Key words
early breast cancer - neoadjuvant - chemotherapy - pathological complete remission
(pCR) - patient preference
Schlüsselwörter
frühes Mammakarzinom - neoadjuvant - Chemotherapie - pathologische Komplettremission
(pCR) - Patientenpräferenz
Introduction
Besides surgery, the therapy concept for patients with primary early breast cancer
comprises chemotherapy in the presence of certain risk factors and, in HER2-positive
cases, a targeted therapy against HER2 [1], [2]. In the mean time we know that chemotherapy for early breast cancer, irrespective
of an adjuvant or a neoadjuvant application, can achieve comparable long-term results
with regard to disease-free survival and overall survival [3], [4]. As little as 15 years ago the rationale for neoadjuvant chemotherapy (NACT) consisted
mainly in the preoperative reduction in volume of large primary tumours to achieve
a better operability or to make a breast-preserving operation possible [5]. There is still a significant indication for NACT in cases of inflammatory tumours
because of their aggressiveness and of inoperable or large operable breast cancers
in dependence on their tumour biology/characteristics. But furthermore, today neoadjuvant
and adjuvant chemotherapies are equal alternative therapeutic options for early breast
cancer [1], [2], [6], [7].
Basically NACT can be carried out when the indication for chemotherapy exists even
before an operation [1], [2], [7]. Apart from improving the chances for a breast-preserving operation, it opens up
further options. In contrast to adjuvant chemotherapy, it allows direct in vivo chemosensitivity
testing and thus an evaluation of the response of the tumour and axillary lymph nodes
to the systemic therapy in real time. In addition, it enables an individualisation
of the therapy on the basis of response controls, on the one hand, by changing the
chemotherapy and, on the other hand, by a risk-adapted post-neoadjuvant management
[1]. Finally the response of the patient to a neoadjuvant therapy can provide additional
prognostic information [1], [8].
The most important parameter for evaluating the success of a therapy is the pathological
complete remission (pCR) [9]. This endpoint can directly reflect the effect of a neoadjuvant treatment. It is
important to consider how the respective pCR is defined. The most frequently used
definitions are ypT0 ypN0 (no residual invasive and in situ tumour remains in the
breast and axilla), ypT0/is ypN0 (no residual invasive tumour remains in the breast
and axilla, independent of an in situ component also known as tpCR) and ypT0/is (no
residual invasive tumour remains in the breast, independent of axillary lymph node
attack or an in situ carcinoma).
In various trials and meta-analyses a relationship at the patient level between the
achievement of pCR and a longer overall survival has been demonstrated [4], [10], [11]. Also in a large meta-analysis of 12 trials initiated by the US drug registration
authority FDA with a total of 11 955 patients, an association between achievement
of pCR and an improved, event-free survival was observed. Here, the prognostic significance
was highest for triple negative breast cancers and HER2-positive, HR-negative tumours
treated with trastuzumab. A formal validation of pCR as a surrogate marker for an
improved, event-free survival and overall survival was not possible at the trial level
in this very heterogeneous patient collective [12].
Therapists and the American as well as the European drug registration authorities
consider pCR to be an independent and relevant endpoint for evaluating the clinical
efficacy of neoadjuvant therapy. The present study is intended to investigate which
treatment targets of neoadjuvant therapy the patients themselves consider to be relevant.
Materials and Methods
Patient recruitment
Patient recruitment was by personal contact, internet advertisements, information
flyers to sport groups and apothecaries, direct enrolment from a database and contact
through an already enrolled person.
Interviewers
The interviews were carried out by a total of four female interviewers. They were
permanent employees of a market research company with 5–10 years of experience in
conducting interviews with patients and physicians.
Qualitative survey
Nation-wide in Germany in-depth interviews with 8 patients were conducted with the
aim to confirm the completeness and unambiguous definition of the relevant preferences
and to exclude any overlap between the preferences.
The patients had all undergone neoadjuvant therapy in the form of chemotherapy and,
in the case of HER2-positive disease, a targeted antibody therapy against HER2. Further
prerequisites that had to be confirmed by the patients themselves were an interval
of > 12 months since the primary diagnosis of breast cancer and between > 12 and < 36
months since tumour surgery.
