inhibition of PARP results in production of double-strand breaks of DNA. Human DNA repair mechanisms largely rely on one intact copy of the gene; cells with a double-strand break are usually targeted for cell death. This susceptibility of BRCA-deficient or BRCA-mutant cells to PARP inhibition has spurred the clinical development of this class of agents.
Sensitivity to platinum compounds is a feature of HRD, and a population of platinum-sensitive patients is expected to be HRD-enriched and most likely to benefit from PARP inhibition.
Several of these drugs have been studied in ovarian cancer as monotherapy or drug combinations and have demonstrated activity in the recurrent setting, with olaparib, rucaparib, and niraparib achieving U.S. Food and Drug Administration approval with varying indications. (Refer to the Recurrent or Persistent Ovarian Epithelial Cancer, FTC, and PPC Treatment section for more information.) Phase III studies are ongoing with these three agents following first-line treatments, and also with a fourth drug, veliparib, which is not otherwise commercially available.

Information about ongoing clinical trials is available from the NCI website.

Treatment Options for Patients with Recurrent or Persistent Ovarian Epithelial Cancer, FTC, and PPC
1. Platinum-sensitive recurrence: For patients whose disease recurs more than 6 months after cessation of the induction, re-treatment with a platinum or platinum-containing combination, such as carboplatin, should be considered (see Table 8).
2. Platinum-refractory or platinum-resistant recurrence: For patients who progress before cessation of induction therapy (platinum refractory) or within 6 months after cessation of induction therapy (platinum resistant), (platinum therapy is generally not useful as part of the treatment plan). Clinical trials should be considered.

Other agents that have shown activity in phase II trials are listed in Table 10 and may also be used alone or in combination with other drugs, but such treatments are best done in prospective trials.

Cytoreduction may be used; this intervention is being studied in the setting of a randomized clinical trial (GOG-0213).

The role of radiation therapy in patients with recurrent ovarian cancer has not been defined.

Platinum-sensitive recurrence
Platinum-containing chemotherapy regimens

The Table shows the chemotherapy regimens used in first relapse for the treatment of platinum-sensitive recurrent ovarian cancer.

^a Trabectedin has been approved for use in treating recurrent ovarian cancer in Europe and Canada.
^b OS data were not mature at the time the manuscript was published.
^c P < .0001.

On the basis of improved survival with etoposide or 5-fluorouracil, carboplatin was approved in 1987 for the treatment of patients with ovarian cancer whose disease recurred after treatment with cisplatin.
In a randomized phase II trial of paclitaxel, a currently used second-line drug, the cisplatin-containing combination of cisplatin plus doxorubicin plus cyclophosphamide (CAP), yielded a superior survival outcome. This study and subsequent studies (see Table) have reinforced the use of carboplatin as the treatment core for patients with platinum-sensitive recurrences.
Cisplatin is occasionally used, particularly in combination with other drugs, because of its lesser myelosuppression, but this advantage over carboplatin is counterbalanced by greater patient intolerance.

Oxaliplatin, initially introduced with the hope that it would overcome platinum resistance, has activity mostly in platinum-sensitive patients but has not been compared with carboplatin alone or in combinations.

With all platinum agents, outcome is generally better the longer the initial interval without recurrence from the initial platinum-containing regimens. Therefore, on occasion, patients with platinum-sensitive recurrences relapsing within 1 year have been included in trials of nonplatinum drugs. In one such trial, comparing the pegylated liposomal doxorubicin to topotecan, the subset of patients who were platinum sensitive had better outcomes with either drug (and in particular with pegylated liposomal doxorubicin) relative to the platinum-resistant cohort.

Several randomized trials have addressed whether the use of a platinum in combination with other chemotherapy agents is superior to single agents (see Table).

