- © 2008 by American Society of Clinical Oncology
Patient-Reported Outcomes of Patients With Advanced Biliary Tract Cancers Receiving Gemcitabine Plus Capecitabine: A Multicenter, Phase II Trial of the Swiss Group for Clinical Cancer Research
- Dieter Koeberle,
- Piercarlo Saletti,
- Markus Borner,
- Daniela Gerber,
- Daniel Dietrich,
- Clemens B. Caspar,
- Walter Mingrone,
- Kurt Beretta,
- Florian Strasser,
- Thomas Ruhstaller,
- Oreste Mora and
- Richard Herrmann
- From the Kantonsspital, St Gallen; Istituto Oncologico della Svizzera Italiana, Bellinzona; Inselspital; Quality of Life Unit and Statistical Unit, Swiss Group for Clinical Cancer Research Coordinating Center, Bern; Kantonsspital, Baden; Kantonsspital, Aarau; Rheinfelden; and University Hospital, Basel, Switzerland
- Corresponding author: Dieter Koeberle, MD, Department of Internal Medicine, Division Oncology/Hematology, Kantonsspital St. Gallen, CH-9007 St. Gallen, Switzerland; e-mail: dieter.koeberle{at}kssg.ch
Abstract
Purpose To evaluate the effects of palliative chemotherapy with gemcitabine plus capecitabine (GemCap) on patient-reported outcomes measured using clinical benefit response (CBR) and quality-of-life (QOL) measures in patients with advanced biliary tract cancer.
Patients and Methods Patients had to manifest symptoms of advanced biliary tract cancer and have at least one of the following: impaired Karnofsky performance score (60 to 80), average analgesic consumption ≥ 10 mg of morphine equivalents per day, and average pain intensity score of ≥ 20 mm out of 100 mm. Treatment consisted of oral capecitabine 650 mg/m2 twice daily on days 1 through 14 plus gemcitabine 1,000 mg/m2 as a 30-minute infusion on days 1 and 8 every 3 weeks until progression. The primary end point was the number of patients categorized as having a CBR or stable CBR (SCBR) during the first three treatment cycles.
Results Forty-four patients were enrolled (bile duct cancer, n = 36; gallbladder cancers, n = 8). The main grade 3 or 4 adverse events included hematologic toxicity and fatigue. After three cycles, 36% of patients achieved a CBR, and 34% achieved an SCBR. Over the full course of treatment, 57% of patients achieved a CBR, and 18% achieved an SCBR. Improved QOL was observed in patients with a CBR or SCBR. The objective response rate was 25%. Median time to progression and overall survival times were 7.2 months and 13.2 months, respectively.
Conclusion Chemotherapy with GemCap is well tolerated and effective and leads to a high CBR rate. Patient-reported outcomes are useful for evaluating the effects of palliative chemotherapy in patients with biliary tract cancer.
INTRODUCTION
Biliary tract carcinomas are relatively rare tumors with a dismal prognosis.1 They are usually seen in middle-aged or elderly people, with most patients aged between 50 and 70 years. The role of nonsurgical oncologic treatment remains a matter of debate. Despite widespread use of palliative chemotherapy in patients with these tumors, there is no standard therapy for advanced disease. Patients who receive palliative chemotherapy are usually treated with single-agent gemcitabine2,3 or with combination regimens including mitomycin,4,5 a platinum compound,6-9 or fluoropyrimidines.10,11 Response rates with these treatments range from 10% to 35%, and median overall survival (OS) time varies between 5 and 12 months.12 However, the data are difficult to interpret because of the relatively small number of published trials and the considerable variability in the clinical characteristics of enrolled patients.
Patients with advanced biliary tract carcinoma suffer from a variety of symptoms, including pain, anorexia/cachexia, GI symptoms, and psychological distress. Prospective data on patient-reported outcomes in this tumor type are scarce.13 Likewise, the available literature on quality of life (QOL) in patients with biliary tract cancers is limited,14 and there is a lack of valid and reliable assessment instruments specifically designed for this tumor type.
In pancreatic cancer, clinical benefit response (CBR; a composite end point comprising assessment of pain, performance status, and weight) was created to assess the impact of therapy on tumor-related symptoms.15,16 Gemcitabine was the first chemotherapeutic agent approved for the treatment of pancreatic cancer on the basis of CBR. In the present trial, we applied the same concept to patients with biliary tract cancer. Similarities between pancreatic and biliary tract cancer can be assumed regarding disease-related symptoms used to define CBR. There is evidence from a randomized phase III trial for advanced pancreatic and biliary cancer17 that chemotherapy affects both cancers and patient QOL in similar ways.
