SEARCH

SEARCH BY CITATION

Keywords:

  • ovarian neoplasms;
  • intraperitoneal;
  • drug therapy;
  • chemotherapy;
  • systematic review

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Because women with advanced ovarian cancer have poor outcomes, it is imperative to continue exploring for novel therapies. The opportunity for intraperitoneal treatment, especially in the subgroup of patients with minimal residual disease, in which the intraperitoneal approach may have a biologic rationale for benefit over and above the standard intravenous route, needs to be explored to the fullest extent. The MEDLINE, EMBASE, and Cochrane Library databases were searched up to January 2006 for randomized trials that compared first-line intraperitoneal-containing chemotherapy with first-line intravenous chemotherapy in the treatment of women with stage III epithelial ovarian cancer. Seven randomized, controlled trials were identified, including 3 large Phase III trials and 4 smaller randomized trials. The 3 large Phase III trials detected statistically significant overall survival benefits with intraperitoneal cisplatin-containing chemotherapy compared with intravenous chemotherapy alone. The improvements in survival were 8 months, 11 months, and 16 months, respectively. Pooled analysis from 6 of the 7 randomized trials confirmed the survival effect with intraperitoneal chemotherapy compared with intravenous chemotherapy alone (relative risk, 0.88; 95% confidence interval, 0.81–0.95). Severe adverse events and catheter-related complications with intraperitoneal chemotherapy were significantly more common and often were dose-limiting. The results from this review indicated that cisplatin-containing intraperitoneal chemotherapy should be offered to patients on the basis of significant improvements in overall survival. The appropriate clinical and institutional multidisciplinary facilities are needed for the safe delivery of this treatment in optimally debulked patients. Further research is needed concerning specific aspects of the treatment, such as optimal agent, dose, and scheduling. Cancer 2007;109:692–702. © 2007 American Cancer Society.

In North America, ovarian cancer occurs in 10.7 of every 100,000 women and has a mortality rate of 6.1 per 100,000 women.1 Although patients with low-risk, stage I disease can expect an approximate survival rate ≥90% at 5 years,2 patients who are diagnosed with advanced regional disease (stage III) or distant disease (stage IV) can expect a 5-year survival rate of approximately 30% to 50% and 13%, respectively.2 Unfortunately, approximately 75% of patients present with advanced disease (stage III or IV) at diagnosis.3, 4

Standard primary treatment for women with advanced ovarian cancer typically involves cytoreductive surgery and intravenous chemotherapy with a platinum agent administered either alone or in combination with a taxane. Despite an initial response rate of approximately 75% with this treatment approach, most patients who have advanced disease develop a recurrence and ultimately will die from complications of the malignancy within 5 years.3

The peritoneal cavity is a common site of ovarian cancer presentation or recurrence, and tumors tend to remain confined to the peritoneal cavity for most of their history.5 Because the disease initially presents on the peritoneum and local patterns of recurrence of this disease involve the peritoneum, the use of intraperitoneal therapy was explored.

The history of intraperitoneal therapy has both a clinical and pharmacologic basis. In the 1950s, intraperitoneal chemotherapy was used for controlling malignant ascites. In the late 1960s, intraperitoneal access options were explored. In the late 1970s, studies showed the slow clearance of chemotherapy drugs from the peritoneal cavity. This biochemical advantage could be exploited in ovarian cancer when it was possible to administer prolonged concentrations of cisplatin into the peritoneal cavity 10- to 20-fold higher than with the conventional intravenous route.5 Because intraperitoneal cisplatin was capable of penetrating small-volume tumors (1–3 mm), maximum chemotherapeutic benefit could be derived for patients with microscopic residual disease or very small-volume, macroscopic disease.5 In addition, the use of large doses of intraperitoneal cisplatin meant not only that the surface of the tumor could be exposed to high concentrations of cisplatin but also that, with a sufficient amount of drug leaking into the circulation, the amount of drug reaching the tumor through capillary flow was approximately doubled compared with a maximally tolerated dose of cisplatin delivered intravenously.6

Given the poor outcome of women with advanced ovarian cancer, it is imperative to continue to explore for novel therapies. The opportunity for intraperitoneal treatment, especially in the subgroup of patients who have minimal residual disease, for whom the intraperitoneal approach may have the biologic rationale of benefit over and above the standard intravenous route, needs to be explored to the fullest extent. There remains a critical need to examine innovative strategies that ultimately may have an impact on outcome in this disease.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Literature Search Strategy

The literature was searched by using MEDLINE (Ovid, 1966 through January 2006), EMBASE (Ovid, 1988 through January 2006), the Cochrane Library (Ovid, Issue 4, 2005), the Physician Data Query database, the Canadian Medical Association Infobase, and the National Guideline Clearinghouse. In addition, abstracts published in the proceedings of the meetings of the American Society of Clinical Oncology (1997–2005) and the European Society for Medical Oncology (2002–2004) were searched for evidence relevant to this review. Reference lists of related literature and recent review articles also were scanned for additional citations. The literature was searched for randomized, controlled trials by using the disease-specific terms ovarian neoplasms/ or ovar.ti and cancer.ti or carcinoma.ti or neoplasms/ combined with the treatment-specific terms intraperitoneal.ti or ip.ti or peritoneal.ti.