Evidence (platinum in combination with other chemotherapy agents):
1. In an analysis of data examining jointly the results of three trials performed by the MRC/Arbeitsgemeinschaft Gynaekologische Onkologie (MRC/AGO) and the International Collaborative Ovarian Neoplasm (ICON) investigators (ICON4), the following results were observed:[Level of evidence: 1iiA]
    ◦ A platinum-plus-paclitaxel combination yielded superior response rates, progression-free survival (PFS), and overall survival (OS), compared with carboplatin as a single agent or other platinum-containing combinations as controls.
    ◦ Platinum plus paclitaxel was compared with several control regimens, although 71% used carboplatin as a single agent in the control, and 80% used carboplatin plus paclitaxel. Prolonged PFS (HR, 0.76; 95% CI, 0.66–0.89; P = .004) and OS (HR, 0.82; 95% CI, 0.69–0.97; P = .023) were improved in the platinum-plus-paclitaxel arm.[Level of evidence: 1iiA]
    ◦ The AGO had previously compared the combination of epirubicin plus carboplatin with carboplatin alone and had not found significant differences in outcome.
    ◦ A meta-analysis of five trials (three of which are in Table 8), with four reviewing independent patient data, supports the use of platinum agents in combination with other active agents rather than carboplatin alone for patients with platinum-sensitive recurrent ovarian cancer.

2. Another trial by European and Canadian groups compared gemcitabine plus carboplatin with carboplatin.
    ◦ The PFS of 8.6 months with the combination was significantly superior to 5.8 months for the carboplatin alone (HR, 0.72; 95% CI, 0.58–0.90; P = .003).[Level of evidence: 1iiDiii]
    ◦ The study was not powered to detect significant differences in OS, and the median survival for both arms was 18 months (HR, 0.96; CI, 0.75–1.23; P = .73).

3. In a phase III trial, carboplatin plus pegylated liposomal doxorubicin was compared with carboplatin plus paclitaxel in patients with platinum-sensitive recurrence (>6 months). The primary endpoint was PFS.
    ◦ The median PFS for the carboplatin-plus-pegylated-liposomal-doxorubicin arm was 11.3 months versus 9.4 months for the carboplatin-plus-paclitaxel arm (HR, 0.823; 95% CI, 0.72–0.94; P = .005).[Level of evidence: 1iiDiii]
    ◦ Long-term follow-up revealed no difference in OS rates between the two arms (30.7 months for carboplatin plus pegylated liposomal doxorubicin vs. 33.0 months for carboplatin plus paclitaxel).
    ◦ The carboplatin-plus-paclitaxel arm was associated with increased severe neutropenia, alopecia, neuropathy, and allergic reaction. The carboplatin-plus-pegylated-liposomal-doxorubicin arm was associated with increased severe thrombocytopenia, nausea, and hand-foot syndrome.
Given its toxicity profile and noninferiority to the standard regimen, carboplatin plus pegylated liposomal doxorubicin is an important option for patients with platinum-sensitive recurrence.

Carboplatin plus paclitaxel has been considered the standard regimen for platinum-sensitive recurrence in the absence of residual neurological toxic effects. The GOG-0213 trial is comparing this regimen with the experimental arm that adds bevacizumab to carboplatin plus paclitaxel.

Bevacizumab, other targeted drugs, and PARP inhibitors with or without chemotherapy

Evidence (bevacizumab with gemcitabine-carboplatin chemotherapy):
1. The Ovarian Cancer Study Comparing Efficacy and Safety of Chemotherapy and Anti-Angiogenic Therapy in Platinum-Sensitive Recurrent Diseases (OCEANS [NCT00434642]), assessed the role of bevacizumab in the treatment of platinum-sensitive recurrence (see Table 8).
In this double-blind, placebo-controlled, phase III trial of chemotherapy (gemcitabine + carboplatin) with or without bevacizumab for recurrent ovarian epithelial cancer, FTC, or PPC, 242 patients were randomly assigned to each arm. In contrast to the first-line studies, treatment was allowed to continue beyond six cycles to ten cycles in responding patients, but there was no maintenance therapy.
    ◦ A subsequent analysis will appear when additional survival data become mature; however, at the time of publication, differences in median OS were not apparent, and crossover from a placebo to bevacizumab had occurred in 31% of the patients.
    ◦ Median PFS for patients receiving bevacizumab was 12.4 months versus 8.4 months for those receiving a placebo.
    ◦ The HR for the effect of bevacizumab on disease progression in patients assigned to the bevacizumab arm compared with placebo was 0.484 (95% CI, 0.388–0.605; P <     .0001).
    ◦ Objective responses to chemotherapy were increased when combined with bevacizumab (78.5% vs. 57.4%; P < .0001).
    ◦ Bevacizumab-associated toxicities such as hypertension and proteinuria were more prominent than in the first-line trials, but feared safety issues such as gastrointestinal perforations did not occur during the study.
    ◦ Discontinuing treatment because of adverse events was more common with bevacizumab (n = 55 vs. n = 12 for placebo), but fewer patients discontinued treatment because of disease progression (n = 104 for bevacizumab vs. n = 160 for placebo).

