Download Estimating Typical, Best-Case, and Worst

VOLUME
29
䡠
NUMBER
4
䡠
FEBRUARY
1
2011
JOURNAL OF CLINICAL ONCOLOGY
R E V I E W
A R T I C L E
How Long Have I Got? Estimating Typical, Best-Case, and
Worst-Case Scenarios for Patients Starting First-Line
Chemotherapy for Metastatic Breast Cancer: A Systematic
Review of Recent Randomized Trials
Belinda E. Kiely, Yu Yang Soon, Martin H.N. Tattersall, and Martin R. Stockler
See accompanying editorial on page 347
From the National Health and Medical
Research Council Clinical Trials Centre
and Sydney Medical School, University
of Sydney; Sydney Cancer Centre-Royal
Prince Alfred and Concord Repatriation
General Hospitals, Sydney, New South
Wales, Australia.
Submitted May 6, 2010; accepted
October 22, 2010; published online
ahead of print at www.jco.org on
December 28, 2010.
Supported by a National Breast Cancer
Foundation PhD Scholarship (B.E.K.), a
National Health and Medical Research
Council Clinical Trials Centre PhD Scholarship, and a Cancer Institute New
South Wales Research Scholar Award.
Authors’ disclosures of potential conflicts of interest and author contributions are found at the end of this
article.
Corresponding author: Martin R. Stockler,
MD, NHMRC Clinical Trials Centre,
Locked Bag 77, Camperdown, NSW,
1450, Australia; e-mail: martin.stockler@
sydney.edu.au.
© 2010 by American Society of Clinical
Oncology
0732-183X/11/2904-456/$20.00
DOI: 10.1200/JCO.2010.30.2174
A
B
S
T
R
A
C
T
Purpose
To estimate scenarios for survival for women with metastatic breast cancer (MBC) who are
starting chemotherapy.
Patients and Methods
We sought randomized, first-line chemotherapy trials for MBC published from 1999 to 2009.
We recorded median progression-free survival (PFS) and median overall survival (OS) and
extracted the following percentiles (represented scenario) from each OS curve: 90th (worstcase), 75th (lower-typical), 25th (upper-typical), and 10th (best-case). We also estimated these
scenarios for each OS curve by multiplying its median by four simple multiples: 0.25 (worst-case),
0.5 (lower-typical), 2 (upper-typical), and 3 (best-case). Estimates were deemed accurate if they
were within 0.75 to 1.33 times the actual value.
Results
From 36 trials (13,083 women), the mean for median PFS was 7.6 months (interquartile range
[IQR], 6.0 to 9.0 months), the mean for median OS was 21.7 months (IQR,18.2 to 24.0 months),
and the mean for the ratio of median OS to median PFS was 3.0 (IQR, 2.4 to 3.5). The mean for
each OS scenario was worst-case, 6.3 months (IQR, 4.8 to 7.5 months); lower-typical, 11.9
months (IQR, 9.9 to 13.2 months); upper-typical, 36.2 months (IQR, 31.1 to 41.3 months); and
best-case, 55.8 months (IQR, 47.5 to 60.2 months). Simple multiples of the median gave accurate
estimates of the worst-case scenario in 73% of OS curves, lower-typical in 97%, upper-typical in
95%, and best-case in 96%. OS was longer in trials with higher proportions of estrogen
receptor–positive tumors (P ⫽ .001) and in trials of trastuzumab-treated human epidermal growth
factor receptor 2–positive tumors (P ⫽ .001).
Conclusion
Simple multiples of an OS curve’s median can accurately estimate typical (half to double the
median), best-case (triple the median), and worst-case (one quarter of the median) scenarios
for survival.
J Clin Oncol 29:456-463. © 2010 by American Society of Clinical Oncology
INTRODUCTION
Metastatic breast cancer (MBC) is a heterogeneous
disease that is essentially incurable. Most women
with MBC receive cytotoxic chemotherapy at some
stage of their illness. When starting chemotherapy
for MBC, women require an understanding of how
their cancer and its treatment will affect their life
expectancy. Surveys of patients with advanced cancer suggest that all want information about their
future,1 most want information about their likely
survival time, and many want specific estimates of
456
best-case, average, and worst-case scenarios for survival.2 Information about how to estimate or describe these scenarios is scarce, so it is no surprise
that cancer clinicians are often reluctant to estimate
and discuss life expectancy.3,4
Survival data from groups of patients with similar characteristics (demographic, disease, and treatment) would be an ideal source of prognostic
information but those data are rare. Clinical trials
are readily available, easily accessible, and often provide detailed information about survival. Overall
survival (OS) curves from clinical trials are usually
© 2010 by American Society of Clinical Oncology
Information downloaded from jco.ascopubs.org and provided by at UNIVERSITY LIBRARY UTRECHT on May 2, 2011 from
Copyright © 2011 American Society
of Clinical Oncology. All rights reserved.
143.121.239.115
Estimating Scenarios for Survival for Metastatic Breast Cancer
summarized by their median, and although median survival may be
the best single-number description of a survival curve, it is probably
not the best way of explaining prognosis to a patient. Providing patients with a single-point estimate of survival based on the median
implies unwarranted precision and leaves little room for hope. We
previously proposed that providing estimates for the typical range,
best-case scenarios, and worst-case scenarios for OS might be a better
way of communicating life expectancy and that the percentiles from
an OS curve are a useful basis for estimating these scenarios.5
The purpose of this study was to find and summarize survival
data from recent, randomized trials of first-line chemotherapy for
MBC that would help clinicians estimate and describe prognosis for
their patients in this situation.
