Health-Related Quality of Life in Women With Metastatic
`Breast Cancer Treated With Trastuzumab (Herceptin)
`David Osoba and Michael Burchmore
`
`The measurement of health-related quality of life
`(HRQL) was an Important component of clinical trials
`of trastuzumab (Herceptin; Genentech, San Francisco.
`CA) In women with progressive HERZ-overexpressing
`metastatic breast cancer who may or may not have had
`prior chemotherapy. Health-related quality of life was
`measured at baseline and specified intervals during
`therapy using the European Organization for Research
`and Treatment of Cancer core Quality of Life Ques-
`tionnaire (QLQ-CSO, version l.0). Five domains were
`chosen a priorl for analysis: global quality of life, phys-
`ical, role and social functioning, and fatigue. In the
`phase II study, in which trastuzumab was given alone.
`there was no change In on-treatment QLQ-C30 scores
`compared with baseline. These results suggest that
`trastuzumab does not adversely affect HRQL during
`therapy. In the phase III study, the effects of trastu-
`zumab plus chemotherapy were compared with those
`of chemotherapy alone. A comparison of QLQ-C30
`scores during treatment did not show statistically sig-
`nificant differences between the two groups at the pre-
`set level of P = .0] . HoWever. comparison of on-treat-
`ment
`scores with baseline
`In patients
`receiving
`chemotherapy alone indicated mild worsening of phys-
`lcai and role functioning and of fatigue throughout the
`duration of treatment, whereas a similar comparison
`of those receiving chemotherapy with trastuzumab re-
`vealed mlid worsening of role functioning at weeks 8
`and 20 and of fatigue only at week 8. These results
`suggest that trastuzumah may be associated with an
`amelioration of the deleterious effects of chemother-
`apy alone. In summary, in the doses and schedules used
`In these studies, trastuxumab is not associated with
`worsening of HRQL.
`Semin Oncol 26 (suppi l2):84—88. Copyright 6) I 999 by
`W.B. Saunders Company.
`
`HE TESTING of any new treatments for
`their efficacy in women with metastatic
`breast cancer should include a concomitant dssess—
`ment of health—related quality of life (HRQL).
`Without such assessments, the benefits of treat—
`ment and effect of adverse events on patients'
`well-being will be only partially understood. If the
`
`From the Department of Medicine, University of British Colum-
`bin, Quality of Life Consulting, Vancouver, BC; and Genentech,
`inc, South San Francisco. CA.
`Dr Osoba has received consulting fees from Genentech, inc.
`Address reprint requests to David Osoba, MD, QOL Consult-
`ing: 4939 Edcndale Court, Vancouver, BC, Canada V7W 3H7.
`Copyright © 1999 by WB. Saunders Company
`009317754/99/26041i207$iO.OO/O
`
`B4
`
`rislcs and benefits to HRQL posed by treatment are
`unknown, health care professionals are hampered
`in their ability to fully explain them to patients
`who may receive similar treatment in the future.
`In the work reported here, HRQL was measured
`as a secondary outcome in two clinical trials of
`women who received trastuzumab (Herccptin; Ge‘
`nentech, San Francisco, CA) for treatment of
`metastatic breast cancer in which HERZ is over-
`expressed. As the results of these clinical trials are
`reported in detail elsewhere in this supplement,1
`this report is limited to an overview of the results
`of the HRQL assessment.
`
`PATIENTS AND METHODS
`
`Women with progressive HERZ-overexpressing metastatic
`breast cancer who either had been previously treated with
`chemotherapy (phase II study) or who had not received previv
`ous cytotoxic chemotherapy (phase [11 study) were treated with
`trastuzumab, as described elsewhere.| In the phase II study,
`HRQL assessments Were scheduled for week i (at baseline, ic,
`before the initiation of tiastuzumab). weeks 12, 24, 36, and 48,
`and every 12 weeks thereafter, as well as on withdrawal from
`study or study termination.l In the phase III study, the assess—
`ments were performed at weeks I, 8, 20, and 32, and every 12
`weeks thereafter.l All assessments were to be completed in the
`clinic before the administration of trastuzumab and/or chemo!
