Gemcitabine Plus Paclitaxel Bests Paclitaxel Alone for Anthracycline- Pretreated Metastatic Breast Cancer

August 1, 2003

This special supplement to Oncology News International includes 28 reportswith updated information on clinical trials investigating capecitabine and other agents inthe treatment of advanced colorectal and breast cancers, and other solid tumors.The reports summarize selected presentations from the 39th Annual Meeting of theAmerican Society of Clinical Oncology (ASCO) and related educational symposiaheld in conjunction with ASCO.

DALLAS-"Interim results of aphase III trial of gemcitabine (Gemzar)plus paclitaxel compared with paclitaxelalone in anthracycline-pretreated metastaticbreast cancer patients show statisticallysignificant improvements in time toprogression and objective response ratedue to the addition of gemcitabine topaclitaxel," Joyce O'Shaughnessy, MD, reported(ASCO abstract 25)."There was a trend to improved painscores, reduced analgesic use, and increasesin overall quality of life in the combinationof gemcitabine plus paclitaxel, whichwas well tolerated with manageable toxicity.Because of the favorable risk-benefitprofile reported in this controlled clinicaltrial, gemcitabine plus paclitaxel is a newtreatment option for metastatic breast cancerpatients who may benefit from combinationchemotherapy," she continued. Dr.O'Shaughnessy is co-director of breastcancer research at Baylor-Sammons CancerCenter and US Oncology, Dallas."We enrolled only patients who hadnot been treated with chemotherapy formetastatic disease, but had received anthracycline-based chemotherapy as adjuvantor neoadjuvant therapy, or a nonanthracycline-based adjuvant regimen ifanthracyclines were contraindicated," shesaid. "Prior hormonal therapy was permitted.Patients had at least one site ofbidimensionally measurable disease."Other eligibility criteria included unresectable,locally recurrent, or metastaticbreast cancer and Karnofsky performancestatus of at least 70.98 Centers ParticipatePatients at 98 participating centers in19 countries were randomized to receivegemcitabine/paclitaxel or paclitaxel alone.The doses of gemcitabine/paclitaxel werechosen based on phase I and II data thathad shown acceptable safety and promisingantitumor activity. Standard gemcitabinepremedications were utilized onboth study arms.Patients were randomized to receiveeither:

  • paclitaxel 175 mg/m2 given over3 hours followed by gemcitabine 1,250mg/m2 on day 1, and then gemcitabineagain, 1,250 mg/m2 on day 8;
  • paclitaxel alone at 175 mg/m2 over3 hours.

Both regimens were given every 21days.There were 257 patients in each treatmentgroup, well-matched for age andethnicity and 97% of patients had metastaticbreast cancer at study entry.This was a patient population with arelatively significant tumor burden with73% of patients on both arms of the studyhaving visceral metastases. On the combinedarm, 43% of patients had three ormore sites of metastatic disease, comparedwith 41% on the paclitaxel-alone arm.Ninety-seven percent of the gemcitabine/paclitaxel patients and 96% of the paclitaxelpatients had been previously treatedwith an anthracycline-based adjuvant chemotherapy.Hormone receptor status waswell balanced on the two arms of thestudy. Approximately 50% of patients onboth study arms were previously treatedwith a hormonal agent.All sites of disease were assessed every8 weeks and treatment was continued untildocumented disease progression.

Longer Time to Progression

The interim analysis was conductedwhen at least 400 patients had developedprogressive disease and provided 75%power to detect a hazard ratio of 0.75 fortime to progression with a two-sided significancelevel of 0.028. The primary objectivefor the final analysis, which will beconducted in late 2004, is overall survival.The study sample size provides 80% powerto detect a hazard ratio for overall survivalof 0.75, with a two-sided significancelevel of 0.03.The median delivered dose of gemcitabinewas 1,134 mg/m


, which was 85% ofthe intended dose. Patients on both armsof the study received a median delivereddose of paclitaxel of 175 mg/m


. Only asmall minority of intended gemcitabinedoses were omitted or reduced.With 424 events having occurred atthe time of the interim analysis, the mediantime-to-progression with combinedgemcitabine/paclitaxel was 5.4 monthscompared with 3.5 months with paclitaxelalone. At 6 months, 44% of the gemcitabine/paclitaxel patients were progression-free, compared with 30% of paclitaxelpatients. The hazard ratio for progressionwith gemcitabine/paclitaxel vs paclitaxelalone was 0.73, which was statistically significantwith a two-sided P value of .0013.The objective rate with gemcitabine/paclitaxelwas 39.3% compared with 25.6%for paclitaxel and the difference betweenthese two was statistically significant.The median duration of response was8.8 months with gemcitabine/paclitaxelcompared with 7.2 months with paclitaxel,a nonsignificant difference.

Pain Improvement

An analysis by cycle of mean brief paininventory scores for symptomatic patientsrevealed a nonsignificant trend towardpain improvement over time with gemcitabine/paclitaxel compared with paclitaxelalone. This improvement in painscores was associated with a decreasedanalgesic requirement in 25% of patients,compared with 15% of patients on thepaclitaxel alone arm. In cycles five and six,patients on the gemcitabine/paclitaxel armhad a statistically significant improvementin their quality of life, compared withtheir own baseline scores showing no adverseeffect of the combination on thismeasure of quality of life.

Mild Toxicities

Grade 3/4 hematologic toxicities andtransfusions were relatively mild on bothstudy arms. Of patients treated with gemcitabine/paclitaxel, 17% developed grade 4 neutropenia compared with 7% withpaclitaxel alone. Of the gemcitabine/paclitaxeltreated patients, 5% developedgrade 3 thrombocytopenia. Febrile neutropeniaor sepsis was observed in 5% ofgemcitabine/paclitaxel treated patientscompared with 2% of paclitaxel patients.Red blood cell transfusions were administeredto 10% of the gemcitabine/paclitaxel-treated patients compared with 4%of paclitaxel alone) patients, althoughgrade 3/4 anemia was uncommon. Forgrade 3/4 nonhematologic toxicity, therewere modest differences in the incidenceof fatigue as well as asymptomatic increasesin hepatic transaminases in thegemcitabine/paclitaxel arm of the study.At this interim analysis, the percent ofpatients who died on study or during the30-day follow-up period was similar onthe two arms, as was the incidence ofdrug-related deaths, which was one patienton each arm. Of patients treatedwith gemcitabine/paclitaxel, 6% discontinuedtherapy due to drug-related adverseevents compared with 2% with paclitaxelalone.