- Email: [email protected]
High Efficacy of Gemcitabine and Cisplatin Plus Trastuzumab in Patients with HER2-overexpressing Metastatic Breast Cancer: A Phase II Study
Clinical Oncology (2005) 17: 630–635 doi:10.1016/j.clon.2005.06.010
Original Article High Efficacy of Gemcitabine and Cisplatin Plus Trastuzumab in Patients with HER2-overexpressing Metastatic Breast Cancer: A Phase II Study H. J. Stemmler*, S. Kahlerty, O. Brudlerz, M. Behax, S. Mu¨llerk, B. Stauch{, V. Heinemann* *Department of Medicine III, Klinikum Großhadern, University of Munich, Germany; yDepartment of Gynaecology and Obstetrics, Klinikum Großhadern, University of Munich, Germany; zOncological Practice, Augsburg, Germany; xOncological Practice, Amberg, Germany; kOncological Practice, Ansbach, Germany; {Oncological Practice, Kronach, Germany ABSTRACT: Aims: Effective and tolerable regimens are sought specifically in women who have been pre-treated with anthracyclines and taxanes. Gemcitabine and cisplatin plus trastuzumab has shown synergistic activity in vitro, and provides a new mechanism of drug interaction. This multicentre phase II study aimed to evaluate the efficacy and tolerability of gemcitabine and cisplatin plus trastuzumab in previously treated patients with metastatic breast cancer (MBC). Materials and methods: Previously treated patients with human epidermal growth factor receptor 2 (HER2) overexpressing MBC were enrolled in a multicentre phase II study (DAKO Hercep Test 3C). Treatment consisted of gemcitabine (750 mg/m2), cisplatin (30 mg/m2) given on days 1 and 8 every 3 weeks, and trastuzumab (4 mg/kg loading dose, 2 mg/kg weekly). Results: Twenty patients were recruited, all of whom had previously received chemotherapy (12 pre-treated with taxanes, 18 pre-treated with anthracyclines seven pre-treated with taxanes and trastuzumab). A median of six cycles of the study treatment was delivered. There were eight partial responses, for an overall response rate of 40% (95% confidence interval 16.5–63.5%). The clinical benefit rate (complete response plus partial response plus stable disease) was 80% (95% CI 54.2–95.8%). The response rate in patients who had already received a trastuzumab-based regimen for MBC was 57.1% (95% CI 7.7–100%). Median time to progression was 10.2 months, and median overall survival was 18.8 months. Main toxicities were leukopenia (grade 3 in 55% of patients) and thrombocytopenia (grade 3 in 35% and grade 4 in 5% of patients). Non-haematological toxicity was rarely severe. Conclusions: Combination chemotherapy with gemcitabine and cisplatin plus trastuzumab is well tolerated and active in patients with HER2 overexpressing MBC, even after prior exposure to anthracyclines and taxanes. Stemmler, H. J. et al. (2005). Clinical Oncology 17, 630–635 Ó 2005 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. Key words: Anthracycline resistance, cisplatin, gemcitabine, HER2 overexpression, metastatic breast cancer, trastuzumab Received: 10 May 2005 Revised: 9 June 2005
Introduction
Treatment for metastatic breast cancer (MBC) is guided by biologic features of the tumour. Women with human epidermal growth factor receptor 2 (HER2) overexpressing tumours are candidates for treatment with trastuzumab. Compared with chemotherapy alone (doxorubicine, cyclophosphamide or paclitaxel), chemotherapy (doxorubicine, cyclophosphamide or paclitaxel) with trastuzumab improves clinical outcome. The addition of trastuzumab improves the Author for correspondence: Dr H. J. Stemmler, Medical Department III, University of Munich, Klinikum Großhadern, Marchioninistrasse, 15, 81377 Munich, Germany. Tel: C49-89-7095-2208; Fax: C49-89-70958828; E-mail: [email protected] 0936-6555/05/000000C06 $35.00/0
Accepted: 14 June 2005
progression-free survival (median time to tumour progression 7.4 vs 4.6 months; P ! 0.001), the response rate (50% vs 32%; P ! 0.001), and the overall survival (median survival 25.1 vs 20.3 months; P Z 0.046) [1]. Chemotherapy strategies are confounded by the increasing exposure of patients to chemotherapy, including anthracycline and taxane-based regimen in the neoadjuvant and adjuvant settings. Two general strategies are apparent and should be followed: improving treatment efficacy by exploring new drugs and drug combinations, and ensuring that efficacy is improved at the lowest cost to quality of life. Several considerations support the use of gemcitabine and cisplatin in salvage therapy, and make it a particularly attractive treatment option for MBC: (1) this combination has shown additive or synergistic activity in vitro, which
Ó 2005 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
GEMCITABINE AND CISPLATIN PLUS TRASTUZUMAB FOR METASTATIC BREAST CANCER
was most pronounced in cisplatin-resistant cell lines, and was found to be caused by an increased formation and an impaired repair of platinum-DNA adducts [2,3]; (2) gemcitabine and cisplatin are usually not included into adjuvant or neoadjuvant chemotherapy; therefore, resistance to either drug is less likely to occur; (3) there is minimal overlapping toxicity for gemcitabine and cisplatin, suggesting acceptable toxicity even in intensively pretreated patients; (4) finally, the addition of trastuzumab to gemcitabine and cisplatin might form an effective triplet combination [4,5]. Naruse et al. [6] have shown evidence for an enhanced anti-tumour effect of trastuzumab in combination with cisplatin. Gemcitabine and cisplatin alone has proven efficacy in studies for pre-treated patients, with response rates ranging from 26% to 63% [7–13]. In a small phase II study (n Z 31), Ruiz et al. [14] reported a favourable high overall response rate of 80% (complete response 12.4%; partial response 67.6%) in patients who had received gemcitabine or cisplatin as first-line treatment for metastatic breast cancer. The present multicentre phase II study was aimed to evaluate the efficacy and tolerability of gemcitabine and cisplatin plus trastuzumab in previously treated patients with metastatic breast cancer.
