The Oncologist Symptom Management and Supportive Care Health-Related Quality of Life Impact of Bevacizumab When Combined with Irinotecan, 5-Fluorouracil, and Leucovorin or 5-Fluorouracil and Leucovorin for Metastatic Colorectal Cancer FAIROOZ F. KABBINAVAR, a JOEL F. WALLACE, b ERIC HOLMGREN, b JING YI, b DAVID CELLA, c KATHLEEN J. YOST, c HERBERT I. HURWITZ d a University of California at Los Angeles (UCLA) School of Medicine, Los Angeles, California, USA; b Genentech, Inc., South San Francisco, California, USA; c Center on Outcomes, Research and Education, Evanston Northwestern Healthcare Research Institute, Evanston, Illinois, USA; d Duke University, Durham, North Carolina, USA Key Words. Colorectal cancer Quality of life Fluorouracil Monoclonal antibodies Bevacizumab Angiogenesis inhibitors Disclosure: Employment/leadership position: Joel F. Wallace, Genentech; Eric Holmgren, Genentech; Jing Yi, Genentech; Intellectual property rights/inventor: None; Consultant/advisory role: David Cella, Genentech; Herbert I. Hurwitz, Genentech, Roche, Sanofi-Aventis, Bristol-Myers Squibb, Exelixis; Honoraria: None; Research funding: The two bevacizumab clinical trials described in this report were sponsored by Genentech, Inc., South San Francisco, CA; analysis support was provided by Genentech, Inc., South San Francisco, CA. Ownership interest: Jing Yi, Genentech; Eric Holmgren, Genentech; Joel F. Wallace, Genentech; Expert testimony: None; Other: None. The content of this article has been reviewed by independent peer reviewers to ensure that it is balanced, objective, and free from commercial bias. No financial relationships relevant to the content of this article have been disclosed by the authors, planners, independent peer reviewers, or staff managers. LEARNING OBJECTIVES After completing this course, the reader should be able to: 1. Assess benefit as measured through validated health-related quality of life instruments as an important component of overall therapeutic impact. 2. Add bevacizumab to a 5-fluorouracil based chemotherapy regimen for the treatment of metastatic colorectal cancer to lengthen the time to disease progression and overall survival time without compromising patients health-related quality of life. 3. Justify why the addition of an effective therapeutic agent to an existing regimen with no worsening of healthrelated quality of life is an additional benefit. CME This article is available for continuing medical education credit at CME.TheOncologist.com. ABSTRACT Purpose. To compare the time to deterioration in healthrelated quality of life (HRQoL) in patients with previously untreated metastatic colorectal cancer receiving a 5-fluorouracil (5-FU)-based chemotherapy regimen Correspondence: Fairooz Kabbinavar, M.D., Division of Hematology/Oncology, UCLA School of Medicine, 10945 LeConte Avenue, Suite 2333D, Los Angeles, California 90095-7059, USA. Telephone: 310-206-3921; Fax: 310-267-0151; e-mail: fkabbina@mednet. ucla.edu Received January 2, 2008; accepted for publication July 18, 2008. AlphaMed Press 1083-7159/2008/$30.00/0 doi: 10.1634/ theoncologist.2008-0003 The Oncologist 2008;13:000 000 www.theoncologist.com
2 HRQoL Impact of Bevacizumab Plus 5-FU Chemotherapy with or without the addition of bevacizumab (BV) in two randomized, placebo-controlled studies. Patients and Methods. Prespecified HRQoL endpoints in the phase II (Study 2192) and phase III (Study 2107) studies were time to deterioration in HRQoL, measured by the Functional Assessment of Cancer Therapy Colorectal (FACT-C) Colorectal Cancer Subscale (CCS), Trial Outcome Index (TOI-C), and FACT-C total score. Time to deterioration in HRQoL was evaluated for patients with baseline and postbaseline assessments, using the stratified log-rank test. Results. In the pivotal phase III trial, HRQoL baseline and postbaseline CCS scores were available for 127 patients receiving irinotecan, 5-FU, and leucovorin (LV) (IFL) and 122 patients receiving IFL plus INTRODUCTION Colorectal cancer (CRC) is the third most commonly diagnosed cancer in the U.S., and is the second leading cause of U.S. cancer-related deaths [1]. However, recently available therapies for CRC have provided significant improvements in survival. Bevacizumab (BV) (Avastin ; Genentech, Inc., South San Francisco, CA) is a recombinant, humanized monoclonal antibody directed against vascular endothelial growth factor. In a pivotal, placebo-controlled, phase III trial in patients with metastatic CRC (Genentech Study 2107) [2], adding BV to irinotecan, 5-fluorouracil (5-FU), and leucovorin (LV) (IFL) resulted in a significantly longer survival time (20.3 versus 15.6 months; hazard ratio [HR], 0.66; p.0001) and progression-free survival (PFS) time (10.6 