Phase II study of paclitaxel plus carboplatin in patients with advanced carcinoma of the urothelium and renal dysfunction (E2896)

A trial of the Eastern Cooperative Oncology Group

Abstract

BACKGROUND

Chemotherapy options for the patients with advanced urothelial carcinoma and renal dysfunction are limited. The authors performed a Phase II trial of paclitaxel plus carboplatin in patients with advanced carcinoma of the urothelium and renal dysfunction.

METHODS

Forty-two patients were accrued; 37 eligible patients were treated. Patients received paclitaxel 225 mg/m² over three hours followed by carboplatin targeted area under the concentration-time curve = 6 mg/mL · minute every three weeks for up to six cycles.

RESULTS

The median number of cycles received was four (range, one to six). The objective response rate was 24.3% (95% confidence interval, 11.9–41.7%). The median progression free survival was 3 months and the median overall survival was 7.1 months. The number of poor prognostic risk factors (Eastern Cooperative Oncology Group performance status > 1 or lung, liver, or bone metastases) significantly predicted for survival. The most common ≥ 1 Grade 3 toxicities included granulocytopenia (60%) and neurotoxicity (35%).

CONCLUSIONS

Paclitaxel/carboplatin is a chemotherapy option for patients with advanced urothelial carcinoma and renal dysfunction. Future trials in chemotherapy development for this patient population are warranted. Cancer 2002;95:1022–7. © 2002 American Cancer Society.

DOI 10.1002/cncr.10782

An estimated 56,500 new cases of bladder carcinoma will be diagnosed in the United States in 2002, resulting in 12,600 deaths.1 Cisplatin-based combination chemotherapy regimens are the standard treatment for patients with metastatic carcinoma of the urothelium. A Phase III trial comparing M-VAC (methotrexate, vinblastine, doxorubicin, cisplatin) with single agent cisplatin showed that the combination regimen was superior both in terms of response rate (39% versus 12%) and survival (12.5 months versus 8.2 months).2 Recently, the doublet of gemcitabine/cisplatin was compared to M-VAC. This trial showed similar survival but a more favorable toxicity profile for gemcitabine/cisplatin.3 However, patients with advanced urothelial carcinoma may have impairment in renal function due to age, comorbid conditions, and/or malignancy-associated factors, making the utilization of cisplatin-based chemotherapy regimens problematic.

Although cisplatin has traditionally been considered the preferred platinum analog in urothelial carcinoma, pooled results of Phase II studies in which 274 patients with advanced urothelial carcinoma were treated with single-agent carboplatin showed a response rate of 14%.4 A prospective randomized comparison of single agent cisplatin versus carboplatin in this disease site has not been performed, nor have direct randomized combination trials. The lack of nephrotoxicity and the ability to dose carboplatin based upon glomerular filtration rate utilizing the Calvert formula5 are potentially important advantages of this agent, especially in the subgroup of patients that may have age- and disease-related alterations in renal function. Carboplatin-based combination chemotherapy regimens for urothelial carcinoma have previously been described. For example, Small et al. reported that the combination of methotrexate/vinblastine/mitoxantrone/carboplatin resulted in a response rate of 57% and a median survival of 10 months in patients with advanced urothelial carcinoma.6

The anti-microtubule agent paclitaxel shows significant activity in bladder carcinoma. The Eastern Cooperative Oncology Group (ECOG) reported a Phase II trial of paclitaxel 250 mg/m² over 24 hours every 21 days in patients with previously untreated advanced transitional cell carcinoma of the urothelium.7 A response rate of 42% (95% confidence interval [CI], 23–63%) and a complete response rate of 27% (95% CI, 12–48%) was reported. Dreicer et al. showed that paclitaxel may be utilized safely in the setting of renal dysfunction in patients with advanced urothelial carcinoma.8 Four of six patients with a median serum creatinine of 2.25 mg/dL achieved objective response with single agent paclitaxel. Several trials of the paclitaxel/carboplatin combination regimen in patients with advanced urothelial carcinoma have been previously reported.9–12 These trials showed a response rate ranging from 22.5 to 72%. In October 1996 the Eastern Cooperative Oncology Group initiated a Phase II trial of paclitaxel/carboplatin in patients with previously untreated advanced urothelial carcinoma and abnormal renal function. To the best of our knowledge, this is the first cooperative group urothelial carcinoma trial focusing specifically on this patient population.

