A Randomized Trial of Carboplatin Versus Iproplatin in Untreated Advanced Ovarian Cancer By C. Trask, A. Silverstone, C.M. Ash, H. Earl, C. Irwin, A. Bakker, J.S. Tobias, and R.L. Souhami Between August 1984 and October 1987, 120 patients with stage IC to IV epithelial ovarian cancer were randomly assigned to receive carboplatin (400 mg/m') or iproplatin (300 mg/m') every 4 weeks as initial treatment. Stratification was made according to International Federation of Gynecology and Obstetrics (FIGO) stage and according to size of residual disease after surgery. Response was evaluated after six courses when patients were restaged, with laparoscopy or laparotomy in clinical complete responders or those with no assessable disease. Treatment was then stopped in surgically proven complete responders. Patients with partial (PR) or minor response (MR) received a further six courses of their original drug at a reduced dose (carboplatin 300 mg/m', iproplatin 225 mg/m2). Patients with stable (SD), progressive (PD), or recurrent disease were treated with cyclophospha-
CISPLATIN
is currently one of the most active chemotherapeutic agents available for the treatment of ovarian cancer,1'2 producing survival rates in excess of 40% at 2 years for advanced disease.' 4 At doses of 100 mg/m 2, toxicity is considerable5'6 with severe emesis, renal impairment, peripheral neuropathy, high-frequency hearing loss, and anemia. Carboplatin (diamine-1,1-cylobutane dicarboxylase platinum II) and iproplatin (cis-dichloro-trans-dihydroxy-bis [isopropylamine] platinum IV) are two cisplatin analogs with fewer marked side effects at doses equivalent to 100
mg/m 2 of cisplatin. Studies in patients with advanced ovarian cancer have shown encouraging
results with both analogs in previously treated patients,7 and carboplatin achieved an equal response rate to cisplatin in randomized studies
involving previously untreated ovarian cancer patients.' 0"'. Myelosuppression was the most significant side effect and was dose-limiting in both the new analogs.
At the time of inception of this study, carboplatin and iproplatin were the most likely successors to cisplatin. While considerable experience has been obtained with carboplatin, much less is known of the relative effectiveness of iproplatin.
This study was planned to compare the response rates and toxicities of both agents, and to investi-
2 mide (1 g/m ). The response rates were 63% (95% confidence interval [CI], 50% to 74%) for carboplatin and 38% (95% CI, 26%to 51%) for iproplatin. Fifteen patients were not assessable for response. The median survival was 114 weeks (95% CI, 82 to 233 weeks) for carboplatin patients and 68 weeks (95% CI, 48 to 82 weeks) for iproplatin patients (P = .008). The amount of residual disease after initial laparotomy was a prognostic factor for survival. Myelosuppression was the main toxicity and was greater with iproplatin. This study shows carboplatin to be more active than iproplatin in the treatment of ovarian cancer and less toxic. Few responses to cyclophosphamide occurred following either drug, implying resistance to the alkylating agent. J Clin Oncol 9:1131-1137. © 1991 by American Society of Clinical Oncology.
gate the efficacy of second-line chemotherapy using cyclophosphamide on relapse. METHODS Between August 1984 and October 1987, 120 patients with International Federation of Gynecology and Obstetrics (FIGO) stage IC to IV epithelial ovarian carcinoma were entered onto the study. None had received prior chemotherapy or radiotherapy, and all had undergone a laparotomy to define accurately the extent of disease and to remove as much of the tumor as possible. At the initial laparotomy, the surgeon was required to designate the amount of residual disease into one of three categories: (1) no residual macroscopic disease, (2) maximum tumor masses less than 1.5 cm, and (3) maximum tumor mass greater than 1.5 cm. To be eligible for the study, the patients' Karnofsky performance status had to be 60% or greater and their life expectancy in excess of 3 months. Hematologic and renal function parameters had to be within the institutions'
From the Southend Hospital,Essex; and University College Hospital,London, England. Submitted May 17, 1990; accepted January25, 1991. Supported in part by a grantfrom Bristol Myers Company, PharmaceuticalResearchand Development Division (protocol no. 118-008-001). Address reprintrequests to C. Trask, MD, Southend Hospital, Prittlewell Chase, Westcliff on Sea, Essex SSO ORY, England. © 1991 by American Society of Clinical Oncology. 0732/183X/91/0907-0022$3.00/0
Journal of Clinical Oncology, Vol 9, No 7 (July), 1991: pp 1131-1137
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 6, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
1131
1132
TRASK ET AL
normal ranges and age between 18 and 80 years. Subjects were excluded from the study if there was a history of previous malignancy, brain metastasis, serious intercurrent infection, or chronic disease requiring regular treatment. Before entering the trial, each patient underwent a full clinical examination, chest x-ray, pelvic and abdominal ultrasound, biochemical tests of liver function, audiometry, and creatinine or pentetic acid (DTPA) clearance. CA 125 estimations were not performed. Patients were randomized within 1 month of original laparotomy by the trials coordinator using computer-generated stratified randomization sheets. Informed consent was obtained from each patient. The protocol was approved by the Ethical Committee of each participating institution. Patients were stratified according to FIGO stage and amount of residual disease and allocated to receive either carboplatin (400 mg/m2) or iproplatin (300 mg/m2) at 4 weekly intervals. Both drugs were diluted in 250 mL 5% dextrose and infused intravenously over 30 minutes without pre- or posthydration. Each center adopted the same antiemetic policy for the two drugs, but the regimen was left to the discretion of the participating centers and was usually a regimen containing high-dose metaclopramide, dexamethasone, and lorazepam. If the tumor failed to respond at any time after four cycles of therapy, or if tumor progression occurred at any time after two cycles, the patient was then 2 treated with cyclophosphamide at a dose of 1 g/m intravenously repeated every 3 to 4 weeks. Full clinical reassessment was made in responding patients after six courses, and complete clinical responders were subjected to a secondlook laparoscopy, or laparotomy when appropriate (and the patient consented). Patients with pathologic complete response (CR) (laparoscopy or laparotomy negative) were followed up at 2 monthly intervals, and no further chemotherapy was given until relapse at which time cyclophosphamide was given. Partial and minor responders continued with a further six cycles of the same analog at a reduced 2 dosage (carboplatin 300 mg/m2, iproplatin 225 mg/m ), and those patients with stable (SD) or progressive disease (PD) 2 were treated with cyclophosphamide also at I g/m . The patients who continued with a further six courses of carboplatin or iproplatin were again fully reassessed in the same way after their twelfth course. Complete responders then stopped therapy; partial responders or less were offered second-line treatment with cyclophosphamide. The objectives of the study were to compare response rates and toxicity. The number of patients required was determined before the trial began and was based on the ability to detect a 25% difference in response rate with a set 9 at 0.05 and p at 0.20. Available data from other studies" indicated that iproplatin appeared to be associated with more toxicity. The study was therefore designed to investigate whether iproplatin produced a higher response rate than carboplatin since, otherwise, carboplatin would be the preferred agent. A one-sided test was therefore chosen for analysis, using an estimated response rate of 50% for 12 carboplatin,l' and required 53 patients on each treatment. A total of 120 patients were recruited to allow for nonassessable responses. Survival curves were constructed according to the method of Kaplan and Meier" and statistical signifi4 cance evaluated by the log-rank test.'