Study Selection Criteria

Articles were to be selected for inclusion in the systematic review of the evidence if they were published reports or published abstracts of randomized trials that compared patients with advanced (stage III) epithelial ovarian cancer who received first-line treatment that involved intraperitoneal-containing chemotherapy versus first-line treatment that involved intravenous chemotherapy only. Trials were to report data on some or all of the outcomes of interest: response, survival, toxicity, catheter-related complications, and/or quality of life.

Articles were excluded if treatment included immunotherapy, intraperitoneal radioactive phosphorus (32P), or hyperthermia. Trials also were excluded if they were reported in a language other than English and data could not be extracted.

Synthesizing the Evidence

Combining the results across trials added power for detecting the efficacy of the treatment and improves the reliability or confidence of the point estimate. Survival data were pooled using Review Manager 4.0.3 (Metaview© Update Software, Oxford, UK), obtained through the Cochrare Collaboration (www.cochrare.org). Results are expressed as the RR with 95% confidence intervals (CI), where an RR less than 1.0 favors the experimental treatment and an RR greater than 1.0 favors control. The random effects model was chosen over the fixed effects model as the more conservative estimate.7

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Literature Search Results

Seven randomized, controlled trials8–14 met the inclusion criteria. Three were larger Phase III trials,8–10 and 4 were smaller randomized trials.11–14 Table 1 lists the included studies and selected trial characteristics sorted by trial size and design. An additional report15 that included further information on Gynecologic Oncology Group (GOG) Trial 172 by Armstrong et al8 also was identified, and data on catheter-related outcomes were extracted from that article.15

Table 1. Literature Search Results and Trial Characteristics
Reference/studyNo. of PtsTrt ArmTrt regimenReceived all cycles, %*Residual disease
MacroscopicMicroscopicNoneVolume
AgentDose, mg/m2RouteDayCycle≤ 1 cm>1 cm
  • Pts indicates patients; Trt, treatment; GOG, Gynecologic Oncology Group; IV, intravenous; IP, intraperitoneal; q21d, every 21 days; AUC, area under the concentration-time curve; SWOG, Southwest Oncology Group; Cyclo., cyclophosphamide; NR, not reported; q28d, every 28 days; Epidox., epidoxorubicin.

  • *

    Values shown are the percentages of eligible patients who received all cycles of the assigned treatment.

  • Patients who received all cycles of the planned cisplatin dose.

  • Patients with residual disease ≤0.5 cm.

  • §

    Patients also may have received 50 mg/m2 of doxorubicin or epirubicin, depending on their cardiovascular condition.

Phase III randomized trials
 Armstrong et al., 20068210IVPaclitaxel135IV1q21d × 68364 361000
IVCisplatin75IV2q21d × 6
 GOG 172205IPPaclitaxel135IV1q21d × 64262 381000
IPCisplatin100IP2q21d × 6
IPPaclitaxel60IP8q21d × 6
 Markman et al, 20019227IVPaclitaxel135IV1q21d × 686643601000
IVCisplatin75IV2q21d × 6
 GOG 114235IPCarboplatinAUC = 9IV1q28d × 271653501000
IPPaclitaxel135IV1q21d × 6
IPCisplatin100IP2q21d × 6
 Alberts et al, 199610279IVCyclo.600IV1q21d × 658742672NR 
IVCisplatin100IV1q21d × 6
 SWOG/GOG 104267IPCyclo.600IV1q21d × 658752573NR 
IPCisplatin100IP1q21d × 6
Small randomized trials
 Yen et al, 20011163IVCyclo.§500IV1q21d × 632NRNRNR1000
IVCisplatin50IVNRq21d × 6
55IPCyclo.§500IV1Rq21d × 625NRNRNR1000
IPCisplatin100IPNRq21d × 6
 Gadducci et al, 20001256IVCisplatin50IVNRq28d × 69679912NRNR
IVEpidox.60IVNRq28d × 6
IVCyclo.600IVNRq28d × 6
57IPCisplatin50IPNRq28d × 664711811NRNR
IPEpidox.60IVNRq28d × 6
IPCyclo.600IVNRq28d × 6
 Polyzos et al, 19991346IVCarboplatin350IVNRq21d × 6NRNRNRNRNRNR
IVCyclo.600IVNRq21d × 6
44IPCarboplatin350IPNRq21d × 6NRNRNRNRNRNR
IPCyclo.600IVNRq21d × 6
 Kirmani et al, 19941433IVCisplatin100IVNRq21d × 660NRNRNR4552
IVCyclo.600IVNRq21d × 6
29IPCisplatin200IPNRq28d × 676NRNRNR6238
IPEtoposide350IPNRq28d × 6