The interviews were held in the patientʼs home and were based on a structured interview
guide. The guideline consisted of 12 main questions with 0–5 subitems and was split
into 2 chapters. Information about the patientʼs medical history and about the relevance
of various features of the neoadjuvant therapy was recorded. The interview lasted
about 60 minutes.
Evaluation of the in-depth interviews
The qualitative interviews were recorded digitally and transcribed from the record.
Evaluation of the interview protocol was done analogously to that of the summarised
qualitative content analysis [13].
Quantitative survey
41 patients were interviewed nation-wide in Germany. The age of the patients was initially
recorded in cumulative categories (under 45, 45–60, 61–70, over 70 years). In a second
step the exact age was recorded. Each patient had to have undergone a neoadjuvant
therapy in the form of chemotherapy as well as, in the case of HER2-positive disease,
a targeted antibody therapy against HER2. Further inclusion criteria that had to be
confirmed by the patients were an interval of > 6 months since the primary diagnosis
of breast cancer and between > 12 and < 36 months since tumour surgery. The required
minimum interval between interview and primary diagnosis was reduced from 12 to 6
months on account of the slow recruitment of the first 8 patients for the qualitative
survey. The interviews were computer-assisted (computer-aided personal interview),
held in the patientʼs home and lasted for about 30 minutes.
The patientsʼ preferences were determined with the help of analytic hierarchy process
(AHP) methods [14], [15], [16]. This multicriterial decision method structures complex decision-making processes.
The hierarchical levels and criteria are presented in [Table 1]. The thematic classification of therapeutic attributes to hierarchical levels was
made on the basis of the qualitative interviews. The criteria were described in analogy
to the terms used by the patients in the qualitative interviews and not in the form
of specialist medical terms.
Table 1 Criteria of the used AHP designs.
|
Hierarchical level 0
|
Hierarchical level 1
|
Hierarchical level 2
|
|
The criteria were determined in 8 qualitative patient interviews and formulated in
the words used by the patients instead of specialist medical terms.
* “Destruction of tumour cells” = pathological complete response (pCR), 2: “Minimisation
of the risk for tumour recurrence” = disease-free survival (DFS), 3: “No reduction
of life expectancy due to the disease” = overall survival (OS).
|
|
1. Efficacy of the neoadjuvant therapy
|
1. Destruction of tumour cells*
2. Minimisation of the risk for tumour recurrence*
3. No reduction of life expectancy due to the disease*
4. Possibility for breast-preserving operation
|
|
|
2. Avoidance of side effects of the neoadjuvant therapy
|
5. Side effects that are stressful for the body
|
1. Fever
2. Diarrhoea
3. Nausea
4. Fatigue
|
|
6. Side effects that cause bodily changes
|
5. Loss of hair
6. Weight gain
|
First of all the patients had to compare the criteria of hierarchical level 1 pair-wise
with one another and then the subcriteria of hierarchical level 2 within the defined
hierarchical model ([Table 1]). In order to increase the attentiveness of the participants and to avoid answer
patterns, 15 pair comparisons per level were doubly randomised. This means that the
order in which the pair comparisons were presented to the patients and the position
of the criterion to be compared were arbitrary. In the process the participant stated
which of the two criteria to be compared was the more important for her and to what
extent this criterion was more important than the other one for her. The evaluation
of importance was done in a semiquantitative manner using a nine-point scale where
1 point is the indifference point and 9 points corresponds to the highest scoring.
The scale steps with points 1, 3, 5, 7 and 0 were verbally designated as identical,
somewhat larger, considerably larger, very much larger and absolutely dominating.
In the data set each pair comparison A vs. B was assigned a value by the answer which
could be 1/9 (= criterion B absolutely dominating) for a large significance for criterion
B, 1/8, 1/7, 1/6, 1/5, 1/4, 1/3 or 1/2, down to 1/1 for equal significance of criteria
A and B. The answers of the 41 female participants were summarised to a sample mean
score. For each pair comparison the geometrical mean for all 41 answers was derived.
In this mathematical process as well as in all other methodological aspects, the study
was oriented on the stipulations of the pilot project from the Institute for Quality
and Economy in the Health Service (Institut für Qualität und Wirtschaftlichkeit im
Gesundheitswesen [IQWiG]) regarding the analytic hierarchy process [17].
The results of the 2 AHP hierarchical levels were evaluated separately. For this the
sample mean scores of all 30 pair comparisons were transferred to two AHP matrices.