Dose Adjustments and Scoring of Toxicity Three patients began treatment at 90% of the starting dose because they were aged older than 75 years. All patients had weekly blood counts during treatment to assess hematologic toxicity. Dose adjustments were made according to the nadir full blood count following the previous cycle (Table 1). Drug administration was postponed by 1 week if there had not been hematologic recovery (WBC > 3.0 x 107/ L or platelets > 100 x 10'/L). Renal toxicity was assessed by serum urea, creatinine, and electrolyte estimations before each treatment and by creatinine/DTPA clearances after every third course. If the serum creatinine rose above 120 l.mol/L and/or significant deteriorating renal function was confirmed by DTPA clearance, therapy was stopped. Ototoxicity was assessed by comparing serial audiograms (every third course) with pretreatment ones. Diarrhea, mucositis, infection, peripheral neuropathy, nausea, and vomiting episodes were recorded at each course using World Health Organization (WHO) criteria and duration of nausea and vomiting measured in hours.
Response Criteria For evaluation of response the WHO criteria were adopted." The response rate recorded was the best response achieved by a patient at any time during their treatment with carboplatin or iproplatin. Independent second-party review of responses was not made.
RESULTS
The characteristics of the 120 patients entered onto the study are summarized in Table 2 and were well balanced in both arms of the study. Median age was 59 years (range, 29 to 79), and about 60% achieved a Karnofsky status greater than 80% at the start of chemotherapy. Over 85% of patients had FIGO stage III or IV disease. Total hysterectomy and bilateral salpingooopharectomy was performed in 63% of patients receiving carboplatin and in 74% receiving iproplatin. In 48% of carboplatin patients and in 44% of
iproplatin patients poorly differentiated tumors were ascertained. Fifteen patients were nonassessable for response for the following reasons: five patients with no macroscopic residual disease the Table 1. Dose-Modification Schedule With Respect to Nadir Blood Counts Following the Previous Course WBCs
Platelets
> 100 75-99 50-74 25-49 < 25
and and/or and/or and/or and/or
> 4.0 3-3.9 2-2.9 1-1.9 < 1.0
% Previous Dose
110* 100 90 75 50
*Maximum dose, carboplatin 400 mg/m2, iproplatin 300 mg/ m2
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 6, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
1133
CARBOPLATIN v IPROPLATIN IN OVARIAN CANCER Table 2. Patient Characteristics at Presentation According to Assigned Treatment Carboplatin (n = 62)
No. Age (years) Range Median Karnofsky status (%) 100 90 80 70 60 Initial surgery TAH BSO O' TAH BSO Biopsy only Other Residual disease Nil FIGO stage IC 2 II 4 III 3 IV 0
Iproplatin (n = 58)
%
No.
29-79 59 30 13 9 6 4 16 23 10 13 < 1.5 0 2 12 0
%
34-75 59 48 21 15 10 6
17 17 12 5 7
26 37 16 21 > 1.5 Nil 0 1 27 11
1 3 3 0
29 29 21 9 12
15 26 28 48 10 17 5 9 < 1.5 > 1.5 0 1 11 0
0 2 26 11
Abbreviations: TAH, total abdominal hysterectomy; BSO, bilateral salpingo-oophorectomy; 0', omentectomy; Nil, no residual disease.
status of the peritoneal washings at their initial surgery was not recorded, three patients in clinical CR (CCR) could not be reexamined laparoscopically, three patients died of unrelated causes (two had cerebrovascular accidents, one after her first cycle, the other after her second, one patient died of hypercalcemia after her first cycle), two patients withdrew (one after five courses and one after one course), one patient developed a fatal second malignancy (carcinoma of cecum) after her sixth course, and one patient was taken off study because of impaired renal function after one course of iproplatin. Of the remaining 105 patients, three carboplatin and four iproplatin patients subsequently withdrew from treatment after their sixth
responders with carboplatin, 12 were confirmed by laparoscopy and eight by laparotomy. Of the 11 pathologic complete responders with iproplatin, five were confirmed by laparoscopy and six by laparotomy. PD on therapy developed in 43% of iproplatin and 19% of carboplatin patients, giving an increase in PD of 24% with iproplatin (95% CI of difference, 8% to 40%). After six courses of chemotherapy, 34 patients (20 carboplatin, 14 iproplatin) had achieved a PR and continued with treatment up to a maximum of 12 courses. Only one carboplatin patient and two iproplatin patients achieved a further improvement in response by continuing chemotherapy. Twenty-four patients who had received carboplatin and 23 who had received iproplatin went on to receive chemotherapy with cyclophosphamide. There were two CCRs to cyclophosphamide, one PR and five SD responses after carboplatin; 14 patients progressed and two were nonassessable. In the iproplatin arm, there were no CRs to cyclophosphamide, one patient achieved a PR, five had SD, 15 had PD, and two were nonassessable. The total response rate with cyclophosphamide was only 9% in the whole group, with 21% achieving SD. Twenty-one patients who progressed on firstline therapy received no relapse chemotherapy. Nine received palliative radiotherapy alone and 10 received other chemotherapy (eg, chlorambucilor melphalan-containing regimens) without response. Survival All 120 patients were analyzed for survival, although this study was not specifically designed to compare survival. In the survival analyses, three patients (two carboplatin, one iproplatin) were censored at the point of diagnosis of a second Table 3. Response Rates According to Assigned Treatment
course.
Response rates are shown in Table 3. A clinical partial response (CPR) or greater was achieved in 63% of carboplatin and 38% of iproplatin patients, giving a 25% improvement in response with carboplatin (95% confidence interval [CI] of difference, 8% to 42%). A pathologic CR was achieved in 32% of carboplatin patients and 19% of iproplatin patients, giving an improvement in response with carboplatin of 13% (95% CI of difference, 2% to 29%). Of the 20 pathologic complete
Carboplatin (n= 62)
Pathologic CR Pathologic PR CCR
20 12
52%
2
11%
CPR
5
Pathologic SD
4
Clinical SD
1
Pathlogic PD Clinical PD Nonassessable
2 10 6
Iproplatin (n = 58)
63%
11 8 2
8%
2 0
19%
3 222 9
10%
33% 38% 5% 3 3%
43 16%
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 6, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
1134
TRASK ET AL
malignancy, two patients (both carboplatin) were censored at the time of their death, since the causes of their deaths were not cancer-related, one patient (iproplatin) was censored due to loss of follow-up, and one patient (iproplatin) was censored at the time when taken off study due to impaired renal function. The median follow-up of the study was 35 months. The median progressionfree interval (time from starting chemotherapy until progression) was 73 weeks (95% CI, 51 to 109 weeks) for the carboplatin patients, and 38 weeks (95% CI, 24 to 57 weeks) for iproplatin patients (P = .019; Fig 1). The overall survival (Fig 2) was significantly better for carboplatin patients, 52% (95% CI, 39% to 65%) alive at 2 years compared with only 33% (95% CI, 20% to 45%) of iproplatin patients. The median survival was 114 weeks (95% CI, 82 to 233 weeks) for carboplatin patients versus 68 weeks (95% CI, 48 to 82 weeks) for iproplatin patients (P = .008). The size of residual disease remaining after initial surgery was a significant prognostic factor. The two arms of the study were well balanced with two thirds of patients having residual disease greater than 1.5 cm. At 3 years, 59% of patients with < 1.5 cm of tumor remaining and 62% of patients with no residual disease are alive compared with 26% of patients with greater than 1.5 cm disease (Fig 3). The median survival for the latter group is 68 weeks (95% CI, 59 to 82 weeks); for the combined better prognostic groups the median survival is 233 weeks (P < .001). The FIGO stage was also an important determinant of survival with FIGO IC and II patients
c u
M L
1 T E
S U
V
G
TIME ( MONTHS )
Fig 2. Overall survival for all patients randomized to carboplatin (n = 62; median survival, 114 weeks) or iproplatin (n = 58, median survival, 68 weeks). Carboplatin: observed events 32, expected events 43.21; iproplatin: observed events 41, expected events 29.79; ) = 7.12, P = .008.