`therapy.
`Health—related quality of life was assessed by the self—admin—
`istered European Organization for Research and Treatment of
`Cancer (EORTC) core Quality of Life Questionnaire (QLQ‘
`C30. version 1.0).2 This questionnaire contains 30 Items,
`grouped into a global quality of life (QL) scale, five functioning
`scales (physical, role, emotional. social, and cognitive), three
`symptom scales (fatigue, pain, and nausea/vomiting), and six
`single items dealing with common symptomslproblcms that
`patients with cancer may experience. Additional modules con-
`sisting of study—specific items were also used, but analysis of
`these data is still incomplete and will not be reported here.
`The QLQ-CEIO responses were scored and analyzed using
`algorithms in the scoring manual supplied by the EORTC
`Study Group on Quality of Life.3 All raw scores were trans
`formed to a. scale ranging from O to 100. Over time, higher
`scores in the functioning scales indicate an improvement,
`whereas in the symptom scales and items, higher scores indicate
`a worsening of symptoms.
`In the phase Ii study, the purpose of the HRQL assessments
`was to determine whether treatment with rrastuzumab was
`associated with a decline in HRQL. To determine changes in
`HRQL scores, the values obtained at various time points while
`on treatment and at disease progression were subtracted from
`the basallne scores. The purpose of the analysis in the phase II]
`study was to test the null hypothesis of no difference in the
`
`Seminars in Oncology, Vol 26, No 4. Suppl I2 (August). I999: pp 94—88
`
`(cid:43)(cid:82)(cid:86)(cid:83)(cid:76)(cid:85)(cid:68)(cid:3)(cid:89)(cid:17)(cid:3)(cid:42)(cid:72)(cid:81)(cid:72)(cid:81)(cid:87)(cid:72)(cid:70)(cid:75)
`Hospira v. Genentech
`(cid:44)(cid:51)(cid:53)(cid:21)(cid:19)(cid:20)(cid:26)(cid:16)(cid:19)(cid:19)(cid:27)(cid:19)(cid:24)(cid:3)
`IPR2017-00805
`(cid:42)(cid:72)(cid:81)(cid:72)(cid:81)(cid:87)(cid:72)(cid:70)(cid:75)(cid:3)(cid:40)(cid:91)(cid:75)(cid:76)(cid:69)(cid:76)(cid:87)(cid:3)(cid:21)(cid:19)(cid:22)(cid:25)
`Genentech Exhibit 2036
`
`

`

`HEALTH-RELATED QUALITY OF LIFE
`
`85
`
`change of HRQL scores from baseline between treatment
`groups at a significance level of P = .01. in both studies, five
`QLQ—C30 scales were prospectively identified as the primary
`variables of interest, ie, global QL, physical, role and social
`functioning. and fatigue. Only the results obtained for these
`variables will be presented here.
`Time windows were created for each predetermined mea-
`suremcnt time. All questionnaires completed within 4 weeks of
`the specified times were accepted for scoring and analysis. in
`cases of missing data. the following imputations were assigned:
`patient death was assigned a value of “0"; if the patient was
`alive but data missing, the last observation was carried forward.
`Missing data at week 8 or 10 were not included in the data
`analysis. No analyses were attempted after week 32 in the phase
`II study and week 36 in the phase 111 study due to attrition of
`data.
`Analyses were performed using a SAS program (SAS Insti—
`tute, Cary, NC). For primary HRQL analysis, a repeated—
`musures ANOVA was used. in the phase "I studyl standard F
`statistics were used to assess the between-subject factors and a
`Huynh—Feldt epsilon adjustment was used for the timetinterac—
`tion terms.
`
`RESULTS
`
`Phase 11 Study
`
`QLQ—C3O completion rates. Baseline QLQ—
`C30 scores were collected for 207 of the 222
`
`patients entered into the study (93%). At subse—
`quent times the percentages of patients for whom
`QLQ—C3O scores were available or analyzed were
`90% at week 12, 97% at week 24, and 94% at
`week 36. These high rates of available scores are
`attributable in part
`to the imputation methods
`used for assigning scores to missing values. The
`ratios of completed questionnaires to those ex—
`pected from patients who were alive and still on
`study were as follows: 110 of 122 (92%) patients at
`week 12, 83 of 86 (97%) patients at week 24, and
`31 of 33 (94%) patients at week 36. One hundred
`fifty—four patients (74%) had complete scores,
`
`which included measurements at baseline and at
`
`week 12, and at progression before 12 weeks.