versus 6.2 months; HR, 0.54; p.0001) than with IFL plus placebo. Grade 3 hypertension was more common during treatment with IFL plus BV than with IFL plus placebo (11% versus 2%). This trial supported the U.S. approval of BV in combination with 5-FU based chemotherapy as first-line treatment for metastatic CRC in 2004. In a placebo-controlled, phase II trial (Genentech Study 2192) [3], adding BV to 5-FU plus LV resulted in a significantly longer PFS time than with 5-FU and LV plus placebo in patients with metastatic CRC who were unsuitable candidates for first-line therapy with irinotecan (9.2 versus 5.5 months; HR, 0.50; p.0002). There was also a trend toward a longer survival time in patients receiving 5-FU, LV, and BV (16.6 versus 12.9 months; HR, 0.79; p.16). Similar to the phase III trial, grade 3 hypertension was more common during treatment with 5-FU/LV plus BV than with 5-FU and LV plus placebo (16% versus 3%) but was controlled with oral medication and did not cause study drug discontinuation. Full efficacy and safety data for Studies 2107 and 2192 have been previously reported [2, 3]. BV. The time to deterioration in HRQoL did not differ significantly between treatment groups as measured by the CCS, TOI-C, or FACT-C total score. In the phase II study, baseline and postbaseline CCS scores were available for 77 and 89 patients receiving 5-FU and LV and 5-FU and LV plus BV, respectively. In that study, the time to deterioration in HRQoL was similar between groups as measured by the CCS and TOI-C scores, but was significantly longer in the 5-FU and LV plus BV arm than in the 5-FU and LV plus placebo arm for the FACT-C total score. Conclusions. When added to 5-FU chemotherapy, BV significantly prolonged overall survival and progression-free survival without compromising HRQoL. The Oncologist 2008;13:000 000 Because patients with metastatic CRC are surviving longer, health-related quality of life (HRQoL) is an increasingly important outcome. It is commonly accepted that HRQoL refers to the extent to which one s usual or expected physical, emotional, and social well-being are affected by a disease or its treatment [4]. Several diseasespecific instruments, such as the Functional Assessment of Cancer Therapy Colorectal (FACT-C) [5, 6], have been developed to measure general HRQoL and concerns specific to patients with CRC. A secondary efficacy objective of Study 2107 and Study 2192 was the evaluation of changes in the HRQoL associated with multiple administrations of BV combined with 5-FU based chemotherapy in patients with metastatic CRC. This manuscript reports the results of that evaluation. MATERIALS AND METHODS The methods used in Study 2107 and Study 2192 have been described in detail elsewhere [2, 3], and are summarized briefly below. All patients provided written informed consent before receiving any study treatment. The studies were approved by the institutional review boards of all participating study sites and were conducted in compliance with the Declaration of Helsinki, the U.S. Food and Drug Administration Good Clinical Practice, and local ethical and legal requirements. Study Design and Patients Both studies were conducted in patients with histologically confirmed, bidimensionally measurable, previously untreated metastatic CRC. Patients assigned to BV-containing regimens received BV at a dose of 5 mg/kg by i.v. infusion once every 2 weeks. Patients on the IFL-containing regimens received their chemotherapy on the bolus schedule
Kabbinavar, Wallace, Holmgren et al. 3 described by Saltz et al. [7]; the 5-FU plus LV chemotherapy was administered in a weekly bolus schedule described by Petrelli et al. [8]. Study 2107 (Fig. 1A) was a randomized, placebo-controlled, multicenter, phase III trial in a total of 923 patients who had an Eastern Cooperative Oncology Group (ECOG) performance status score of 0 or 1, a life expectancy of at least 3 months, and adequate hematologic, liver, and kidney function [2]. Patients were randomly assigned to receive IFL plus placebo (n 411), IFL plus BV (n 402), or 5-FU and LV plus BV (n 110) until the occurrence of disease progression or unacceptable adverse events or for a maximum of 96 weeks. The FACT-C data were collected for the first 300 subjects enrolled across the three arms of the study. After approximately 100 patients per study arm had been treated, the independent data monitoring committee determined that IFL plus BV had an acceptable safety profile and enrollment to the 5-FU and LV plus BV treatment arm was therefore suspended. Only the two fully enrolled treatment arms (IFL plus placebo and IFL plus BV) are considered in the present paper. Study 2192 (Fig. 1B) was a randomized, multicenter, placebo-controlled, phase