PATIENTS AND METHODS
We searched the specialized register maintained by the Secretariat of the
Cochrane Breast Cancer Group (CBCG)6 for references coded as “advanced
breast cancer” and “chemotherapy” published from January 1999 to June 2009
inclusive. The results were supplemented by searches of MEDLINE and the
reference list from a recently published systematic review of MBC chemotherapy trials.7
Two authors independently screened the references and selected trials
meeting the following inclusion criteria: randomized trial comparing at least
two different chemotherapy regimens being given as first-line chemotherapy
for women with MBC, including at least 90 patients enrolled per treatment
arm, and Kaplan-Meier curves for OS.
We recorded the year of publication, number of treatment arms, names
and schedules of the chemotherapy agents, number of participants, median
follow-up, patient demographics, and tumor characteristics for each trial. We
recorded the median progression-free survival (PFS) and median OS for each
treatment group.
Each OS curve was independently traced by two authors by using UNSCAN-IT graph digitizing software (Silk Scientific, Orem, UT).8 The median
and following percentiles (represented scenarios) were extracted from each OS
curve: 90th (worst-case), 75th (lower-typical), 25th (upper-typical), and 10th
(best-case; Fig 1). One-, 2-, and 5-year survival percentages were also recorded.
Discrepancies were resolved by repeated measurement and discussion.
We expected the shape of each OS curve to be approximately exponential
allowing its percentiles to be estimated by simple multiples of its median. On
the basis of our previous work5,9 and the fact that clinical trial patients are
relatively more homogeneous than an unselected clinical population that
Proportion Surviving
90th percentile - worst-case scenario
75th percentile
0.8
0.6
Typical
scenario
Median
0.4
25th percentile
10th percentile best-case
scenario
0.2
0
1
2
3
4
5
Time Since Study Entry (years)
Fig 1. Survival curve percentiles and their corresponding scenarios.
www.jco.org
6
includes a variety of cancers, we hypothesized a priori that the appropriate
simple multiples of the median for each percentile would be 0.25 for the 90th,
0.5 for the 75th, 2 for the 25th, and 3 for the 10th. These values were less
extreme than those for a true exponential distribution in which the 90th
percentile is 0.17 times the median, and the 10th percentile is 3.3 times the
median. We also decided a priori that an estimate would be deemed accurate if
it was within 0.75 to 1.33 times the actual value, our arbitrary range for
reasonable accuracy.
We hypothesized that longer median OS would be observed in trials with
the following characteristics: higher proportion of women with estrogen receptor (ER) –positive tumors, lower proportion of premenopausal women,
lower proportion with visceral metastases, higher proportion with Eastern
Cooperative Oncology Group (ECOG) performance status ⬍ 2, lower proportion who had received prior adjuvant chemotherapy, use of a taxane, use of
trastuzumab in women with human epidermal growth factor receptor 2
(HER2) –positive tumors, and more recent year of publication. Associations
between these factors and median OS were assessed by using simple linear
regression. The association between median OS and trastuzumab-treated
HER2-positive disease was explored by comparing trials in which all women
were HER2-positive and received trastuzumab with trials in which HER2
status was not checked and no women received trastuzumab. The association
between median OS and use of a taxane was explored by comparing trials in
which all women received a taxane with trials in which no women received
a taxane.
RESULTS
The CBCG register search identified 284 references, and 212 of these
related to randomized chemotherapy trials in MBC. Because many of
these studies were not about first-line therapy, were small, or were
abstracts only with no survival curves, only 27 trials were eligible for
this review. The reasons for exclusion are summarized in Figure 2.
Supplementary searches of MEDLINE and reference lists identified an
additional nine trials. The 36 selected trials included 78 survival curves
and 13,083 patients. Thirty-four studies were published in the peerreviewed literature,10-43 and two were published as abstracts only.44,45
A statistically significant difference in PFS was reported in 13 trials
(36%),10-13,17,22,30,31,34,37-39,43 and a significant difference in OS was
reported in seven trials (19%).11,17,23,27,36,38,43
The characteristics of the trials, including the chemotherapy
agents used, are summarized in Table 1. The median follow-up for the
36 trials ranged from 10 months to 102 months with a median of 29
months. Most trials had two treatment groups, and the majority of
women enrolled were postmenopausal with visceral metastases and an
ECOG performance status of 0 or 1. Approximately 50% of women in
these trials had ER-positive tumors. In six trials, trastuzumab was
given with chemotherapy to patients with HER2-positive tumors in at
least one treatment group, and there were two trials in which all
patients enrolled were HER2-positive and all received trastuzumab
with chemotherapy. One trial had not reached median OS in one of its
treatment groups at the time of reporting.45 Similarly, the follow-up
was insufficient for the 25th percentile and 10th percentile to be
obtained from the OS curves in 13 and 27 trials, respectively.
The mean value and interquartile range (IQR) for median PFS
was 7.6 months (IQR, 6.0 to 9.0 moths), for median postprogression
survival was 14.0 months (IQR, 10.8 to 15.6 months), for median OS
was 21.7 months (IQR, 18.2 to 24.0 months), and for the ratio of
median OS to median PFS was 3.0 (range, 2.4 to 3.5). The mean value
for 1-year survival was 73% (IQR, 69% to 78%), and the mean value
for 2-year survival was 45% (IQR, 38% to 50%). The mean value for
© 2010 by American Society of Clinical Oncology
Information downloaded from jco.ascopubs.org and provided by at UNIVERSITY LIBRARY UTRECHT on May 2, 2011 from
Copyright © 2011 American Society
of Clinical Oncology. All rights reserved.
