Systemic chemotherapy is the standard treatment for metastatic epithelial ovarian cancer. Early studies utilized single alkylating agents such as melphalan or combinations of alkylating agents and antimetabolites. After the introduction of cisplatin in the latter half of the 1970s, platinum-based combination chemotherapy became the most frequently used treatment regimen in the United States. Cisplatin was initially used in combination with cyclophosphamide and doxorubicin, but the modest contribution of doxorubicin to the efficacy of the combination prompted simplification of the standard regimen to cyclophosphamide and cisplatin. By the early 1990s a combination of carboplatin and cyclophosphamide was found to be as effective as cisplatin and cyclophosphamide, but substantially less toxic, with less nausea, renal toxicity and neurotoxicity. Paclitaxel was incorporated into combination chemotherapy in the 1990s.  Over the last decade, a substantial amount of clinical investigation has centered on (1) whether a combination of a platinum compound and a taxane is superior to an optimal dose of a platinum compound alone, (2) the choice of the optimal platinum compound and taxane, and (3) the most convenient and effective schedule for taxane administration. Recent studies have focused on the introduction of a other compounds into a “triplet” or a “sequential doublet” to eliminate drug resistant cancer cells.

Single Alkylating Agents
For many years, oral single-agent alkylating therapy was used. The standard dose for the single alkylating agent, melphalan, is 0.2 mg/kg/day given orally for 5 consecutive days every 28 days. In three separate GOG studies of suboptimal stage III ovarian cancer, 193 patients were treated with this regimen. Sixty-two patients (33%) had a clinical response; with a 16% complete response rate and a 17% partial response rate. However, the median duration of response was only 7 months and median survival was 12 months. The use of single-agent chemotherapy for metastatic epithelial ovarian cancer is generally reserved for patients whose overall physical condition precludes the use of more toxic therapy.

A variety of regimens containing combinations of cytotoxic drugs have been tested in the treatment of advanced epithelial ovarian cancer. Combination chemotherapy has been shown to be superior to single-agent therapy in most patients with advanced epithelial ovarian cancer.

Platinum Compounds
For the last two decades, platinum compounds have been the most active agents against ovarian cancer. In early studies from England, cisplatin was found superior to cyclophosphamide as a single agent. Concurrently, cisplatin was tested in combination with a variety of different drugs. Platinum-containing regimens have proven superior to regimens that lacked platinum compounds. In a meta-analysis performed on 37 randomized studies involving 5,667 patients with advanced-stage disease, those patients given cisplatin-containing combination chemotherapy were compared with those treated with regimens that did not include cisplatin. Platinum based chemotherapy was superior to non-platinum based chemotherapy. A trend favored platinum combinations over single agent platinum. In studies of cisplatinum containing regimens, several trials have compared cyclophosphamide and cisplatin (CP) with cisplatin, doxorubicin and cyclophosphamide (PAC).  No study showed a significant difference in survival between treatment arms. The GOG’s randomized prospective comparison of equitoxic doses of PAC versus PC showed no benefit to the inclusion of doxorubicin in the combination. While a meta-analysis of the combined data from these four trials showed a 7% survival advantage at 6 years for those patients treated with the doxorubicin-containing regimen, the survival curves converge at 8 years. Consequently, in recent years doxorubicin has not been incorporated in regimens for epithelial ovarian cancer in the United States. It is, however, possible that a small subset of ovarian cancer patients benefit substantially from treatment with anthracyclines. In the future, new technologies may permit the identification of those patients and permit individualization of their treatment.

In the 1980s, the second-generation platinum analog carboplatin became available, exhibiting greater myelotoxicity, but substantially less nephrotoxicity, ototoxicity, peripheral neurotoxicity, and emetogenic potential than did cisplatin. Initial studies showed that carboplatin and cisplatin had approximately a 4:1 ratio of potency. Thus a standard single-agent dose of about 400 mg/M₂ has been used in most Phase II trials. The dose is best calculated by using the probable area under the curve (AUC) based on the glomerular filtration rate (GFR) according to the Calvert formula. When used as a single agent, a target AUC of 7 may be appropriate for untreated patients with ovarian cancer. When used in combination with other myelotoxic drugs, AUCs of 5 to 6 have been used more frequently.

Meta-analysis has suggested that cisplatin and carboplatin are equally effective against epithelial ovarian cancer. Direct comparison of cisplatin and cyclophosphamide to carboplatin and cyclophosphamide demonstrated comparable response rates and survival, but significantly less toxicity for the carboplatin containing regimen.