Patients and Methods Patient Population
Twenty patients with histologically and immunohistochemically confirmed HER2 overexpressing MBC (DA KO3C; DAKO Hercep TestÔ) were recruited between March 2001 and April 2004 on a treatment protocol approved by the Local Ethics Committee (Ethics Committee University of Munich, Germany; #018/00). All patients were required to give written informed consent before study entry. Prior treatment with chemotherapy, hormonal therapy, immunotherapy (including trastuzumab) or local radiotherapy was permitted. Patients were required to have at least one bi-dimensionally measurable lesion outside a previous radiation port. Other eligibility criteria included age from 18 to 70 years, Karnofsky performance status 70% or over, life expectancy of 12 weeks, and adequate haematological, renal, cardiac and hepatic function (leukocyte count R 3.0 ! 109/l or absolute neutrophil count R 2 ! 109/l; platelet count R 100 ! 109/l; haemoglobin R 8 g/dl; total serum bilirubin % 1.25 ! upper limit of normal 9 [ULN] in the absence of liver metastasis or %3.0 ! ULN in the presence of liver metastasis; transaminase [ALT, AST] level % 3 ! ULN in the absence of liver metastasis or %5 ! ULN in the presence of liver metastasis; and alkaline phosphatase level % 2.5 ! ULN). Creatinine clearance was required to exceed 60 ml/min, and cardiac ejection fraction has to be normal. Patients were not eligible for study enrolment if they had the following: a history of cardiac dysfunction; an impaired left-ventricular ejection fraction; if they were pregnant, lactating or refused effective contraception; if they had
631
bone metastases alone, known brain metastases or a secondary malignancy; a history of another primary malignant disease other than carcinoma in situ of the uterine cervix or adequately treated basal cell skin cancer; and active infection or any other concomitant severe clinical condition making implementation of the protocol including prehydration difficult. Patient Assessment
Patients were evaluated on a regular basis during treatment. The following assessments were carried out before each 3week cycle: physical examination, full blood count, serum chemistry and assessment of toxicities. During the initial phase of treatment, complete blood counts were carried out twice weekly to determine the nadir values. If the haematological values had not recovered by the time of scheduled treatment, the full blood count was repeated every week until recovery of leukocyte count to 3.0 ! 109/l and platelets to R 100 ! 109/l. For all patients, tumours were measured by physical examination, imaging procedures as ultrasound, computed tomography or magnetic resonance imaging within 14 days before entry into the study and, subsequently, after every two cycles of treatment applying the initially used imaging procedure. Patient response was assessed by standard WHO criteria, as follows: complete response was defined as the disappearance of all known disease, as determined by two observations not less than 4 weeks apart, whereas partial response was defined as a decrease by at least 50% of the sum of the products of the largest perpendicular diameters of all measurable lesions, as determined by two observations not less than 4 weeks apart. Stable disease, lasting for at least 6 weeks from the start of study (i.e. first drug administration), was defined as less than 50% decrease and less than 25% increase in the sum of the products of the largest perpendicular diameters of all measurable lesions. Progressive disease was a greater than 25% increase in the size of at least one bi-dimensionally or unidimensionally measurable lesion, or the appearance of a new lesion. The occurrence of pleural effusion was considered to be a sign of progression if it could be substantiated by positive cytology. Cardiac function surveillance was repeated after every two courses of therapy, applying echocardiography. In addition, time to tumour progression (time from start of therapy to first evidence of progressive disease or last follow-up), and survival (time from start of therapy to death), were measured (intent-to-treat). Toxicity was graded according to the National Cancer Institute Common Toxicity Criteria [15]. Treatment Schedule
Treatment consisted of trastuzumab given on a weekly basis (2 mg/kg) after an initial loading dose of 4 mg/kg. Gemcitabine 750 mg/m2 was given as a 30-min infusion followed by cisplatin 30 mg/m2 given as a 1-h infusion on days 1 and 8 of a 3-week treatment cycle. Before infusion
632
CLINICAL ONCOLOGY
of cisplatin, patients received hydration with 1 l/m2 over 2 h or oral application of 2–3 l mineral water over 8 h. Immediately before cisplatin, patients received intravenous furosemide 20 mg. Posthydration was carried out 2–24 h after cisplatin administration using an oral application of at least 1 l/m2. All patients received routinely 5HT3 antagonists and 8 mg dexamethasone before the start of treatment. Treatment was continued until disease progression or the occurrence of unacceptable toxicity. Dose adjustments were made on the basis of leukocyte and platelet counts on the day of treatment and clinical assessments of non-haematological toxicities. The doses of both drugs were reduced by 25% if the leukocyte count was between 2.5 and 3.0 ! 109/l, whereas the platelet count exceeded 100 ! 109/l; if the leukocyte count was less than 2.5 ! 109/l or the platelet count was less than 100 ! 109/l, the doses of both drugs were omitted. Omitted day-8 doses were not replaced, but the next cycle was given as planned but at reduced doses. If any toxicity of grade 3 or above (except nausea, vomiting or alopecia) occurred, drug doses were reduced by 50%. If the patient tolerated the dosemodified treatment well, a re-increase of dosage could be attempted in the following cycle. If the creatinine clearance fell below 60 ml/min, cisplatin administration was discontinued, and, in the case of any grade 4 non-haematological toxicity, the study chemotherapy was discontinued. The use of haematopoietic growth factors was allowed in patients with protracted myelosuppression. Administration of other cytotoxic, immune or hormonal agents or radiation therapy was not permitted during the study, with the exception of steroids given as analgetic or antiemetic treatment, or local palliative radiation. Statistical Analysis
The primary objective of the study was to determine the objective response rate to the study treatment. Secondary end points included time to progression, survival and toxicity. Simon’s [16] optimal two-stage design was used to ensure that the number of patients exposed to this therapy was minimised should the therapy prove ineffective. The 95% confidence interval for response rates were calculated using methods for exact binominal confidence intervals [17]. Time-to-event end points were calculated according to the method of Kaplan and Meier using the STATISTICA software [18]. Patients who received at least one treatment cycle were evaluable for toxicity, and those receiving at least two treatment cycles or those who progressed after the first cycle were evaluable for response.