143.121.239.115
457
Kiely et al
Table 1. Characteristics of the 36 Included Trials
Trials With
Information
on Patient
Characteristics
References
(n = 284)
Trials
Characteristic
Insufficient information
(n = 30; 10.6%)
Not randomized
chemotherapy trials in MBC
(n = 67; 23.6%)
Duplicate references for the
same study
(n = 22; 7.7%)
Not first line
(n = 71; 25%)
< 90 patients/arm
(n = 48; 16.9%)
No OS curve
(n = 19; 6.7%)
Total included
(n = 27; 9.5%)
Fig 2. Reasons for exclusion of trials. MBC, metastatic breast cancer; OS,
overall survival.
5-year survival was 12% (IQR, 7% to 17%), but this information was
available in only 14 trials (39%).
The distribution of the survival times for each of the selected
percentiles (scenarios) is shown in Figure 3. Each distribution was
roughly symmetrical with a skew to the right (toward longer survival
times). Most values for each scenario fell within a relatively narrow
range, indicating the remarkable similarity of survival times among
these trials. The mean value for each of the scenarios was worst-case,
6.3 months (IQR, 4.8 to 7.5 months); lower-typical, 11.9 months
(IQR, 9.9 to 13.2 months); upper-typical, 36.2 months (IQR, 31.1 to
458
© 2010 by American Society of Clinical Oncology
Trial characteristics
Year of publication
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
No. of treatment groups compared per trial
2
3
4
Agents used in the 78 treatment groups
Anthracycline
Taxane
Cyclophosphamide
Fluorouracil
Trastuzumab
Carboplatin
Methotrexate
Gemcitabine
Vinorelbine
Lapatinib
Bevacizumab
Capecitabine
Patient characteristics
No. of patients per group
Median
Range
Age, years
Median
Range
Follow-up, months
Median
Range
Adjuvant chemotherapy, %
Median
Range
ECOG performance status ⬍ 2, %
Median
Range
ER-positive tumors, %
Median
Range
Visceral metastases, %
Median
Range
Premenopausal patients, %
Median
Range
Palliative endocrine therapy, %
Median
Range
No.
%
3
3
6
3
3
5
3
3
1
4
2
8
8
17
8
8
14
6
8
3
11
6
31
4
1
86
11
3
50
40
27
18
12
5
3
2
2
1
1
1
64
51
35
23
15
6
4
3
3
1
1
1
No.
%
36
100
32
89
23
64
32
89
30
83
28
78
27
75
18
50
11
31
143
90-385
55
50-60
29
10-102
41
4-100
91
66-100
47
16-73
65
40-83
31
0-54
31
4-54
Abbreviations: ECOG, Eastern Cooperative Oncology Group; ER, estrogen receptor.
JOURNAL OF CLINICAL ONCOLOGY
Information downloaded from jco.ascopubs.org and provided by at UNIVERSITY LIBRARY UTRECHT on May 2, 2011 from
Copyright © 2011 American Society
of Clinical Oncology. All rights reserved.
143.121.239.115
Estimating Scenarios for Survival for Metastatic Breast Cancer
Frequency
Distribution of Worst Case (90th Percentile) From 78 Curves
30
24
18
12
6
0
0
6
12
18
24
30
36
42
48
54
60
66
72
78
84
90
78
84
90
Time (months)
Frequency
Distribution of Lower Typical (75th Percentile) From 78 Curves
30
24
18
12
6
0
0
6
12
18
24
30
36
42
48
54
60
66
72
Time (months)
Frequency
Distribution of Median Survival From 77 Curves
30
24
18
12
6
0
Fig 3. Distribution
for survival.
0
6
12
18
24
30
36
42
48
54
60
66
72
78
84
90
78
84
90
78
84
90
of
scenarios
Time (months)
Frequency
Distribution of Upper Typical (25th Percentile) Survival From 58 Curves
30
24
18
12
6
0
0
6
12
18
24
30
36
42
48
54
60
66
72
Time (months)
Frequency
Distribution of Best Case (10th Percentile) Survival From 26 Curves
30
24
18
12
6
0
0
6
12
18
24
30
36
42
48
54
60
66
72
Time (months)
41.3 months); and best-case, 55.8 months (IQR, 47.5 to 60.2 months).
The median and mean times for each percentile (scenario) were almost identical.
Simple multiples of the median provided accurate estimates of
the selected percentiles (scenarios) for most survival curves: 0.25 ⫻
median accurately estimated the 90th percentile (worst-case) in 73%
of curves, 0.5 ⫻ median accurately estimated the 75th percentile
(lower-typical) in 97% of curves, 2 ⫻ median accurately estimated the
25th percentile (upper-typical) in 95% of curves, and 3 ⫻ median
accurately estimated the 10th percentile (best-case) in 96% of curves.
www.jco.org
The accuracy of using simple multiples of the median survival was
independent of the duration of median survival.
Table 2 summarizes associations between the characteristics
of each trial and its median OS. Longer median OS was associated
with using trastuzumab to treat women with HER2-positive tumors, with using taxane-based chemotherapy, and with having a
higher proportion of women with ER-positive tumors. The median OS in the two trials of women with HER2-positive tumors
receiving trastuzumab was 16.2 months longer than the median OS
in the 29 trials of women with tumors of unknown HER2 status not
© 2010 by American Society of Clinical Oncology
Information downloaded from jco.ascopubs.org and provided by at UNIVERSITY LIBRARY UTRECHT on May 2, 2011 from
Copyright © 2011 American Society
of Clinical Oncology. All rights reserved.