With the help of the AHP function of the PMR package in statistics software R, the
criterion weightings and the consistency of the AHP matrix were calculated [18]. Using the consistency ratio according to Saaty [15], [16], we checked whether the preferences revealed by the comparisons were consistent
with the other comparisons [17].
From the weightings of the hierarchical levels 1 and 2 the results of the entire –
comprising three levels (0, 1, 2) – AHP model were determined. For this the hierarchical
level 0 was derived from the sum of the weightings of hierarchical level 1. In the
entire model the individual subcriteria of hierarchical level 2 were contained in
the relative weightings to each other and their sum gave the weighting of hierarchical
level 2.
Results
Qualitative survey
Eight patients aged between 30 and 78 years took part in the qualitative interviews.
Two of these patients had had an HER2-positive breast cancer.
In the interviews it was clear that the efficacy of the neoadjuvant therapy represented
the most important therapeutic criterion for the patients. The patients mentioned
the destruction of tumour cells, reduction of the risk of recurrence, no shortening
of life expectancy due to the disease, and the possibility for breast-preserving surgery
as criteria for the efficacy. The endpoints pCR, disease-free survival and overall
survival from clinical studies are reflected in the qualitative patient survey by
the criteria “destruction of tumour cells”, “minimisation of the risk for tumour recurrence”
and “no shortening of life expectancy due to the disease”.
The side effects of the therapy were also important for the patients, although they
were widely accepted as part of the price for an effective drug. Even so they were
considered to be a burden that the patients would happily do without. The patients
also assessed their quality of life mainly in dependence of the presence of side effects.
Quantitative survey
The results of the quantitative interviews of 41 patients, among whom 15 (36.6 %)
had had HER2-positive disease, were available. The patients came from 13 federal states:
7 from Baden-Württemberg, 3 from Bavaria, 1 from Bremen, 4 from Hamburg, 1 from Hesse,
1 from Mecklenburg-Western Pomerania, 4 from Lower Saxony, 9 from North Rhine-Westphalia,
3 from Saxony, 2 from Saxony-Anhalt, 1 from Schleswig-Holstein and 2 from Thuringia.
27 patients (65.8 %) were between 45 and 70 years old. 14.6 % were older than 70 years
and 19.5 % younger than 45 years. For seven patients we only knew that they were either
younger than 45 years (n = 1), between 45 and 60 years (n = 5) or, respectively, between
61 and 70 years old (n = 1). Apart from this, the interviewers did not get any more
detailed information about their ages. For the other 34 participants in the study
the exact ages were known with a median value of 50 years (range 29–76 years). For
9 patients (22.0 %) the primary diagnosis was made 6–12 months prior to the interview,
for 13 (31.7 %) 13–24 months earlier and for 19 (46.3 %) more than 24 months before
the interview ([Table 2]).
Table 2 Demographic data.
|
Characteristic
|
|
n (%)
|
|
* Median age and range for 34 patients. For 7 patients it was merely known that their
ages were either under 45 years (n = 1), between 45 and 60 years (n = 5) or, respectively,
between 61 and 70 years (n = 1).
|
|
Age*
|
Median (Range), years
|
50 (29–76)
|
|
Interval between primary diagnosis and interview
|
6–12 months
|
9 (22.0)
|
|
13–24 months
|
13 (31.7)
|
|
≥ 24 months
|
19 (46.3)
|
|
HER2 status
|
positive
|
15 (36.6 %)
|
|
negative
|
26 (63.4 %)
|
The efficacy of the neoadjuvant therapy was more important for the patients than the
avoidance of side effects. On comparison of the criteria in the second hierarchical
level the group weightings of all efficacy criteria were above those of side effects.
The elimination of tumour cells, or in medical specialist terminology the pathological
complete remission represented for the patients the most important therapeutic target
on account of its highest group weighting (0.316). The reduction of the risk of recurrence
and thus disease-free survival (0.256) and no shortening of life expectancy due to
the disease (0.241) followed in second and third places before the possibility for
a breast-preserving operation. Avoidance of side effects was considered to be the
least important criterion: group weighting of 0.050 for side effects that cause a
burden for the body, and 0.038 for those that cause bodily changes ([Table 3]).