surviving significantly longer than stage III (P = .012) or IV patients (P = .003) (Fig 4). Stage III patients had a median survival of 82 weeks (95% CI, 65 to 114 weeks), stage IV patients, 61 weeks (95% CI, 38 to 93 weeks); at 3 years, 75% (95% CI, 48% to 100%) of stage IC and II patients are alive. Histologic differentiation of the primary tumor was not related to prognosis. Toxicity Toxicity was analyzed only in the first six courses of carboplatin and iproplatin. Myelosuppression, nausea, and vomiting were the most significant side effects. The most serious hematologic toxicity was thrombocytopenia, reaching WHO grade 4
c U
P R 0 G R S S
N F E E
TIME ( MONTHS ) TIME ( MONTHS )
Fig 1. Progression-free interval for all patients randomized to carboplatin (n = 62; median interval, 73 weeks) or iproplatin (n = 58; median interval, 38 weeks). Carboplatin: observed events 37, expected events 47.10; iproplatin: observed events 41, expected events 30.90; = 5.47, P = .019.
?x
Fig 3. Overall survival for all patients randomized according to the amount of residual disease left after diagnostic laparotomy. Nil, no residual macroscopic disease (n = 16; median survival, 183 weeks); < 1.5 cm, residual disease < 1.5 cm (n = 26); > 1.5 cm, residual disease > 1.5 cm (n = 78; median survival, 68 weeks).
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 6, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
1135
CARBOPLATIN v IPROPLATIN IN OVARIAN CANCER J
.
14 Z
13 12 11 10
m Fn 0
0-1O
z
8
O
7 6 5 4
z
w
C) Z
0
w LU
3 2 1 0
2 4 TIME ( MONTHS )
Fig 4. Overall survival for all patients randomized according to initial FIGO stage. FIGO IC and II n = 16, median survival not ascertained; FIGO Ill n = 82, median survival 82 weeks; FIGO IV n = 22, median survival 61 weeks.
11NNN nn
I-. z
toxicity on four occasions in the carboplatin arm and 22 times in the iproplatin arm. Platelet transfusions were required on 14 occasions for 13 iproplatin patients, whereas none were needed in the carboplatin arm. Half the platelet transfusion episodes occurred before the second course of chemotherapy with iproplatin. Grade 4 leukopenia only occurred once in an iproplatin patient. There were seven episodes of infection (WHO grade 1 or 2 only) on carboplatin and three on iproplatin. Nine patients experienced mucositis on carboplatin and three on iproplatin. Blood transfusions were given nine times in eight carboplatin patients and 14 times in 11 iproplatin patients; the hemoglobin fell below 6.5 g/dL only once with both drugs. The cumulative hematologic toxicities for the first six courses for hemoglobin, WBC, and platelets are shown in Fig 5. Dose reductions due to myelosuppression were greater with iproplatin, although carboplatin caused more leukopenia than iproplatin (Table 4). By course 6, the mean dose of iproplatin was reduced to 69% of 300 mg/m2, and the mean
D o. I
0 z
10 COURSE NUMBER Fig 5. Cumulative hematologic toxicities showing nadir hemoglobin (g/dL), WBC count (x 10'/L), and platelets (x 109/ L) experienced by all patients with the first six courses, with SEM. (0) Iproplatin, (I) carboplatin.
for a mean of 13 hours with both carboplatin and iproplatin. Chemotherapy-induced diarrhea was a rare event with carboplatin (only three episodes in two Table 4. Reasons for Dose Modifications and Course Delays According to Assigned Treatment
carboplatin dose was 82% of 400 mg/m2 with a
mean difference of 13% (95% CI, 5% to 20%). Most patients experienced WHO grade 2 or 3 nausea and vomiting, although no patient experienced grade 4 intractable vomiting. The mean duration of nausea and vomiting is shown in Fig 6. The duration of nausea was longer with carboplatin: an overall mean of 58 hours compared with 48 hours with iproplatin with a mean difference of 10 hours (95% CI, 1 to 19 hours). Vomiting lasted
Dose adjustment reasons Leukopenia Thrombocytopenia Both Course delay reasons Leukopenia Thrombocytopenia Both
Carboplatin
Iproplatin
(n = 279)
(n = 239)
59 11 16
30 51 13
12 1 0
10 8 3
NOTE. Data are total number of cycles.
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 6, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
1136
TRASK ET AL 80
NAU 70
0, 60-
Z 0
504030-
Z 2 I
VOMITING 20o-
*
10-
1
2
3
4
5
6
COURSE NUMBER
Fig 6. Mean duration (with SEM) in hours of nausea and vomiting experienced by all patients with the first six courses. (0) Iproplatin, (0) carboplatin.