II trial conducted in 209 patients who were not considered by their treating physician to be optimal candidates for first-line irinotecan-based therapy, and who in addition were required to have at least one of the following features: advanced age ( 65 years old), low serum albumin level ( 3.5 g/dl), poor ECOG performance status (score of 1 or 2), or a history of abdominal/pelvic radiotherapy [3]. The patients were randomly assigned to receive 5-FU and LV plus placebo (n 105) or 5-FU and LV plus BV (n 104). The study treatment was continued until disease progression or for a maximum of 96 weeks. HRQoL Measures The FACT-C comprises four subscales measuring general HRQoL domains physical well-being (seven items), social/family well-being (seven items), emotional well-being (six items), and functional well-being (seven items) and a fifth subscale, the Colorectal Cancer Subscale (CCS), measuring seven concerns specific to patients with CRC (appetite, bowel control, stoma-related problems, weight loss, body image, and abdominal swelling or cramping). All FACT-C questions have five-category response options ranging from 0 not at all to 4 very much. Higher scores indicate better HRQoL [5, 6]. Three HRQoL measures were used in Study 2107 and Study 2192: the CCS, the Trial Outcome Index (TOI-C), which is the sum of the physical and functional well-being and the CCS, and the FACT-C total, which is the sum of four general subscales and the CCS. Each of these scales www.theoncologist.com Figure 1. Study schema. (A): Study schema of the phase III trial of BV plus IFL versus IFL plus placebo (Study 2107). (B): Study schema of the phase II trial of BV plus 5-FU and LV versus 5-FU and LV plus placebo (Study 2192). Abbreviations: 5-FU, 5-fluorouracil; BV, bevacizumab; IFL, irinotecan, 5-FU, and LV; LV, leucovorin. has demonstrated reliability, validity, and responsiveness to change [6]. Study Endpoints The primary HRQoL endpoint was the time to deterioration in HRQoL measured by the CCS score. Prespecified secondary HRQoL endpoints included the time to deterioration in HRQoL measured by the TOI-C and FACT-C total scores. HRQoL deterioration was determined prospectively based on clinically meaningful decreases in scores: 3 points for the CCS ( 3), 7 points for the TOI-C ( 7), and 9 points for the FACT-C total ( 9). The minimally important difference is defined as the smallest difference in score in the domain of interest that patients perceive as important, either beneficial or harmful, and that would lead the clinician to consider a change in the patient s management [9]. The minimally important difference estimates were developed based on distribution- and anchor-based methods, reflecting a growing consensus on the best approach for determining a minimally important difference [10]. Distribution- and anchor-based methods have also been used to estimate minimally important differences for scales and subscales from the Functional Assessment of Chronic Illness Therapy measurement system [5] for non-small cell lung cancer [11], fatigue or anemia [11, 12], and most recently breast cancer [13, 14] and biological response mod-
4 HRQoL Impact of Bevacizumab Plus 5-FU Chemotherapy ifiers [15]. Additional analyses for the purpose of refining the preliminary minimally important differences were completed after the prespecified analyses associated with the clinical studies discussed here and resulted in the recommended minimally important differences of 2 3 points for the CCS score, 4 6 points for the TOI-C score, and 5 8 points for the FACT-C total score [16]. Sensitivity analyses were conducted to assess the overall study conclusions in light of the one-point differences in recommended ranges for the TOI-C and FACT-C total scores and the prespecified endpoints. The FACT-C was self-administered at the beginning of study visits at baseline and during treatment, every 6 weeks for the first 24 weeks and then every 12 weeks. Assessments were completed prior to other study procedures, including tumor assessments, being performed on that day. Statistical Methods For analyses of overall survival and PFS, patient discontinuations and missing data were handled by treating the patient as a censored observation at the date he or she was last known to be alive. For objective response, patients with no postbaseline tumor evaluations were treated as nonresponders. Patients without a HRQoL assessment at screening were not included in any HRQoL analyses. HRQoL scores were not imputed for visits at which the questionnaire was not administered. Scores for scales with missing responses were determined as described in the instructions for computing the FACT-C scores [5]. If a patient progressed or died before HRQoL