Eligibility Criteria

Women with pathologically confirmed, uni- or bidimensionally measurable, HER-2–positive MBC were eligible. Patients were required to be ≥ 18 years of age, have an Eastern Cooperative Oncology Group performance status of 0 to 2, and provide written informed consent. Women of childbearing potential had to have a negative pregnancy test before entry and practice appropriate contraception while on study.

Cancer tissue specimens were evaluated for HER-2 overexpression initially by immunohistochemistry (IHC; HercepTest; Dako Corp, Carpinteria, CA) performed centrally, with an IHC staining score of 2+ or 3+ required for enrollment. An amendment (after 60 patients enrolled) permitted initial HER-2 assessment to be performed locally using either the HercepTest or fluorescent in situ hybridization (FISH; PathVysion assay kit; Vysis Inc, Downers Grove, IL). Patients with samples scoring 3+ by IHC or positive by FISH were eligible, whereas those with a 2+ IHC score also had to have a positive result by FISH. For 2+ patients only, specimens were forwarded to the central reference laboratory for FISH evaluation.

Patients could not have received prior chemotherapy for MBC, although prior adjuvant or neoadjuvant chemotherapy was permitted provided a taxane had not been used and the patient’s cumulative doxorubicin exposure was ≤ 360 mg/m². Concurrent immunotherapy or hormonal therapy was not allowed. Patients with brain metastasis not responding to treatment were excluded. At least 4 weeks must have elapsed between last surgery or radiation therapy with full recovery, and indicator lesions had to be outside prior radiation fields. Patients with a history of congestive heart failure, prior malignancy within the last 5 years that might affect breast cancer diagnosis or assessment, or any other medical or psychiatric condition that could compromise study participation were excluded.

A pretreatment evaluation performed ≤ 2 weeks before study entry included a complete history and physical examination, CBCs, serum chemistries, and tumor assessment. Laboratory requirements at study entry included an absolute neutrophil count ≥ 1,500/μL, platelet count ≥ 100,000/μL, hemoglobin ≥ 9 g/dL, serum creatinine ≤ 2.0 mg/dL, bilirubin ≤ 1.5 mg/dL, AST level less than 1.5× institutional upper limit of normal, and a normal left ventricular ejection fraction. Imaging tests to evaluate indicator lesions were performed ≤ 3 weeks before study entry.

Treatment

The first cycle of TPC consisted of trastuzumab 4 mg/kg on day 1, paclitaxel 175 mg/m² and carboplatin area under the time-concentration curve AUC = 6 (mg/mL·min) on day 2, and trastuzumab 2 mg/kg on days 8 and 15. Subsequent cycles administered trastuzumab 2 mg/kg on days 1, 8, and 15, with paclitaxel 175 mg/m² and carboplatin area under the time-concentration curve = 6 (mg/mL·min) on day 2. The first cycle of TP consisted of trastuzumab 4 mg/kg on day 1, paclitaxel 175 mg/m² on day 2, and trastuzumab 2 mg/kg on days 8 and 15. Subsequent cycles administered trastuzumab 2 mg/kg on days 1, 8, and 15, with paclitaxel 175 mg/m² on day 2. Both regimens were repeated every 3 weeks for at least six cycles. After chemotherapy, patients received trastuzumab 2 mg/kg/wk until disease progression or other discontinuation event.

Chemotherapy dose reductions were allowed in the event of serious hematologic, neurologic, or GI toxicity and were defined by the protocol. Hematopoietic growth factors were to be used only when clinically indicated and not prophylactically.

Evaluation of Response and Toxicity

Tumor assessments were performed by physical examination before every cycle, with imaging studies evaluating indicator lesions repeated every other cycle. Toxicities were evaluated before every cycle and graded using National Cancer Institute Common Toxicity Criteria.

Responses were assessed using standardized criteria.²⁴ Complete response (CR) required the disappearance of all known disease for ≥ 4 weeks. Partial response (PR) required a ≥ 50% decrease in the sum of products of the largest perpendicular diameters of all bidimensionally measurable lesions or a decrease ≥ 50% in the sum of diameters of all unidimensionally measurable lesions for ≥ 4 weeks, with no new lesions. Stable disease was defined as a less than 50% decrease and ≤ 25% increase in the size of bidimensionally measurable lesions, or a less than 50% decrease and ≤ 25% increase in the size of unidimensionally measurable lesions for ≥ 4 weeks, with no new lesions. Progressive disease was a more than 25% increase in the size of any measurable lesion or the appearance of a new lesion.

Study Design and End Points

Patients were stratified by IHC score (2+ or 3+) before random assignment. The primary study end point was ORR, with duration of response, PFS, and OS time evaluated as secondary end points. An estimated total of 98 patients per treatment arm were needed to provide 80% power to detect a 20% difference in ORR between TPC (70%) and TP (50%) at a .05 two-sided significance level. Analysis was performed using SAS version 8.0 software (SAS Institute, Cary, NC).

All patients receiving at least two cycles of therapy and with at least one follow-up tumor assessment were assessable for response. Patients who discontinued before two cycles because of toxicity were assessable as those who had experienced treatment failure, whereas those who experienced progression before completing two cycles of treatment were classified as having progressive disease. Duration of response was defined as the interval of time between the date of onset of CR or PR and the date of progression or the date of last follow-up. PFS was defined as the interval between the date of first dose and the date of progression or death as a result of any cause. Patients without progression who received subsequent therapy were censored at the start date of the new treatment. Patients alive and without evidence of disease progression and not receiving new therapy were censored at the last date of contact. Survival was defined as the interval between the date of first dose and the date of death.