The primary objective of our phase II trial was to evaluate the ability of palliative chemotherapy to maintain or improve tumor-related symptoms as measured by CBR in patients with advanced biliary tract cancer. We selected the combination of gemcitabine and capecitabine (GemCap) as our preferred chemotherapy regimen based on its promising antitumor activity and mild toxicity profile as demonstrated in clinical studies of patients with biliary tract10,11 and pancreatic cancer.18-20
PATIENTS AND METHODS
Eligibility Criteria
Patients eligible for the study had histologically or cytologically confirmed, unresectable, locally advanced or metastatic biliary tract cancer. Measurable disease was not required. No prior chemotherapy for advanced disease was allowed. Patients were required to be symptomatic and have at least one of the following characteristics at baseline: Karnofsky performance score (KPS) between 60 and 80, average analgesic consumption of ≥ 10 mg of morphine equivalents per day, and average pain intensity score of ≥ 20 mm out of a maximum of 100 mm based on a linear analog self-assessment (LASA) scale. Other eligibility criteria included age ≥ 18 years and adequate bone marrow, hepatic, and renal function indicated by an absolute neutrophil count ≥ 1,500/μL, platelets ≥ 100,000/μL, total bilirubin ≤ 4× the upper limit of normal, serum ALT ≤ 5× the upper limit of normal, and calculated creatinine clearance ≥ 50 mL/min.
Exclusion criteria included known CNS metastases; a history of other primary malignancy within 5 years, except for nonmelanomatous skin cancer or adequately treated in situ cervical cancer; any psychiatric disorder, cognitive dysfunction, or language problems that would prevent the patient from completing trial questionnaires or diaries; or any serious underlying medical condition that could impair the ability of the patient to participate in the trial or to take oral medication. The protocol was approved by the local ethics review boards of all participating institutions, and all patients gave written informed consent before enrollment.
Treatment Plan
Treatment consisted of oral capecitabine 650 mg/m2 twice daily on days 1 through 14 and gemcitabine 1,000 mg/m2 intravenously over 30 minutes on days 1 and 8 every 3 weeks, as described by Hess et al.18 Treatment was continued for a maximum of eight cycles or until disease progression, unacceptable toxicity, or withdrawal of patient consent.
Adverse events were recorded according to the National Cancer Institute Common Toxicity Criteria (version 2.0). The occurrence of grade 3 or 4 hematologic adverse events (except anemia) or grade 2, 3, or 4 nonhematologic adverse events necessitated delay or interruption of treatment until toxicities had resolved to grade 0 or 1. The gemcitabine dose was reduced by 25% for grade 3 or 4 hematologic toxicities at the discretion of the investigator, and the doses of both drugs were reduced by 25% for grade 3 and by 50% for grade 4 nonhematologic toxicities. Treatment cycles could be delayed by up to 3 weeks.
The best possible palliative and supportive care for disease-related symptoms was offered to all patients. Hematopoietic growth factors were used to treat symptomatic neutropenia but were not administered prophylactically. All patients underwent a pain stabilization period for at least 4 days before registration. Analgesics were adjusted so that patients received an opioid in a fixed regimen that aimed to provide adequate pain control with less than 30% changes in the daily opioid dose.
Assessments
CBR.
For the assessment of CBR, we applied the definitions of Burris et al,16 which required predefined improvements in pain (ie, intensity or analgesic consumption), KPS, or weight for at least 4 consecutive weeks, without worsening of any of these symptoms (Fig 1). Pain and KPS were the primary criteria, and weight gain was a secondary criterion. A new category of stable CBR (SCBR) was introduced in this trial to extend the CBR criteria to the subset of patients with stable responses in all measurements (Fig 1). As a consequence, new definitions for positive, negative, and stable weight change were made.
Summary of assessment of clinical benefit response.
A weekly patient diary card, including the LASA indicator for pain intensity from the Memorial Pain Assessment Card21 and information on pain treatment, was completed by each patient. The cards were completed for at least 4 days before registration as a baseline assessment and then daily for 24 consecutive weeks (eight cycles) after study entry. KPS was determined weekly by the treating oncologist during the consultation. Duration of CBR was defined as the largest number of consecutive weeks during which there was a positive change in at least one parameter (if multiple parameters were positive) or as the duration of positive change (if only one parameter was positive). Time to CBR was measured from baseline to the first onset of an improvement in at least one of the CBR parameters.