patients) compared with 56 episodes with iproplatin in 27 patients. Only one (iproplatin) patient experienced significant deterioration of renal function, and another patient experienced a deterioration of acoustic acuity as judged by serial audiograms, possibly attributable to iproplatin. This patient developed tinnitus after six courses of cyclophosphamide following three courses of iproplatin. Three patients showed WHO grade I peripheral neuropathic changes (all carboplatin), and in each case, the symptoms developed at the fourth course. Three patients experienced anaphylactic reactions (two carboplatin, one iproplatin), all during the eighth course. One iproplatin patient developed a skin reaction. The relationship between pretreatment creatinine clearance and overall toxicity was analyzed by relating clearance to maximum dose reduction by course 6, since dose reduction was on the basis of hematologic, renal, or gastrointestinal toxicity. No relationship was shown for either iproplatin or carboplatin (Pearson's correlation coefficient .08 and .203, respectively). DISCUSSION
Previous data have suggested that iproplatin may have a greater response rate than carboplatin8"9 in the treatment of ovarian cancer. The toxicity of iproplatin was expected to be greater than carboplatin. This study was therefore designed to be able to detect a 25% difference in response rate in favor of iproplatin, necessitating an accrual of at least 100 eligible patients. Our results confirm that the hematologic toxicity of
iproplatin was indeed greater than carboplatin. However, the response rate for iproplatin was only 38% compared with a response rate of 63% for carboplatin. Survival was worse in the iproplatin group (median survival, 68 weeks). Thus, at doses that are more toxic than carboplatin, iproplatin is less effective and, therefore, has no role in the treatment of ovarian carcinoma. Carboplatin proved to be a safe drug to use at a dose of 400 mg/m2, with acceptable toxicity. Acute gastrointestinal side effects were transient; vomiting lasted for a mean of 13 hours after the infusion, but the onset was usually delayed. Nausea persisted for a mean of 58 hours; however, the symptoms were never dose-limiting and were controllable on an outpatient basis. Diarrhea was unusual. Other studies have shown less nausea and vomiting toxicity for carboplatin than reported here." This probably reflects differences in antiemetic regimes or the length of inpatient stay. However, if carboplatin is to be used on an outpatient basis, an effective antiemetic regime must be instituted and the patient must be warned to expect a degree of nausea and vomiting. Myelosuppression was mild provided the dose adjustments were made with reference to weekly blood counts: by course 6, the dose of carboplatin had been decreased to 82% of the starting dose. Thrombocytopenia was sufficient to cause dose adjustments in 10% of courses, but no platelet transfusions were necessary. Leukopenia was never sufficiently profound to cause infections requiring intravenous antibiotics. As in previous studies,'0,1116 there was no significant degree of renal
toxicity, ototoxicity, or neurotoxicity, and this study again confirms that there is no need for any pre- or posthydration, implying that carboplatin is eminently suitable for use as an outpatient regimen. This study also confirmed that the size of residual disease after surgery is an excellent predictor of survival. This has been suggested by a number of retrospective analyses,'-. 1 but this study confirms prospectively that the size of residual disease after surgery is as good a predictor of survival as the FIGO stage itself. Patients with no residual disease or disease less than 1.5 cm in diameter show a 60% (95% CI, 43% to 78%) 3-year survival, whereas those with residual disease greater than 1.5 cm in diameter have only a 26% 3-year survival (95% CI, 15% to 37%) (P < .001).
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 6, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
CARBOPLATIN v IPROPLATIN IN OVARIAN CANCER
1137
Resistance to carboplatin and iproplatin appears to be accompanied by a considerable degree of resistance to cyclophosphamide. This may indicate that there will be little advantage in combining the two drugs. There would also seem to be little advantage in continuing chemotherapy beyond six cycles as only three of 34 patients in both arms improved their response rate by continuing therapy after six courses. The overall response rate to carboplatin was high, with an encouragingly long median survival of 26 months. These results compare favorably with other carboplatin single-agent studies... The median survival appears comparable with that obtained with cisplatin at a dose of 100 mg/m 2,
which is reported to be between 19 and 23 months in various studies.2,",' 6 Carboplatin can be administered as an outpatient drug with little acute toxicity. Since these results are comparable with those achieved in many studies of combination chemotherapy,3'4' 20°- a randomized study comparing carboplatin alone with combination chemotherapy would be valuable with an assessment of toxicity, survival, and cost. ACKNOWLEDGMENT We gratefully acknowledge the support of the following consultant gynecologists: Jane Crow, John Kersey, Michael Martin, Heulwen Morgan, Raymond Pilkington, Tim Pocock, Albert Singer, John Ward, and Chris Welch.
REFERENCES 1. Ozols RF, Young RC: Chemotherapy of ovarian cancer. Semin Oncol 11:251-263, 1984 2. Sessa C: European studies with cisplatin and cisplatin analogs in advanced ovarian cancer. Eur J Cancer Clin Oncol 22:1271-1277, 1986 3. Gruppo Interregionale Cooperativo Oncologico Ginecologico: Randomised comparison of cisplatin with cyclophosphamide/cisplatin and cyclophosphamide/cisplatin/ doxorubicin in advanced ovarian cancer. Lancet 2:353-359, 1987 4. Bruckner HW, Cohen CJ, Goldberg JD: Cisplatin regimes and improved prognosis in patients with poorly differentiated ovarian cancer. Am J Obstet Gynecol 145:653658, 1983 5. Barker GH, Wiltshaw E: Use of high dose cisplatin following failure on previous chemotherapy for advanced carcinoma of the ovary. Br J Obstet Gynaecol 88:1192-1199, 1981 6. Ozols RF, Ostchega Y, Myers CE, et al: High-dose cisplatin in hypertonic saline in refractory ovarian cancer. J Clin Oncol 3:1246-1250, 1985 7. Calvert AH, Wiltshaw E, Bakers JW, et al: Phase II trial of cis diamine-1,1-cyclo-butrane dicarboxylate platinum II (CBDCA-JM8) in patients with carcinoma of the ovary not previously treated with cisplatin. 4th Int Symp Platinum Coordination Complexes in Cancer Chemother C-18, 1983 (abstr) 8. Bramwell V, et al: Phase I study of cis-dichloro-transdihydroxy-bis-isopropylamine platinum IV (CHIP). 2nd European Conference Clin Oncol, 1983 (abstr) 9. Creaven PJ, Madajewicz S, Pendyala L, et al: Phase I study of a new anti-neoplastic platinum analogue cisdichloro-trans-dihydroxy-bis-isopropylamine platinum IV (CHIP). Proc Am Soc Clin Oncol C-86, 1982 (abstr) 10. Wiltshaw E, Evans B, Jones AC, et al: JM8, successor to cisplatin in advanced ovarian carcinoma? Lancet 1:587, 1983 (letter)
11. Adams M, Kerby IJ, Recher I, et al: A comparison of the toxicity and efficacy of cisplatin and carboplatin in advanced ovarian cancer. Acta Oncologica 28:57-60, 1989 12. Fleiss JL: Statistical methods for rates and proportions (ed 3). New York, NY, Wiley, 1981 13. Kaplan EL, Meier P: Non-parametric estimation from incomplete observations. J Am Stat Assoc 53:457-481, 1958 14. Peto R, Pike MC, Armitage P, et aI: Design and analysis of randomised clinical trials requiring prolonged observation of each patient. Br J Cancer 35:1-39, 1977 15. World Health Organization: Handbook for Reporting Results of Cancer Treatment. WHO offset publication vol 48, World Health Organization, Geneva, 1979 16. Wiltshaw E: Ovarian trials at the Royal Marsden. Cancer Treat Rev 12:67-71, 1985 (suppl A) 17. Griffiths CT: Surgical resection of tumour bulk in the primary treatment of ovarian cancer. NCI Monog 42:101104, 1975 18. Munnell EW: The changing prognosis and treatment in cancer of the ovary-A report of 235 patients with primary ovarian cancer 1952-1961. Am J Bot Gynecol 100:790-805, 1968 19. Mendiola C, Aranburo P, Cortes-Fines H, et al: Prognostic and therapeutic value of 2nd look surgery in advanced epithelial ovarian carcinoma. 4th European Conference on Clinical Oncology and Cancer Nursing. 832, 1987 (abstr) 20. Vogl SE, Pagaro M, Kaplan BH, et al: Cisplatin based combination chemotherapy for advanced ovarian cancer. Cancer 51:2024-2030, 1983 21. Williams CJ, Mead B, Arnold A, et al: Chemotherapy of advanced ovarian carcinoma. Cancer 49:1778-1783, 1982 22. Cohen CJ, Goldberg JD, Holland JF, et al: Improved therapy with cisplatin regimens for patients with ovarian cancer as measured by surgical end staging. Am J Obstet Gynecol 145:955-967, 1983
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 6, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
CISPLATIN
is currently one of the most active chemotherapeutic agents available for the treatment of ovarian cancer,1'2 producing survival rates in excess of 40% at 2 years for advanced disease.' 4 At doses of 100 mg/m 2, toxicity is considerable5'6 with severe emesis, renal impairment, peripheral neuropathy, high-frequency hearing loss, and anemia. Carboplatin (diamine-1,1-cylobutane dicarboxylase platinum II) and iproplatin (cis-dichloro-trans-dihydroxy-bis [isopropylamine] platinum IV) are two cisplatin analogs with fewer marked side effects at doses equivalent to 100
mg/m 2 of cisplatin. Studies in patients with advanced ovarian cancer have shown encouraging
results with both analogs in previously treated patients,7 and carboplatin achieved an equal response rate to cisplatin in randomized studies
involving previously untreated ovarian cancer patients.' 0"'. Myelosuppression was the most significant side effect and was dose-limiting in both the new analogs.