declined by the prespecified value, the time to deterioration in HRQoL for that patient was analyzed as being the time of death or disease progression. Patients who did not die or experience documented deterioration in HRQoL or disease progression/death were treated as censored observations at the time of the last HRQoL assessment. Patients who discontinued from the study without a postbaseline HRQoL assessment or disease progression were treated as censored observations on the date of randomization. The stratified log-rank test was used to compare the two treatment arms for time to deterioration in HRQoL, and the stratification factors were ECOG performance status score (0 or 1), number of metastatic disease sites (1 or 1), location of primary tumor (colon or rectum), and census region for randomization site in Study 2107 and ECOG performance status score (0 or 1), number of metastatic disease sites (1 or 1), and location of primary tumor (colon or rectum) in Study 2192. Kaplan Meier methodology was used to estimate the median time to deterioration in HRQoL. The Brookmeyer and Crowley method was used to calculate 95% confidence intervals (CIs) [17]. Estimation of the HR was determined using a stratified Cox model, and the same stratification factors mentioned above were used. An HR of one indicates that there is no difference between two arms in the hazard of deterioration in HRQoL during first-line therapy. An HR less than one indicates that the time to deterioration in HRQoL during first-line therapy is longer for patients in the BV arm than for patients in the controltreated arm. No adjustments were made for multiple testing. Two additional analyses were performed to evaluate the effect of adding BV to chemotherapy on HRQoL. The first analysis consisted of a comparison of the change from baseline to the last available assessment using the Wilcoxon rank-sum test (implemented using the Mantel-Haenszel test with modified ridit scores). The second analysis addressed the possible source of bias resulting from the difference between the treatment arms in patterns of patient discontinuation. This analysis consisted of a comparison between the two treatment arms of the last available assessment, which was stratified according to the time on study at the time of the last assessment (e.g., 6 weeks, 12 weeks, and 18 weeks), using the Mantel-Haenszel test with modified ridit scores. This analysis addressed the question of how HRQoL compared between BV- and control-treated patients who were on study for similar lengths of time. This type of evaluation is relevant when assessing whether a longer PFS time in the BV arm was accompanied by a worsening in HRQoL. RESULTS Demographics Demographic characteristics for the participants randomized in each of the studies and who also completed instruments at baseline and one additional time point are summarized in Table 1. The demographic characteristics were similar in the HRQoL analysis patient subset to those of the entire study population in each study (data not shown). In Study 2107, a longer mean duration of metastatic disease in the HRQoL analysis subset was observed in the placebo arm (5.1 months versus 4 months). In Study 2192, a higher incidence of patients with an ECOG performance status score of 0 in the BV arm (32.6% versus 28.8%) was observed. Study 2107 Time to deterioration in HRQoL results in Study 2107 are summarized in Tables 2 and 3. The baseline and at least one postbaseline CCS score were available for 127 patients in the IFL plus placebo arm and for 122 patients in the IFL plus BV arm; the baseline and at least one postbaseline TOI-C score were available for 125 patients in the IFL plus placebo arm and 122 patients in the IFL plus BV arm; and the baseline and at
Kabbinavar, Wallace, Holmgren et al. 5 Table 1. Characteristics of participants with health-related quality-of-life scores (Colorectal Cancer Subscale) at baseline and at least one postbaseline assessment for Studies 2107 and 2192 Study 2107 Study 2192 IFL/placebo IFL/BV 5-FU/LV/placebo 5-FU/LV/BV n 127 122 77 89 Mean age, years (SD) 58.9 (11.12) 59.4 (12.36) 69.3 (9.23) 70.6 (9.39) Range 35 81 23 84 41 90 35 89 Gender Female 49 (38.6%) 52 (42.6%) 35 (45.5%) 37 (41.6%) Male 78 (61.4%) 70 (57.4%) 42 (54.5%) 52 (58.4%) ECOG performance status score 0 69 (54.3%) 71 (58.2%) 21 (27.3%) 29 (32.6%) 1 58 (45.7%) 51 (41.8%) 56 (72.7%) 60 (67.4%) Duration of metastatic disease (months) Mean (SD) 5.1 (13.2) 4.4 (10.8) 3.7 (6.1) 4.6 (10.2) Range 1 125 1 91 1 46 1 75 Location of primary tumor Colon 99 (78.0%) 95 (77.9%) 60 (77.9%) 73 (82.0%) Rectum 28 (22.0%) 27 (22.1%) 17 (22.1%) 16 (18.0%) None of the differences between the two treatment arms for each of the trials were