Trial Characteristics

To be eligible to participate in the randomized trials, patients had to have newly diagnosed epithelial ovarian cancer that was either stage III8–11, 13 or ranged from stage II to stage IV.12, 14 In 1 trial, approximately 12% of patients had stage III primary peritoneal cancer.8 The median patient age ranged from a low of 52.8 years to a high of 61.0 years.8–14 Patients had to have adequate performance status, blood counts, and renal and hepatic function8–14 and could not have received any prior chemotherapy or radiotherapy.8, 9, 12, 14

Across the trials, the majority of patients were diagnosed with the histologic subtype serous adenocarcinoma.8–14 Trials were conducted in the United States,8–10, 14 Taiwan,11 Italy,12 and Greece.13 Trial participants were predominantly white (≥90%) in 3 trials,8–10 whereas the remaining trials did not report explicit information on ethnicity.11–13, 14 When it was reported, patients were stratified by performance status,10, 13 the presence of residual disease,8 the amount of residual disease,10, 13, 14 tumor grade,12 whether second-look surgery was selected,8 the time of enrolment,10 and/or cooperative group.10 Minimal residual disease was categorized as tumor that measured ≤0.5 cm in 1 trial,10 ≤1 cm in 4 trials,8, 9, 11, 14 <2 cm in 1 trial,12 and ≤2 cm in 2 trials.10, 13 Patients decided whether to undergo second-look laparotomy at registration in 1 trial,8 and all patients with negative disease at the completion of treatment were requested to undergo a second-look laparotomy in 6 trials.9–14

With the exception of 1 trial,14 all of the studies compared combination chemotherapy that included either cisplatin or carboplatin delivered through intravenous versus intraperitoneal injection.8–13 In that 1 trial, intraperitoneal etoposide was combined with intraperitoneal cisplatin as part of the investigational arm.14 In another trial, patients in both treatment arms also may have received 50 mg/m2 of doxorubicin or epirubicin, depending on their cardiovascular condition.11 Most of the treatment regimens were scheduled for 6 cycles of chemotherapy, with the exception of 1 trial that delivered 8 cycles in the investigational arm only.9

Protocols for treatment modification included cycle delay,8–10, 13 dose reduction,8 or the addition of granulocyte-stimulating factor.8 Patients were removed from the study if treatment delay was >3 weeks8 or was caused by excess toxicity.10 Dose reductions were not allowed in 1 trial9 and were not reported in 5 trials.11–14 Patient crossover because of excess toxicity or catheter-related complications was reported in 3 trials8, 9, 12 and was not reported in 4 trials.10, 11, 13, 14 One group of investigators reported that, if toxicity from cisplatin in either treatment arm was intolerable, then patients were switched to intravenous carboplatin.8

The type of catheters reported was implantable in 1 trial,8 Tenckhoff in 2 trials,9, 11 Port-A-Cath in 2 trials,12, 14 and temporary in 2 trials12, 13 and was not reported in 1 trial.10 Across the 7 trials, the proportion of patients who received all cycles of the assigned chemotherapy ranged from 32% to 96% of patients in the control arms and from 25% to 76% of patients in the treatment arms that involved intraperitoneal-containing chemotherapy.8–14

Primary outcomes were the pathologic response rate,9–11, 14 progression-free survival,8, 9 and/or overall survival.8–10, 12 Quality of life was assessed by using the Functional Assessment of Cancer Therapy-Ovarian (FACT-O) instrument in 1 trial8 and was not reported in the remaining trials.9–14

Trial Quality

All of the identified trials were nonblinded, randomized, controlled trials.8–14 In reports from 4 trials, information about the randomization procedure was provided,8, 9, 11, 12 whereas reports from the remaining trials did not provide that information.10, 13, 14 Five trials8–12 reported patient accrual with sufficient power to detect significant differences between treatment groups with 1-sided8, 9 or 2-sided significance10–12 testing at an α level of .05.8–12 Two trials did not report power calculations.13, 14 Two trials were stopped early,12, 14 and 1 trial was extended for increased patient accrual.10 In the trial by Gadduci et al,12 patient accrual was stopped early after 113 patients entered the study. In that trial, only 60% of eligible patients actually were randomized.12 Another study was closed early because of poor patient accrual with the introduction of carboplatin and the shift in community practice toward carboplatin-based chemotherapy.14 In the trial by Alberts et al,10 a separate analysis of patients who had disease that measured ≤0.5 cm was added as part of the study design. In that trial, patient accrual continued for an additional year to achieve a sufficiently large sample size appropriate for analysis.10