143.121.239.115
459
Kiely et al
Table 2. Associations Between Trial Characteristics and Median Survival
Trial Characteristic
Parameter
Estimateⴱ
All patients were HER2-positive and receiving trastuzumab†
All patients receiving taxane chemotherapy‡
Percentage ER positive
Percentage ECOG performance status ⬍ 2
Year of publication
Percentage of patients with adjuvant chemotherapy
Percentage of premenopausal patients
Percentage of patients with visceral disease
Median age of patients, years
16.2
5.8
0.3
0.2
0.5
0.1
0.1
⫺0.05
0.02
95% CI
Pearson’s
Correlation
Coefficient
P
7.4 to 25.1
1.7 to 9.9
0.1 to 0.4
⫺0.1 to 0.5
⫺0.2 to 1.3
⫺0.05 to 0.2
⫺0.1 to 0.4
⫺0.3 to 0.2
⫺1.0 to 0.9
0.57
0.50
0.58
0.27
0.24
0.24
0.29
⫺0.09
0.01
.001
.007
.001
.15
.16
.19
.24
.66
.97
Abbreviations: HER2, human epidermal growth factor receptor 2; ER, estrogen receptor; ECOG, Eastern Cooperative Oncology Group.
ⴱ
Number of months difference in median overall survival associated with a one-unit increase in trial characteristic.
†Compared with trials in which all patients had HER2 status of unknown/negative and received no trastuzumab.
‡Compared with trials in which no patients received taxanes.
receiving trastuzumab (95% CI, 7.4 to 25.1 months; P ⫽ .001).
Median OS in 14 trials in which all patients received a taxane was
5.8 months longer than in 14 trials in which no patients received a
taxane (95% CI, 1.7 to 9.9 months; P ⫽ .007). For every 10%
increase in the proportion of women with ER-positive tumors, the
median OS increased by 2.5 months (95% CI, 1.1 to 3.8 months;
P ⫽ .001). No other trial characteristics were significantly associated with survival. There were insufficient data to perform reliable
multivariable analyses.
DISCUSSION
OS curves were remarkably similar in these trials of first-line chemotherapy for MBC published in the last 10 years. For most of the OS
curves, simple multiples of each curve’s median provided accurate
estimates of its percentiles, corresponding to worst-case, typical, and
best-case scenarios. This meant that for most OS curves, the survival
time for the 10% of patients in the worst-case scenario could be
accurately estimated as about one quarter of its median or less, the
survival of the middle 50% of patients (75th to 25th percentiles, typical
scenario) was accurately estimated as ranging from about half to
double its median, and the survival time for the 10% of patients in the
best-case scenario was accurately estimated as about three times its
median or more. In these trials of first-line chemotherapy for MBC,
the median OS was typically about three times the median PFS. These
observations provide a basis for simple rules of thumb that doctors can
use to help estimate and explain life expectancy to women with MBC
who are starting chemotherapy.
On the basis of prior work,5 we expected and found that the
survival curves in this review were flatter at the ends and steeper in the
middle than a true exponential curve, and the less extreme multiples of the median were thus appropriate for estimating the 10th
and 90th percentiles.
We expressed the accuracy of using simple multiples of the
median to estimate other percentiles of a given survival curve in
terms of how often the estimate was within 0.75 to 1.33 times the
actual value. This criterion for accuracy is arbitrary but was specified a priori on the basis of clinical judgment and is symmetrical on
a logarithmic scale, as is appropriate for a ratio of two values
460
© 2010 by American Society of Clinical Oncology
(estimated:observed). Accuracy was lowest for the worst-case scenario in which multiplying the median by 0.25 tended to underestimate survival. This underestimation probably reflects the
exclusion from trials of patients with an obviously poor prognosis,
for example by requiring a life expectancy of at least 3 months as an
inclusion criterion. The simple multiple of 0.25 is probably still
reasonable for routine clinical practice in which patients with a
shorter life expectancy are not excluded.
This review has important clinical implications. Questions about
life expectancy are frequently asked and are difficult to answer, and
many patients have a poor understanding of their prognosis.46 There
is ample literature providing guidance on how to communicate about
prognosis2,47-49 but little on how to estimate life expectancy. This lack
of guidance and information is one of the reasons that has been
proposed for doctors’ reluctance to discuss prognosis.4,50 On the basis
of this review and our previous work, we suggest that questions such as
“How long have I got?” are best answered by providing worst-case,
typical, and best-case scenarios that are based on simple multiples of
the estimated median survival of a group of similar patients.
The ideal source of an estimated median survival for a group of
similar patients would be a population-based study of similar patients
using similar treatments. In the absence of such ideal data, pertinent
clinical trials provide a pragmatic and readily available starting point.
However, patients in clinical trials are not representative of those seen
in routine clinical practice, so adjustments may be required to account
for differences that might influence their outcomes. These adjustments may only need to be small for the majority of patients who
would meet all or most eligibility criteria for clinical trials. For example, the distribution of median survival times in this review was quite
narrow, with 90% of values falling between 16 and 28 months. Patients
with characteristics associated with a poor prognosis, such as brain
metastases or poor performance status, are often excluded from clinical trials and would require larger adjustments, or preferably different
sources of survival data on which to base their survival estimates.
Additional studies are needed to determine how survival data from
trials should be adjusted for individual patients.
Although surveys suggest that the majority of patients with advanced cancer want specific information on the best-case, worst-case,
and typical scenarios for their survival,2 patient understanding and
JOURNAL OF CLINICAL ONCOLOGY
Information downloaded from jco.ascopubs.org and provided by at UNIVERSITY LIBRARY UTRECHT on May 2, 2011 from
Copyright © 2011 American Society
of Clinical Oncology. All rights reserved.
143.121.239.115
Estimating Scenarios for Survival for Metastatic Breast Cancer
satisfaction when prognostic information is presented this way has not
been assessed, and research in this area is needed. Many patients want
realism and honesty from their doctor when discussing prognosis but,
at the same time, most also want hope and optimism.48,51,52 We
believe estimating and explaining typical, best-case, and worst-case
scenarios for survival is preferable to presenting a single-point estimate of the median survival and can help clinicians convey realistic
hope. The typical scenario is explained as the most likely outcome,
representative of the middle 50% of patients. The best-case scenario is
less likely, representing the 10% of patients with the best prognosis,
but it is still a possible outcome that can be hoped for. This hope is
balanced by presenting the equally possible worst-case scenario, which
allows the patient to prepare for the worst. Research is needed to test
this belief.