Paclitaxel
Paclitaxel was shown to be a very active agent against ovarian cancer. The overall response rates for paclitaxel in phase II trials was 36% in previously treated patients, which is a higher rate than was seen for cisplatin when it was first tested.

Three large concurrently controlled randomized trials in previously untreated patients with ovarian cancer have compared paclitaxel and a platinum compound to different platinum containing regimens that did not contain paclitaxel. The Gynecologic Oncology Group randomized 410 women with suboptimally cytoreduced stage III-IV ovarian cancer to six cycles of chemotherapy with a combination of cisplatin (75 mg/M₂) and paclitaxel (135 mg/M₂ over 24 hr) or with cisplatin (75 mg/M₂) and cyclophosphamide (750 mg/M₂) (GOG-111). Patients who received the paclitaxel combination had a superior overall response rate (73% vs 60%), clinical complete response rate (51% vs 31%), prolongation of disease-free survival (18 months vs 14 months), and prolongation of overall survival (36 months vs. 24 months).

The superiority of paclitaxel-cisplatinum was confirmed in a trial (OV-10) conducted jointly by the European Organization for the Research and Treatment of Cancer (EORTC), the Nordic Ovarian Cancer Study Group (NOCOVA), and the National Cancer Institute of Canada (NCIC) in which 680 women with both optimal and suboptimal disease in stages IIB-IV were treated. Patients were randomized to to six cycles of cisplatin (75 mg/M₂) and cyclophosphamide (750 mg/M₂) or to six cycles of cisplatin (75 mg/M₂) with a slightly higher dose of paclitaxel (175 mg/M₂) over a shorter interval (3 h) than in GOG-111. In this study, the paclitaxel-containing arm produced a significant improvement in median progression-free interval (15.5 vs 11.5 months) and overall median overall survival (35.6 vs 25.8 months) that extended to both optimal and suboptimal groups. Greater neurotoxicity was observed with the combination of paclitaxel and cisplatin when the paclitaxel was infused over 3 hours in OV-10 compared to infusion over 24 h in GOG-111.

A third study, the International Collaborative Ovarian Neoplasm-3 (ICON-3), was conducted as 4 parallel trials across 130 centers in eight countries in Europe to compare carboplatin (AUC 5) and paclitaxel (175 mg/M₂ over 3 h) to regimens that did not contain paclitaxel. Prior to randomization, each patient and physician could choose a control arm that included either carboplatin (AUC 5) alone or a combination of cisplatin (50 mg/M₂), doxorubicin (50 mg/M₂) and cyclophosphamide (500 mg/M₂). Of the women on the control arm, 69% received carboplatin. Dose escalation was permitted and occurred in approximately half of patients. Overall, 2,074 women participated with ovarian cancer of all stages, including 20% of women with stage I-II disease. A central review of pathology, surgical staging, cytoreduction or protocol adherence was not performed. No significant difference was found in overall median survival between the paclitaxel-carboplatin group (36.1 months) and the control group (35.4 months). In comparing the three trials, the outcome was no worse in the paclitaxel-carboplatin arm of ICON-3 than in the paclitaxel-cisplatin arms of GOG-111 and OV-10, but the control group in ICON-3 enjoyed a better outcome. The better survival of controls in ICON-3 might relate to the fraction of early stage patients or possibly to the optimization of carboplatin dosage permitted by dose escalation. A trend (p = .22) was noted toward a benefit of paclitaxel-carboplatin in patients who had > 2cm residual disease, patients comparable to those entered in GOG-111.

The GOG had undertaken a comparison of sequential and simultaneous paclitaxel-cisplatin (GOG-132). A three-arm comparison of equitoxic doses of paclitaxel (200 mg/M₂ over 24 h) versus cisplatin (100 mg/M₂) versus a combination of paclitaxel (135 mg/M₂ over 24 h) and cisplatin (75 mg/M₂) was carried out in 648 suboptimal stage III and IV patients. Crossover was permitted and the three groups exhibited similar median overall survival of 26 to 30 months. The simultaneous regimen was better tolerated than the sequential use of agents at these optimized doses. The overall response rate to cisplatin alone or to the paclitaxel-cisplatin combination was 67%, whereas the response to paclitaxel alone was 42% (p < .001). This observation suggests that more than half of patients who are treated arbitrarily with the combination of paclitaxel and a platinum derivative may not benefit from the taxane. To date, laboratory studies do not support any super-additive interaction of platinum compounds and taxanes, arguing against synergy in the clinic. In the future, using molecular markers it may be possible to identify those patients who would respond or not respond to paclitaxel, providing another opportunity to individualize therapy.