Evidence (bevacizumab with paclitaxel-carboplatin chemotherapy):
1. NRG/Gynecologic Oncology Group GOG-0213 (NCT00565851) assessed both the role of surgical debulking and the addition of bevacizumab induction and maintenance in women with platinum-sensitive recurrences of ovarian cancer. The nonsurgical portion of GOG-0213 had 81% power for a true hazard ratio (HR) of 0.75; it enrolled 674 women from December 2007 to August 2011, and the published analysis took place after a median follow-up exceeding 4 years.
    ◦ OS was not significantly different: 37.3 months (95% CI, 32.6–39.7) versus 42.2 months (95% CI, 37.7–46.2).][Level of evidence: 1iiA]
    ◦ The secondary endpoint of median PFS was significantly in favor of the addition of bevacizumab: 10.4 months (95% CI, 9.7–11) for chemotherapy alone versus 13.8 months (95% CI, 13.0–14.7).
    ◦ Bevacizumab (15 mg/kg every 3 weeks) with chemotherapy and its use in maintenance led to an excess of grade 3 and 4 adverse events (8% for chemotherapy alone vs. 30%), any bleeding (12% vs. 42%), and any hypertension (3% vs. 41%).

Evidence (PARP inhibitors with or without anti-angiogenic agents):
PARP is a family of enzymes involved in base-excision repair of DNA single-strand breaks. In patients with homologous recombination deficiency (HRD), including patients with germline BRCA1 or BRCA2 (gBRCA) mutations or with nongermline HRD–positive tumors, inhibition of PARP results in production of double-strand breaks of DNA. Human DNA repair mechanisms largely rely on one intact copy of the gene; cells with a double-strand break are usually targeted for cell death. This susceptibility of BRCA-deficient or BRCA-mutant cells to PARP inhibition,[19,20] has spurred the clinical development of this class of agents. Sensitivity to platinum compounds is a feature of HRD, and a population of platinum-sensitive patients is expected to be HRD-enriched and most likely to benefit from PARP inhibition.
1. In a randomized, double blind, placebo-controlled phase II trial of olaparib maintenance therapy, eligible patients had platinum-sensitive, high-grade serous ovarian cancer. Patients were randomly assigned to receive olaparib (400 mg bid) or placebo. Having a gBRCA1 or gBRCA2 mutation was not required for eligibility; however, 23% of patients in the experimental group and 22% of patients in the placebo group had a known BRCA1 or BRCA2 mutation. The primary endpoint was PFS.[21][Level of evidence: 1iiDiii] ◦ PFS was longer in the olaparib arm; median 8.4 months versus 4.8 months (HR, 0.35; 95% CI, 0.25–0.49; P < .001).[Level of evidence: 1iiDiii]
    ◦ OS was not different between the two groups, as noted in an updated report.
    ◦ The more common adverse events in the olaparib group were nausea, fatigue, vomiting, and anemia.

2. Olaparib tablets (as opposed to the previous capsule formulation) underwent evaluation in SOLO2 (NCT01874353), a double-blind, randomized, placebo-controlled phase III trial in patients with high-grade serous or endometrioid, primary peritoneal, or fallopian tube cancer. Patients had platinum-sensitive relapses and were preselected for BRCA 1/2 mutations.[Level of evidence: 1iiDiii] Stratification for response (complete vs. partial) to previous platinum and platinum-free intervals (>6–12 vs. >12) and 2:1 random allocation to olaparib in two 150-mg twice-daily or matching placebo tablets took place. Of 295 eligible patients enrolled, 196 were assigned to olaparib, and 99 were assigned to a placebo.
    ◦ The primary endpoint was PFS and significantly favored olaparib 19.1 m (95% CI, 16.3–25.7) over placebo (5.5 m [5.2–5.8]; HR, 0.30 [95% CI, 0.22–0.41]; P < .0001).
    ◦ Serious adverse events were experienced by 18% of patients on olaparib and by 8% on placebo. The adverse events consisted mostly of anemia, abdominal pain, and intestinal obstruction.