We have not attempted to estimate survival from the time
of diagnosis of breast cancer or from the time of diagnosis of
metastatic disease but acknowledge that these are two other
common times for patients to request life expectancy information. The patients in this review will be quite different in terms
of their survival time and treatments before beginning palliative
chemotherapy, but we have found that after starting chemotherapy, survival times are quite similar.
We found that median OS was about three times median PFS
(IQR, 2.4 to 3.5 months), as did Saad et al53 in their recent review. This
is an observation about summary data from groups of patients, but it
raises the hypothesis that doubling the observed time to progression
might provide a useful estimate of postprogression survival. If corroborated, a rule of thumb based on this observation might facilitate
discussion of prognosis after progression on first-line chemotherapy.
The relationship between PFS and OS might also be helpful for planning future studies.
This review was not designed to determine individual prognostic
and predictive factors in MBC, and individual patient data would be
required for this. Prognostic indexes including clinical variables and
other measures such as circulating tumor cells can provide more
precise information on life expectancy for an individual patient; however, such indexes will never be 100% accurate. Our suggestion to
provide patients with estimates for three scenarios could supplement
such measures. There will always be best-case, worst-case, and typical
scenarios, and it is important to explain these to patients rather than
presenting them with a one-point estimate from a prognostic index. It
is possible that rules of thumb based on simple multiples could be used
to estimate these scenarios from a prognostic index estimate rather
than from the median.
We did not seek to determine the most effective first-line treatment for MBC; this would require randomized comparisons. Our
indirect comparisons of trial characteristics and survival times raise
some interesting hypotheses and may help doctors select the most
appropriate reference groups on which to base their survival estimates.
The treatments associated with the longest OS were trastuzumab in
HER2-positive tumors and taxanes. To improve the accuracy of survival estimates for women about to begin trastuzumab and chemotherapy or taxane-based chemotherapy, we suggest doctors start with
an OS curve from a pertinent trial that includes these treatments.
Although our observations are interesting, they should not be taken as
definitive evidence that trastuzumab or taxanes prolong OS. Other
characteristics of the patients, treatments, and trials may be responsible, and data from prospective randomized trials are required for such
www.jco.org
conclusions. There are more appropriate sources of data for determining the effects of trastuzumab and taxanes on survival. The randomized phase III trial by Slamon et al39 reported that the addition of
trastuzumab to chemotherapy for women with HER2-positive MBC
improved median OS by approximately 5 months. The Cochrane
meta-analysis of taxanes in MBC54 reported that taxane-containing
regimens significantly prolonged OS; however, when the analysis was
restricted to trials of first-line chemotherapy, as in our review, there
was no longer a survival benefit.
Trials with a higher proportion of women with ER-positive
MBC were also associated with longer median OS. Because these
patients were not receiving endocrine therapy and many would
have exhausted their options for endocrine therapy, this observation supports ER-positive tumors having a more favorable biology
regardless of treatment.
Whether using simple multiples of the median to calculate estimates of scenarios for survival can be applied to other advanced
cancers needs to be determined, but we are encouraged by similar
results for advanced non–small-cell lung cancer using the same simple
multiples (0.25, 0.5, 2, and 3).9 The accuracy of these rules of thumb
will soon be tested by using clinicians’ predictions of survival at baseline in a randomized trial.55
The strengths of this review are that it is comprehensive and
applies to contemporary treatment for MBC. It provides simple rules
of thumb for doctors to use as a basis to estimate survival times.
Although survival estimates will need to be adjusted to account for
individual characteristics of the patient, tumor, and treatment, such
adjustments probably need only be small for the majority of patients.
The main limitation of this review is that it includes only patients
in clinical trials who have a better prognosis than patients treated
similarly in routine clinical practice. There are few data available on
the survival of patients outside clinical trials. We suggest that doctors
use median OS from a pertinent clinical trial as a starting point and
then adjust that median according to the characteristics of the individual patient at hand. Data on the upper-typical and best-case scenarios
were not included in many trials because they were published before
these results were mature. This was particularly so in the most recent
trials with the longest survival times such as those in women with
HER2-positive MBC receiving trastuzumab. Publication of survival outcomes after longer follow-up is needed to provide better
founded estimates of long-term survival but, unfortunately, such
publications are rare. We did not record information on the durations of chemotherapy used in these trials or on second- or
subsequent-line chemotherapy used. These factors may affect OS,
but a recent meta-analysis of first-line chemotherapy duration
suggests effects on survival time are likely to be modest.56 Finally,
systematic reviews of this sort are unable to estimate the survival
benefits attributable to chemotherapy.
Another potential limitation of this review is that the treatments
used may not reflect today’s reality, largely because of changes in
adjuvant therapy in the last 10 years. Anthracyclines and taxanes are
less likely to be used as first-line therapy today because many patients
will have received these treatments in the adjuvant setting. Similarly,
more targeted treatment options are becoming available and will
hopefully lead to longer OS times in subgroups of patients. Even if the
median survival times increase with improved treatments, we expect
the multiples of the median to remain accurate for estimating scenarios for survival. This is because it was accurate for most of the 78
© 2010 by American Society of Clinical Oncology
Information downloaded from jco.ascopubs.org and provided by at UNIVERSITY LIBRARY UTRECHT on May 2, 2011 from
Copyright © 2011 American Society
of Clinical Oncology. All rights reserved.