At the time of inception of this study, carboplatin and iproplatin were the most likely successors to cisplatin. While considerable experience has been obtained with carboplatin, much less is known of the relative effectiveness of iproplatin.
This study was planned to compare the response rates and toxicities of both agents, and to investi-
2 mide (1 g/m ). The response rates were 63% (95% confidence interval [CI], 50% to 74%) for carboplatin and 38% (95% CI, 26%to 51%) for iproplatin. Fifteen patients were not assessable for response. The median survival was 114 weeks (95% CI, 82 to 233 weeks) for carboplatin patients and 68 weeks (95% CI, 48 to 82 weeks) for iproplatin patients (P = .008). The amount of residual disease after initial laparotomy was a prognostic factor for survival. Myelosuppression was the main toxicity and was greater with iproplatin. This study shows carboplatin to be more active than iproplatin in the treatment of ovarian cancer and less toxic. Few responses to cyclophosphamide occurred following either drug, implying resistance to the alkylating agent. J Clin Oncol 9:1131-1137. © 1991 by American Society of Clinical Oncology.
gate the efficacy of second-line chemotherapy using cyclophosphamide on relapse. METHODS Between August 1984 and October 1987, 120 patients with International Federation of Gynecology and Obstetrics (FIGO) stage IC to IV epithelial ovarian carcinoma were entered onto the study. None had received prior chemotherapy or radiotherapy, and all had undergone a laparotomy to define accurately the extent of disease and to remove as much of the tumor as possible. At the initial laparotomy, the surgeon was required to designate the amount of residual disease into one of three categories: (1) no residual macroscopic disease, (2) maximum tumor masses less than 1.5 cm, and (3) maximum tumor mass greater than 1.5 cm. To be eligible for the study, the patients' Karnofsky performance status had to be 60% or greater and their life expectancy in excess of 3 months. Hematologic and renal function parameters had to be within the institutions'
From the Southend Hospital,Essex; and University College Hospital,London, England. Submitted May 17, 1990; accepted January25, 1991. Supported in part by a grantfrom Bristol Myers Company, PharmaceuticalResearchand Development Division (protocol no. 118-008-001). Address reprintrequests to C. Trask, MD, Southend Hospital, Prittlewell Chase, Westcliff on Sea, Essex SSO ORY, England. © 1991 by American Society of Clinical Oncology. 0732/183X/91/0907-0022$3.00/0
Journal of Clinical Oncology, Vol 9, No 7 (July), 1991: pp 1131-1137
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 6, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
1131
1132
TRASK ET AL
normal ranges and age between 18 and 80 years. Subjects were excluded from the study if there was a history of previous malignancy, brain metastasis, serious intercurrent infection, or chronic disease requiring regular treatment. Before entering the trial, each patient underwent a full clinical examination, chest x-ray, pelvic and abdominal ultrasound, biochemical tests of liver function, audiometry, and creatinine or pentetic acid (DTPA) clearance. CA 125 estimations were not performed. Patients were randomized within 1 month of original laparotomy by the trials coordinator using computer-generated stratified randomization sheets. Informed consent was obtained from each patient. The protocol was approved by the Ethical Committee of each participating institution. Patients were stratified according to FIGO stage and amount of residual disease and allocated to receive either carboplatin (400 mg/m2) or iproplatin (300 mg/m2) at 4 weekly intervals. Both drugs were diluted in 250 mL 5% dextrose and infused intravenously over 30 minutes without pre- or posthydration. Each center adopted the same antiemetic policy for the two drugs, but the regimen was left to the discretion of the participating centers and was usually a regimen containing high-dose metaclopramide, dexamethasone, and lorazepam. If the tumor failed to respond at any time after four cycles of therapy, or if tumor progression occurred at any time after two cycles, the patient was then 2 treated with cyclophosphamide at a dose of 1 g/m intravenously repeated every 3 to 4 weeks. Full clinical reassessment was made in responding patients after six courses, and complete clinical responders were subjected to a secondlook laparoscopy, or laparotomy when appropriate (and the patient consented). Patients with pathologic complete response (CR) (laparoscopy or laparotomy negative) were followed up at 2 monthly intervals, and no further chemotherapy was given until relapse at which time cyclophosphamide was given. Partial and minor responders continued with a further six cycles of the same analog at a reduced 2 dosage (carboplatin 300 mg/m2, iproplatin 225 mg/m ), and those patients with stable (SD) or progressive disease (PD) 2 were treated with cyclophosphamide also at I g/m . The patients who continued with a further six courses of carboplatin or iproplatin were again fully reassessed in the same way after their twelfth course. Complete responders then stopped therapy; partial responders or less were offered second-line treatment with cyclophosphamide. The objectives of the study were to compare response rates and toxicity. The number of patients required was determined before the trial began and was based on the ability to detect a 25% difference in response rate with a set 9 at 0.05 and p at 0.20. Available data from other studies" indicated that iproplatin appeared to be associated with more toxicity. The study was therefore designed to investigate whether iproplatin produced a higher response rate than carboplatin since, otherwise, carboplatin would be the preferred agent. A one-sided test was therefore chosen for analysis, using an estimated response rate of 50% for 12 carboplatin,l' and required 53 patients on each treatment. A total of 120 patients were recruited to allow for nonassessable responses. Survival curves were constructed according to the method of Kaplan and Meier" and statistical signifi4 cance evaluated by the log-rank test.'