3. Rucaparib underwent phase II evaluation in ARIEL2 (NCT01891344), an open-label study enrolling 206 patients, 204 of whom were actually receiving the drug (192 were actually in classifiable subgroups) and had high-grade platinum-sensitive recurrences between October 2013 and November 2014.[24] There were three predefined homologous recombination deficiency subgroups on the basis of tumor mutational analysis:
    ◦ BRCA mutant (deleterious genetic or somatic) (n = 40).
    ◦ BRCA wild type and high loss of heterozygosity (LOH) quantified by next-generation sequencing analysis (LOH high) (n = 82).
    ◦ BRCA wild type and low LOH (LOH low) (n = 70).
The drug was given orally at 600 mg twice daily, and patients were treated until disease progression or other reasons for discontinuation. Median duration of treatment for the 204 patients was 5.7 months.
    ◦ Median PFS after start of rucaparib treatment for patients with deleterious BRCA mutations was 12.8 months (95% CI, 9.0–14.7); for those with LOH, high was 5.7 months (5.3–7.6) and low was 5.2 months (3.6–5.5).
    ◦ The study also showed that mutation and methylation status of BRCA and other homologous recombination-related genes, such as RAD51C, can be associated with high genomic LOH in BRCA wild-type tumors, conferring higher rates of response to rucaparib than seen in patients with low genomic LOH.[24][Level of evidence: 3iiDiii]

4. In a double blind, placebo-controlled phase III trial, 533 patients with platinum-sensitive, predominantly high-grade serous ovarian cancer were randomly assigned in a 2:1 ratio to maintenance with oral niraparib or placebo and followed for the primary endpoint of PFS. Patients were categorized according to the presence or absence of gBRCA or non-BRCA HRD-positive ovarian cancer or non-BRCA HRD-negative ovarian cancer, based on BRCA Analysis testing (Myriad Genetics) from tumor and blood samples.
    a. Patients on niraparib had significantly longer median PFS duration compared with a placebo.[Level of evidence: 1iiDiii] Comparisons across categories ranged from HR, 0.27 for gBRCA cancer (21.0 months vs. 5.5 months), HR, 0.38 for non-BRCA cancer, HRD-positive cancer (12.9 months vs. 3.8 months), and HR, 0.45 for non-BRCA, HRD-negative cancer (9.3 months vs. 3.9 months).
    b. OS data were not mature at the time of this report, but deaths during the study occurred in 16.1% of patients on niraparib and 19.3% of patients on placebo.
    c. One-third to nearly one-half of the patients had received at least three previous lines of therapy that included the following:
        ■ Grade 3 or 4 adverse events that were managed with dose modifications while patients were on niraparib included thrombocytopenia (in 33.8% of patients), anemia (in 25.3%), and neutropenia (in 19.6%).
        ■ Other excess severe toxicities while patients were on niraparib that occurred at starting doses of 300 mg once daily included fatigue (in 30 patients vs. 1 patient on the placebo), hypertension (in 30 patients vs. 4 on the placebo), nausea (in 11 patients vs. 2 on the placebo), and vomiting (in 7 patients vs. 1 on the placebo).
    d. A phase III, randomized, double-blind, placebo-controlled study of niraparib maintenance in patients with HRD-positive advanced ovarian cancer following response to front-line platinum-based chemotherapy (NCT01847274) is closed to patient accrual and results are pending.
    e. Other PARP inhibitor trials have been exploring their role in platinum-resistant disease and their role in combination with other agents.