143.121.239.115
461
Kiely et al
survival curves in this review, even though the treatments used in each
trial varied. To improve accuracy, we suggest that clinicians select an
OS curve from a clinical trial in which patients are randomly assigned
to receive similar treatment to that being planned for the patient in
question and use the median from this curve as the basis for estimating scenarios.
OS distributions from recent randomized trials of first-line chemotherapy in MBC were remarkably similar, approximately exponential, and had a median OS of nearly 2 years, on average. Simple
multiples of an OS curve’s median gave good estimates of its percentiles corresponding to worst-case (0.25 times the median), typical (0.5
to 2.0 times the median), and best-case (3.0 times the median) scenarios for nearly all OS curves. Presenting women with estimates for a
range of scenarios for survival should provide prognostic information
that is more accurate, realistic, and hopeful than a single-point estimate and hopefully easier to understand.
REFERENCES
1. Innes S, Payne S: Advanced cancer patients’
prognostic information preferences: A review. Palliat Med 23:29-39, 2009
2. Hagerty RG, Butow PN, Ellis PA, et al: Cancer
patient preferences for communication of prognosis in
the metastatic setting. J Clin Oncol 22:1721-1730, 2004
3. Fallowfield LJ, Jenkins VA, Beveridge HA:
Truth may hurt but deceit hurts more: Communication in palliative care. Palliat Med 16:297-303, 2002
4. Hancock K, Clayton JM, Parker SM, et al:
Truth-telling in discussing prognosis in advanced
life-limiting illnesses: A systematic review. Palliat
Med 21:507-517, 2007
5. Stockler MR, Tattersall MH, Boyer MJ, et al:
Disarming the guarded prognosis: Predicting survival in newly referred patients with incurable cancer. Br J Cancer 94:208-212, 2006
6. The Cochrane Collaboration. http://www
.cochrane.org/
7. Wilcken N, Dear R: Chemotherapy in metastatic breast cancer: A summary of all randomised
trials reported 2000-2007. Eur J Cancer 44:22182225, 2008
8. UN-SCAN-IT Graph Digitizing Software (ed 6).
Orem, UT, Silk Scientific, 1988-2009
9. Alam M, Blinman P, Stockler M: Useful
estimates of survival for patients starting first line
chemotherapy for advanced non-small-cell lung
cancer based on a systematic review of recent
randomised trials. 13th World Conference on
Lung Cancer, San Francisco, CA, July 31-August
4, 2009
10. Ackland SP, Anton A, Breitbach GP, et al:
Dose-intensive epirubicin-based chemotherapy is
superior to an intensive intravenous cyclophosphamide, methotrexate, and fluorouracil regimen in
metastatic breast cancer: A randomized multinational study. J Clin Oncol 19:943-953, 2001
11. Albain KS, Nag SM, Calderillo-Ruiz G, et al:
Gemcitabine plus paclitaxel versus paclitaxel monotherapy in patients with metastatic breast cancer
and prior anthracycline treatment. J Clin Oncol 26:
3950-3957, 2008
12. Andersson M, Madsen EL, Overgaard M, et
al: Doxorubicin versus methotrexate both combined
with cyclophosphamide, 5-fluorouracil and tamoxifen in postmenopausal patients with advanced
breast cancer: A randomised study with more than
10 years follow-up from the Danish Breast Cancer
462
© 2010 by American Society of Clinical Oncology
AUTHORS’ DISCLOSURES OF POTENTIAL CONFLICTS OF
INTEREST
The author(s) indicated no potential conflicts of interest.
AUTHOR CONTRIBUTIONS
Conception and design: Belinda E. Kiely, Martin H.N. Tattersall,
Martin R. Stockler
Collection and assembly of data: Belinda E. Kiely, Yu Yang Soon
Data analysis and interpretation: Belinda E. Kiely, Yu Yang Soon,
Martin H.N. Tattersall, Martin R. Stockler
Manuscript writing: Belinda E. Kiely, Yu Yang Soon, Martin H.N.
Tattersall, Martin R. Stockler
Final approval of manuscript: Belinda E. Kiely, Yu Yang Soon, Martin
H.N. Tattersall, Martin R. Stockler
Cooperative Group—Danish Breast Cancer Cooperative Group (DBCG). Eur J Cancer 35:39-46, 1999
13. Epirubicin-based chemotherapy in metastatic
breast cancer patients: Role of dose-intensity and
duration of treatment. J Clin Oncol 18:3115-3124,
2000
14. Batist G, Ramakrishnan G, Rao CS, et al:
Reduced cardiotoxicity and preserved antitumor
efficacy of liposome-encapsulated doxorubicin
and cyclophosphamide compared with conventional doxorubicin and cyclophosphamide in a randomized, multicenter trial of metastatic breast
cancer. J Clin Oncol 19:1444-1454, 2001
15. Biganzoli L, Cufer T, Bruning P, et al: Doxorubicin and paclitaxel versus doxorubicin and cyclophosphamide as first-line chemotherapy in
metastatic breast cancer: The European Organization for Research and Treatment of Cancer 10961
Multicenter Phase III Trial. J Clin Oncol 20:31143121, 2002
16. Bishop JF, Dewar J, Toner GC, et al: Initial
paclitaxel improves outcome compared with CMFP
combination chemotherapy as front-line therapy in
untreated metastatic breast cancer. J Clin Oncol
17:2355-2364, 1999
17. Bontenbal M, Creemers GJ, Braun HJ, et al:
Phase II to III study comparing doxorubicin and
docetaxel with fluorouracil, doxorubicin, and cyclophosphamide as first-line chemotherapy in patients