Dose Adjustments and Scoring of Toxicity Three patients began treatment at 90% of the starting dose because they were aged older than 75 years. All patients had weekly blood counts during treatment to assess hematologic toxicity. Dose adjustments were made according to the nadir full blood count following the previous cycle (Table 1). Drug administration was postponed by 1 week if there had not been hematologic recovery (WBC > 3.0 x 107/ L or platelets > 100 x 10'/L). Renal toxicity was assessed by serum urea, creatinine, and electrolyte estimations before each treatment and by creatinine/DTPA clearances after every third course. If the serum creatinine rose above 120 l.mol/L and/or significant deteriorating renal function was confirmed by DTPA clearance, therapy was stopped. Ototoxicity was assessed by comparing serial audiograms (every third course) with pretreatment ones. Diarrhea, mucositis, infection, peripheral neuropathy, nausea, and vomiting episodes were recorded at each course using World Health Organization (WHO) criteria and duration of nausea and vomiting measured in hours.
Response Criteria For evaluation of response the WHO criteria were adopted." The response rate recorded was the best response achieved by a patient at any time during their treatment with carboplatin or iproplatin. Independent second-party review of responses was not made.
RESULTS
The characteristics of the 120 patients entered onto the study are summarized in Table 2 and were well balanced in both arms of the study. Median age was 59 years (range, 29 to 79), and about 60% achieved a Karnofsky status greater than 80% at the start of chemotherapy. Over 85% of patients had FIGO stage III or IV disease. Total hysterectomy and bilateral salpingooopharectomy was performed in 63% of patients receiving carboplatin and in 74% receiving iproplatin. In 48% of carboplatin patients and in 44% of
iproplatin patients poorly differentiated tumors were ascertained. Fifteen patients were nonassessable for response for the following reasons: five patients with no macroscopic residual disease the Table 1. Dose-Modification Schedule With Respect to Nadir Blood Counts Following the Previous Course WBCs
Platelets
> 100 75-99 50-74 25-49 < 25
and and/or and/or and/or and/or
> 4.0 3-3.9 2-2.9 1-1.9 < 1.0
% Previous Dose
110* 100 90 75 50
*Maximum dose, carboplatin 400 mg/m2, iproplatin 300 mg/ m2
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 6, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
1133
CARBOPLATIN v IPROPLATIN IN OVARIAN CANCER Table 2. Patient Characteristics at Presentation According to Assigned Treatment Carboplatin (n = 62)
No. Age (years) Range Median Karnofsky status (%) 100 90 80 70 60 Initial surgery TAH BSO O' TAH BSO Biopsy only Other Residual disease Nil FIGO stage IC 2 II 4 III 3 IV 0
Iproplatin (n = 58)
%
No.
29-79 59 30 13 9 6 4 16 23 10 13 < 1.5 0 2 12 0
%
34-75 59 48 21 15 10 6
17 17 12 5 7
26 37 16 21 > 1.5 Nil 0 1 27 11
1 3 3 0
29 29 21 9 12
15 26 28 48 10 17 5 9 < 1.5 > 1.5 0 1 11 0
0 2 26 11
Abbreviations: TAH, total abdominal hysterectomy; BSO, bilateral salpingo-oophorectomy; 0', omentectomy; Nil, no residual disease.
status of the peritoneal washings at their initial surgery was not recorded, three patients in clinical CR (CCR) could not be reexamined laparoscopically, three patients died of unrelated causes (two had cerebrovascular accidents, one after her first cycle, the other after her second, one patient died of hypercalcemia after her first cycle), two patients withdrew (one after five courses and one after one course), one patient developed a fatal second malignancy (carcinoma of cecum) after her sixth course, and one patient was taken off study because of impaired renal function after one course of iproplatin. Of the remaining 105 patients, three carboplatin and four iproplatin patients subsequently withdrew from treatment after their sixth
responders with carboplatin, 12 were confirmed by laparoscopy and eight by laparotomy. Of the 11 pathologic complete responders with iproplatin, five were confirmed by laparoscopy and six by laparotomy. PD on therapy developed in 43% of iproplatin and 19% of carboplatin patients, giving an increase in PD of 24% with iproplatin (95% CI of difference, 8% to 40%). After six courses of chemotherapy, 34 patients (20 carboplatin, 14 iproplatin) had achieved a PR and continued with treatment up to a maximum of 12 courses. Only one carboplatin patient and two iproplatin patients achieved a further improvement in response by continuing chemotherapy. Twenty-four patients who had received carboplatin and 23 who had received iproplatin went on to receive chemotherapy with cyclophosphamide. There were two CCRs to cyclophosphamide, one PR and five SD responses after carboplatin; 14 patients progressed and two were nonassessable. In the iproplatin arm, there were no CRs to cyclophosphamide, one patient achieved a PR, five had SD, 15 had PD, and two were nonassessable. The total response rate with cyclophosphamide was only 9% in the whole group, with 21% achieving SD. Twenty-one patients who progressed on firstline therapy received no relapse chemotherapy. Nine received palliative radiotherapy alone and 10 received other chemotherapy (eg, chlorambucilor melphalan-containing regimens) without response. Survival All 120 patients were analyzed for survival, although this study was not specifically designed to compare survival. In the survival analyses, three patients (two carboplatin, one iproplatin) were censored at the point of diagnosis of a second Table 3. Response Rates According to Assigned Treatment
course.
Response rates are shown in Table 3. A clinical partial response (CPR) or greater was achieved in 63% of carboplatin and 38% of iproplatin patients, giving a 25% improvement in response with carboplatin (95% confidence interval [CI] of difference, 8% to 42%). A pathologic CR was achieved in 32% of carboplatin patients and 19% of iproplatin patients, giving an improvement in response with carboplatin of 13% (95% CI of difference, 2% to 29%). Of the 20 pathologic complete
Carboplatin (n= 62)
Pathologic CR Pathologic PR CCR
20 12
52%
2
11%
CPR
5
Pathologic SD
4
Clinical SD
1
Pathlogic PD Clinical PD Nonassessable
2 10 6
Iproplatin (n = 58)
63%
11 8 2
8%
2 0
19%
3 222 9
10%
33% 38% 5% 3 3%
43 16%
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 6, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
1134
TRASK ET AL
malignancy, two patients (both carboplatin) were censored at the time of their death, since the causes of their deaths were not cancer-related, one patient (iproplatin) was censored due to loss of follow-up, and one patient (iproplatin) was censored at the time when taken off study due to impaired renal function. The median follow-up of the study was 35 months. The median progressionfree interval (time from starting chemotherapy until progression) was 73 weeks (95% CI, 51 to 109 weeks) for the carboplatin patients, and 38 weeks (95% CI, 24 to 57 weeks) for iproplatin patients (P = .019; Fig 1). The overall survival (Fig 2) was significantly better for carboplatin patients, 52% (95% CI, 39% to 65%) alive at 2 years compared with only 33% (95% CI, 20% to 45%) of iproplatin patients. The median survival was 114 weeks (95% CI, 82 to 233 weeks) for carboplatin patients versus 68 weeks (95% CI, 48 to 82 weeks) for iproplatin patients (P = .008). The size of residual disease remaining after initial surgery was a significant prognostic factor. The two arms of the study were well balanced with two thirds of patients having residual disease greater than 1.5 cm. At 3 years, 59% of patients with < 1.5 cm of tumor remaining and 62% of patients with no residual disease are alive compared with 26% of patients with greater than 1.5 cm disease (Fig 3). The median survival for the latter group is 68 weeks (95% CI, 59 to 82 weeks); for the combined better prognostic groups the median survival is 233 weeks (P < .001). The FIGO stage was also an important determinant of survival with FIGO IC and II patients
c u
M L
1 T E
S U
V
G
TIME ( MONTHS )
Fig 2. Overall survival for all patients randomized to carboplatin (n = 62; median survival, 114 weeks) or iproplatin (n = 58, median survival, 68 weeks). Carboplatin: observed events 32, expected events 43.21; iproplatin: observed events 41, expected events 29.79; ) = 7.12, P = .008.