Table 3 Group weightings in hierarchical level 1.
|
Therapeutic target of neoadjuvant therapy
|
Weighting
|
Consistency ratio
|
|
The weightings were determined from the sample mean values of the pair-wise comparison
of the criteria in hierarchical level 1 by all 41 patients. The consistency ratio
according to Saaty [17] checks whether the individual pair-wise comparisons are consistent with all the
other comparisons.
|
|
Sum of group weightings
|
1.00
|
0.007
|
|
Destruction of tumour cells
|
0.316
|
|
|
Minimisation of the risk for tumour recurrence
|
0.256
|
|
|
No shortening of life expectancy due to the disease
|
0.241
|
|
|
Possibility for breast-preserving surgery
|
0.099
|
|
|
Avoidance of side effects causing bodily burdens
|
0.050
|
|
|
Avoidance of side effects causing bodily changes
|
0,038
|
|
With a weighting of 0.912 the superordinate therapeutic target efficacy of neoadjuvant
therapy gains a clear preference over the superordinate therapeutic target avoidance
of side effects with a weighting of 0.088 ([Table 4]).
Table 4 Group weightings of all hierarchical levels.
|
Therapeutic targets of neoadjuvant therapy
|
Group weightings according to hierarchical level
|
Consistency ratio
|
|
0
|
1
|
2
|
|
|
From the weightings of the hierarchical levels 1 and 2, the results of the entire
AHP model encompassing all three levels were derived. Here the hierarchical level
0 is determined from the sums of the weightings of hierarchical level 1. The individual
subcriteria of the hierarchical level 2 are obtained in their relative weightings
to each other in the entire model and give in their sum the weighting of hierarchical
level 2.
|
|
Efficacy
|
0.912
|
|
|
0.007
|
-
Destruction of tumour cells
-
Minimising the risk for tumour recurrence
-
No shortening of life expectancy due to the disease
-
Possibility for breast-preserving operation
|
|
0.316
0.256
0.241
0.099
|
|
|
|
Avoidance of side effects
|
0.088
|
|
|
0.005
|
|
Side effects that cause bodily burdens
|
|
0.050
|
|
|
-
Fever
-
Diarrhoea
-
Nausea
-
Fatigue
|
|
|
0.011
0.008
0.015
0.016
|
|
|
Side effects that cause bodily changes
|
|
0.038
|
|
|
|
|
|
|
0.022
0.016
|
|
On comparison of the subcriteria in hierarchical level 2, the patients gave the top
priority to the avoidance of fatigue (0.225) followed by avoidance of nausea (0.202)
and loss of hair (0.181) ([Table 5]).
Table 5 Group weightings of hierarchical level 2.
|
Therapeutic target of neoadjuvant therapy
|
Weighting
|
Consistency ratio
|
|
The weightings were determined from the sample mean values of the pair-wise comparisons
of the criteria in hierarchical level 1 by all 41 patients. The consistency ratio
according to Saaty [17] checks whether the individual pair-wise comparisons are consistent with all the
other comparisons.
|
|
Sum of group weights
|
1.00
|
0.005
|
|
Avoidance of fatigue
|
0.225
|
|
|
Avoidance of nausea
|
0.202
|
|
|
Avoidance of loss of hair
|
0.181
|
|
|
Avoidance of fever
|
0.148
|
|
|
Avoidance of weight gain
|
0.138
|
|
|
Avoidance of diarrhoea
|
0.106
|
|
Discussion
There are studies on patient preferences for chemotherapy. However, to the best of
our knowledge as yet no specific investigation of the preferences of patients with
regard to neoadjuvant therapy has yet been published.
We chose the AHP method to determine patient preferences. This is an instrument developed
by the mathematician Saaty more than 30 years ago that also finds use in decision-making
processes in the health-care services [19], [20]. According to a pilot project of IQWiG for patients with the diagnosis “depression”,
it was shown to be a very structured and easily implemented procedure for the quantitative
measurement of patient preferences and it also provides important tips about suitable
endpoints for clinical trials [17]. The method has also been used to deduce the preferences of patients with cancer
diseases [21], [22], [23].
The patients in our study on the whole assigned more importance to the efficacy of
a neoadjuvant therapy as well as to all four efficacy parameters than to the avoidance
of side effects. This result is reasonable on account of the early stage of their
disease and the curative therapy options. This observation is also in accord with
those of other studies in which the effectiveness of a therapy is of the greatest
importance for patients with breast cancer [24] and where the maintenance of an as high as possible quality of life becomes more
important in advanced stages [25].