Two randomized, prospective clinical studies have compared the combination of paclitaxel and carboplatin to that of paclitaxel and cisplatin. In both studies, response rates and duration of survival are similar, but the carboplatin-containing regimens have more acceptable toxicity. In the first trial conducted by the GOG (GOG-158), patients were randomized to carboplatin (AUC 7.5) and paclitaxel (175 mg/M₂ over 3 h versus the previous standard of cisplatin (75 mg/M₂) and paclitaxel (135 mg/M₂ over 24 h). Progression-free survival of the carboplatin-containing arm was 22 months versus 21.7 months for the control arm. The gastrointestinal and neurotoxicity of the carboplatin arm were appreciably lower than that of the cisplatin arm. In addition, paclitaxel administration over 3 h is substantially more convenient than administration over 24 h. A similar result was obtained in a large randomized trial in Germany, in which carboplatin (AUC 6) and paclitaxel (185 mg/M₂ over 3 h) were compared to paclitaxel (135 mg/M₂ over 24 h) and cisplatin (75 mg/M₂). Thus, the best established regimen in patients with advanced stage disease is a combination of carboplatin and paclitaxel over 3 h.

Docetaxel
Docetaxel is a semi-synthetic second generation taxane with properties that differ from paclitaxel. Docetaxel is a more potent promoter of microtubule assembly and stabilization than paclitaxel. Docetaxel is taken up, bound and retained more effectively by cancer cells than paclitaxel. Docetaxel has produced a 23% to 28% overall response rate in platinum resistant ovarian cancer. A combination of docetaxel and cisplatin or carboplatin has achieved a 66% to 81% overall response rate in Phase II trials. Consequently, a combination of docetaxel (75 mg/M₂ over 1 h) and carboplatin (AUC 5) has been compared to paclitaxel (175 mg/M₂) and carboplatin (AUC 5) in the SCOTROC trial. Similar efficacy was observed, but docetaxel-carboplatin was associated with significantly less neurotoxicity.

Other Doublets and Triplets
Use of platinum compounds and taxanes has improved median and overall survival, but the outcome in patients with advanced ovarian cancer is still disappointing. Drug resistance ultimately develops in the majority. A number of drugs have exhibited activity against recurrent disease including liposomal doxorubicin, topotecan and gemcitabine. The GOG has conducted a series of Phase I pilot studies in previously untreated patients to define combinations that are suitable for a group wide Phase III trial. A five arm study has now been initiated, administering 8 cycles of chemotherapy to women with newly diagnosed stage III/IV ovarian cancer. The control arm consists of paclitaxel (175 mg/M₂ over 3h on day 1) and carboplatin (AUC 6 on day1). Two additonal arms contain triplets. The gemcitabine triplet includes paclitaxel (175 mg/M₂ over 3 h on day 1), carboplatin (AUC 5 on day 1) and gemcitabine (800 mg/M₂ / day on day1 and 8). The doxil triplet administers paclitaxel (175 mg/M₂ over 3 h on day 1), carboplatin (AUC 5 on d1) and doxil (30 mg/M₂ on day 1 every other cycle). Two arms contain sequential doublets. The topocetan doublet includes 4 cycles of carboplatin (AUC 5 on day 3) with topotecan (1.25 mg/M₂/ day on day1 to 3) followed by 4 cycles of paclitaxel and carboplatin. Finally, a gemcitabine doublet includes 4 cycles of carboplatin (AUC 6 on day 8) with gemcitabine (1000 mg/M₂/day on days 1 and 8) followed by 4 cycles of paclitaxel and carboplatin.

Dose Intensification with Intravenous Chemotherapy
Higher or more frequent doses of chemotherapy might be more effective, if tolerated. The issue of dose-intensification of cisplatin was examined in a prospective trial conducted by the GOG. In this study, 243 patients with suboptimal ovarian cancer were randomized to receive either 50 mg/M₂ or 100 mg/M₂ cisplatin plus 500 mg/M₂ cyclophosphamide. There was no difference in response rates in those patients with measurable disease and the overall survival was identical. As one might anticipate, there was greater toxicity associated with the high-dose regimen. A group in Scotland performed a similar study and found that patients who received 100 mg/M₂ cisplatin plus 750 mg/M₂ cyclophosphamide had a significantly longer median survival compared with those who received 50 mg/M₂ cisplatin plus the same dose of cyclophosphamide. The overall median survival time was 114 weeks in the high-dose group and 69 weeks in the low-dose group (p = .0008), but this difference disappeared with longer follow-up. Thus, doubling the dose of cisplatin does not improve long term survival. Dose escalation of paclitaxel and carboplatin require G-CSF due to their combined myelosuppressive effects, but there is no evidence to support more intensive administration of either agent.