Results Patient Characteristics
Twenty eligible patients with HER2 overexpressing MBC (DAKO3C) were recruited. All patients were analysed for efficacy and toxicity (intent-to-treat). Patient characteristics are shown in Table 1. Median age was 52.5 years (range 39–68 years). All patients had previously received
Table 1 – Baseline patient characteristics Patients (n Z 20) Variable Age (years) Median Range Menopausal status Premenopausal Postmenopausal
Number
%
52.5 39–68 4 16
20 80
12
60
8
40
HER2 status 3C (DAKO Hercep TestÔ)
20
100
Number of metastatic sites 1 2 R3
10 8 2
50 40 10
Sites of metastases Visceral disease Soft tissue or nodal disease only Bone metastases only
18 1 1
90 5 5
Prior treatment Any chemotherapy Palliative chemotherapy Adjuvant chemotherapy Trastuzumab C taxanes (metastatic)
20 12 8 7
100 60 40 35
Number of prior chemotherapy regimens for metastatic breast cancer 0 1 2 R3
8 3 3 6
40 15 15 30
Sensitivity to prior anthracyclines Not exposed Sensitive Resistant* Refractoryy
2 9 6 3
10 45 30 10
Hormone receptor status Oestrogen receptor, progesterone receptor positive, or both Oestrogen receptor and progesterone receptor negative
*Patients were considered resistant if they relapsed within 12 months after completing anthracycline-based adjuvant chemotherapy, or if they initially responded to anthracycline-based chemotherapy for metastatic disease but progression occurred while still on therapy. yPatients were considered refractory if they never achieved an objective response to anthracyclinebased palliative chemotherapy and progressed while on therapy.
chemotherapy, and 12 of them had received up to five prior chemotherapy regimens for metastatic disease. Eighteen patients (90%) had previously received anthracyclines, and 11 patients (55%) had received an anthracycline- and taxane-based regimen. Nearly half of the patients were considered either resistant or refractory to anthracyclines (n Z 9; 45%). Seven patients had received previous trastuzumab-based regimens for metastatic disease (35%).
GEMCITABINE AND CISPLATIN PLUS TRASTUZUMAB FOR METASTATIC BREAST CANCER
633
All patients had visceral disease, and 10 patients (50%) had more than one metastatic site. Treatment Delivery
A total of 140 cycles of the study medication were delivered. Patients received a median number of six cycles (range: 1–18). Dose reductions, delays and omissions occurred in eight (5.7%), three (2.1%) and two (1.4%) cycles, respectively. Seven patients received maintenance therapy with trastuzumab after a median of 6.5 cycles of combination therapy. Response and Survival
There were no complete responses, but eight patients (40%) achieved a partial response, for an objective response rate of 40% (95% CI 16.5–63.5%). Eight additional patients (40%) had stable disease, and tumour progression occurred in three patients (15%). Overall disease control (objective response plus stable disease) was obtained in 80% of patients (95% CI 54.2–95.8%). A response analysis by baseline characteristics is shown in Table 2. The median time to progression was 10.2 months (range: 1.9–19.7 months). The median overall survival amounts to 18.8 months (range: 5.8–22.9 months). Time to progression and the overall survival curve is shown in Fig. 1. Toxicity
Haematological and non-haematological toxicities are shown in Table 3. The predominant haematological toxicity Table 2 – Response rates by baseline characteristics
Variable
Overall response Total number Number of patients of patients % 95% CI
Number of prior cycles of chemotherapy for metastatic breast cancer 0 (first-line) 8 3 37.5 5.7–80.8 R 1 (beyond first-line) 12 5 45.5 8.9–74.4 Number of metastatic sites 1 R2
10 10
3 5
30 50
Anthracycline sensitivity Sensitive or not exposed Resistant or refractory
11 9
3 5
27.3 4.1–58.7 55.6 15.0–96.1
7
4
57.1
7.7–100
12
5
41.7
8.9–74.4
8
3
37.5
5.7–80.8
Trastuzumab pre-treated plus taxanes Hormone receptor status Oestrogen receptor, progesterone receptor positive, or both Oestrogen receptor and progesterone receptor negative
3.7–53.7 12.3–87.7
Fig. 1 – Time to progression (TTP) and overall survival (OS).
was leukopenia that reached WHO grade 3 in 17.1% of cycles. Grade 3 thrombocytopenia occurred in 12.8% of cycles and reached grade 4 in 0.7% of cycles. Anaemia grade 3/4 was less frequent, and occurred in 2.1% of cycles. Non-haematological toxicity was mild to moderate in most patients.
Discussion
With the increasing use of the anthracycline- and taxanebased regimen in the adjuvant and neoadjuvant settings, and their established application in the treatment of the advanced and metastatic stages of breast cancer, there is a need for non-cross-resistant regimens. The rationale for a gemcitabine-based combination treatment is supported by its unique mechanism of action allowing synergistic cytotoxic interaction with cisplatin. Gemcitabine acts as an effective inhibitor of DNA repair, and thereby may overcome drug resistance induced by an upregulation of DNA repair. The combination of gemcitabine and cisplatin has proven efficacy in studies for pre-treated patients with metastatic breast cancer, with response rates ranging from 26% to 63%, and an increased response rate of up to 80% in a single study for first-line patients. As shown in pre-clinical studies, the addition of trastuzumab enhances the effectiveness of cisplatin [6]. Moreover, the combination of trastuzumab and gemcitabine has proven efficacy in HER2 overexpressing MBC [19]. O’Shaughnessy et al. [19] reported an overall response rate of 38% and a median time to progression of 5.8 months for pre-treated patients with metastatic breast cancer. Christodoulou et al. [20] have found a comparable high overall response rate of 36% and a comparable time to tumour progression of 7.8 months in pre-treated patients. The addition of trastuzumab to gemcitabine plus cisplatin might, therefore, form an effective and triplet combination for patients with HER2 overexpressing MBC [4,5]. In view of the pre-treatment of many patients, the efficacy of the triplet combination in this phase II trial was
634
CLINICAL ONCOLOGY
Table 3 – Toxicity profile (number and % of cycles and patients) WHO grade 1 Number
2
3
4
%
Number
%
Number
%
Number
%
Haematological toxicity (n and % of cycles) Leukopenia 23 Thrombocytopenia 35 Anaemia 55
16.4 24.9 39.3
57 29 35
40.7 20.7 24.9
24 17 3
17.1 12.1 2.1
– 1 –
– 0.7 –
Non-haematological toxicity (n and % of patients) Alopecia (grade 1 and 2) 6 Nausea and vomiting 7 Asthenia 3 Neurotoxicity 2 Creatinine 1
30 35 15 10 5
3 4 2 1 –
15 20 10 5 –
– 1 1 1 –
– 5 5 5 –