statistically significant. Abbreviations: 5-FU, 5-fluorouracil; BV, bevacizumab; ECOG, Eastern Cooperative Oncology Group; LV, leucovorin; SD, standard deviation. least one postbaseline FACT-C total score were available for 124 patients in the IFL plus placebo arm and 121 patients in the IFL plus BV arm. There was no statistically significant difference between treatment arms in time to deterioration in HRQoL as measured by the CCS score ( 3; HR, 0.92; 95% CI, 0.68 1.26; p.6190), TOI-C score ( 7; HR, 1.13; 95% CI, 0.82 1.55; p.4618), or FACT-C total score ( 9; HR, 0.96; 95% CI, 0.69 1.33; p.8116). The median (95% CI) time to deterioration in HRQoL as measured by the CCS score ( 3) was 2.76 (2.46 3.88) months in the IFL plus placebo arm and 2.89 (2.60 4.17) months in the IFL plus BV arm. The median (95% CI) time to deterioration in HRQoL as measured by the TOI-C score ( 7) was 3.48 (2.69 4.50) months in the IFL plus placebo arm and 2.79 (2.60 4.01) months in the IFL plus BV arm. The median (95% CI) time to deterioration in HRQoL as measured by the FACT-C total score ( 9) was 3.94 (2.79 5.22) months in the IFL plus placebo arm and 3.98 (2.76 5.42) months in the IFL plus BV arm. The additional analyses suggested that there were no statistical differences between the IFL plus placebo and IFL plus BV arms in change from baseline to last assessment of HRQoL according to the CCS, TOI-C, and FACT-C total scores. The change from baseline score stratified by the time of the last visit was not significant for the CCS (p.6584), TOI-C (p.5012), or FACT-C total (p.7116) score. Sensitivity analyses using the upper end of the published recommended minimally important differences of 6 points for the TOI-C score and 8 points for the FACT-C total score [16] resulted in a shorter time to deterioration in HRQoL compared with the prespecified provisional minimally important difference of 7 points for the TOI-C score and no difference in time to deterioration in HRQoL compared with the prespecified provisional minimally important difference of 9 points for the FACT-C total score within each of the treatment arms, and did not change the results between treatment arms. Study 2192 Results for time to deterioration in HRQoL in Study 2192 are summarized in Tables 4 and 5. FACT-C scores were available at baseline and at least one postbaseline visit for 77 patients in the 5-FU and LV plus placebo arm and for 89 patients in the 5-FU and LV plus BV arm. There was no statistically significant difference between treatment arms in time to deterioration in HRQoL as measured by the CCS score ( 3; HR, 0.81; 95% CI, 0.57 1.14; p.2176) or TOI-C score ( 7; HR, 0.71; 95% CI, 0.51 1.01; p.0537). The median (95% CI) time to deterioration in HRQoL as measured by the CCS score ( 3) was 3.02 (1.58 3.22) months in the 5-FU and LV plus placebo arm and 3.12 (2.07 3.65) months in the 5-FU and LV plus BV arm. The median (95% CI) time to deterioration in www.theoncologist.com
6 HRQoL Impact of Bevacizumab Plus 5-FU Chemotherapy Table 2. Study 2107: Time to deterioration in Colorectal Cancer Subscale score during first-line therapy IFL/placebo IFL/BV n a 127 122 Patients with an event b 111 (87.4%) 102 (83.6%) Decline in score ( 3) 75 (59.1%) 76 (62.3%) Disease progression 35 26 Death 1 0 Censored observations 16 (12.6%) 20 (16.4%) Time to deterioration, c months Median (95% CI) 2.76 (2.46 3.88) 2.89 (2.60 4.17) 25th 75th percentile 1.35 5.55 1.35 7.10 Range 1.15 22.37 e 0.26 e 22.08 e Stratified analysis d Hazard ratio (95% CI) NA 0.922 (0.68 1.26) p-value (log rank) NA.6190 a Number of patients with a baseline assessment. Earliest contributing event is shown. c Summary statistics are from a Kaplan Meier analysis; the 95% CI was calculated using the method of Brookmeyer and Crowley [17]. d Relative to the IFL/placebo arm (estimated by Cox regression). Strata are: Eastern Cooperative Oncology Group performance status score (0 or 1), number of metastatic disease sites (1 or 1), location of primary tumor (colon or rectum), and census region for randomization site. e Indicates a censored value. Abbreviations: BV, bevacizumab; CI, confidence interval; IFL, irinotecan, 5-fluorouracil, and leucovorin; NA, not applicable. Table 3. Study 2107: Time to deterioration in TOI-C score (change from baseline 7) and FACT-C total score (change from baseline 9) during first-line therapy IFL/placebo IFL/BV TOI-C n 125 122 Median (95% CI), months 3.48 (2.69 4.50) 2.79 (2.60 4.01) Hazard ratio a (95% CI) NA 1.13 (0.82 1.55) p-value (log rank) NA.4618 FACT-C n 124 121 Median (95% CI), months 3.94 (2.79 5.22) 3.98 (2.76 5.42) Hazard ratio a (95% CI) NA 0.96 (0.69 1.33) p-value (log rank) NA.8116 a p-value is from stratified log-rank test. Strata are: Eastern Cooperative Oncology Group performance status score (0 or 1), number of metastatic disease sites (1 or 1), location of primary tumor (colon or rectum), and census region for randomization site. Abbreviations: BV, bevacizumab; CI, confidence interval; FACT-C, Functional Assessment of Cancer Therapy Colorectal; IFL, irinotecan, 5-fluorouracil, and leucovorin; NA, not applicable; TOI-C, Trial Outcome Index. HRQoL as measured by the TOI-C score ( 7) was 3.02 (1.61 3.22) months in the 5-FU and LV plus placebo arm and 3.22 (3.02 4.73) months in the 5-FU and LV plus BV arm. Time to deterioration in HRQoL as measured by the FACT-C total score was significantly longer in the 5-FU and LV plus BV arm than in the 5-FU and LV plus placebo arm (FACT-C total, 9; HR, 0.69; 95% CI, 0.49 0.98; p.0396). The median (95% CI) time to deterioration in HRQoL as measured by the FACT-C total score ( 9) was 3.02 (1.61 3.58) months in the 5-FU and LV plus pla-
Kabbinavar, Wallace, Holmgren et al. 7 Table 4. Study 2192: Time to deterioration in Colorectal Cancer Subscale score during first-line therapy 5-FU/LV/placebo 5-FU/LV/BV n a 77 89 Patients with an event b 68 (88.3%) 77 (86.5%) Decline in score ( 3) 48 (62.3%) 57 (64.0%) Disease progression 19 20 Death 1 0 Censored observations 9 (11.7%) 12 (13.5%) Time to deterioration, c months Median (95% CI) 3.02 (1.58 3.22) 3.12 (2.07 3.65) 25th 75th percentile 1.35 4.93 1.38 6.87 Range 1.18 14.59 1.22 22.77 e Stratified analysis d Hazard ratio (95% CI) NA 0.808 (0.57 1.14) p-value (log rank) NA.2176 a Number of patients with a baseline assessment and at least one postbaseline assessment. Includes one patient in the 5-FU/ LV/placebo arm whose baseline assessment was obtained after receiving the first chemotherapy dose and before the first study drug dose. b Earliest contributing event is shown. c Summary statistics are from a Kaplan Meier analysis; the 95% CI was calculated using the method of Brookmeyer and Crowley [17]. d Relative to the 5-FU/LV/placebo arm (estimated by Cox regression). Strata are: Eastern Cooperative Oncology Group performance status score (0 or 1), number of metastatic disease sites (1 or 1), and location of primary tumor (colon or rectum). e Indicates a censored value. Abbreviations: 5-FU, 5-fluorouracil; BV, bevacizumab; CI, confidence interval; LV, leucovorin; NA, not applicable. Table 5. Study 2192: Time to deterioration in TOI-C score (change from baseline 7 points) and FACT-C total score (change from baseline 9 points) during first-line therapy 5-FU/LV/placebo 5-FU/LV/BV TOI-C n 77 89 Median (95% CI), months 3.02 (1.61 3.22) 3.22 (3.02 4.73) Hazard ratio a (95% CI) NA 0.713 (0.51 1.01) p-value (log rank) NA.0537 FACT-C n 77 89 Median (95% CI), months 3.02 (1.61 3.58) 3.58 (3.09 4.93) Hazard ratio a (95% CI) NA 0.694 (0.49 0.98) p-value (log rank) NA.0396 a p-value is from stratified log-rank test. Strata are: Eastern Cooperative Oncology Group performance status score (0 or 1), number of metastatic disease sites (1 or 1), and location of primary tumor (colon or rectum). Abbreviations: 5-FU, 5-fluorouracil; BV, bevacizumab; CI, confidence interval; FACT-C, Functional Assessment of Cancer Therapy Colorectal; LV, leucovorin; TOI-C, Trial Outcome Index. cebo arm and 3.58 (3.09 4.93) months in the 5-FU and LV plus BV arm. The additional analyses of changes from baseline in FACT-C scores indicated no statistical differences between the treatment arms in change from baseline to last assessment of HRQoL as assessed by the CCS, TOI-C, and FACT-C total scores. The change from baseline scores stratified by the time of the last visit were not significant for the CCS (p.6510), TOI-C (p.1248), or FACT-C total (p.0959) scores. Sensitivity analyses using the upper end of the pub- www.theoncologist.com
8 HRQoL Impact of Bevacizumab Plus 5-FU Chemotherapy lished recommended minimally important differences of 6 points for the TOI-C score and 8 points for the FACT-C total score [16] indicated a shorter time to deterioration in HRQoL compared with the prespecified provisional minimally important difference of 7 points for the TOI-C score and no difference in time to deterioration in HRQoL compared with the prespecified provisional minimally important difference of 9 points for the FACT-C total score within each of the treatment arms, and did not change the results between treatment arms. DISCUSSION The efficacy of chemotherapy in advanced malignancies is usually assessed through traditional endpoints such as response rates, toxicity, disease-free survival, and overall survival. However, these endpoints do not take into account the functional condition of the patients or the quality of their survival, and hence may not truly assess the overall