Baseline characteristics generally were similar across treatment groups in all of the 7 trials8–14; however, statistical comparisons between treatment arms were reported in only 2 trials.11, 13 Both of those trials reported no statistically significant differences in baseline characteristics between treatment groups.11, 13

Across 6 trials, completeness of follow-up was >80%.8–13 In 1 trial, the planned analysis included only evaluable patients.14 In that trial of 87 patients who were randomized to receive either intravenous or intraperitoneal chemotherapy, results for 62 evaluable patients were reported. Five trials reported an intention-to-treat principle either on eligible patients8–11 or on the whole randomized study population.12 The proportion of patients deemed ineligible after randomization ranged from 0% to 17% of the patient populations.8–12, 14 In 1 trial, the number of patients deemed ineligible was not reported.13 One trial9 began as a 3-arm study, but the third arm of cisplatin and cyclophosphamide was discontinued. Only 66 patients were accrued in that arm, and no results were reported for those patients.

Outcomes

Table 2 displays response and survival results for the 7 randomized trials that were identified in the review of the evidence. Response, progression-free survival, and median survival were the primary outcomes of interest. Five-year overall survival generally also was reported or data were available through extraction from survival curves; however, statistical estimates of differences between treatment groups were not provided in any of the randomized trials. One trial also reported disease-free survival as a study endpoint.14 Disease-free survival rates from that study14 are presented in Table 2 under progression-free survival.

Table 2. Response Rates and Survival Outcomes
ReferenceNo. of ptsTrt ArmSecond-look SurgeryPFSOS
No. of ptsCR, %Median, months5-Year, %*Median, months5-Year, %*
  • PFS indicates progression-free survival; OS, overall survival; Pts., patients; Trt., treatment; CR, complete response; IV, intravenous; IP, intraperitoneal; NA, not applicable; RR, relative risk; 95% CI, 95% confidence interval; NR, not reported; HR, hazard ratio; NS, not significant.

  • *

    Where data were not available, 5-year survival was extracted by the reviewer from the survival curves and is denoted as an approximation.

  • One tailed significance testing.

  • The clinical CR rate was based on the intention to treat population of 128 patients.

  • §

    The clinical CR rate was based on the entire study population.

  • Disease-free survival.

  • #

    Data were extracted by the reviewer from survival curves and are denoted as approximations.

Phase III randomized trials
 Armstrong et al, 20068210IV724118215037
205IP695724276645
  P value   NA.05NR.03NR
  RR (95% CI)    0.80 (0.64–1.00) 0.75 (0.58–0.97) 
 Markman et al, 20019227IV193NR22255245
235IP182NR28326354
  P value   NA.01NR.05NR
  RR (90% CI)    0.78 (0.66–0.94) 0.81 (0.65–1.00) 
 Alberts et al, 199610279IV15836NRNR4138
267IP13947NRNR4945
  P value   NA  .02NR
  HR (95% CI)      0.76 (0.61–0.96) 
Small randomized trials
 Yen et al, 20011163IV7338NRNR48˜25
55IPTotal36NRNR43˜26
  P value   .99  .47NR
  HR (95% CI)      1.13 (0.69–1.86) 
 Gadducci et al, 20001256IV343925˜3451˜44
57IP334342˜3967˜58
  P value   NS.13NR.14NR
 Polyzos et al, 19991346IV648§19NR25NR
44IP445§18NR26NR
  P value   NSNSNRNSNR
 Kirmani et al, 19941433IV195814˜20˜36#˜50
29IP165612˜20˜32#˜28
  P value   NS.46NR.45NR
Response rates after second-look surgery

No statistically significant differences in clinical or pathologically confirmed complete response rates were reported between any of the treatment arms in any of the trials.8–14 Of the 3 Phase III studies, Armstrong et al8 reported a 16% difference in the complete pathologic response rate in favor of intraperitoneal versus intravenous chemotherapy, but only a small proportion of their patients actually underwent second-look surgery, and response was not a planned study endpoint. In the remaining 2 trials,9, 10 results from second-look surgery were a planned endpoint but were deemed unreliable because of low procedure rates and the disproportionate number of procedures performed between treatment groups. Markman et al9 did not report results for complete response, whereas Alberts et al10 reported an 11% difference in the complete response rate in favor of patients who received intraperitoneal chemotherapy. Across all of the studies, complete pathologic response rates ranged from 36% to 58% with intravenous chemotherapy and from 36% to 57% with intraperitoneal chemotherapy.8, 10–14

Progression-free survival

Table 2 shows that 2 of the Phase III trials detected statistically significant 6-month improvements in median progression-free survival in favor of patients who received intraperitoneal-containing chemotherapy versus intravenous chemotherapy.8, 9 In 1 trial,9 after adjusting for important prognostic factors (residual disease status, cell type, and histologic grade), the relative risk (RR) estimate decreased slightly from 0.78 (90% confidence interval [90% CI], 0.66–0.94) to 0.75 (90% CI, not reported). Clinically, the longer median progression-free survival for patients who received intraperitoneal-containing chemotherapy reflects an approximate 20% reduction in the risk of progression compared with patients who received intravenous chemotherapy.