with metastatic breast cancer: Results of a Dutch
Community Setting Trial for the Clinical Trial Group
of the Comprehensive Cancer Centre. J Clin Oncol
23:7081-7088, 2005
18. Cassier PA, Chabaud S, Trillet-Lenoir V, et al:
A phase-III trial of doxorubicin and docetaxel versus
doxorubicin and paclitaxel in metastatic breast cancer: Results of the ERASME 3 study. Breast Cancer
Res Treat 109:343-350, 2008
19. Conte PF, Guarneri V, Bruzzi P, et al: Concomitant versus sequential administration of epirubicin
and paclitaxel as first-line therapy in metastatic
breast carcinoma: Results for the Gruppo Oncologico Nord Ovest randomized trial. Cancer 101:704712, 2004
20. Costanza ME, Weiss RB, Henderson IC, et al:
Safety and efficacy of using a single agent or a phase
II agent before instituting standard combination chemotherapy in previously untreated metastatic breast
cancer patients: Report of a randomized study—
Cancer and Leukemia Group B 8642. J Clin Oncol
17:1397-1406, 1999
21. Di Leo A, Gomez HL, Aziz Z, et al: Phase III,
double-blind, randomized study comparing lapatinib
plus paclitaxel with placebo plus paclitaxel as firstline treatment for metastatic breast cancer. J Clin
Oncol 26:5544-5552, 2008
22. Ejlertsen B, Mouridsen HT, Langkjer ST, et al:
Phase III study of intravenous vinorelbine in combination with epirubicin versus epirubicin alone in
patients with advanced breast cancer: A Scandinavian Breast Group Trial (SBG9403). J Clin Oncol
22:2313-2320, 2004
23. Fountzilas G, Dafni U, Dimopoulos MA, et al:
A randomized phase III study comparing three
anthracycline-free taxane-based regimens, as first
line chemotherapy, in metastatic breast cancer: A
Hellenic Cooperative Oncology Group study. Breast
Cancer Res Treat 115:87-99, 2009
24. Fountzilas G, Kalofonos HP, Dafni U, et al:
Paclitaxel and epirubicin versus paclitaxel and carboplatin as first-line chemotherapy in patients with
advanced breast cancer: A phase III study conducted by the Hellenic Cooperative Oncology
Group. Ann Oncol 15:1517-1526, 2004
25. Harris L, Batist G, Belt R, et al: Liposomeencapsulated doxorubicin compared with conventional doxorubicin in a randomized multicenter trial
as first-line therapy of metastatic breast carcinoma.
Cancer 94:25-36, 2002
26. Heidemann E, Stoeger H, Souchon R, et al: Is
first-line single-agent mitoxantrone in the treatment
of high-risk metastatic breast cancer patients as
effective as combination chemotherapy? No difference in survival but higher quality of life were found
in a multicenter randomized trial. Ann Oncol 13:
1717-1729, 2002
27. Jassem J, Pieńkowski T, Płuzańska A, et al:
Doxorubicin and paclitaxel versus fluorouracil, doxorubicin, and cyclophosphamide as first-line therapy
for women with metastatic breast cancer: Final
results of a randomized phase III multicenter trial.
J Clin Oncol 19:1707-1715, 2001
28. Katsumata N, Watanabe T, Minami H, et al:
Phase III trial of doxorubicin plus cyclophosphamide
(AC), docetaxel, and alternating AC and docetaxel
as front-line chemotherapy for metastatic breast cancer: Japan Clinical Oncology Group trial (JCOG9802).
Ann Oncol 20:1210-1215, 2009
29. Langley RE, Carmichael J, Jones AL, et al:
Phase III trial of epirubicin plus paclitaxel compared
with epirubicin plus cyclophosphamide as first-line
chemotherapy for metastatic breast cancer: United
JOURNAL OF CLINICAL ONCOLOGY
Information downloaded from jco.ascopubs.org and provided by at UNIVERSITY LIBRARY UTRECHT on May 2, 2011 from
Copyright © 2011 American Society
of Clinical Oncology. All rights reserved.
143.121.239.115
Estimating Scenarios for Survival for Metastatic Breast Cancer
Kingdom National Cancer Research Institute trial
AB01. J Clin Oncol 23:8322-8330, 2005
30. Miller K, Wang M, Gralow J, et al: Paclitaxel
plus bevacizumab versus paclitaxel alone for metastatic breast cancer. N Engl J Med 357:2666-2676,
2007
31. Nabholtz JM, Falkson C, Campos D, et al:
Docetaxel and doxorubicin compared with doxorubicin and cyclophosphamide as first-line chemotherapy for metastatic breast cancer: Results of
a randomized, multicenter, phase III trial. J Clin
Oncol 21:968-975, 2003
32. Namer M, Soler-Michel P, Turpin F, et al:
Results of a phase III prospective, randomised trial,
comparing mitoxantrone and vinorelbine (MV) in
combination with standard FAC/FEC in front-line
therapy of metastatic breast cancer. Eur J Cancer
37:1132-1140, 2001
33. O’Brien ME, Wigler N, Inbar M, et al: Reduced
cardiotoxicity and comparable efficacy in a phase
III trial of pegylated liposomal doxorubicin HCl
(CAELYX/Doxil) versus conventional doxorubicin for
first-line treatment of metastatic breast cancer. Ann
Oncol 15:440-449, 2004
34. Paridaens R, Biganzoli L, Bruning P, et al:
Paclitaxel versus doxorubicin as first-line singleagent chemotherapy for metastatic breast cancer: A
European Organization for Research and Treatment
of Cancer Randomized Study with cross-over. J Clin
Oncol 18:724-733, 2000
35. Parnes HL, Cirrincione C, Aisner J, et al:
Phase III study of cyclophosphamide, doxorubicin,
and fluorouracil (CAF) plus leucovorin versus CAF for
