DALLAS-"Interim results of a phase III trial of gemcitabine(Drug information on gemcitabine) (Gemzar) plus paclitaxel(Drug information on paclitaxel) compared with paclitaxel alone in anthracycline-pretreated metastatic breast cancer patients show statistically significant improvements in time to progression and objective response rate due to the addition of gemcitabine to paclitaxel," Joyce O'Shaughnessy, MD, reported (ASCO abstract 25). "There was a trend to improved pain scores, reduced analgesic use, and increases in overall quality of life in the combination of gemcitabine plus paclitaxel, which was well tolerated with manageable toxicity. Because of the favorable risk-benefit profile reported in this controlled clinical trial, gemcitabine plus paclitaxel is a new treatment option for metastatic breast cancer patients who may benefit from combination chemotherapy," she continued. Dr. O'Shaughnessy is co-director of breast cancer research at Baylor-Sammons Cancer Center and US Oncology, Dallas. "We enrolled only patients who had not been treated with chemotherapy for metastatic disease, but had received anthracycline- based chemotherapy as adjuvant or neoadjuvant therapy, or a nonanthracycline- based adjuvant regimen if anthracyclines were contraindicated," she said. "Prior hormonal therapy was permitted. Patients had at least one site of bidimensionally measurable disease." Other eligibility criteria included unresectable, locally recurrent, or metastatic breast cancer and Karnofsky performance status of at least 70. 98 Centers Participate Patients at 98 participating centers in 19 countries were randomized to receive gemcitabine/paclitaxel or paclitaxel alone. The doses of gemcitabine/paclitaxel were chosen based on phase I and II data that had shown acceptable safety and promising antitumor activity. Standard gemcitabine premedications were utilized on both study arms. Patients were randomized to receive either:

• paclitaxel 175 mg/m² given over 3 hours followed by gemcitabine 1,250 mg/m² on day 1, and then gemcitabine again, 1,250 mg/m² on day 8;
• paclitaxel alone at 175 mg/m² over 3 hours.

Both regimens were given every 21 days. There were 257 patients in each treatment group, well-matched for age and ethnicity and 97% of patients had metastatic breast cancer at study entry. This was a patient population with a relatively significant tumor burden with 73% of patients on both arms of the study having visceral metastases. On the combined arm, 43% of patients had three or more sites of metastatic disease, compared with 41% on the paclitaxel-alone arm. Ninety-seven percent of the gemcitabine/ paclitaxel patients and 96% of the paclitaxel patients had been previously treated with an anthracycline-based adjuvant chemotherapy. Hormone receptor status was well balanced on the two arms of the study. Approximately 50% of patients on both study arms were previously treated with a hormonal agent. All sites of disease were assessed every 8 weeks and treatment was continued until documented disease progression. Longer Time to Progression The interim analysis was conducted when at least 400 patients had developed progressive disease and provided 75% power to detect a hazard ratio of 0.75 for time to progression with a two-sided significance level of 0.028. The primary objective for the final analysis, which will be conducted in late 2004, is overall survival. The study sample size provides 80% power to detect a hazard ratio for overall survival of 0.75, with a two-sided significance level of 0.03. The median delivered dose of gemcitabine was 1,134 mg/m², which was 85% of the intended dose. Patients on both arms of the study received a median delivered dose of paclitaxel of 175 mg/m². Only a small minority of intended gemcitabine doses were omitted or reduced. With 424 events having occurred at the time of the interim analysis, the median time-to-progression with combined gemcitabine/paclitaxel was 5.4 months compared with 3.5 months with paclitaxel alone. At 6 months, 44% of the gemcitabine/ paclitaxel patients were progression- free, compared with 30% of paclitaxel patients. The hazard ratio for progression with gemcitabine/paclitaxel vs paclitaxel alone was 0.73, which was statistically significant with a two-sided P value of .0013. The objective rate with gemcitabine/paclitaxel was 39.3% compared with 25.6% for paclitaxel and the difference between these two was statistically significant. The median duration of response was 8.8 months with gemcitabine/paclitaxel compared with 7.2 months with paclitaxel, a nonsignificant difference. Pain Improvement An analysis by cycle of mean brief pain inventory scores for symptomatic patients revealed a nonsignificant trend toward pain improvement over time with gemcitabine/ paclitaxel compared with paclitaxel alone. This improvement in pain scores was associated with a decreased analgesic requirement in 25% of patients, compared with 15% of patients on the paclitaxel alone arm. In cycles five and six, patients on the gemcitabine/paclitaxel arm had a statistically significant improvement in their quality of life, compared with their own baseline scores showing no adverse effect of the combination on this measure of quality of life. Mild Toxicities Grade 3/4 hematologic toxicities and transfusions were relatively mild on both study arms. Of patients treated with gemcitabine/ paclitaxel, 17% developed grade 4 neutropenia compared with 7% with paclitaxel alone. Of the gemcitabine/paclitaxel treated patients, 5% developed grade 3 thrombocytopenia. Febrile neutropenia or sepsis was observed in 5% of gemcitabine/paclitaxel treated patients compared with 2% of paclitaxel patients. Red blood cell transfusions were administered to 10% of the gemcitabine/paclitaxel- treated patients compared with 4% of paclitaxel alone) patients, although grade 3/4 anemia was uncommon. For grade 3/4 nonhematologic toxicity, there were modest differences in the incidence of fatigue as well as asymptomatic increases in hepatic transaminases in the gemcitabine/paclitaxel arm of the study. At this interim analysis, the percent of patients who died on study or during the 30-day follow-up period was similar on the two arms, as was the incidence of drug-related deaths, which was one patient on each arm. Of patients treated with gemcitabine/paclitaxel, 6% discontinued therapy due to drug-related adverse events compared with 2% with paclitaxel alone.