QOL indicators.
Patients completed QOL assessments before the oncologists' consultations at baseline, weekly during cycles 1 to 3, and on days 1 and 8 of cycle 4 and subsequent cycles. The QOL form consisted of 10 single-item LASA indicators. We used global LASA indicators for physical well-being,22 mood,22,23 coping effort,22,24 overall treatment burden,25 and functional performance,26 which are sensitive to the wide spectrum of reactions seen in patients on chemotherapy. The indicators for physical well-being, mood, coping effort, and functional performance are sensitive to tumor response in metastatic colorectal cancer,27 and the scales for mood and coping effort are also sensitive to mood disorders and psychosocial dysfunction.28 Overall treatment burden showed a discriminating capacity regarding chemotherapy adverse effects in patients with advanced gastric carcinoma.29 In addition, symptom-specific LASA indicators for pain, tiredness, itching, appetite, and nausea/vomiting were included.
All QOL LASA indicators had a scale range of 0 to 100, with higher values indicating a better condition for all indicators. QOL data were analyzed descriptively; means and SEs of the untransformed data were calculated. We compared the QOL assessment of each assessment time point with the baseline QOL assessment and considered a mean change from baseline of 8 points or more in a LASA scale as clinically meaningful.30
Other measures of efficacy.
Tumor response was assessed using the Response Evaluation Criteria in Solid Tumors,31 based on computed tomography scans performed at baseline and every 9 weeks during treatment. Objective responses (complete or partial) were confirmed after a minimum of 4 weeks. Time to disease progression (TTP) was calculated from study entry to disease progression, and OS was calculated from study entry to death.
Statistics
The sample size was estimated for the proportion of patients categorized as CBR or SCBR after three cycles of trial treatment (primary end point). Simon's optimal two-stage phase II design was used.32 The treatment was considered of no further interest if the proportion of patients with CBR and SCBR was ≤ 15% and as promising if the proportion was ≥ 35%. For a significance level of 5% and a power of 90%, a total of 44 patients were needed, with 19 patients in stage I and a further 25 patients in stage II. Descriptive statistics and frequency tables were performed. TTP and OS were estimated using the Kaplan-Meier method. All patients were included in the analyses for efficacy end points. Only patients who received at least one dose of study medication were included in the analyses for safety end points. Statistical analyses were performed using SAS 9.1 (SAS Institute, Cary, NC) and S-Plus 7.0 (Insightful Corporation, Seattle, WA).
RESULTS
Patient Characteristics
A total of 44 patients were enrolled from six centers of the Swiss Group for Clinical Cancer Research in Switzerland between May 2003 and June 2006. All patients were eligible and assessable. Table 1 lists the demographic data and baseline disease and pretreatment characteristics for all patients.
Baseline Patient Characteristics
Treatment Administration
Patients received a total of 266 cycles of GemCap, with a median of eight cycles (range, one to eight cycles) administered per patient. The median dose for capecitabine was 631 mg/m2 (range, 102 to 838 mg/m2) twice daily, and the median dose for gemcitabine was 978 mg/m2 (range, 412 to 1,042 mg/m2). Dose reductions of capecitabine and gemcitabine were necessary in 6% and 9% of all cycles, respectively.
Safety
The incidence of toxicity per patient for events of any grade and for grade 3 or 4 events is listed in Table 2. Chemotherapy with GemCap was generally well tolerated. The most common adverse events were anemia, leukopenia, and fatigue. Most events were grade 1 or 2 in severity. Grade 3 or 4 events generally occurred in ≤ 7% of patients, with the exception of leukopenia (11%) and fatigue (11%). No patient experienced neutropenic fever, and there were no treatment-related deaths.
Major Toxicities (N = 44)
Efficacy
CBR evaluation.
All 44 patients completed baseline evaluations and were assessable for CBR evaluation. While on treatment, 771 (95%) of 812 expected diaries were completed.
After three cycles, 16 patients (36%; 95% CI, 22% to 52%) achieved a CBR, and an additional 15 patients (34%) had an SCBR. Over the full course of treatment, 25 patients (57%; 95% CI, 41% to 75%) achieved a CBR, and eight further patients (18%) had an SCBR. The median time to CBR was 4 weeks (range, 1 to 14 weeks), and the median duration of CBR was 17 weeks (range, 4 to 25 weeks). Of the 25 patients achieving a CBR, 13 had a positive change in KPS, six had a positive change in KPS and pain intensity, three had a positive change in pain intensity while receiving a stable dose of analgesics, and one had a positive change in KPS and decreased need for analgesics. Two patients achieved a CBR based on an improvement in weight.