PATIENTS AND METHODS

Patients

Patients were enrolled at participating institutions of the Eastern Cooperative Oncology Group. Patients with histologically confirmed metastatic or regionally progressive carcinoma of the urothelium (renal pelvis, ureter, bladder, or urethra) were eligible. All histologic subtypes were allowed. Patients could not have received chemotherapy for metastatic disease, but adjuvant or neoadjuvant chemotherapy at least 12 months prior to enrollment was allowed. All patients had bidimensionally measurable disease. Patients could not have received prior radiation therapy. Patients needed an ECOG performance status of ≤ 2. Initial laboratory requirements included absolute granulocyte count ≥ 1800/mm³, platelet count ≥ 100,000/mm³, AST ≤ 2.5 × normal, and bilirubin ≤ 2.0 mg/dL. Patients were required to have serum creatinine of 1.6–4.0 mg/dL (initially the lower limit was 1.5 mg/dL, but the protocol was amended after one patient was treated). Patients with active cardiac disease or infections were ineligible. The study was approved by the Institutional Review Boards of the participating institutions and all patients signed informed consent.

Treatment Plan

Patients received paclitaxel 225 mg/m² by intravenous infusion over three hours followed by carboplatin dosed to a targeted area under the concentration-time curve (area under the curve) [AUC] = 6 mg/mL · minute over 30 minutes every 3 weeks. The dose of carboplatin was calculated by the Calvert formula: mg = targeted AUC × (GFR + 25).5 The GFR was assumed to be the estimated creatinine clearance as calculated by the Jelliffe formula.13

• 

The estimated creatinine clearance was recalculated at each cycle. Actual body weight was utilized in calculating doses. Patients received dexamethasone 20 mg orally 14 and 7 hours prior to treatment; in addition, dexamethasone 20 mg, diphenyhydramine 25 mg, and cimetidine 300 mg (or equivalent) were administered intravenously 60 minutes prior to treatment. Patients were eligible to continue therapy to a maximum of six cycles if stable or responding disease was shown and toxicity was acceptable.

Dose Modification

Toxicity was graded based upon the National Cancer Institute Common Toxicity Criteria. Dose modifications were planned based upon hematologic and nonhematologic toxicity. Cycles were administered if day of treatment absolute granulocyte count was ≥ 1800/mm³ and platelet count ≥ 100,000/mm³. If counts were not adequate, therapy was delayed for up to three weeks, after which the patient would go off study. During Cycle 1, granulocyte-colony-stimulating factor (G-CSF) was not utilized if patients had received no prior systemic chemotherapy. For patients with prior adjuvant/neoadjuvant treatment, or patients who developed granulocytopenic fever, G-CSF 5 μg/kg was administered subcutaneously starting 24 hours after chemotherapy and continuing daily until the absolute granulocyte count was ≥ 10,000/mm³. If febrile granulocytopenia recurred, the doses of drugs were decreased by 25%. Any patient experiencing bleeding with a platelet count ≤ 40,000/mm³ or ≤ 20,000/mm³ without bleeding had doses of paclitaxel and carboplatin reduced by 25%. For Grade 3–4 mucositis or diarrhea, therapy was held until resolution of symptoms, and then doses were reduced by 25%. For Grade 3 neuropathy, patients had chemotherapy held until recovery to ≤ Grade 2; subsequent cycles were reduced by 50%. For any other Grade 3 or 4 nonhematologic toxicity, such as neuropathy or myalgia, treatment was held until patients recovered completely or to Grade 1 status, then resumed at 50% paclitaxel and carboplatin doses.

Laboratory and Radiographic Data

Before study entry, all patients underwent complete history and physical examination. Patients' height, weight, and ECOG performance status were recorded. Pretreatment complete blood count, serum creatinine, AST, bilirubin, and urinalysis were performed. All patients underwent a pretreatment electrocardiogram. A chest radiograph was performed, and abdominopelvic computerized tomography or magnetic resonance imaging was utilized as indicated to document measurable disease. Patients underwent complete blood counts weekly, and serum chemistries were repeated at the start of each treatment cycle. Tumor measurements were scheduled to be repeated every two cycles of treatment.

Response Criteria

Responses were defined using ECOG Solid Tumor Response Criteria. Complete response was defined as the complete disappearance of all clinically detectable disease for a period of at least four weeks. Partial response required a ≥ 50% decrease in tumor area (multiplication of longest diameter by the greatest perpendicular diameter) or a ≥ 50% decrease in the sum of the products of the perpendicular diameters of multiple lesions in the same organ for at least four weeks. Patients with a partial response could not have an increase in the size of any area of known malignant disease of ≥ 25% or the appearance of any new areas of metastatic disease.