The analysis was run on the intent-to-treat population, which included all registered patients. The efficacy analysis included only eligible patients, responses were calculated for all assessable patients, and all patients who received at least one dose of study drug were included in the analysis of toxicity. Kaplan-Meier methods were used to assess OS and PFS; log-rank tests were used to compare the differences between the treatment arms.

The study protocol and informed consent were reviewed and approved by the appropriate institutional review boards.

Response

Ninety-two of the 98 women (94%) randomly assigned to TPC were assessable for response; three patients discontinued therapy in the first cycle because of toxicity (n = 1 each: neutropenia, infusion reaction, and pleural effusion), two patients were registered and later deemed ineligible (because of study treatment before registration and an inadequate left ventricular ejection fraction), and one patient was removed from study because of protocol deviations (no computed tomography scans performed to determine response). In the TP arm, 94 of 98 patients (96%) were assessable for response; one patient withdrew from the study during cycle 1 because of toxicity and three patients enrolled but were later deemed ineligible (n = 1 each: disease was HER-2 1+, patient took tamoxifen off study, and study treatment before registration). The ORR was significantly higher with TPC (52%; 95% CI, 42% to 62%; 10% CR; 42% PR) versus TP (36%; 95% CI, 26% to 46%; 3% CR; 33% PR; P = .04; Table 2)

Improved response rates were most evident in the cohort of IHC 3+ patients, with an ORR of 57% with TPC (95% CI, 45% to 70%) versus 36% for TP (95% CI, 25% to 48%; P = .03). The ORRs in IHC 2+ patients (42% with TPC [95% CI, 25% to 59%] and 35% with TP [95% CI, 19% to 52%]) were not statistically different in this relatively small cohort.

Median duration of response for all patients was 13.0 months (range, 1.6 to 46.3 months) in the TPC arm compared with 11 months (range, 1.4 to 47.1 months) in the TP arm. Median PFS was significantly longer in patients receiving TPC compared with TP: 10.7 (range, < 1 to 50.4 months) v 7.1 months (range, < 1 to 55.3 months; P = .03), respectively (HR = 0.66; 95% CI, 0.59 to 0.73; Fig 1). Median PFS was also significantly longer in IHC 3+ patients receiving TPC compared with TP: 13.8 (range, < 1 to 50.4 months) v 7.6 months (range, < 1 to 55.3 months; P = .005), respectively (HR = 0.55; 95% CI, 0.48 to 0.62; Fig 2).

For TPC, the clinical benefit rate (CBR; CR + PR + stable disease ≥ 6 months) was 55%; for TP, the CBR was 40% (P = .06). For IHC3+ TPC patients, the CBR was 61%, whereas in TP patients, it was 40% (P = .03).

At last follow-up, half of the patients in each arm had died, primarily from disease progression. There were no significant differences in cause or number of deaths between study arms. The 4-year OS rates were 38% with TPC versus 31% with TP. There was a trend toward improved survival with TPC; the median OS time was 35.7 (range, < 1 to 56.8 months) v 32.2 months (range, < 1 to 55.3 months) with TP (HR = 0.9; 95% CI, 0.88 to 0.92), but this difference was not statistically significant (Fig 3). In IHC 3+ patients, there was a greater difference in median survival times: 41.5 months (range, < 1 to 51.8 months) with TPC and 30.6 months (range, < 1 to 55.3 months) with TP (HR = 0.74; 95% CI, 0.46 to 0.64), but again the difference was not significant (Fig 4). The estimated 4-year survival rates in this cohort were 39% v 29%, respectively.

At the time this trial was designed and initiated, IHC was the only approved assay method to evaluate HER-2 status in routine clinical specimens to select patients for trastuzumab therapy. During the course of the study, evaluation of HER2 gene amplification by FISH was also approved, leading to inclusion of this method for determination of initial patient eligibility. An attempt was made to evaluate retrospectively all specimens using FISH and compare these findings with clinical outcomes; however, FISH results were only available for approximately 70% of the patients. Although we noted trends toward improved clinical response and PFS with TPC compared with TP, in the subset of FISH-positive patients, the incomplete data set and the retrospective nature of the evaluation prevent any more conclusive statement.

Toxicity

One hundred ninety-five patients received at least one dose of study drug and were assessable for toxicity. Table 3 summarizes grade 3 to 4 treatment-related toxicities. Overall, both regimens were well tolerated; however, there was a higher incidence of grade 4 neutropenia in the TPC arm compared with the TP arm (36% v 12%, respectively; P = .0001). In both treatment arms, neutropenia generally was uncomplicated and the overall incidence of febrile neutropenia was low (3% in the TPC arm and 1% in the TP arm). As expected, grade 3 thrombocytopenia occurred more frequently with TPC compared with TP (9% v 1%, respectively); however, no patient developed grade 4 thrombocytopenia. Congestive heart failure developed in two TP patients, was managed, and both patients were alive as of December 2003. Other grade 3 or 4 toxicities occurred infrequently, with no differences between the study arms.