surviving significantly longer than stage III (P = .012) or IV patients (P = .003) (Fig 4). Stage III patients had a median survival of 82 weeks (95% CI, 65 to 114 weeks), stage IV patients, 61 weeks (95% CI, 38 to 93 weeks); at 3 years, 75% (95% CI, 48% to 100%) of stage IC and II patients are alive. Histologic differentiation of the primary tumor was not related to prognosis. Toxicity Toxicity was analyzed only in the first six courses of carboplatin and iproplatin. Myelosuppression, nausea, and vomiting were the most significant side effects. The most serious hematologic toxicity was thrombocytopenia, reaching WHO grade 4
c U
P R 0 G R S S
N F E E
TIME ( MONTHS ) TIME ( MONTHS )
Fig 1. Progression-free interval for all patients randomized to carboplatin (n = 62; median interval, 73 weeks) or iproplatin (n = 58; median interval, 38 weeks). Carboplatin: observed events 37, expected events 47.10; iproplatin: observed events 41, expected events 30.90; = 5.47, P = .019.
?x
Fig 3. Overall survival for all patients randomized according to the amount of residual disease left after diagnostic laparotomy. Nil, no residual macroscopic disease (n = 16; median survival, 183 weeks); < 1.5 cm, residual disease < 1.5 cm (n = 26); > 1.5 cm, residual disease > 1.5 cm (n = 78; median survival, 68 weeks).
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 6, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
1135
CARBOPLATIN v IPROPLATIN IN OVARIAN CANCER J
.
14 Z
13 12 11 10
m Fn 0
0-1O
z
8
O
7 6 5 4
z
w
C) Z
0
w LU
3 2 1 0
2 4 TIME ( MONTHS )
Fig 4. Overall survival for all patients randomized according to initial FIGO stage. FIGO IC and II n = 16, median survival not ascertained; FIGO Ill n = 82, median survival 82 weeks; FIGO IV n = 22, median survival 61 weeks.
11NNN nn
I-. z
toxicity on four occasions in the carboplatin arm and 22 times in the iproplatin arm. Platelet transfusions were required on 14 occasions for 13 iproplatin patients, whereas none were needed in the carboplatin arm. Half the platelet transfusion episodes occurred before the second course of chemotherapy with iproplatin. Grade 4 leukopenia only occurred once in an iproplatin patient. There were seven episodes of infection (WHO grade 1 or 2 only) on carboplatin and three on iproplatin. Nine patients experienced mucositis on carboplatin and three on iproplatin. Blood transfusions were given nine times in eight carboplatin patients and 14 times in 11 iproplatin patients; the hemoglobin fell below 6.5 g/dL only once with both drugs. The cumulative hematologic toxicities for the first six courses for hemoglobin, WBC, and platelets are shown in Fig 5. Dose reductions due to myelosuppression were greater with iproplatin, although carboplatin caused more leukopenia than iproplatin (Table 4). By course 6, the mean dose of iproplatin was reduced to 69% of 300 mg/m2, and the mean
D o. I
0 z
10 COURSE NUMBER Fig 5. Cumulative hematologic toxicities showing nadir hemoglobin (g/dL), WBC count (x 10'/L), and platelets (x 109/ L) experienced by all patients with the first six courses, with SEM. (0) Iproplatin, (I) carboplatin.
for a mean of 13 hours with both carboplatin and iproplatin. Chemotherapy-induced diarrhea was a rare event with carboplatin (only three episodes in two Table 4. Reasons for Dose Modifications and Course Delays According to Assigned Treatment
carboplatin dose was 82% of 400 mg/m2 with a
mean difference of 13% (95% CI, 5% to 20%). Most patients experienced WHO grade 2 or 3 nausea and vomiting, although no patient experienced grade 4 intractable vomiting. The mean duration of nausea and vomiting is shown in Fig 6. The duration of nausea was longer with carboplatin: an overall mean of 58 hours compared with 48 hours with iproplatin with a mean difference of 10 hours (95% CI, 1 to 19 hours). Vomiting lasted
Dose adjustment reasons Leukopenia Thrombocytopenia Both Course delay reasons Leukopenia Thrombocytopenia Both
Carboplatin
Iproplatin
(n = 279)
(n = 239)
59 11 16
30 51 13
12 1 0
10 8 3
NOTE. Data are total number of cycles.
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 6, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
1136
TRASK ET AL 80
NAU 70
0, 60-
Z 0
504030-
Z 2 I
VOMITING 20o-
*
10-
1
2
3
4
5
6
COURSE NUMBER
Fig 6. Mean duration (with SEM) in hours of nausea and vomiting experienced by all patients with the first six courses. (0) Iproplatin, (0) carboplatin.
patients) compared with 56 episodes with iproplatin in 27 patients. Only one (iproplatin) patient experienced significant deterioration of renal function, and another patient experienced a deterioration of acoustic acuity as judged by serial audiograms, possibly attributable to iproplatin. This patient developed tinnitus after six courses of cyclophosphamide following three courses of iproplatin. Three patients showed WHO grade I peripheral neuropathic changes (all carboplatin), and in each case, the symptoms developed at the fourth course. Three patients experienced anaphylactic reactions (two carboplatin, one iproplatin), all during the eighth course. One iproplatin patient developed a skin reaction. The relationship between pretreatment creatinine clearance and overall toxicity was analyzed by relating clearance to maximum dose reduction by course 6, since dose reduction was on the basis of hematologic, renal, or gastrointestinal toxicity. No relationship was shown for either iproplatin or carboplatin (Pearson's correlation coefficient .08 and .203, respectively). DISCUSSION
Previous data have suggested that iproplatin may have a greater response rate than carboplatin8"9 in the treatment of ovarian cancer. The toxicity of iproplatin was expected to be greater than carboplatin. This study was therefore designed to be able to detect a 25% difference in response rate in favor of iproplatin, necessitating an accrual of at least 100 eligible patients. Our results confirm that the hematologic toxicity of
iproplatin was indeed greater than carboplatin. However, the response rate for iproplatin was only 38% compared with a response rate of 63% for carboplatin. Survival was worse in the iproplatin group (median survival, 68 weeks). Thus, at doses that are more toxic than carboplatin, iproplatin is less effective and, therefore, has no role in the treatment of ovarian carcinoma. Carboplatin proved to be a safe drug to use at a dose of 400 mg/m2, with acceptable toxicity. Acute gastrointestinal side effects were transient; vomiting lasted for a mean of 13 hours after the infusion, but the onset was usually delayed. Nausea persisted for a mean of 58 hours; however, the symptoms were never dose-limiting and were controllable on an outpatient basis. Diarrhea was unusual. Other studies have shown less nausea and vomiting toxicity for carboplatin than reported here." This probably reflects differences in antiemetic regimes or the length of inpatient stay. However, if carboplatin is to be used on an outpatient basis, an effective antiemetic regime must be instituted and the patient must be warned to expect a degree of nausea and vomiting. Myelosuppression was mild provided the dose adjustments were made with reference to weekly blood counts: by course 6, the dose of carboplatin had been decreased to 82% of the starting dose. Thrombocytopenia was sufficient to cause dose adjustments in 10% of courses, but no platelet transfusions were necessary. Leukopenia was never sufficiently profound to cause infections requiring intravenous antibiotics. As in previous studies,'0,1116 there was no significant degree of renal
toxicity, ototoxicity, or neurotoxicity, and this study again confirms that there is no need for any pre- or posthydration, implying that carboplatin is eminently suitable for use as an outpatient regimen. This study also confirmed that the size of residual disease after surgery is an excellent predictor of survival. This has been suggested by a number of retrospective analyses,'-. 1 but this study confirms prospectively that the size of residual disease after surgery is as good a predictor of survival as the FIGO stage itself. Patients with no residual disease or disease less than 1.5 cm in diameter show a 60% (95% CI, 43% to 78%) 3-year survival, whereas those with residual disease greater than 1.5 cm in diameter have only a 26% 3-year survival (95% CI, 15% to 37%) (P < .001).