Statistical Design

The study was designed to proceed in two stages. The design provided a 95% probability of rejecting the treatment if the true response rate was ≤ 10%, and ≤ 10% probability of rejecting the regimen if the true response rate was 30% or more. Planned accrual was for 35 evaluable patients. If 3 responses were observed among the initial 18 evaluable patients, the study would proceed to the second stage of accrual.

RESULTS

Between October 1996 and December 1999, 42 patients were registered. Four patients were determined to be ineligible; three of these patients received treatment. Reasons for ineligibility include prior systemic chemotherapy, incorrect histologic diagnosis, history of a cardiac arrythmia requiring medication, and normal serum creatinine. Additionally, one patient withdrew prior to receiving treatment. Pretreatment characteristics of the 37 eligible treated patients are shown in Table 1. The median age of patients treated was 70 years (range, 34–82). Thirty five were males and two were females. The median serum creatinine was 1.7 mg/dL (range, 1.5–3.0). The median creatinine clearance calculated by the Jeliffe formula13 was 35.2 mL/min (range, 21.5–58.6). One patient had previously received systemic adjuvant chemotherapy. Thirty-five patients had transitional-cell histology; two had nontransitional histology. Poor prognostic risk factors14 were defined as ECOG performance status > 1 or lung, liver, or bone metastases. Fifteen (35%) patients had zero risk factors; 19 patients (51%) had one risk factor; and 3 (8%) patients had two risk factors.

Table 1. Patient Characteristics (37 Eligible Patients)

Characteristic	No. of patients	Percent
ECOG: Eastern Cooperative Oncology Group.
Age in years
Median	70
Range	34–82
Gender
Male	35	95
Female	2	5
Serum creatinine (mg/dL)
Median	1.7
Range	1.5–3.0
Calculated creatinine clearance (mL/minute)
Median	35.2
Range	21.5–58.6
ECOG performance status
0	13	35
1	20	54
2	4	11
Prior chemotherapy
None	36	97
Adjuvant	1	3
Histology
Transitional cell	35	95
Nontransitional histology	2	5
Sites of metastases
Node only	10	27
Liver	9	24
Bone	5	14
Lung	17	46
Number of poor prognosis risk factors
0	15	41
1	19	51
2	3	8

The median number of cycles of treatment received was four (range, one to six). Five patients were unevaluable for response assessment. Four patients were unevaluable due to early death during Cycle 1 of treatment, and one patient had inadequate baseline imaging. Three complete responses (CR) were shown (CR rate 8.1%; 95% CI, 1.7–21.9%) and six partial responses (PR) were shown. Among the responders, five patients had node only disease. Six responders had no poor prognosis risk factors, and three had one risk factor. The overall response rate was 24.3% (95% CI, 11.9–41.7%). Median response duration was 6.6 months for patients attaining CR and 4.2 months for patients with PR. When stratified by serum creatinine (< 1.7 mg/dL versus ≥ 1.8 mg/dL) or by calculated creatinine clearance (< 35 mL/minute versus ≥ 35 mL/minute), no statistically significant differences in response rate were shown.

The median progression free survival was 3.0 months. The median overall survival was 7.1 months. When stratified for the number of poor prognostic risk factors, the median survival for patients with zero risk factors was 15.6 months, for patients with one risk factor 5.2 months, and for patients with two risk factors 1.7 months (log rank test P = 0.009; Fig. 1). When stratified by serum creatinine (< 1.7 mg/dL versus ≥ 1.8 mg/dL) or by calculated creatinine clearance (< 35 mL/minute versus ≥ 35 mL/minute), no statistically significant differences in survival were shown.

Figure 1. Overall survival stratified by the number of poor prognostic risk factors Log rank test P = 0.009.

Four patients experienced early death during Cycle 1. One patient died of granulocytopenic sepsis. One patient died of a gastrointestinal tract bleed (without thrombocytopenia). One patient died during sleep of unknown cause. One patient died of a pulmonary embolus. Toxicity data for all 40 treated patients (37 eligible and 3 ineligible) are summarized in Table 2. Eighteen patients (45%) had Grade 4 granulocytopenia. One patient had Grade 4 thrombocytopenia (1%). The most common Grade 3 toxicities were neurotoxicity (n = 14), granulocytopenia (n = 6), anemia (n = 6), and metabolic disorder (n = 6). When stratified by serum creatinine (< 1.7 mg/dL versus ≥ 1.8 mg/dL) or by calculated creatinine clearance (< 35 mL/minute versus ≥ 35 mL/minute), no statistically significant differences in worst grade toxicity were shown. As well, no significant differences in worst grade toxicity were shown when patients were stratified by the number of poor prognostic risk factors.