QOL.
QOL form completion rates (ie, completed v expected forms) were 100% at baseline and 92% (560 of 607 forms) during treatment, with 90% of forms (300 of 332 forms) completed during cycles 1 to 3. The QOL scores at baseline were in the upper half of the scale, with the worst values recorded for coping effort (61) and tiredness (65) and the best values recorded for nausea/vomiting (88) and itching (93). Scores for itching and nausea/vomiting remained consistently high during the whole assessment period. For all other QOL indicators, clinically meaningful improvements compared with baseline (ie, 8-point increase) were reported. These improvements generally started during cycle 3 and were maintained thereafter.
Patients with a treatment duration of ≤ three cycles reported worse QOL scores for physical well-being, mood, pain, tiredness, appetite, and functional performance at baseline and during the first two cycles, with a tendency to improve thereafter compared with patients receiving more than three cycles (Fig 2). Patients receiving more than three cycles of treatment reported stable QOL scores over time (Fig 2).
Patient-rated quality-of-life linear analog self-assessment indicators by treatment duration (≤ three cycles [yellow curves] v > three cycles [blue curves]). Data shown are mean values over time (range, 0 to 100), with higher values indicating a better condition.
There was no difference between patients with a CBR and those with an SCBR with regard to QOL. Between responders (CBR and SCBR) and nonresponders, differences were observed in physical well-being, mood, appetite, coping effort, and functional performance, with the worse QOL levels at baseline and during the first three cycles for nonresponders.
Other measures of efficacy.
Four patients had assessable, but not measurable, disease at baseline. One patient (2%) achieved a complete response, and 10 patients (23%) achieved partial responses, for an overall tumor response rate of 25% (95% CI, 13% to 40%). There were 24 patients (55%) with stable disease lasting for ≥ 8 weeks.
Median TTP was 7.2 months (95% CI, 4.5 to 8.1 months; Fig 3A), and median OS time was 13.2 months (95% CI, 8.9 to 17.1 months; Fig 3B). The median follow-up time in the seven patients who remain alive is 19 months.
Kaplan-Meier curve (yellow) and 95% CIs (blue) for (A) time to progression and (B) overall survival.
In a post hoc analysis, we observed a correlation between CBR and objective tumor response (Table 3). Of the 11 patients who achieved an objective tumor response, seven were rated as CBR, three were rated as SCBR, and one was rated as not assessable. Only one patient rated as a non–clinical benefit responder (NCBR) achieved an objective tumor response. Kaplan-Meier curves for OS were plotted to compare patients with a CBR, SCBR, or NCBR in the first three cycles. Although this analysis was not planned before the trial, it was included to explore the association between both measures. In this exploratory analysis, patients with a CBR and an SCBR had a similar median OS (14.2 months; 95% CI, 8.9 to 22.5 months) compared with NCBR patients (10.5 months; 95% CI, 2.6 to 22.6 months).
Objective Tumor Response and Clinical Benefit
DISCUSSION
In the treatment of advanced cancers, symptom palliation is of paramount importance and is considered to be the main determinant for patients' QOL. Therefore, QOL issues need to be balanced against treatment efficacy. Clinical benefit, in terms of alleviation of cancer-related symptoms, has been a key criterion in introducing and approving gemcitabine for advanced pancreatic cancer.15,16,33 Chemotherapy-induced palliative effects on disease-related symptoms have been documented by prospectively defining the criteria for clinical benefit, which consist of pain severity, performance status, and weight. Patients are considered to have a CBR only when they experience a marked and sustained improvement in any of these parameters without worsening of any of the others. In two studies, treatment with gemcitabine resulted in CBR rates of 24% and 27% in patients with advanced pancreatic cancer.15,16 The CBR criteria were also applied in a phase II trial evaluating the palliative effect of capecitabine in advanced pancreatic cancer, with a CBR rate of 24%.34
In the present trial, a new category of SCBR was used to extend the CBR criteria to the subset of patients with stable responses in all measurements. This concept takes into account the fact that symptom stabilization can be an important effect of palliative chemotherapy. Results of this trial demonstrated that a high rate (75%) of CBR or SCBR is achieved with chemotherapy in patients with advanced biliary tract cancer. The CBR rate (57%) was substantially higher than in pancreatic cancer patients. This difference can be explained, at least in part, by the higher chemotherapy sensitivity of biliary tract cancer. In our trial, the objective response rate of 25%, with an additional 55% of patients achieving stable disease, also contrasts with the inferior results achieved generally in pancreatic cancer.16,20
Patients with a CBR or SCBR showed similar results with regard to tumor response and QOL. The benefit of symptom control was reflected by an improvement in QOL in patients with a CBR or SCBR. Our QOL results suggest that poorer QOL levels at baseline and during the first three treatment cycles may be associated with an NCBR or early treatment discontinuation. In addition, we found preliminary evidence to suggest that patients achieving either a CBR or an SCBR tend to have improved tumor response compared with patients with an NCBR. Biliary tract cancer patients with progressive and nonresponsive tumor symptoms under palliative first-line chemotherapy thus deserve best available supportive treatment and early evaluation of new treatment choices.