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 6, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
CARBOPLATIN v IPROPLATIN IN OVARIAN CANCER
1137
Resistance to carboplatin and iproplatin appears to be accompanied by a considerable degree of resistance to cyclophosphamide. This may indicate that there will be little advantage in combining the two drugs. There would also seem to be little advantage in continuing chemotherapy beyond six cycles as only three of 34 patients in both arms improved their response rate by continuing therapy after six courses. The overall response rate to carboplatin was high, with an encouragingly long median survival of 26 months. These results compare favorably with other carboplatin single-agent studies... The median survival appears comparable with that obtained with cisplatin at a dose of 100 mg/m 2,
which is reported to be between 19 and 23 months in various studies.2,",' 6 Carboplatin can be administered as an outpatient drug with little acute toxicity. Since these results are comparable with those achieved in many studies of combination chemotherapy,3'4' 20°- a randomized study comparing carboplatin alone with combination chemotherapy would be valuable with an assessment of toxicity, survival, and cost. ACKNOWLEDGMENT We gratefully acknowledge the support of the following consultant gynecologists: Jane Crow, John Kersey, Michael Martin, Heulwen Morgan, Raymond Pilkington, Tim Pocock, Albert Singer, John Ward, and Chris Welch.
REFERENCES 1. Ozols RF, Young RC: Chemotherapy of ovarian cancer. Semin Oncol 11:251-263, 1984 2. Sessa C: European studies with cisplatin and cisplatin analogs in advanced ovarian cancer. Eur J Cancer Clin Oncol 22:1271-1277, 1986 3. Gruppo Interregionale Cooperativo Oncologico Ginecologico: Randomised comparison of cisplatin with cyclophosphamide/cisplatin and cyclophosphamide/cisplatin/ doxorubicin in advanced ovarian cancer. Lancet 2:353-359, 1987 4. Bruckner HW, Cohen CJ, Goldberg JD: Cisplatin regimes and improved prognosis in patients with poorly differentiated ovarian cancer. Am J Obstet Gynecol 145:653658, 1983 5. Barker GH, Wiltshaw E: Use of high dose cisplatin following failure on previous chemotherapy for advanced carcinoma of the ovary. Br J Obstet Gynaecol 88:1192-1199, 1981 6. Ozols RF, Ostchega Y, Myers CE, et al: High-dose cisplatin in hypertonic saline in refractory ovarian cancer. J Clin Oncol 3:1246-1250, 1985 7. Calvert AH, Wiltshaw E, Bakers JW, et al: Phase II trial of cis diamine-1,1-cyclo-butrane dicarboxylate platinum II (CBDCA-JM8) in patients with carcinoma of the ovary not previously treated with cisplatin. 4th Int Symp Platinum Coordination Complexes in Cancer Chemother C-18, 1983 (abstr) 8. Bramwell V, et al: Phase I study of cis-dichloro-transdihydroxy-bis-isopropylamine platinum IV (CHIP). 2nd European Conference Clin Oncol, 1983 (abstr) 9. Creaven PJ, Madajewicz S, Pendyala L, et al: Phase I study of a new anti-neoplastic platinum analogue cisdichloro-trans-dihydroxy-bis-isopropylamine platinum IV (CHIP). Proc Am Soc Clin Oncol C-86, 1982 (abstr) 10. Wiltshaw E, Evans B, Jones AC, et al: JM8, successor to cisplatin in advanced ovarian carcinoma? Lancet 1:587, 1983 (letter)
11. Adams M, Kerby IJ, Recher I, et al: A comparison of the toxicity and efficacy of cisplatin and carboplatin in advanced ovarian cancer. Acta Oncologica 28:57-60, 1989 12. Fleiss JL: Statistical methods for rates and proportions (ed 3). New York, NY, Wiley, 1981 13. Kaplan EL, Meier P: Non-parametric estimation from incomplete observations. J Am Stat Assoc 53:457-481, 1958 14. Peto R, Pike MC, Armitage P, et aI: Design and analysis of randomised clinical trials requiring prolonged observation of each patient. Br J Cancer 35:1-39, 1977 15. World Health Organization: Handbook for Reporting Results of Cancer Treatment. WHO offset publication vol 48, World Health Organization, Geneva, 1979 16. Wiltshaw E: Ovarian trials at the Royal Marsden. Cancer Treat Rev 12:67-71, 1985 (suppl A) 17. Griffiths CT: Surgical resection of tumour bulk in the primary treatment of ovarian cancer. NCI Monog 42:101104, 1975 18. Munnell EW: The changing prognosis and treatment in cancer of the ovary-A report of 235 patients with primary ovarian cancer 1952-1961. Am J Bot Gynecol 100:790-805, 1968 19. Mendiola C, Aranburo P, Cortes-Fines H, et al: Prognostic and therapeutic value of 2nd look surgery in advanced epithelial ovarian carcinoma. 4th European Conference on Clinical Oncology and Cancer Nursing. 832, 1987 (abstr) 20. Vogl SE, Pagaro M, Kaplan BH, et al: Cisplatin based combination chemotherapy for advanced ovarian cancer. Cancer 51:2024-2030, 1983 21. Williams CJ, Mead B, Arnold A, et al: Chemotherapy of advanced ovarian carcinoma. Cancer 49:1778-1783, 1982 22. Cohen CJ, Goldberg JD, Holland JF, et al: Improved therapy with cisplatin regimens for patients with ovarian cancer as measured by surgical end staging. Am J Obstet Gynecol 145:955-967, 1983
Downloaded from ascopubs.org by University of Newcastle Upon Tyne on April 6, 2019 from 128.240.208.034 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