Of the remaining studies, 2 underpowered trials12, 13 reported no statistically significant differences between treatment groups, and 3 trials did not report results for that outcome.10, 11, 14 One trial10 reported that the amount of residual disease at the start of treatment (≤0.5 cm vs >0.5 cm to 2 cm) was the most important factor in determining response regardless of the type of treatment administered (P = .005). One trial that reported disease-free survival as an outcome of interest reported no statistically significant differences between treatment arms.14

Although it was not a study endpoint in any of the randomized trials, there was sufficient information from 3 randomized trials8, 9, 12 to extract data for pooling progression-free survival at 5 years. The remaining trials did not report data sufficient for pooling.10, 11, 13, 14 Figure 1 shows that, with an RR of 0.91 (95% CI, 0.85–0.98), the pooled data indicate that treatment involving intraperitoneal chemotherapy extends progression-free survival to a greater extent than intravenous chemotherapy alone. The beneficial effect was similar when only the Phase III trails8, 9 were pooled (RR, 0.92; 95% CI, 0.85–0.99). Thus, after 5 years, the clinical benefit with intraperitoneal-containing chemotherapy still was present with an approximate 10% reduction in the risk of progression compared with patients who received intravenous chemotherapy.

thumbnail image

Figure 1. Pooled analysis of 5-year progression survival. RR indicates relative risk; 95% CI, 95% confidence interval; Ch2, chi-square test; df, degrees of freedom.

Download figure to PowerPoint

Survival

Three large trials reported statistically significant improvements in median survival with intraperitoneal-containing chemotherapy compared with intravenous chemotherapy.8–10 The improvements in median survival were 16 months, 11 months, and 8 months in the 3 trials, respectively.8–10 The longer median survival for patients who received intraperitoneal-containing chemotherapy reflects an approximate 20% to 25% reduction in the risk of death compared with patients who received intravenous chemotherapy: a clinically significant benefit.

In the trial by Armstrong et al,8 there were no significant differences in outcomes between patients with and without visible residual disease, nor were there differences when results were analyzed by the number of eligible patients versus the number of randomized patients. In that trial, patients in the treatment arm received an extra dose of intraperitoneal paclitaxel on Day 8.

In the trial by Markman et al,9 the results were similar when adjusted for disease volume, histology, and age.9 In that trial, patients in the treatment arm received 8 cycles of chemotherapy, compared with 6 cycles of chemotherapy in the intravenous arm.

In the trial by Alberts et al,10 the factors associated with improved survival were absence of gross disease (P < .001), younger age (P < .001), type of tumor other than clear cell or mucinous (P < .001), and timing of enrolment after surgery (P < .001). The size of residual tumor (≤0.5 cm or from >0.5 cm to 2 cm) did not affect overall survival outcomes significantly.

Although the other 4 small trials reported no significant differences between treatment groups,11–14 1 underpowered trial did report a nonstatistically significant 16-month improvement in survival with intraperitoneal versus intravenous chemotherapy.12 Five-year overall survival was not reported as a study endpoint in any of the randomized trials; however, there was sufficient information from 6 of the 7 randomized trials8–12, 14 to extract data for pooling results across trials. The remaining trial did not report data sufficient for pooling.13 Figure 2 shows that the RR of 0.88 (95% CI, 0.85–0.98) indicates a greater treatment effect with intraperitoneal chemotherapy compared with intravenous chemotherapy alone. The effect was similar when only the Phase III trails were pooled8–10 (RR, 0.85; 95% CI, 0.78–0.94). The pooled results indicate that the clinical benefit with intraperitoneal-containing chemotherapy after 5 years still was present with an approximate 12% to 15% reduction in the risk of death compared with the risk for patients who received intravenous chemotherapy alone.

thumbnail image

Figure 2. Pooled analysis of 5-year overall survival. RR indicates relative risk; 95% CI, 95% confidence interval; Ch2, chi-square test; df, degrees of freedom.