5. Olaparib was also evaluated as a single agent in a multicenter phase II trial for patients with documented BRCA1- or BRCA2-germline mutations.[Level of evidence: 3iiiDiv] This trial was open to patients with platinum-resistant ovarian cancer, breast cancer treated with three or more previous regimens, pancreatic cancer with previously administered gemcitabine, or prostate cancer previously treated with hormonal therapy and one systemic therapy. Olaparib was given at 400 mg twice a day. The primary endpoint was response rate. A total of 298 patients were included.
    ◦ The overall response rate was 26.2%; the response rate was 31.1% in patients with ovarian cancer.[Level of evidence: 3iiiDiv]
The data from this trial were used by the U.S. Food and Drug Administration to approve olaparib for patients with ovarian cancer, who have known BRCA1 or BRCA2 mutations and have failed three previous regimens.

6. Several other trials have combined olaparib with either cytotoxic chemotherapy or other biologic therapy.See Table.
    ◦ Extension in PFS, but not in OS, has been noted.

PARP inhibitor trials as maintenance after platinum-based responses are ongoing (see Table).Olaparib Combinations

AUC = area under the curve;   NR = not reported

Platinum-refractory or platinum-resistant recurrence

Chemotherapy
Clinical recurrences that take place within 6 months of completion of a platinum-containing regimen are considered platinum-refractory or platinum-resistant recurrences. Anthracyclines (particularly when formulated as pegylated liposomal doxorubicin), taxanes, topotecan, and gemcitabine are used as single agents for these recurrences on the basis of activity and their favorable therapeutic indices relative to agents listed in Table. The long list underscores the marginal benefit, if any, of these agents. Clinical trials should be considered for patients with platinum-resistant disease.

Drugs used to treat platinum-refractory or platinum-resistant recurrences include the following:

• Paclitaxel.
Treatment with paclitaxel historically provided the first agent with consistent activity in patients with platinum-refractory or platinum-resistant recurrences.
Patients generally received paclitaxel in front-line induction regimens. Re-treatment with paclitaxel, particularly in weekly schedules, had activity comparable with that of other drugs. Residual neuropathy upon recurrence may shift the choice of treatment towards other agents.

• Topotecan.
Randomized studies have indicated that the use of topotecan achieved results that were comparable with those achieved with paclitaxel.

Evidence (topotecan):
1. Topotecan was compared with pegylated liposomal doxorubicin in a randomized trial of 474 patients and demonstrated similar response rates, PFS, and OS at the time of the initial report. Responses occurred primarily in the platinum-resistant subsets.

2. In phase II studies, topotecan administered intravenously on days 1 to 5 of a 21-day cycle yielded objective response rates ranging from 13% to 16.3% and other outcomes that were equivalent or superior to paclitaxel.
    ■ Objective responses were reported in patients with platinum-refractory disease.
    ■ Substantial myelosuppression followed administration. Other toxic effects included nausea, vomiting, alopecia, and asthenia. A number of schedules and oral formulations to reduce toxicity are under evaluation.

3. In a phase III study, 235 patients who did not respond to initial treatment with a platinum-based regimen, but who had not previously received paclitaxel or topotecan, were randomly assigned to receive either topotecan as a 30-minute infusion daily for 5 days every 21 days or paclitaxel as a 3-hour infusion every 21 days.[Level of evidence: 1iiDiii]
    ■ The overall objective response rate was 20.5% for patients who were randomly assigned to treatment with topotecan and 13.2% for patients who were randomly assigned to treatment with paclitaxel (P = .138).
    ■ Both groups experienced myelosuppression and gastrointestinal (GI) toxic effects. Nausea and vomiting, fatigue, and infection were observed more commonly after treatment with topotecan, whereas alopecia, arthralgia, myalgia, and neuropathy were observed more commonly after treatment with paclitaxel.

4. The combination of weekly topotecan and biweekly bevacizumab was evaluated in a phase II study.
    ■ Results showed an objective response rate of 25% (all partial responses) in a platinum-resistant patient population.
    ■ The most common grade 3 and grade 4 toxicities were hypertension, neutropenia, and GI toxicity, though no bowel perforations occurred.
• Pegylated liposomal doxorubicin
Evidence (pegylated liposomal doxorubicin):
1. In a phase II study encapsulated doxorubicin was given intravenously (IV) once every 21 to 28 days.
    ■ Results demonstrated one complete response and eight partial responses in 35 patients with platinum-refractory or paclitaxel-refractory disease (response rate, 25.7%).
    ■ In general, liposomal doxorubicin has few acute side effects other than hypersensitivity. The most frequent toxic effects (stomatitis and hand-foot syndrome) were usually observed after the first cycle, and were more pronounced after dose rates exceeded 10 mg/m2 per week. Neutropenia and nausea were minimal, and alopecia rarely occurred.