`Health—related quality of life results. Overall,
`there was no apparent worsening of QLQ—C3O
`scores in the five predetermined scales (global QL,
`physical, role and social functioning, and fatigue;
`Table 1). Individual patient scores for all
`five
`scales were calculated for each time point and
`baseline scores were subtracted from values at
`
`weeks 12, 24, and 36 to determine the change in
`each. For patients with missing data after week 12,
`the mean changes in scores are likely to underes—
`timate the true changes, since after week 12 the
`last available observation was carried forward.
`
`These results indicate that patients who were able
`to continue therapy with trastuzumab did not ex’
`perience deleterious effects in their HRQL. HOWr
`ever, attrition rates over time were high, and these
`results must be interpreted with some caution
`since surviving patients remaining on treatment
`are more likely to report better HRQL than those
`who must discontinue treatment because of disease
`progression.
`
`Phase III Study
`
`QLQ’C3O completion rates. At baseline, 431 of
`469 (92%) patients completed the questionnaire.
`At subsequent time points, the numbers of regu—
`larly scheduled questionnaires completed were 369
`of390 (95%) at week 8, 282 of320 (88%) at week
`20, and 160 of 181 (88%) at week 32.
`Health—related quality of life results. Analysis of
`the five predetermined scales did not show any
`statistically significant differences between those
`receiving trastuzumab with chemotherapy and
`those receiving chemotherapy only, at the preset
`
`Table I. Mean Change In QLQ-CJO Scores From Basellno to Week 36 (Phase II Study)
`
`A negative number indicates worsening for global QL physical. role and social functioning. and improvement for fatigue.
`
`Global QL
`Physicai function
`Social function
`Role function
`Fatigue
`
`Baseiino Score
`
`62.2 t-2|.0
`75.7 i 24.6
`70.4 :9: 29.5
`67.7 1 NJ
`33.9 t 23.4
`
`4.3 1' 20.1
`0 I 22.8
`6.5 I 24.8
`ID i 32.8
`—0.l it 23.3
`
`Mean Change From Baseline*
`Week 24
`
`H i 2L3
`“0.3 i' 24.4
`4.8 I 26.0
`0 :t 33.7
`0.4 1' 23.0
`
`—0.5 i 2L7
`-2.2 t 24.4
`L9 i 25.3
`- L4 1 32.5
`2.0 t 23.2
`
`* Mean change from baseline (mean 1 SD) for all patients who completed baseline and week l2 assessments: includes disease progression.
`
`

`

`86
`
`level of P = .01 (Table 2). As in the phase II
`study, the changes in scores are likely to be under—
`estimates of the true values due to use of the
`
`carry—forward method to impute scores after the
`week 8 assessment.
`I
`
`We have developed another approach to intern
`preting the significance of changes, in which the
`degree of change in QLQvC30 scores perceived by
`patients is categorized as “ little,” “moderate,” or
`“very much."4 In this method, changes between 5
`and 10 in QLQ—C30 scores would indicate “little”
`change,
`11 to 20 would indicate “moderate”
`change, and more than 20 would indicate “very
`much” change. Using this interpretation, the tras—
`tuzumab plus chemotherapy group would be
`judged to have experienced a little worsening in
`social and role functioning and in fatigue scores at
`week 8, in only role functioning at week 20, and
`no change from baseline at week 32. The chemOr
`therapy—only group experienced a little worsening
`in physical and role functioning and in fatigue,
`which persisted throughout the entire 32 weeks of
`observation.