therapeutic impact. The benefit of chemotherapy in incurable cancers needs to be assessed directly through validated HRQoL instruments, rather than inferred from response rates, survival benefits, and other traditional endpoints. As patients are living longer with newer therapies, physicians now have to take into account potential long-term toxicities when choosing therapies. Cancer patients also would like to have this information to help them while making decisions regarding therapies. The addition of BV to 5-FU based chemotherapy regimens for the treatment of metastatic CRC has been demonstrated to increase time to disease progression and prolong overall survival. With these clinical gains, it is also important to evaluate the impact of adding BV to chemotherapy on patients HRQoL. The endpoints for our HRQoL analyses were predefined in all trials, so that the HRQoL analysis is hypothesis testing rather than hypothesis generating. The analyses of Study 2107 and Study 2192 are the first comparisons of patients HRQoL with the addition of BV to chemotherapy. In the clinical setting, a benefit can be defined as an improvement in baseline symptoms. In the case of adding an additional agent to a treatment regimen, benefit can also be defined as a lack of shortening the time to progression of symptoms when compared with the original treatment regimen alone. The primary HRQoL endpoint in these studies was the time to deterioration in HRQoL as measured by the minimally important difference ( 3 points) for the CCS score. There was not a statistically significant difference between treatment arms in time to deterioration as measured by the CCS or TOI-C score in either study, suggesting that the addition of BV did not add to patients treatment burden. Measures of time to deterioration in HRQoL using the FACT-C total score were also similar between treatment arms in Study 2107 but statistically different in Study 2192, with patients receiving BV plus chemotherapy having a longer time to deterioration than patients receiving chemotherapy alone. There are some limitations to the analyses. First, the analyses were planned on a subset of patients in Study 2107. However, the demographic characteristics of the patients from whom questionnaires were available were comparable with those of the overall study population. Second, HRQoL assessments were missing at some visits for patients included in the analyses. This may have affected the results when missing data appear prior to deterioration in HRQoL. However, multiple missing assessments before deterioration were relatively rare and the proportion of missing assessments was comparable between the two treatment arms in each study. None of the differences in missing assessments were statistically significant. In conclusion, this prospective HRQoL analysis supports the clinical benefit of the addition of BV to 5-FU based chemotherapy in improving time to disease progression and prolonging overall survival, without compromising patients HRQoL. ACKNOWLEDGMENT This work was presented as a poster presentation at the 2005 Annual Meeting of the American Society of Clinical Oncology 2005 (abstract no. 3564). AUTHOR CONTRIBUTIONS Conception/design: Fairooz F. Kabbinavar, David Cella, Herbert I. Hurwitz Provision of study materials or patients: Fairooz F. Kabbinavar, Herbert I. Hurwitz Collection/assembly of data: Fairooz F. Kabbinavar Data analysis and interpretation: Fairooz F. Kabbinavar, Joel F. Wallace, Eric Holmgren, Jing Yi, David Cella, Kathleen J. Yost, Herbert I. Hurwitz Manuscript writing: Fairooz F. Kabbinavar, Joel F. Wallace, Eric Holmgren, David Cella, Kathleen J. Yost, Herbert I. Hurwitz Final approval of manuscript: Fairooz F. Kabbinavar, Joel F. Wallace, Eric Holmgren, Jing Yi, David Cella, Kathleen J. Yost, Herbert I. Hurwitz The authors take full responsibility for the content of the paper but thank Linda Phillips, Ph.D., from Genentech, Inc., for her assistance in copyediting and formatting the manuscript. The authors thank Anita Chawla, Ph.D., Betty Nelson, Ph.D., and William Novotny, M.D., from Genentech, Inc., for their contributions to Studies 2107 and 2192 and review of the earlier drafts of the manuscript. REFERENCES 1 American Cancer Society. Detailed Guide: Colon and Rectum Cancer. March 7, 2006. Available at http://www.cancer.org/docroot/cri/ CRI_2_3x.asp?dt 10. Accessed June 30, 2006. 2 Hurwitz H, Fehrenbacher L, Novotny W et al. Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med 2004;350:2335 2342. 3 Kabbinavar FF, Schulz J, McCleod M et al. Addition of bevacizumab to bolus fluorouracil and leucovorin in first-line metastatic colorectal cancer: Results of a randomized phase II trial. J Clin Oncol 2005;23:3697 3705.