– – – – –
– – – – –
quite satisfactory, with an overall response rate of 40% (95% CI 16.5–63.5%) and a clinical benefit rate (complete response plus partial response plus stable disease) of 80% (95% CI 54.2–95.8%). It seems particularly noteworthy that the response rate had not decreased in patients with anthracycline-resistant or refractory disease compared with those who were sensitive or not exposed to anthracyclines (55.6% vs 27.3%). Moreover, four out of seven patients who had progressive disease during trastuzumab-based pretreatment (in combination with paclitaxel) responded to the schedule (57.1%), and two other patients achieved a stabilisation of the disease (clinical benefit rate 85.7%). Despite a higher rate of responses in pre-treated patients compared with first-line patients (45.5 vs 37.5%), the sample size of this study is insufficient to draw any conclusions about the efficacy of sub-groups. Most responses and disease stabilisations were durable, with a median time to progression of 10.2 months and a median overall survival of 18.8 months. Compared with studies with gemcitabine and trastuzumab alone for pretreated patients, the time to tumour progression of 10.2 months seems to be prolonged [19,20]. Additionally, these results demonstrate the potentially beneficial effect of trastuzumab, which enhances the duration of response and the overall survival of HER2 overexpressing patients with MBC compared with gemcitabine–cisplatin schedules commonly used in anthracycline-pre-treated patients with MBC [7–14]. Clinical experience of the toxicity of a triplet combination for pre-treated metastatic breast cancer is limited. Toxicity data are available from a study given this triplet for non-small-cell lung cancer [21]. Here the data suggest no cumulative toxicity by the addition of trastuzumab to gemcitabine plus cisplatin. Although severe haematological toxicity (grade 3 and 4), mainly leukopenia and thrombocytopenia, was observed in 16.7% and 13.5% of cycles, respectively, the treatment-associated toxicity profile was generally acceptable. In our study, a median of six cycles could be administered, and dose reductions, delays and omissions occurred in eight (6.3%), three (2.4%) and two (1.6%) cycles, respectively. This is particularly
noteworthy, as half of our patients had received two or more prior chemotherapy regimens, including taxanes and anthracyclines. Symptomatic adverse events, such as nausea and vomiting or asthenia were generally mild to moderate. Conclusion
In conclusion, the combination of trastuzumab and gemcitabine plus cisplatin proved to be a well tolerated and effective regimen that provided sustained disease control in patients with HER2 overexpressing MBC. Specifically, after previous exposure to anthracyclines, taxanes, or both, this regimen should be recommended.
References 1 Slamon DJ, Leyland-Jones B, Shak S, et al. Use of chemotherapy plus a monoclonal antibody against HER2 for metastatic breast cancer that overexpresses HER2. N Engl J Med 2001;344:783–792. 2 Carmichael J. The role of gemcitabine in the treatment of other tumours. Br J Cancer 1998;78(suppl 3):21–25. 3 van Moorsel CJ, Pinedo HM, Veerman G, et al. Mechanisms of synergism between cisplatin and gemcitabine in ovarian and non-smallcell lung cancer cell lines. Br J Cancer 1999;80:981–990. 4 Heinemann V, Stemmler HJ, Wohlrab B, et al. Gemcitabine and cisplatin C/ trastuzumab (herceptin) in intensively pretreated metastatic breast cancer (MBC) [abstract]. Jahrestagung der DGHO 2002, Munich 27–30 October, 2002. 5 Stemmler HJ, Heinemann V, Brehm G. Response to a triple combination of gemcitabine, cisplatin and herceptin in metastatic breast cancer previously refractory to gemcitabine and herceptin: a case report [abstract]. Jahrestagung der DGHO 1999, Jena 3–6 October, 1999. 6 Naruse I, Fukumoto H, Saijo N, Nishio K. Enhanced anti-tumor effect of trastuzumab in combination with cisplatin. Jpn J Cancer Res 2002; 93:574–581. 7 Chaudry S, Abdel-Rahman HA, Patil R. Prospective phase II study of weekly cisplatin–gemcitabine in refractory metastatic breast cancer (MBC) [abstract 430]. Proc Am Soc Clin Oncol 2000;19:111a. 8 de la Pena HF, Tellez E, Bastarrachea J. Gemcitabine plus cisplatin as a palliative treatment in metastatic or refractory breast cancer (MRBC) [abstract]. Proc Am Soc Clin Oncol 1998;2001:20.
GEMCITABINE AND CISPLATIN PLUS TRASTUZUMAB FOR METASTATIC BREAST CANCER
9 Doroshow JH, Tetef M, Margolin K. Significant activity of gemcitabine (Gem) and cisplatin (Ddp) in both heavily (H) and minimally (M) pretreated metastatic breast cancer (Mbc) patients (Pts): a California Cancer Consortium/Loyola University of Chicago trial [abstract 609H]. Proc Am Soc Clin Oncol 2000;19:155a. 10 Galvez CA, Calmarini F, Curie M. Monthly cisplatin (C) and gemcitabine (G) as second-line chemotherapy for patients with advanced breast cancer [abstract]. Breast Cancer Res Treat 2000;64:81. 11 Nagourney RA, Link JS, Blitzer JB, Forsthoff C, Evans SS. Gemcitabine plus cisplatin repeating doublet therapy in previously treated, relapsed breast cancer patients. J Clin Oncol 2000;18:2245–2249. 12 Nagourney RA. Gemcitabine plus cisplatin in breast cancer. Oncology (Huntingt) 2001;15(suppl 3):28–33. 13 van Moorsel CJ, Veerman G, Bergman AM, et al. Combination chemotherapy studies with gemcitabine. Semin Oncol 1997;24(suppl 7):S7. 14 Ruiz GC, de la Pena HF, Meza FA. A phase II study of gemcitabine plus cisplatin in metastatic breast cancer (MBC) [abstract]. Proc Am Soc Clin Oncol 2001;20:60b.
635
15 Ajani JA, Welch SR, Raber MN, Fields WS, Krakoff IH. Comprehensive criteria for assessing therapy-induced toxicity. Cancer Invest 1990; 8:147–159. 16 Simon R. Optimal two-stage designs for phase II clinical trials. Control Clin Trials 1989;10:1–10. 17 Cox DR. The analyses of binary data. London, UK: Methuen, 1970. 18 Kaplan EL, Meier P. Non-parametric estimation from incomplete observations. J Am Stat Assoc 1959;53:457–481. 19 O’Shaughnessy JA, Vukelja S, Marsland T, Kimmel G, Ratnam S, Pippen JE. Phase II study of trastuzumab plus gemcitabine in chemotherapy-pretreated patients with metastatic breast cancer. Clin Breast Cancer 2004;5:142–147. 20 Christodoulou C, Fountzilas G, Razi E. Gemcitabine and trastuzumab combination as salvage treatment in patients with HER-2-positive metastatic breast cancer. Proc Am Soc Clin Oncol 2003;22. 21 Gatzemeier U, Groth G, Butts C, et al. Randomized phase II trial of gemcitabine–cisplatin with or without trastuzumab in HER2-positive non-small-cell lung cancer. Ann Oncol 2004;15:19–27.