Download figure to PowerPoint

Adverse events associated with chemotherapy

Table 3 presents the most common grade 3 or 4 adverse events reported across the 7 trials identified in the search of the literature. On average, more statistically significant adverse events were detected in patients who received intraperitoneal-containing chemotherapy compared with patients who received intravenously administered chemotherapy. The most common severe adverse events reported were hematologic and gastrointestinal. In 2 trials, approximately 75% of patients experienced severe leucopenia with intraperitoneal chemotherapy8, 9; and, in 1 trial, almost 50% of patients experienced severe gastrointestinal adverse events with intraperitoneal chemotherapy.8 In contrast, 3 trials reported significantly less leucopenia with cyclophosphamide and intraperitoneal cisplatin10, 11 or carboplatin13 compared with identical doses delivered intravenously. One trial10 also reported significantly less tinnitus, clinical hearing loss, and neuromuscular toxic effects in patients who received intraperitoneal chemotherapy. In 2 trials,10, 13 grade ≥2 abdominal pain was significantly greater in the intraperitoneal group (P < .01). Significant transient dyspnea also was observed in patients who received intraperitoneal chemotherapy and may have be related to compression of the base of the lung by the fluid-filled intraperitoneal cavity.10, 13

Table 3. Grade 3 or 4 Toxicity
ReferenceNo. of ptsTrt armGrade 3/4 adverse effects, %
LeucoNeutroThromboInfectionGIAnemiaFatigueRenal or GUNeurotoxicityMetabolicPainNo. of Trt-related deaths
  • Pts indicates patients; Trt, treatment; Leuko, leucopenia; Neutro, neutropenia; Thrombo, thrombocytopenia; GI, gastrointestinal; GU, genitourinary; IV, intravenous; NR, not reported; IP, intraperitoneal.

  • *

    Statistically significant differences between treatment groups (P ≤ .05).

  • Grade 4 neutropenia was significantly higher in patients treated with intraperitoneal chemotherapy, however actual data were not reported.

  • Toxicity grade ≥2.

  • §

    Toxicity was reported by cycle of chemotherapy.

Phase III randomized trials
 Armstrong et al, 20068210IV64*NR4*6*24*NR48*2*9*7*1*4
201IP76*NR12*16*46*NR18*7*19*27*11*5
 Markman et al, 20019227IV62NR3*<217*NR<45<9<2*NR2
235IP77NR49*<437*NR1<21210*NR2
 Alberts et al, 199610279IV50*69*9NRNR25NRNR25*NR20
267IP40*56*8NRNR26NRNR15*NR182
Small randomized trials
 Yen et al, 20011163IV33*NR16NRNR19NRNRNRNRNRNR
55IP18*NR13NRNR13NRNRNRNRNR1
 Gadducci et al, 20001256IV19NR2NRNR6NR00NRNRNR
57IP24NR0NRNR9NR00NRNRNR
 Polyzos et al, 19991346IV39*NR22NRNRNRNRNRNRNRNRNR
44IP11*NR7NRNRNRNRNRNRNRNRNR
 Kirmani et al, 19941433IV21§42§5§NR317§NR03NRNRNR
29IP19§42§0§NR193§NR69NRNRNR

Patient death caused by treatment was an infrequent occurrence, and there generally were similar rates detected between treatment groups. The greatest difference was 2 deaths in the intraperitoneal arm and no deaths in the intravenous arm in 1 of the trials.10

On average, adherence to the assigned chemotherapy regimen was greater for patients who received intravenous chemotherapy compared with patients who received intraperitoneal chemotherapy (Table 1). In the trial by Armstrong et al,8 only 42% of patients completed the assigned intraperitoneal chemotherapy regimen. In that trial, of 205 eligible patients in the intraperitoneal chemotherapy arm, 84 patients received ≥1 course(s) of intravenous chemotherapy, and 37 patients received carboplatin instead of cisplatin. Twenty-nine percent of patients discontinued intraperitoneal chemotherapy because of nausea, emesis, dehydration, renal, or metabolic adverse events.15 In another trial,9 18% of patients received ≤2 courses of intraperitoneal chemotherapy therapy primarily because of excessive bone marrow toxicity with initial doses of intravenous carboplatin (area under the curve [AUC], 9). In 1 trial in which toxicity was roughly similar between the 2 treatment arms, 20 patients (32%) switched to intravenous chemotherapy primarily because of patient refusal and/or logistic issues.12 In that trial, 11% of patients crossed over to intravenous chemotherapy prior to receiving any intraperitoneal chemotherapy.12 The remaining trials10, 11, 13, 14 did not report details on toxicity-related treatment discontinuation or cross over to intravenous chemotherapy.

Complications associated with intraperitoneal chemotherapy

The type of catheters employed varied within and among trials; and, as shown in Table 4, complications associated with intraperitoneal chemotherapy were not reported consistently. Of the 24% to 75% of patients who were not able to complete the assigned intraperitoneal chemotherapy regimen, only the authors of GOG Trial 172 explicitly reported the number of patients who discontinued treatment because of catheter-related complications.8, 15 Of the 119 patients who discontinued intraperitoneal treatment, catheter-related complications were either the primary factors (34%) or the contributing factors (8%) behind the decision to cross over to intravenous chemotherapy.8, 15 In 2 other trials,12, 14 ≥10% and ≥16% of patients experienced catheter-related complications that required treatment discontinuation. Catheter-related complications included abdominal pain, bleeding, infection, peritonitis, catheter blockage, leakage, movement, malfunction, and/or access problems.