Kabbinavar, Wallace, Holmgren et al. 9 4 Cella DF. Measuring quality of life in palliative care. Semin Oncol 1995; 22(suppl 3):73 81. 5 Cella D. Manual of the Functional Assessment of Chronic Illness Therapy (FACIT) Measurement System, Version 4. Evanston, IL: Center on Outcomes, Research & Education (CORE), Evanston Northwestern Healthcare and Northwestern University, 1997. 6 Ward WL, Hahn EA, Mo F et al. Reliability and validity of the Functional Assessment of Cancer Therapy Colorectal (FACT-C) quality of life instrument. Qual Life Res 1999;8:181 195. 7 Saltz LB, Cox JV, Blanke C et al. Irinotecan plus fluorouracil and leucovorin for metastatic colorectal cancer. Irinotecan Study Group. N Engl J Med 2000;343:905 914. 8 Petrelli N, Douglass HO Jr, Herrera L et al. The modulation of fluorouracil with leucovorin in metastatic colorectal carcinoma: A prospective randomized phase III trial. Gastrointestinal Tumor Study Group. J Clin Oncol 1989;7:1419 1426. 9 Guyatt GH, Osoba D, Wu AW et al.; Clinical Consensus Meeting Group. Methods to explain the clinical significance of health status measures. Mayo Clin Proc 2002;77:371 383. 10 Cella D, Eton DT, Fairclough DL et al. What is a clinically meaningful change on the Functional Assessment of Cancer Therapy Lung (FACT-L) questionnaire? Results from Eastern Cooperative Oncology Group (ECOG) Study 5592. J Clin Epidemiol 2002;55:285 295. 11 Cella D, Eton DT, Lai JS et al. Combining anchor and distribution-based methods to derive minimal clinically important differences on the Functional Assessment of Cancer Therapy (FACT) anemia and fatigue scales. J Pain Symptom Manage 2002;24:547 561. 12 Patrick DL, Gagnon DD, Zagari MJ et al. Assessing the clinical significance of health-related quality of life (HrQOL) improvements in anaemic cancer patients receiving epoetin alfa. Eur J Cancer 2003;39:335 345. 13 Eton DT, Cella D, Yost KJ et al. A combination of distribution- and anchorbased approaches determined minimally important differences (MIDs) for four endpoints in a breast cancer scale. J Clin Epidemiol 2004;57:898 910. 14 Yost KJ, Yount SE, Eton DT et al. Validation of the Functional Assessment of Cancer Therapy Breast Symptom Index (FBSI). Breast Cancer Res Treat 2005;90:295 298. 15 Yost KJ, Sorensen MV, Hahn EA et al. Using multiple anchor- and distribution-based estimates to evaluate clinically meaningful change on the Functional Assessment of Cancer Therapy Biologic Response Modifiers (FACT-BRM) instrument. Value Health 2005;8:117 127. 16 Yost KJ, Cella D, Chawla A et al. Minimally important differences were estimated for the Functional Assessment of Cancer Therapy Colorectal (FACT-C) instrument using a combinatin of distribution- and anchor-based approaches. J Clin Epidemiol 2005;58:1241 1251. 17 Brookmeyer R, Crowley J. A confidence interval for the median survival time. Biometrics 1982;38:29 41. www.theoncologist.com