2. Pegylated liposomal doxorubicin and topotecan have been compared in a randomized trial of 474 patients with recurrent ovarian cancer.[35][Level of evidence: 1iiA]■ Response rates (19.7% vs. 17.0%; P = .390), PFS (16.1 weeks vs. 17.0 weeks; P = .095), and OS (60 weeks vs. 56.7 weeks; P = .341) did not differ significantly between the pegylated liposomal doxorubicin and topotecan arms.[Level of evidence: 1iiA]
    ■ Survival was longer for the patients with platinum-sensitive disease who received pegylated liposomal doxorubicin.

• Docetaxel.
This drug has shown activity in paclitaxel-pretreated patients and is a reasonable alternative to weekly paclitaxel in the recurrent setting.

• Gemcitabine.

Evidence (gemcitabine):
1. Several phase II trials of gemcitabine as a single-agent– administered IV on days 1, 8, and 15 of a 28-day cycle have been reported.
    ■ The response rate ranges from 13% to 19% in evaluable patients.
    ■ Responses have been observed in patients whose disease is platinum refractory and/or paclitaxel refractory as well as in patients with bulky disease.
    ■ Leukopenia, anemia, and thrombocytopenia were the most common toxic effects. Many patients reported transient flu-like symptoms and a rash after drug administration. Other toxic effects, including nausea, were usually mild.

2. A randomized trial of gemcitabine versus pegylated liposomal doxorubicin showed noninferiority and no advantage in therapeutic index of one drug over the other.

• Pemetrexed.
Evidence (pemetrexed):
1. A randomized, double-blinded phase II European trial with 102 patients evaluated pemetrexed at two doses: standard-dose (500 mg/m2) versus high-dose (900 mg/m2) IV every 3 weeks.
    ■ The response rate was 9.3% for the standard dose and 10.4% for the high dose.
    ■ The toxicity profile favored the standard dose, with fatigue, nausea, and vomiting as the most common severe toxicities.

2. A phase II study by the Gynecologic Oncology Group utilized pemetrexed (900 mg/m2) IV every 3 weeks in 51 patients with platinum-resistant recurrent disease.
    ■ The response rate was 21% in a heavily pretreated population in which 39% of the patients had received five or more regimens previously.
    ■ Myelosuppression and fatigue were the most common severe toxicities.

Chemotherapy and/or bevacizumab

• Chemotherapy with or without bevacizumab.
The U.S. Food and Drug Administration has approved the use of bevacizumab in combination with pegylated liposomal doxorubicin, paclitaxel, or topotecan as a result of the OCEANS and AURELIA trials.

OCEANS (NCT00434642) assessed the role of bevacizumab in the treatment of platinum-sensitive recurrences.

Evidence (bevacizumab with chemotherapy):
1. The Avastin Use in Platinum-Resistant Epithelial Ovarian Cancer (AURELIA [NCT00976911]) trial was an open-label randomized trial designed to evaluate the effect of adding bevacizumab to standard chemotherapy in patients with platinum-resistant recurrent ovarian cancer. Eligible patients had platinum-resistant disease (progression within 6 months of finishing a regimen) and no more than two previous regimens. Patients with platinum-refractory disease (those with progression during receipt of a platinum-containing regimen) and those with clinical or radiological signs of bowel involvement were ineligible. Patients were prescribed one of three chemotherapy regimens, on the basis of physician preference:
    a. Pegylated liposomal doxorubicin 40 mg/m2 by IV on day 1 every 4 weeks.
    b. Paclitaxel 80 mg/m2 by IV on days 1, 8, 15, and 22 every 4 weeks.
    c. Topotecan 4 mg/m2 by IV on days 1, 8, and 15 every 4 weeks; or 1.25 mg/m2 by IV on days 1 through 5 every 3 weeks.