Original Article High Efficacy of Gemcitabine and Cisplatin Plus Trastuzumab in Patients with HER2-overexpressing Metastatic Breast Cancer: A Phase II Study H. J. Stemmler*, S. Kahlerty, O. Brudlerz, M. Behax, S. Mu¨llerk, B. Stauch{, V. Heinemann* *Department of Medicine III, Klinikum Großhadern, University of Munich, Germany; yDepartment of Gynaecology and Obstetrics, Klinikum Großhadern, University of Munich, Germany; zOncological Practice, Augsburg, Germany; xOncological Practice, Amberg, Germany; kOncological Practice, Ansbach, Germany; {Oncological Practice, Kronach, Germany ABSTRACT: Aims: Effective and tolerable regimens are sought specifically in women who have been pre-treated with anthracyclines and taxanes. Gemcitabine and cisplatin plus trastuzumab has shown synergistic activity in vitro, and provides a new mechanism of drug interaction. This multicentre phase II study aimed to evaluate the efficacy and tolerability of gemcitabine and cisplatin plus trastuzumab in previously treated patients with metastatic breast cancer (MBC). Materials and methods: Previously treated patients with human epidermal growth factor receptor 2 (HER2) overexpressing MBC were enrolled in a multicentre phase II study (DAKO Hercep Test 3C). Treatment consisted of gemcitabine (750 mg/m2), cisplatin (30 mg/m2) given on days 1 and 8 every 3 weeks, and trastuzumab (4 mg/kg loading dose, 2 mg/kg weekly). Results: Twenty patients were recruited, all of whom had previously received chemotherapy (12 pre-treated with taxanes, 18 pre-treated with anthracyclines seven pre-treated with taxanes and trastuzumab). A median of six cycles of the study treatment was delivered. There were eight partial responses, for an overall response rate of 40% (95% confidence interval 16.5–63.5%). The clinical benefit rate (complete response plus partial response plus stable disease) was 80% (95% CI 54.2–95.8%). The response rate in patients who had already received a trastuzumab-based regimen for MBC was 57.1% (95% CI 7.7–100%). Median time to progression was 10.2 months, and median overall survival was 18.8 months. Main toxicities were leukopenia (grade 3 in 55% of patients) and thrombocytopenia (grade 3 in 35% and grade 4 in 5% of patients). Non-haematological toxicity was rarely severe. Conclusions: Combination chemotherapy with gemcitabine and cisplatin plus trastuzumab is well tolerated and active in patients with HER2 overexpressing MBC, even after prior exposure to anthracyclines and taxanes. Stemmler, H. J. et al. (2005). Clinical Oncology 17, 630–635 Ó 2005 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. Key words: Anthracycline resistance, cisplatin, gemcitabine, HER2 overexpression, metastatic breast cancer, trastuzumab Received: 10 May 2005 Revised: 9 June 2005
Introduction
Treatment for metastatic breast cancer (MBC) is guided by biologic features of the tumour. Women with human epidermal growth factor receptor 2 (HER2) overexpressing tumours are candidates for treatment with trastuzumab. Compared with chemotherapy alone (doxorubicine, cyclophosphamide or paclitaxel), chemotherapy (doxorubicine, cyclophosphamide or paclitaxel) with trastuzumab improves clinical outcome. The addition of trastuzumab improves the Author for correspondence: Dr H. J. Stemmler, Medical Department III, University of Munich, Klinikum Großhadern, Marchioninistrasse, 15, 81377 Munich, Germany. Tel: C49-89-7095-2208; Fax: C49-89-70958828; E-mail: [email protected] 0936-6555/05/000000C06 $35.00/0
Accepted: 14 June 2005
progression-free survival (median time to tumour progression 7.4 vs 4.6 months; P ! 0.001), the response rate (50% vs 32%; P ! 0.001), and the overall survival (median survival 25.1 vs 20.3 months; P Z 0.046) [1]. Chemotherapy strategies are confounded by the increasing exposure of patients to chemotherapy, including anthracycline and taxane-based regimen in the neoadjuvant and adjuvant settings. Two general strategies are apparent and should be followed: improving treatment efficacy by exploring new drugs and drug combinations, and ensuring that efficacy is improved at the lowest cost to quality of life. Several considerations support the use of gemcitabine and cisplatin in salvage therapy, and make it a particularly attractive treatment option for MBC: (1) this combination has shown additive or synergistic activity in vitro, which
Ó 2005 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
GEMCITABINE AND CISPLATIN PLUS TRASTUZUMAB FOR METASTATIC BREAST CANCER
was most pronounced in cisplatin-resistant cell lines, and was found to be caused by an increased formation and an impaired repair of platinum-DNA adducts [2,3]; (2) gemcitabine and cisplatin are usually not included into adjuvant or neoadjuvant chemotherapy; therefore, resistance to either drug is less likely to occur; (3) there is minimal overlapping toxicity for gemcitabine and cisplatin, suggesting acceptable toxicity even in intensively pretreated patients; (4) finally, the addition of trastuzumab to gemcitabine and cisplatin might form an effective triplet combination [4,5]. Naruse et al. [6] have shown evidence for an enhanced anti-tumour effect of trastuzumab in combination with cisplatin. Gemcitabine and cisplatin alone has proven efficacy in studies for pre-treated patients, with response rates ranging from 26% to 63% [7–13]. In a small phase II study (n Z 31), Ruiz et al. [14] reported a favourable high overall response rate of 80% (complete response 12.4%; partial response 67.6%) in patients who had received gemcitabine or cisplatin as first-line treatment for metastatic breast cancer. The present multicentre phase II study was aimed to evaluate the efficacy and tolerability of gemcitabine and cisplatin plus trastuzumab in previously treated patients with metastatic breast cancer.