Table 4. Complications Associated with Intraperitoneal Treatment
ReferenceNo. of PtsTrt ArmReceived <6 cycles of IP chemo, %*Catheter-related complications, %
TotalCatheter-related ComplicationsPainBlockageLeakBleedingDislodgedInfectionMalfunctionAccess. ProblemsPeritonitis
  • IP indicates intraperitoneal; Chemo, chemotherapy; Pts, patients; Trt, treatment; NR, not reported.

  • *

    Percentages reflect the 119 patients who were unable to complete all cycles of intraperitoneal chemotherapy.

  • Patients who received all cycles of the planned cisplatin dose.

Phase III randomized trials
 Armstrong et al, 20068 and Walker et al, 00615205IP583417*8*4*NRNR21*NR7*NR
 Markman et al, 20019235IP29NRNRNRNRNRNRNRNRNRNR
 Alberts et al, 199610267IP42NRNRNRNRNRNRNRNRNRNR
Small randomized trials
 Yen et al, 20011155IP75NR422618131197NR2
 Gadducci et al, 20001257IP3616509NRNRNR2NRNR9
 Polyzos et al, 19991344IPNR11NRNRNRNRNRNRNRNRNR
 Kirmani et al, 19941429IP24≥10NRNRNRNRNRNR14NR3

In 1 trial,8, 15 whether or not the intraperitoneal catheter was placed during the primary surgery did not appear to be significant. On a related note, with an increased incidence of infected catheters at the time of surgery, patients who underwent left colon or rectosigmoid resections were less likely to initiate intraperitoneal chemotherapy than patients who did not undergo those procedures (16% vs 5%; P = .012) and also were less likely to complete all 6 cycles of intraperitoneal chemotherapy (34% vs 44%; P value not reported).

Quality of life

Only 1 of the 7 trials reported quality-of-life outcomes.8 With lower FACT-O scores representing a poorer quality of life, Armstrong et al8 reported significantly lower scores for patients in the intraperitoneal chemotherapy arm versus patients in the intravenous chemotherapy arm prior to randomization (106.4 vs 111.9; P = .03), before the fourth cycle (103.3 vs 114.7; P<.001), and at 3 to 6 weeks after the sixth cycle (110.5 vs 118.4; P = .009). The difference in quality-of-life scores was not significant between treatment groups at 12 months after the completion of the sixth cycle of chemotherapy (125.5 vs 127.2; P = .56).

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

Three large Phase III trials and 4 smaller randomized trials inform the role of intraperitoneal chemotherapy in the first-line treatment of women with ovarian cancer. It is the results from the 3 Phase III trials that were powered sufficiently to detect meaningful survival differences between treatment groups that steer the discussion and inform the conclusions.

Although the phase III trials detected clinically and statistically significant survival advantages with intraperitoneal chemotherapy, there are several important reasons for the uncertainty surrounding the use of intraperitoneal chemotherapy as a standard practice for this patient population. One reason is that the optimal intraperitoneal chemotherapy regimen has yet to be defined. In the trial that produced the greatest survival benefits detected to date by Armstrong et al8, grade 3 or 4 toxicity was much more common in patients who received intraperitoneal chemotherapy, and most patients (58%) were unable to complete the assigned cycles of treatment. It is possible that the increase in toxicity was caused by the intraperitoneal dose of paclitaxel administered on Day 8 after intraperitoneal cisplatin; however, even in the control arm with fewer adverse events, the severe toxicity associated with intravenous paclitaxel and cisplatin remained considerable. The same applies to the trial by Markman et al9; although, in that trial, the increased toxicity in the treatment arm may have arisen from the additional 2 cycles of carboplatin (AUC, 9) administered only to patients in the intraperitoneal treatment arm. As a better-tolerated agent in place of cisplatin, intravenous carboplatin AUC 5 or 6 is administered as first-line treatment in this patient population, and intravenous paclitaxel at a dose of 175 mg/m2 is administered over 3 hours instead of over 24 hours, like in the studies by Armstrong et al8 and Markman et al,9 to reduce neurotoxicity. Historically, in a retrospective study, a lower response rate was detected with 200 mg/m2 to 300 mg/m2 of intraperitoneal carboplatin compared with 100 mg/m2 to 105 mg/m2 of intraperitoneal cisplatin for small-volume, residual cancer16; however; there is more recent literature on the use of higher doses of intraperitoneal carboplatin. In 1 study, a combination of intraperitoneal carboplatin with intravenous paclitaxel at 175 mg/m2 had an acceptable toxicity rate at an AUC of 6 to 7.17 Another study revealed that survival rates appeared to be higher among patients who received an intraperitoneal carboplatin dose >400 mg/m2 compared with patients who received doses less than this threshold.18 There also is some suggestion that intraperitoneal carboplatin may be more efficacious than intravenous carboplatin19; thus, the use of intraperitoneal carboplatin may be an attractive substitution for intraperitoneal cisplatin. However, there have been no randomized comparisons between intraperitoneal cisplatin and carboplatin to date. Furthermore, the 3 trials that showed a survival advantage in the current evidence series used intraperitoneal cisplatin.8–10 Thus, the role and optimal dose of intraperitoneal carboplatin are yet to be determined.