In the present study the patients mentioned treatment targets for the neoadjuvant
therapy that corresponded precisely with the endpoints of clinical trials. This suggests
that the endpoints investigated in this study: pCR, disease-free survival, and overall
survival can be directly classified as being patient relevant. One could argue that
this is not the uninfluenced opinion of the patients but rather the repetition of
acquired knowledge. In the event, the patients had had access to information from
the consultations with the responsible physician in the framework of the decision-making
process for the neoadjuvant therapy as well as other available information material.
The patients could then compare this knowledge with their own experiences in the course
of the neoadjuvant therapy which, in our opinion, rather supports the conclusion about
the patient relevance of the respective endpoints.
More generally, the question arises to what extent the preferences determined in this
study correspond with those of newly diagnosed cancer patients without any neoadjuvant
therapy since the actual experiences with neoadjuvant therapy could have influenced
the perspectives of the interviewed patients. Other studies have shown that the therapeutic
preferences of patients with early breast cancer depend less on experience with chemotherapy
but rather much more on their basic attitudes to chemotherapy prior to commencement
of the therapy [26], [27].
In our study the achievement of pCR was the most highly preferred treatment target
of neoadjuvant therapy. Also for the physician pCR is the most important parameter
to evaluate the success of neoadjuvant therapy [9]. The crucial factor here is the association between pCR and an improved survival
at the patient level as demonstrated in clinical studies and meta-analyses [4], [10], [11], even when the formal validation of pCR as a surrogate marker for survival has not
yet been established due to a lack of evidence at the study level [12]. Our investigation does not allow any conclusions about the extent to which patients
consider pCR as a surrogate marker for disease-free survival or overall survival.
It does show that the patients recognise pCR together with disease-free survival,
overall survival, and breast-preserving operations as independent and personally relevant
endpoints. For the patients, pCR has the highest significance. This can possible be
explained by the fact that the patient directly experiences and can understand her
response to therapy as the thus resulting complete elimination of tumour cells. She
thereby probably associates this also with the destruction of circulating tumour cells
and thus potential focus points for metastasis in her body.
Two other studies on patient preferences in chemotherapy for breast cancer named an
extension of overall survival as the major treatment target of chemotherapy. However,
apart from the preference of the patients for efficacy in chemotherapy, they are only
difficultly comparable. They included breast cancer patients in disease stages I to
IV, for example. In one study in total 11 criteria were evaluated, including merely
overall survival as an efficacy parameter beside 10 side effect criteria [24]. In the second study, prolongation of overall survival was judged to be more important
than delaying disease progression without survival benefit, delaying disease progression
per se, and maintenance of a higher quality of life [25].
The perspectives of the patient towards neoadjuvant therapy have been taken into account
with the new official registration practice. On the basis of the existing evidence,
the FDA now accepts pCR as endpoint for an accelerated approval of therapies for early
breast cancer [8]. The first approval on the basis of this regulation was granted in 2013 for neoadjuvant
therapy using pertuzumab with trastuzumab and chemotherapy [28]. Also the European drug approval authority EMA gave its approval for pertuzumab
as first drug to be used in neoadjuvant therapy for early breast cancer in July 2015
on the basis of the endpoint pCR [29].
The major strength of our study is the dedicated patient interviewing carried out
by means of personally conducted, comprehensive interviews by experienced interviewers.
Our study, that provides a quantitative weighting of patient preferences, also exhibits
some limitations. It acquired its information exclusively from patient interviews
that focussed on the acquisition of quantitative preferences. Detailed information
on patient and tumour characteristics as well as the employed therapy regimes was
thus not systematically collected. Thus it remains unclear if and what influence these
factors had on the patientsʼ evaluation. In this context it would also have been interesting
to learn from a parallel survey of the responsible physicians to what extent the preferences
of the patients agreed with those of their physicians.
Practical Conclusions
For the patients the achievement of pCR represents an independent, relevant and highly
desirable goal of neoadjuvant therapy.
Disclosure
In the preparation of this manuscript the authors received editorial assistance from
Dr. Susanne Hell, supported by an unrestricted grant from Roche Pharma AG. All statements
in the text reflect the opinions of all of the authors.