Patients and Methods Patient Population
Twenty patients with histologically and immunohistochemically confirmed HER2 overexpressing MBC (DA KO3C; DAKO Hercep TestÔ) were recruited between March 2001 and April 2004 on a treatment protocol approved by the Local Ethics Committee (Ethics Committee University of Munich, Germany; #018/00). All patients were required to give written informed consent before study entry. Prior treatment with chemotherapy, hormonal therapy, immunotherapy (including trastuzumab) or local radiotherapy was permitted. Patients were required to have at least one bi-dimensionally measurable lesion outside a previous radiation port. Other eligibility criteria included age from 18 to 70 years, Karnofsky performance status 70% or over, life expectancy of 12 weeks, and adequate haematological, renal, cardiac and hepatic function (leukocyte count R 3.0 ! 109/l or absolute neutrophil count R 2 ! 109/l; platelet count R 100 ! 109/l; haemoglobin R 8 g/dl; total serum bilirubin % 1.25 ! upper limit of normal 9 [ULN] in the absence of liver metastasis or %3.0 ! ULN in the presence of liver metastasis; transaminase [ALT, AST] level % 3 ! ULN in the absence of liver metastasis or %5 ! ULN in the presence of liver metastasis; and alkaline phosphatase level % 2.5 ! ULN). Creatinine clearance was required to exceed 60 ml/min, and cardiac ejection fraction has to be normal. Patients were not eligible for study enrolment if they had the following: a history of cardiac dysfunction; an impaired left-ventricular ejection fraction; if they were pregnant, lactating or refused effective contraception; if they had
631
bone metastases alone, known brain metastases or a secondary malignancy; a history of another primary malignant disease other than carcinoma in situ of the uterine cervix or adequately treated basal cell skin cancer; and active infection or any other concomitant severe clinical condition making implementation of the protocol including prehydration difficult. Patient Assessment
Patients were evaluated on a regular basis during treatment. The following assessments were carried out before each 3week cycle: physical examination, full blood count, serum chemistry and assessment of toxicities. During the initial phase of treatment, complete blood counts were carried out twice weekly to determine the nadir values. If the haematological values had not recovered by the time of scheduled treatment, the full blood count was repeated every week until recovery of leukocyte count to 3.0 ! 109/l and platelets to R 100 ! 109/l. For all patients, tumours were measured by physical examination, imaging procedures as ultrasound, computed tomography or magnetic resonance imaging within 14 days before entry into the study and, subsequently, after every two cycles of treatment applying the initially used imaging procedure. Patient response was assessed by standard WHO criteria, as follows: complete response was defined as the disappearance of all known disease, as determined by two observations not less than 4 weeks apart, whereas partial response was defined as a decrease by at least 50% of the sum of the products of the largest perpendicular diameters of all measurable lesions, as determined by two observations not less than 4 weeks apart. Stable disease, lasting for at least 6 weeks from the start of study (i.e. first drug administration), was defined as less than 50% decrease and less than 25% increase in the sum of the products of the largest perpendicular diameters of all measurable lesions. Progressive disease was a greater than 25% increase in the size of at least one bi-dimensionally or unidimensionally measurable lesion, or the appearance of a new lesion. The occurrence of pleural effusion was considered to be a sign of progression if it could be substantiated by positive cytology. Cardiac function surveillance was repeated after every two courses of therapy, applying echocardiography. In addition, time to tumour progression (time from start of therapy to first evidence of progressive disease or last follow-up), and survival (time from start of therapy to death), were measured (intent-to-treat). Toxicity was graded according to the National Cancer Institute Common Toxicity Criteria [15]. Treatment Schedule
Treatment consisted of trastuzumab given on a weekly basis (2 mg/kg) after an initial loading dose of 4 mg/kg. Gemcitabine 750 mg/m2 was given as a 30-min infusion followed by cisplatin 30 mg/m2 given as a 1-h infusion on days 1 and 8 of a 3-week treatment cycle. Before infusion
632
CLINICAL ONCOLOGY
of cisplatin, patients received hydration with 1 l/m2 over 2 h or oral application of 2–3 l mineral water over 8 h. Immediately before cisplatin, patients received intravenous furosemide 20 mg. Posthydration was carried out 2–24 h after cisplatin administration using an oral application of at least 1 l/m2. All patients received routinely 5HT3 antagonists and 8 mg dexamethasone before the start of treatment. Treatment was continued until disease progression or the occurrence of unacceptable toxicity. Dose adjustments were made on the basis of leukocyte and platelet counts on the day of treatment and clinical assessments of non-haematological toxicities. The doses of both drugs were reduced by 25% if the leukocyte count was between 2.5 and 3.0 ! 109/l, whereas the platelet count exceeded 100 ! 109/l; if the leukocyte count was less than 2.5 ! 109/l or the platelet count was less than 100 ! 109/l, the doses of both drugs were omitted. Omitted day-8 doses were not replaced, but the next cycle was given as planned but at reduced doses. If any toxicity of grade 3 or above (except nausea, vomiting or alopecia) occurred, drug doses were reduced by 50%. If the patient tolerated the dosemodified treatment well, a re-increase of dosage could be attempted in the following cycle. If the creatinine clearance fell below 60 ml/min, cisplatin administration was discontinued, and, in the case of any grade 4 non-haematological toxicity, the study chemotherapy was discontinued. The use of haematopoietic growth factors was allowed in patients with protracted myelosuppression. Administration of other cytotoxic, immune or hormonal agents or radiation therapy was not permitted during the study, with the exception of steroids given as analgetic or antiemetic treatment, or local palliative radiation. Statistical Analysis
The primary objective of the study was to determine the objective response rate to the study treatment. Secondary end points included time to progression, survival and toxicity. Simon’s [16] optimal two-stage design was used to ensure that the number of patients exposed to this therapy was minimised should the therapy prove ineffective. The 95% confidence interval for response rates were calculated using methods for exact binominal confidence intervals [17]. Time-to-event end points were calculated according to the method of Kaplan and Meier using the STATISTICA software [18]. Patients who received at least one treatment cycle were evaluable for toxicity, and those receiving at least two treatment cycles or those who progressed after the first cycle were evaluable for response.