`
`DISCUSSION
`
`The objective for inclusion of the HRQL coma
`ponent in the openvlabel phase II study was to
`describe changes in HRQL for previously treated
`women with HERZ—overexpressing metastatic
`
`OSOBA AND BURCHMORE
`
`breast cancer while on treatment with trastUr
`zumab. The results indicate that HRQL, as mea—
`sured by selected scales of the QLQ—C3O question‘
`naire, did not change appreciably over 36 weeks of
`treatment. However, since the number of patients
`remaining on 'treatment decreased rapidly over
`time (only 33 patients at 36 weeks), those survin
`ing and completing HRQL questionnaires would
`be expected to report higher HRQL scores than
`those in whom treatment was stopped due to dis—
`ease progression. The results at week 12 are of
`interest because at this time there were still 83 of
`
`207 patients on treatment with HRQL data, some
`of whom had progressive disease, yet the HRQL
`scores had not changed appreciably from baseline
`(other than a little worsening in social function-
`ing). A comparison of the scores in patients with
`complete and partial disease responses with those
`for patients with progressive disease would help to
`clarify the true picture. However, in the phase II
`study the number of patients remaining on study at
`week 38 is probably too small for a meaningful
`comparison.
`The HRQL component was included in the
`phase III study to determine whether the addition
`of trastuzumab to chemotherapy with doxorubicin
`and cyclophosphamide or paclitaxel was associated
`with a worsening of HRQL compared with treat—
`ment with chemotherapy alone. In this study,
`
`Table 2. Change in QLQ-C30 Scores From Baseline to Week 32 (Phase III Study)
`
`Baseline Score
`
`Mean Change From Baseline‘l‘
`Week 20
`
`improvement for fatigue
`
`Trastuzumab plus chemotherapy
`Global QL
`Physical function
`Social function
`Role function
`Fatigue
`
`Chemotherapy alone
`Global QL
`Physical function
`Social function
`Role function
`Fatigue
`
`N =207
`59.3: |.8
`7|.5: L9
`68.0 :t 2.!
`64.6 i 2.5
`37.6 1- L9
`
`N = I94
`58.4113
`70.61”
`“J i 2.2
`66.2 1 2.7
`36.9 1 2.0
`
`*Changes represent mean t SE. A negative number indicates a worsening for global QL. physical. role. and social functioning. and
`
`

`

`HEALTH-RELATED QUALITY OF LIFE
`
`87
`
`changes in QLQvC3O scores were not statistically
`significantly different at the preset level of P = .01
`between the two arms of the study. However, in
`the trastuzumab plus chemotherapy group, base
`line QLQ—C3O scores for global QL as well as role
`and social functioning worsened by 8 weeks, then
`returned to baseline by 20 and 32 weeks of the
`study. In contrast, they remained worse than base
`line for global QL and physical and role function—
`ing in the chemotherapy—alone arm for the entire
`32 weeks. As in the phase II study, the comparison
`of HRQL scores at various onrtreatment
`times
`with baseline scores includes patients with varying
`disease response status. Further analysis comparing
`scores of patients with complete and partial rev
`sponses to those with progressive disease may be
`helpful. There appears to be sufficient patients
`remaining at each time point to make such a
`comparison meaningful.
`In previous studies of chemotherapy in breast
`cancer, it has been shown that HRQL scores de—
`teriorate while patients are receiving treatments-5
`In the phase II trial reported here, no significant
`deterioration in HRQL scores was observed in
`patients treated with trastuzumab. In the phase III
`study, deterioration in HRQL scores appeared to
`be shorter in duration in patients treated with
`trastuzumab alone compared with those treated
`with trastuzumab and chemotherapy. These results
`suggest that not only is there is no additional
`deterioration of HRQL associated with trastu—
`zumab therapy, but trastuzumab therapy also may
`be associated with an ameliorative effect on
`
`HRQL. This possibility remains to be explored
`further in future studies.
`
`One difficulty with the interpretation of these
`results from both studies is that calculation of
`
`scores for patients with missing data (due to factors
`other than death) is problematic. In this study we
`used a “last available observation carried forward”
`
`approach. However, in patients with progressive
`disease, this approach tends to overestimate the
`results at the missing time points, since the scores
`from the completions at earlier time points in the
`study, before disease progression, are likely to be
`better (ie, higher functioning scores and lower
`symptom scores) than those from later time points
`at which data are more likely to be missing. This
`“better” score, if carried forward for two or more
`time points, will skew the score in favor of a better
`score for the later time points. If missing data are
`
`associated with worsening disease just preceding
`death or removal of the patient from study (non—
`random missing data), then the direction of the
`incorrect estimate can be known with certainty.