metastatic breast cancer: Cancer and Leukemia
Group B 9140. J Clin Oncol 21:1819-1824, 2003
36. Reyno L, Seymour L, Tu D, et al: Phase III
study of N,N-diethyl-2-[4-(phenylmethyl) phenoxy]
ethanamine (BMS-217380-01) combined with doxorubicin versus doxorubicin alone in metastatic/
recurrent breast cancer: National Cancer Institute of
Canada Clinical Trials Group Study MA.19. J Clin
Oncol 22:269-276, 2004
37. Robert N, Leyland-Jones B, Asmar L, et al:
Randomized phase III study of trastuzumab, paclitaxel, and carboplatin compared with trastuzumab
and paclitaxel in women with HER-2-overexpressing
metastatic breast cancer. J Clin Oncol 24:27862792, 2006
38. Seidman AD, Berry D, Cirrincione C, et al:
Randomized phase III trial of weekly compared with
every-3-weeks paclitaxel for metastatic breast cancer, with trastuzumab for all HER-2 overexpressors
and random assignment to trastuzumab or not in
HER-2 nonoverexpressors: Final results of Cancer
and Leukemia Group B protocol 9840. J Clin Oncol
26:1642-1649, 2008
39. Slamon DJ, Leyland-Jones B, Shak S, et al:
Use of chemotherapy plus a monoclonal antibody
against HER2 for metastatic breast cancer that
overexpresses HER2. N Engl J Med 344:783-792,
2001
40. Sledge GW Jr, Hu P, Falkson G, et al:
Comparison of chemotherapy with chemohormonal therapy as first-line therapy for metastatic,
hormone-sensitive breast cancer: An Eastern Cooperative Oncology Group study. J Clin Oncol
18:262-266, 2000
41. Sledge GW, Neuberg D, Bernardo P, et al:
Phase III trial of doxorubicin, paclitaxel, and the
combination of doxorubicin and paclitaxel as frontline chemotherapy for metastatic breast cancer: An
intergroup trial (E1193). J Clin Oncol 21:588-592,
2003
42. Fountzilas G, Papadimitriou C, Dafni U, et al:
Dose-dense sequential chemotherapy with epirubicin and paclitaxel versus the combination, as firstline chemotherapy, in advanced breast cancer: A
randomized study conducted by the Hellenic Cooperative Oncology Group. J Clin Oncol 19:2232-2239,
2001
43. Marty M, Cognetti F, Maraninchi D, et al:
Randomized phase II trial of the efficacy and safety
of trastuzumab combined with docetaxel in patients
with human epidermal growth factor receptor
2-positive metastatic breast cancer administered as
first-line treatment: The M77001 study group. J Clin
Oncol 23:4265-4274, 2005
44. Lueck H, Minckwitz GV, Du Bois A, et al:
Epirubicin/paclitaxel (EP) vs. capecitabine/paclitaxel
(XP) in first-line metastatic breast cancer (MBC): A
prospective, randomized multicentre phase III study
of the AGO breast cancer study group. J Clin Oncol
24:18s, 2006 (suppl; abstr 517)
45. Forbes JF, Kennedy J, Pienkowski T, et al:
BCIRG 007: Randomized phase III trial of trastuzumab plus docetaxel with or without carboplatin
first line in HER2 positive metastatic breast cancer
(MBC). J Clin Oncol 24:18s, 2006 (suppl; abstr
LBA516)
46. Barnett MM: Does it hurt to know the worst?
Psychological morbidity, information preferences
and understanding of prognosis in patients with
advanced cancer. Psychooncology 15:44-55, 2006
47. Clayton JM, Hancock KM, Butow PN, et al:
Clinical practice guidelines for communicating prognosis and end-of-life issues with adults in the advanced stages of a life-limiting illness, and their
caregivers. Med J Aust 186:S77, S79, S83-S108,
2007
48. Hagerty RG, Butow PN, Ellis PM, et al: Communicating with realism and hope: Incurable cancer
patients’ views on the disclosure of prognosis. J Clin
Oncol 23:1278-1288, 2005
49. Helft PR: Necessary collusion: Prognostic
communication with advanced cancer patients.
J Clin Oncol 23:3146-3150, 2005
50. Lamont EB, Christakis NA: Complexities in
prognostication in advanced cancer: “To help them
live their lives the way they want to.” JAMA 290:
98-104, 2003
51. Kutner JS, Steiner JF, Corbett KK, et al: Information needs in terminal illness. Soc Sci Med 48:
1341-1352, 1999
52. Davey HM, Butow PN, Armstrong BK: Cancer
patients’ preferences for written prognostic information provided outside the clinical context. Br J
Cancer 89:1450-1456, 2003
53. Saad ED, Katz A, Buyse M: Overall survival
and post-progression survival in advanced breast
cancer: A review of recent randomized clinical trials.
J Clin Oncol 28:1958-1962, 2010
54. Ghersi D, Wilcken N, Simes J, et al: Taxane
containing regimens for metastatic breast cancer.
Cochrane Database Syst Rev 2:CD003366, 2005
55. Stockler MR, O’Connell R, Nowak AK, et al:
Effect of sertraline on symptoms and survival in
patients with advanced cancer, but without major
depression: A placebo-controlled double-blind randomised trial. Lancet Oncol 8:603-612, 2007
56. Gennari A, Sormani M, Nanni O, et al: Impact
of first-line chemotherapy duration in metastatic
breast cancer: A systematic review. J Clin Oncol
28:15s, 2010 (suppl; abstr 1023)
■ ■ ■
www.jco.org
© 2010 by American Society of Clinical Oncology
Information downloaded from jco.ascopubs.org and provided by at UNIVERSITY LIBRARY UTRECHT on May 2, 2011 from
Copyright © 2011 American Society
of Clinical Oncology. All rights reserved.
143.121.239.115
463