Results Patient Characteristics
Twenty eligible patients with HER2 overexpressing MBC (DAKO3C) were recruited. All patients were analysed for efficacy and toxicity (intent-to-treat). Patient characteristics are shown in Table 1. Median age was 52.5 years (range 39–68 years). All patients had previously received
Table 1 – Baseline patient characteristics Patients (n Z 20) Variable Age (years) Median Range Menopausal status Premenopausal Postmenopausal
Number
%
52.5 39–68 4 16
20 80
12
60
8
40
HER2 status 3C (DAKO Hercep TestÔ)
20
100
Number of metastatic sites 1 2 R3
10 8 2
50 40 10
Sites of metastases Visceral disease Soft tissue or nodal disease only Bone metastases only
18 1 1
90 5 5
Prior treatment Any chemotherapy Palliative chemotherapy Adjuvant chemotherapy Trastuzumab C taxanes (metastatic)
20 12 8 7
100 60 40 35
Number of prior chemotherapy regimens for metastatic breast cancer 0 1 2 R3
8 3 3 6
40 15 15 30
Sensitivity to prior anthracyclines Not exposed Sensitive Resistant* Refractoryy
2 9 6 3
10 45 30 10
Hormone receptor status Oestrogen receptor, progesterone receptor positive, or both Oestrogen receptor and progesterone receptor negative
*Patients were considered resistant if they relapsed within 12 months after completing anthracycline-based adjuvant chemotherapy, or if they initially responded to anthracycline-based chemotherapy for metastatic disease but progression occurred while still on therapy. yPatients were considered refractory if they never achieved an objective response to anthracyclinebased palliative chemotherapy and progressed while on therapy.
chemotherapy, and 12 of them had received up to five prior chemotherapy regimens for metastatic disease. Eighteen patients (90%) had previously received anthracyclines, and 11 patients (55%) had received an anthracycline- and taxane-based regimen. Nearly half of the patients were considered either resistant or refractory to anthracyclines (n Z 9; 45%). Seven patients had received previous trastuzumab-based regimens for metastatic disease (35%).
GEMCITABINE AND CISPLATIN PLUS TRASTUZUMAB FOR METASTATIC BREAST CANCER
633
All patients had visceral disease, and 10 patients (50%) had more than one metastatic site. Treatment Delivery
A total of 140 cycles of the study medication were delivered. Patients received a median number of six cycles (range: 1–18). Dose reductions, delays and omissions occurred in eight (5.7%), three (2.1%) and two (1.4%) cycles, respectively. Seven patients received maintenance therapy with trastuzumab after a median of 6.5 cycles of combination therapy. Response and Survival
There were no complete responses, but eight patients (40%) achieved a partial response, for an objective response rate of 40% (95% CI 16.5–63.5%). Eight additional patients (40%) had stable disease, and tumour progression occurred in three patients (15%). Overall disease control (objective response plus stable disease) was obtained in 80% of patients (95% CI 54.2–95.8%). A response analysis by baseline characteristics is shown in Table 2. The median time to progression was 10.2 months (range: 1.9–19.7 months). The median overall survival amounts to 18.8 months (range: 5.8–22.9 months). Time to progression and the overall survival curve is shown in Fig. 1. Toxicity
Haematological and non-haematological toxicities are shown in Table 3. The predominant haematological toxicity Table 2 – Response rates by baseline characteristics
Variable
Overall response Total number Number of patients of patients % 95% CI
Number of prior cycles of chemotherapy for metastatic breast cancer 0 (first-line) 8 3 37.5 5.7–80.8 R 1 (beyond first-line) 12 5 45.5 8.9–74.4 Number of metastatic sites 1 R2
10 10
3 5
30 50
Anthracycline sensitivity Sensitive or not exposed Resistant or refractory
11 9
3 5
27.3 4.1–58.7 55.6 15.0–96.1
7
4
57.1
7.7–100
12
5
41.7
8.9–74.4
8
3
37.5
5.7–80.8
Trastuzumab pre-treated plus taxanes Hormone receptor status Oestrogen receptor, progesterone receptor positive, or both Oestrogen receptor and progesterone receptor negative
3.7–53.7 12.3–87.7
Fig. 1 – Time to progression (TTP) and overall survival (OS).
was leukopenia that reached WHO grade 3 in 17.1% of cycles. Grade 3 thrombocytopenia occurred in 12.8% of cycles and reached grade 4 in 0.7% of cycles. Anaemia grade 3/4 was less frequent, and occurred in 2.1% of cycles. Non-haematological toxicity was mild to moderate in most patients.
Discussion
With the increasing use of the anthracycline- and taxanebased regimen in the adjuvant and neoadjuvant settings, and their established application in the treatment of the advanced and metastatic stages of breast cancer, there is a need for non-cross-resistant regimens. The rationale for a gemcitabine-based combination treatment is supported by its unique mechanism of action allowing synergistic cytotoxic interaction with cisplatin. Gemcitabine acts as an effective inhibitor of DNA repair, and thereby may overcome drug resistance induced by an upregulation of DNA repair. The combination of gemcitabine and cisplatin has proven efficacy in studies for pre-treated patients with metastatic breast cancer, with response rates ranging from 26% to 63%, and an increased response rate of up to 80% in a single study for first-line patients. As shown in pre-clinical studies, the addition of trastuzumab enhances the effectiveness of cisplatin [6]. Moreover, the combination of trastuzumab and gemcitabine has proven efficacy in HER2 overexpressing MBC [19]. O’Shaughnessy et al. [19] reported an overall response rate of 38% and a median time to progression of 5.8 months for pre-treated patients with metastatic breast cancer. Christodoulou et al. [20] have found a comparable high overall response rate of 36% and a comparable time to tumour progression of 7.8 months in pre-treated patients. The addition of trastuzumab to gemcitabine plus cisplatin might, therefore, form an effective and triplet combination for patients with HER2 overexpressing MBC [4,5]. In view of the pre-treatment of many patients, the efficacy of the triplet combination in this phase II trial was
634
CLINICAL ONCOLOGY
Table 3 – Toxicity profile (number and % of cycles and patients) WHO grade 1 Number
2
3
4
%
Number
%
Number
%
Number
%
Haematological toxicity (n and % of cycles) Leukopenia 23 Thrombocytopenia 35 Anaemia 55
16.4 24.9 39.3
57 29 35
40.7 20.7 24.9
24 17 3
17.1 12.1 2.1
– 1 –
– 0.7 –
Non-haematological toxicity (n and % of patients) Alopecia (grade 1 and 2) 6 Nausea and vomiting 7 Asthenia 3 Neurotoxicity 2 Creatinine 1
30 35 15 10 5
3 4 2 1 –
15 20 10 5 –
– 1 1 1 –
– 5 5 5 –
– – – – –
– – – – –
quite satisfactory, with an overall response rate of 40% (95% CI 16.5–63.5%) and a clinical benefit rate (complete response plus partial response plus stable disease) of 80% (95% CI 54.2–95.8%). It seems particularly noteworthy that the response rate had not decreased in patients with anthracycline-resistant or refractory disease compared with those who were sensitive or not exposed to anthracyclines (55.6% vs 27.3%). Moreover, four out of seven patients who had progressive disease during trastuzumab-based pretreatment (in combination with paclitaxel) responded to the schedule (57.1%), and two other patients achieved a stabilisation of the disease (clinical benefit rate 85.7%). Despite a higher rate of responses in pre-treated patients compared with first-line patients (45.5 vs 37.5%), the sample size of this study is insufficient to draw any conclusions about the efficacy of sub-groups. Most responses and disease stabilisations were durable, with a median time to progression of 10.2 months and a median overall survival of 18.8 months. Compared with studies with gemcitabine and trastuzumab alone for pretreated patients, the time to tumour progression of 10.2 months seems to be prolonged [19,20]. Additionally, these results demonstrate the potentially beneficial effect of trastuzumab, which enhances the duration of response and the overall survival of HER2 overexpressing patients with MBC compared with gemcitabine–cisplatin schedules commonly used in anthracycline-pre-treated patients with MBC [7–14]. Clinical experience of the toxicity of a triplet combination for pre-treated metastatic breast cancer is limited. Toxicity data are available from a study given this triplet for non-small-cell lung cancer [21]. Here the data suggest no cumulative toxicity by the addition of trastuzumab to gemcitabine plus cisplatin. Although severe haematological toxicity (grade 3 and 4), mainly leukopenia and thrombocytopenia, was observed in 16.7% and 13.5% of cycles, respectively, the treatment-associated toxicity profile was generally acceptable. In our study, a median of six cycles could be administered, and dose reductions, delays and omissions occurred in eight (6.3%), three (2.4%) and two (1.6%) cycles, respectively. This is particularly
noteworthy, as half of our patients had received two or more prior chemotherapy regimens, including taxanes and anthracyclines. Symptomatic adverse events, such as nausea and vomiting or asthenia were generally mild to moderate. Conclusion
In conclusion, the combination of trastuzumab and gemcitabine plus cisplatin proved to be a well tolerated and effective regimen that provided sustained disease control in patients with HER2 overexpressing MBC. Specifically, after previous exposure to anthracyclines, taxanes, or both, this regimen should be recommended.