Second, the optimal patient population still is under debate, because the greatest benefits were observed in patients who had residual disease volumes ≤1 cm in any single area. For patients who have residual disease volumes >1 cm, it has been established that there is poor tumor penetration in bulky disease with intraperitoneal chemotherapy, but it is not known how this impacts on survival outcomes.

Third, there are important catheter-related complications and difficulties with implantation associated with intraperitoneal chemotherapy that patients and clinicians may be reluctant to accept. Alberts et al20 suggested that complications, such as fibrosis-sheath formation and small bowel obstruction or perforation, can be reduced with the use of the Port-A-Cath designed for intravenous injection, and Armstrong et al8 suggested that complications may be reduced with the standardization of the catheter device and of the technique and timing of port implantation. Despite potential future improvements in catheterization devices and implantation, in the trial by Armstrong et al,8 catheter-related complications played a large role in patient's decisions to cross over to intravenous chemotherapy.8, 15 In the review by Markman and Walker,21 practical guidance is provided for the delivery of intraperitoneal chemotherapy. Those authors reported that venous and intraperitoneal access devices may be placed at the time of initial surgery (resection and staging laparotomy) provided that surgery is optimal and without complications or, in the alternative, that such devices also may be placed several weeks after initial surgery. The drug-delivery device should be a fully implantable port attached to a large, single-lumen, venous, silicone catheter to avoid kinks and flow obstructions. In addition, peritoneal catheters with fenestrations and Dacron cuffs should not be used, because they are associated with complications and are not removed easily in the office setting under local anesthesia. Specific issues around the delivery of intraperitoneal chemotherapy, such as flushing the catheter with heparin, mixing and warming the chemotherapy agents, dealing with pain, hydration, and the assessment of peritonitis and/or gastrointestinal injury, are detailed further in that review.21

In conclusion, as front-line therapy for patients with optimally debulked stage III ovarian cancer, intraperitoneal chemotherapy confers an improved survival benefit in a patient population at high risk for disease recurrence. The increased survival benefits observed with intraperitoneal chemotherapy must be weighed against increased toxicity and complication rates related to catheterization. The optimal disease volumes for front-line therapy still are open to debate along with the optimal intraperitoneal chemotherapy regimens for front-line therapy. The greatest survival benefits reported to date were detected in patients who had ≤1 cm of residual disease in any 1 area with 135 mg/m2 of intravenous paclitaxel over 24 hours on Day 1, 100 mg/m2 of intraperitoneal cisplatin on Day 2, and 60 mg/m2 of intraperitoneal paclitaxel on Day 8 repeated every 21 days for 6 cycles. On that schedule, however, only 42% of patients were able to complete all 6 cycles of the assigned treatment.

The role of intraperitoneal chemotherapy in the first-line treatment of women with stage III epithelial ovarian cancer still is emerging. Overall, cisplatin-based, intraperitoneal chemotherapy for the first-line treatment of women with stage III epithelial ovarian cancer is a viable treatment option that should be offered to patients on the basis of significant improvements in overall survival. In the United States, the National Cancer Institute and the Society of Gynecologic Oncologists have endorsed the use of intraperitoneal chemotherapy in recent position papers.22, 23 In Canada, the Society of Gynecologic Oncologists of Canada also has endorsed the use of intraperitoneal chemotherapy.24 In each case, the provision of endorsement was that the appropriate clinical and institutional multidisciplinary facilities were needed for the safe delivery of this treatment in optimally debulked patients—a provision that supports the concerns raised in this review of the evidence. Future research in the randomized setting should focus on the optimal patient population to receive treatment and the best possible intraperitoneal chemotherapy regimen, including the most appropriate agents, infusate volume, schedule, and dose. The utility of intraperitoneal chemotherapy in various other patient populations with ovarian cancer, such as those with early-stage, high-risk disease, also should be explored in a prospective fashion.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES

We thank the members of the Gynecology Cancer Disease Site Group (DSG) for their contributions. For a complete list of DSG members, please visit www.cancercare.on.ca

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. REFERENCES