CBR has become a standard measure for pancreatic cancer patients, although it has been criticized35 (eg, for not taking a wide enough range of QOL domains into account). There is further evidence suggesting that CBR overestimates the beneficial effects of chemotherapy in certain patients and underestimates them in others.36 Overestimation may result if adverse effects of chemotherapy are not appropriately taken into account, whereas failure to consider the clinical relevance of alterations in other symptoms may result in underestimation. Another potential reason for overestimation of CBR in this trial is based on the assessment of KPS, which was rated by the treating oncologist and not by two independent observers, as was done by Burris et al.16 Despite the methodologic complexity of CBR assessment, the paucity of studies presenting patient-reported outcomes in patients with biliary tract cancer, and the potential for heterogeneity in a phase II setting, we must be cautious in making definitive conclusions regarding the magnitude of symptom alleviation found in our trial.
Chemotherapy with GemCap was generally well tolerated and effective, with an objective response rate of 25% and a stable disease rate of 55%, yielding a disease control rate of 80%. The median TTP of 7.2 months and median OS of 13.2 months compare favorably with results of other trials.12 More importantly, we were able to replicate and thus confirm the findings from other investigators that GemCap may benefit patients with advanced biliary tract cancer.10,11
The GemCap regimen warrants further investigation and will be compared with single-agent gemcitabine in a randomized phase III trial conducted by the National Cancer Institute of Canada's Clinical Trial Group in cooperation with the Swiss Group for Clinical Cancer Research. This and other phase III trials will help to define a future reference treatment in biliary tract cancer. The development of clinical trials that include assessments of symptom control and QOL is the main way to improve the care of biliary tract cancer patients.
AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST
The author(s) indicated no potential conflicts of interest.
AUTHOR CONTRIBUTIONS
Conception and design: Dieter Koeberle, Daniel Dietrich, Florian Strasser, Richard Herrmann
Administrative support: Dieter Koeberle, Daniela Gerber, Daniel Dietrich, Richard Herrmann
Provision of study materials or patients: Dieter Koeberle, Piercarlo Saletti, Markus Borner, Clemens B. Caspar, Walter Mingrone, Kurt Beretta, Florian Strasser, Thomas Ruhstaller, Oreste Mora, Richard Herrmann
Collection and assembly of data: Dieter Koeberle, Markus Borner, Daniela Gerber, Daniel Dietrich, Richard Herrmann
Data analysis and interpretation: Dieter Koeberle, Piercarlo Saletti, Daniela Gerber, Daniel Dietrich, Florian Strasser, Richard Herrmann
Manuscript writing: Dieter Koeberle, Piercarlo Saletti, Daniela Gerber, Daniel Dietrich, Richard Herrmann
Final approval of manuscript: Dieter Koeberle, Piercarlo Saletti, Markus Borner, Daniela Gerber, Daniel Dietrich, Clemens B. Caspar, Walter Mingrone, Kurt Beretta, Florian Strasser, Thomas Ruhstaller, Oreste Mora, Richard Herrmann
Acknowledgments
We thank Yvonne Wechsler, PhD, for her contribution to the development of the study design, Doris Lanz for central data management and trial coordination, and Lee Miller for proofreading this article.
Footnotes
-
Supported by Roche Pharma Switzerland, Eli Lilly Switzerland, and the Swiss federal government.
Presented in part at the 14th Annual European Cancer Conference, September 23-27, 2007, Barcelona, Spain.
Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
Clinical Trials repository link available on www.JCO.org.
- Received January 31, 2008.
- Accepted April 23, 2008.