`However,
`if the true score of the missing data
`(were it known) from a later time point is actually
`better than from a previous time point, and the
`data are missing at random (ie, there are additional
`scores after the missing score), then the carry—
`forward score will be an underestimate of the true
`
`the
`score. If the missing data are random but
`disease status cannot be assessed or the effects of
`
`toxicity on HRQL are uncertain, it is impossible to
`determine whether the last available observation
`is an overestimate or an underestimate of the true
`score.
`
`A method that may circumvent some of the
`difficulties associated with imputation of data is to
`determine the mean duration of change (improve—
`ment or deterioration) by a time-to—event analysis
`in the intent—to—treat population(s).7 In this ap
`preach, the event needs to be well defined (eg, a
`score improvement of at least 10 points [on a scale
`of O to 100] compared with baseline, lasting for a
`defined period of time, followed by a deterioration
`of at least 10 points). To obtain the mean duration
`of improvement for the group,
`the duration of
`improvement for each subject is calculated, and
`the individual durations are summed and divided
`
`by the number of patients in the intent—tOvtreat
`population. In phase III studies, the mean duration
`of improvement in each HRQL domain score can
`then be compared in two or more arms of the
`study.
`Furthermore,
`duration—ofvimprovement
`curves can be plotted for various HRQL scales for
`those patients who experienced improvement. If
`the pattern in HRQL scores in a given study
`indicates an initial deterioration in HRQL scores,
`then a similar method could be used to first cal—
`culate the mean duration of deterioration followed
`
`by a calculation of the mean duration of improve-
`ment.
`
`Despite the limitations in the analysis method
`used in the studies reported here, it can be concluded
`that trastuzumab does not appear to adVersely affect
`HRQL as measured by the QLQ—C30. Further anal,
`ysis of the data, using an intent—to—treat approach,
`and the continued measurement of HRQL in pa—
`tients treated with trastuzumab are warranted.
`
`

`

`88
`
`OSOBA AND BURCHMORE
`
`ACKNOWLEDGMENT
`
`The authors thank Genentech, Inc, for supporting these
`studies and ]ill Vardy for assistance with preparation of the
`manuscript.
`
`REFERENCES
`
`‘
`
`1. Shak 5: Overview of the Herceptin anti—HERZ monoclo-
`nal antibody clinical program in HERs—overexpressing meta-
`static breast cancer. Semin Oncol 26:71—77, 1999 (suppl 12)
`2. Aaronson NK. Ahmedzai S. Bergman B. or al: The Euv
`ropean Organization for Research and Treatment of Cancer
`QLQ—C30: A quality—of—life instrument for use in international
`clinical trials in oncology.) Natl Cancer lnst 85:365c376, 1993
`3. Fayers P, Aaronson N, Bjordal K, et al: EORTC QLQ-
`
`C30 Scoring Manual. Brussels, Belgium. Quality of Life Unit,
`EORTC Data Centre, 1995
`4. Osoha D, Rodrigues G, Myles J. et al: Interpreting the
`significance of changes in health—related qualityvof—life scores.
`J Clin Oncol 16:139—144, 1998
`5. Osoba D, Zee B, Pater JL. et a1: Psychometric properties
`and responsiveness of the EORTC Quality of Life Question-
`naire (QLQ—C30)
`in patients with breast, ovarian and lung
`cancer. Qual Life Res 3553—364, 1994
`6. King MT: The interpretation of scores from the EORTC
`quality of life questionnaire QLQ-C30. Qual Life Res 5:555-
`567, 1996
`7. Osoba D, Tannoclt IF, Ernst DS, et al: Health—related
`quality of life in men with prostate cancer treated by pred—
`nisone alone or mitoxanrrone and prednisone. J Clin Oncol
`17:1654-1663, 1999
`
`