LOS ANGELES--Interim analysis of a major German-Austrian trial
comparing cisplatin (Platinol)/paclitaxel (Taxol) with carboplatin
(Paraplatin)/paclitaxel as first-line treatment in ovarian cancer
found significantly less toxicity with carboplatin/paclitaxel, with
no apparent loss of efficacy.
Andreas du Bois, MD, reporting on behalf of the Austrian-German
Oncology (AGO) Study Group, said that the carboplatin/paclitaxel
combination has been recommended as the new "standard
therapy" combination for all future phase III studies by this group.
A total of 798 patients were randomized to first-line treatment with
either cisplatin/paclitaxel or carboplatin/paclitaxel. Paclitaxel
(185 mg/m²) was given as a 3-hour infusion. Cisplatin was given
at 75 mg/m², and carboplatin was given to an area under the
curve (AUC) of 6. Cycles were repeated every 3 weeks for a maximum of
Inclusion criteria were epithelial ovarian cancer of FIGO stage
IIC-IV and no prior radio- or chemotherapy. "No secondary
debulking was allowed, so if patients were scheduled to receive a
second cytoreductive surgery after 3 cycles, they could not be
entered into the study," Dr. du Bois said.
Maintenance therapy (ie, unauthorized second-line treatment) after
completion of primary treatment was not allowed.
Dr. du Bois reported data from 702 evaluable patients. "The
median observation period for these patients is 70 weeks," he
said. At study entry, 209 patients had measurable disease, and
three-fourths of them are evaluable for response. Toxicity data are
based on more than 3,500 courses in 690 patients.
Treatment delays of more than 1 week were uncommon in either group
(6% with carboplatin, 5% with cisplatin). Fewer than 5% of courses
were delivered with reduced doses. Most patients received at least 5
or 6 courses of chemotherapy, although 10% to 15% of patients stopped
before that, mostly due to progressive disease.
Eight patients switched over from the cisplatin arm to the
carboplatin arm due to toxicity. No patients switched from
carboplatin to cisplatin.
Toxicity problems were more common and more significant overall on
the cisplatin/paclitaxel arm (see Table).
"Hematologic toxicity (thrombocytopenia and neutropenia) was
more common in the carboplatin arm, but most of the myelosuppression
did not result in clinically relevant toxicity," Dr. du Bois said.
Nonhematologic toxicity was significantly more frequent in the
cisplatin-treated patients. "When we consider all grade 3-4
toxicities, 56% of patients in the cisplatin group experienced at
least one course with grade 3-4 toxicity, compared to 37% of
carboplatin-treated patients," he said.
The most striking differences were seen in GI toxicity, especially
emesis and nausea. "Grade 3-4 toxicity, which means severe
nausea and vomiting, was threefold increased in the cisplatin group,
compared to carboplatin," he said.
Neuropathy occurred much more frequently in the cisplatin arm.
Overall, significantly fewer of the carboplatin-treated patients had
neurotoxicity, and it occurred a little later, he said.
"Our latest analysis showed that one-third of cisplatin-treated
patients still show neurotoxicity after 1 year of follow-up vs about
10% of carboplatin-treated patients," Dr. du Bois said.
"This is important because a substantial proportion of patients
came back for second-line treatment and those still suffering from
neurotoxicity were not able to receive potentially neurotoxicity-causing
Differences in nonhematologic toxicity translated into differences in
quality of life. "The cisplatin-treated patients had a decreased
quality of life, whereas carboplatin treatment led to maintenance of
quality of life after course 3. This is highly significant," he said.
The investigators wondered why quality of life would improve while
patients were still on chemotherapy. "This may be because the
baseline estimation of quality of life was done about 2 weeks after
radical surgery, 1 or 2 days after communication of the diagnosis to
the patient, and surely reflects depression," Dr. du Bois said.
"Quality of life apparently increases when the patient has time
to accept the diagnosis and recover from surgery; then the
differences become obviously related to therapy."
The efficacy data, although not fully mature, show no significant
differences in response rate, complete remission rate, or
Response rates were slightly but not significantly higher with
cis-platin (80% vs 68% for carboplatin), "but this did not
translate into improved progression-free survival," he said.
Median progression-free survival was 75 weeks for carboplatin and 73
weeks for cisplatin. Overall 1-year survival is about 91% in both groups.
"Our conclusions are that carboplatin/paclitaxel compared to
cisplatin/paclitaxel is less toxic and allows better quality of life
during chemotherapy," Dr. du Bois said.
He noted that although the data are not yet mature enough to state
that the two regimens are equivalent regarding overall survival, the
researchers believe this will prove to be the case.
"Therefore," he said, "the AGO recommends that
carboplatin/paclitaxel be used as standard therapy in its future
phase III trials. It is also an acceptable alternative to
cisplatin/paclitaxel for patients not eligible for ongoing trials."
The discussant for the study, Tate Thigpen, MD, of the University of
Mississippi, said that longer follow-up is necessary before
conclusions regarding efficacy can be made. "We must ensure that
substituting carboplatin for cisplatin is a safe thing to do for our
patients in terms of efficacy," he said.
A new study similar to the AGO trial, GOG-158, has completed accrual
of 817 stage III ovarian cancer patients. This study is powered to
have a 90% probability of detecting a 30% difference in survival
between cisplatin/paclitaxel and carboplatin/paclitaxel.
To date, he said, there have been 253 events in GOG-158, an
insufficient number to allow reporting of survival data. "Our
goal is 382 events," he said. In contrast, AGO with 702
evaluable patients reported on survival with only 222 events.
Based on the AGO data, Dr. Thigpen said, "it is fair to say that
paclitaxel/carboplatin has better patient tolerance than
paclitaxel/cisplatin, but conclusions regarding efficacy are
premature at present and require further follow-up."
Dr. du Bois did not disagree with that statement, but said that
"our response data and progression-free survival data,
especially in the suboptimal disease group, are very close to mature.
We needed data to help in the design of new studies, so for that
purpose, I think it is justified to look at these data at this time."