Table 2. Summary of All Toxicities across Multiple Cycles (40 Treated Patients)

Toxicity	Grade
	1	2	3	4	5
Granulocytopenia	2	5	6	18	0
Thrombocytopenia	17	2	0	1	0
Anemia	10	7	6	1	0
Hemorrhage	4	1	0	0	1
Infection	0	12	3	0	1
Cardiac	3	0	2	0	1
Pulmonary	4	6	2	0	1
Liver	10	3	0	0	0
Skin	5	2	1	0	0
Nausea	9	7	4	0	0
Vomiting	6	5	2	0	0
Diarrhea	6	2	1	0	0
Stomatitis	9	1	0	0	0
Neurologic	12	11	14	0	0
Metabolic	8	4	6	0	0

DISCUSSION

Carcinoma of the urothelium is a chemotherapy-sensitive neoplasm. Combination chemotherapy regimens have been developed for this disease site, including the cisplatin based regimens M-VAC and gemcitabine/cisplatin.2, 3 However, patients with urothelial carcinoma may have age- and disease-related abnormalities in renal function, making cisplatin-based regimens difficult to utilize. Limited data exist concerning the natural history and prognosis of patients with advanced urothelial carcinoma in the setting of renal dysfunction. Similarly, many chemotherapy studies have historically excluded this patient cohort, resulting in limited evidence for clinical decision making concerning the optimal approach to chemotherapy in this population. To the best of our knowledge, this is the first Phase II cooperative group chemotherapy trial specifically designed for this patient population.

Several previous Phase II trials have reported results of the paclitaxel/carboplatin regimen in advanced urothelial carcinoma. Although many patients in these trials had preserved renal function, patients with moderately impaired renal function were also eligible. Redman et al. completed a Phase II trial of paclitaxel 200 mg/m² plus carboplatin AUC 5 every 21 days. This trial enrolled patients with serum creatinine ≤ 2 mg/dL or creatinine clearance ≥ 40 mL/minute. A response rate of 51.5% was reported with a survival of 9.5 months.10 Pycha reported a Phase II trial of paclitaxel 175 mg/m² plus carboplatin AUC 5 every 21 days. A 72% response rate and median survival > 13 months were reported.11 Small et al. reported the Southwest Oncology Group (SWOG) trial of paclitaxel 200 mg/m² plus carboplatin AUC 5 every 21 days. In this trial, a response rate of only 21% was seen, with median survival of 9 months.12 Both studies by Pycha and SWOG required a creatinine clearance ≥ 30 mL/minutes for enrollment. The authors of the SWOG trial have suggested that differences in response rates in these trials may be linked to different patient characteristics, for example, the relative balance of patients with node only metastases versus visceral metastatic disease. In addition, the authors point out that survival in the SWOG trial was similar to that shown in other Phase II trials and suggested that survival rather than response rate be utilized to assess efficacy of chemotherapy regimens for bladder carcinoma.

The current response rate was similar to that reported in the SWOG trial,12 but our median survival of 7.2 months is approximately 2 months less than that seen in other trials of paclitaxel/carboplatin.9, 10, 12 Poor prognostic risk factors in advanced urothelial carcinoma have been previously identified. In a recent study by Bajorin et al. of 203 patients with advanced bladder carcinoma treated with M-VAC, the presence of lung, liver, or bone metastases and a baseline Karnofsky performance status less than 80% predicted independently for a poor outcome.14 Patients that had both of these risk factors had a five year survival of 0% (median survival, 9.3 months). However, if patients had neither risk factor, the five year survival was 33% (median survival, 33 months). This study did not include patients with abnormal renal function, such as those treated in the current trial. When our patient population was stratified by the number of poor prognostic risk factors (0, 1, or 2), a clear survival difference was shown between the groups. The current trial confirms the prognostic model of Bajorin but is the first trial that applies the model to patients with abnormal renal function. However, the median survivals shown for each prognostic group in the current trial are inferior to those reported by Bajorin for patients treated with M-VAC.14 This could be related to the relative effectiveness of the regimens. However, the worse survival in the current trial raises the question of whether renal dysfunction itself may be a poor prognostic risk factor in patients with advanced urothelial carcinoma. Arguing against this hypothesis is the finding that response rate and survival in the current study were not significantly different when patients were stratified based upon renal function. Prospective validation of renal dysfunction as a poor prognostic risk factor is problematic given that these patients are frequently excluded from chemotherapy trials.

Further development of chemotherapy for this patient population is warranted. The ECOG has initiated a followup study of paclitaxel 100 mg/m² plus gemcitabine 600–800 mg/m² on Days 1, 8, and 15 every 28 days. The results of the current trial will be the reference for future trials in patients with advanced urothelial carcinoma and renal dysfunction.