References 1 Slamon DJ, Leyland-Jones B, Shak S, et al. Use of chemotherapy plus a monoclonal antibody against HER2 for metastatic breast cancer that overexpresses HER2. N Engl J Med 2001;344:783–792. 2 Carmichael J. The role of gemcitabine in the treatment of other tumours. Br J Cancer 1998;78(suppl 3):21–25. 3 van Moorsel CJ, Pinedo HM, Veerman G, et al. Mechanisms of synergism between cisplatin and gemcitabine in ovarian and non-smallcell lung cancer cell lines. Br J Cancer 1999;80:981–990. 4 Heinemann V, Stemmler HJ, Wohlrab B, et al. Gemcitabine and cisplatin C/ trastuzumab (herceptin) in intensively pretreated metastatic breast cancer (MBC) [abstract]. Jahrestagung der DGHO 2002, Munich 27–30 October, 2002. 5 Stemmler HJ, Heinemann V, Brehm G. Response to a triple combination of gemcitabine, cisplatin and herceptin in metastatic breast cancer previously refractory to gemcitabine and herceptin: a case report [abstract]. Jahrestagung der DGHO 1999, Jena 3–6 October, 1999. 6 Naruse I, Fukumoto H, Saijo N, Nishio K. Enhanced anti-tumor effect of trastuzumab in combination with cisplatin. Jpn J Cancer Res 2002; 93:574–581. 7 Chaudry S, Abdel-Rahman HA, Patil R. Prospective phase II study of weekly cisplatin–gemcitabine in refractory metastatic breast cancer (MBC) [abstract 430]. Proc Am Soc Clin Oncol 2000;19:111a. 8 de la Pena HF, Tellez E, Bastarrachea J. Gemcitabine plus cisplatin as a palliative treatment in metastatic or refractory breast cancer (MRBC) [abstract]. Proc Am Soc Clin Oncol 1998;2001:20.
GEMCITABINE AND CISPLATIN PLUS TRASTUZUMAB FOR METASTATIC BREAST CANCER
9 Doroshow JH, Tetef M, Margolin K. Significant activity of gemcitabine (Gem) and cisplatin (Ddp) in both heavily (H) and minimally (M) pretreated metastatic breast cancer (Mbc) patients (Pts): a California Cancer Consortium/Loyola University of Chicago trial [abstract 609H]. Proc Am Soc Clin Oncol 2000;19:155a. 10 Galvez CA, Calmarini F, Curie M. Monthly cisplatin (C) and gemcitabine (G) as second-line chemotherapy for patients with advanced breast cancer [abstract]. Breast Cancer Res Treat 2000;64:81. 11 Nagourney RA, Link JS, Blitzer JB, Forsthoff C, Evans SS. Gemcitabine plus cisplatin repeating doublet therapy in previously treated, relapsed breast cancer patients. J Clin Oncol 2000;18:2245–2249. 12 Nagourney RA. Gemcitabine plus cisplatin in breast cancer. Oncology (Huntingt) 2001;15(suppl 3):28–33. 13 van Moorsel CJ, Veerman G, Bergman AM, et al. Combination chemotherapy studies with gemcitabine. Semin Oncol 1997;24(suppl 7):S7. 14 Ruiz GC, de la Pena HF, Meza FA. A phase II study of gemcitabine plus cisplatin in metastatic breast cancer (MBC) [abstract]. Proc Am Soc Clin Oncol 2001;20:60b.
635
15 Ajani JA, Welch SR, Raber MN, Fields WS, Krakoff IH. Comprehensive criteria for assessing therapy-induced toxicity. Cancer Invest 1990; 8:147–159. 16 Simon R. Optimal two-stage designs for phase II clinical trials. Control Clin Trials 1989;10:1–10. 17 Cox DR. The analyses of binary data. London, UK: Methuen, 1970. 18 Kaplan EL, Meier P. Non-parametric estimation from incomplete observations. J Am Stat Assoc 1959;53:457–481. 19 O’Shaughnessy JA, Vukelja S, Marsland T, Kimmel G, Ratnam S, Pippen JE. Phase II study of trastuzumab plus gemcitabine in chemotherapy-pretreated patients with metastatic breast cancer. Clin Breast Cancer 2004;5:142–147. 20 Christodoulou C, Fountzilas G, Razi E. Gemcitabine and trastuzumab combination as salvage treatment in patients with HER-2-positive metastatic breast cancer. Proc Am Soc Clin Oncol 2003;22. 21 Gatzemeier U, Groth G, Butts C, et al. Randomized phase II trial of gemcitabine–cisplatin with or without trastuzumab in HER2-positive non-small-cell lung cancer. Ann Oncol 2004;15:19–27.