Patients were then randomly assigned to receive either chemotherapy alone or chemotherapy with bevacizumab (10 mg/kg every 2 weeks, or 15 mg/kg every 3 weeks if on the 3-week-dosing schedule). Crossover to a bevacizumab-containing regimen was allowed at progression for those patients in the chemotherapy-only arm. PFS was the primary outcome, with response rate, OS, safety, and quality of life used as secondary endpoints. The enrollment included 361 patients with a median follow-up of 13.9 months in the chemotherapy-only arm and 13.0 months in the chemotherapy-plus-bevacizumab arm.
    ■ Patients in the bevacizumab arm exhibited longer PFS (HR, 0.48; 95% CI, 0.38 to 0.60); median PFS was 3.4 months in the chemotherapy alone arm versus 6.7 months in the chemotherapy plus bevacizumab arm.
    ■ The objective response rate was 12.6% in the chemotherapy-alone arm versus 30.9% in the chemotherapy-plus-bevacizumab arm.
    ■ There was no statistically significant difference in OS between the regimens (13.3 months chemotherapy alone vs. 16.6 months chemotherapy plus bevacizumab).
    ■ Patients in the chemotherapy-plus-bevacizumab arm had an increased incidence of hypertension and proteinuria, when compared with patients in the chemotherapy-only arm.
    ■ GI perforation occurred in 2% of those receiving chemotherapy plus bevacizumab, which reflects the study’s stringent exclusion criteria.
    ■ The primary endpoint for the quality-of-life portion of the study was a 15% or greater absolute improvement in the abdominal and GI symptom portion of the assessment modules at week 8 to week 9 of the protocol for patients in the chemotherapy plus bevacizumab arm.[Level of Evidence: 1iC] The study used patient-reported outcomes from the European Organization for Research and Treatment of Cancer (EORTC) Ovarian Cancer Module 28 and the Functional Assessment of Cancer Therapy-Ovarian Cancer symptom index at baseline and every 8 to 9 weeks until disease progression.

Although there were some limitations in study design, more patients on the chemotherapy-plus-bevacizumab arm had 15% or greater improvement in their GI scores when compared with baseline. For the chemotherapy-plus-bevacizumab arm, 34 of 115 patients (29.6%) showed improvement versus 15 of 118 (12.7%) patients who showed improvement on the chemotherapy-alone arm (difference, 16.9%; 95% CI, 6.1%–27.6%; P = .002).

These studies confirm the effect of improving PFS when bevacizumab is added to chemotherapy for ovarian cancer. In the OCEANS trial, the HR for progression was even more prominent than in the first-line trials, and a significant effect was seen when the bevacizumab-chemotherapy combination was extended beyond six cycles until progression.
In summary, the improvement achieved by bevacizumab in relative risk and PFS rates in platinum-sensitive and platinum-resistant recurrences has been consistently more than the improvement achieved with chemotherapy alone; however, bevacizumab-related toxic effects must be considered.

• Bevacizumab alone.
Three phase II studies have shown activity for this antibody to vascular endothelial growth factor (VEGF).
1. The first study (GOG-0170D) included 62 patients who had received only one or two previous treatments. These last patients had received one additional platinum-based regimen because of an initial interval of 12 months or longer after first-line regimens and also had to have a performance status of 0 or 1. Patients received a dose of 15 mg/kg every 21 days.
    ■ There were 2 complete responses and 11 partial responses, a median PFS of 4.7 months, and an OS of 17 months. This activity was noted in both platinum-sensitive and platinum-resistant subsets.

2. The second study included only patients with platinum-resistant disease using an identical dose schedule.
    ■ The study was stopped because 5 of 44 patients experienced bowel perforations, one of them fatal; seven partial responses had been observed. This increased risk of bowel perforations was associated with three or more previous treatments.[Level of evidence: 3iiiDii]
3. The third study (CCC-PHII-45) included 70 patients who received 50 mg of oral cyclophosphamide daily, in addition to bevacizumab (10 mg/kg every 2 weeks).
    ■ Partial responses were observed in 17 patients, and 4 patients had intestinal perforations.

Other drugs used to treat platinum-refractory or platinum-resistant recurrence (efficacy not well defined)

The drugs shown in Table are not fully confirmed to have activity in a platinum-resistant setting, have a less desirable therapeutic index, and have a level of evidence lower than 3iiiDiv.