Download A Prospective Randomized Pilot Study to Evaluate Predictors of

124 Vol. 9, 124 –133, January 2003
Clinical Cancer Research
A Prospective Randomized Pilot Study to Evaluate Predictors of
Response in Serial Core Biopsies to Single Agent Neoadjuvant
Doxorubicin or Paclitaxel for Patients with Locally
Advanced Breast Cancer1
Vered Stearns,2 Baljit Singh, Theodore Tsangaris,
Jeanette G. Crawford, Antonella Novielli,
Mathew J. Ellis, Claudine Isaacs,
Marie Pennanen, Cecilia Tibery, Ahmad Farhad,
Rebecca Slack, and Daniel F. Hayes
Breast Cancer Program, Department of Medicine [V. S., J. G. C.,
A. N., M. J. E., C. I., C. T., A. F., D. F. H.], Department of Pathology
[B. S.], Department of Surgery [T. T., M. P. ], and Biostatistics Unit
[R. S.], Lombardi Cancer Center, Georgetown University School of
Medicine, Washington, DC 20007
ABSTRACT
Introduction: Response to neoadjuvant chemotherapy
for locally advanced breast cancer can be correlated with
long-term outcomes. Surrogate end-point biomarkers may
be used to assess response to the treatment. Most reported
studies assessed the effects of combination chemotherapy.
We assessed the feasibility of obtaining serial core breast
biopsies, and correlated rates of apoptosis, proliferation,
and expression of related proteins at baseline, during, and
after neoadjuvant single agent chemotherapy for locally
advanced breast cancer with response.
Experimental Design: Women with a histologically confirmed unresected T3 or T4 infiltrating carcinoma of the
breast were eligible. The first 20 patients received three
cycles of doxorubicin 90 mg/m2 followed by three cycles of
paclitaxel 250 mg/m2, or the reverse. Nine women received
four cycles of each (doxorubicin 60 mg/m2 and paclitaxel 175
mg/m2). Cycles were administered 14 days apart with filgastrim. End points included: (a) clinical and pathological
response; (b) serial apoptotic [terminal deoxynucleotidyl
transferase (Tdt)-mediated nick end labeling] and prolifer-
Received 6/5/02; accepted 7/24/02.
The costs of publication of this article were defrayed in part by the
payment of page charges. This article must therefore be hereby marked
advertisement in accordance with 18 U.S.C. Section 1734 solely to
indicate this fact.
1
Supported in part by 1K12CA76903 (to V. S.), CRTG-99-250-01CCE (to V. S.), CI-3 of the Cancer Research Fund of the Damon
Runyon-Walter Winchell Foundation (to V. S.), investigator-initiated
clinical grants from Bristol-Myers Squibb Oncology and Amgen, Inc.,
and by a grant from the Fashion Footwear Foundation/QVC Presents
Shoes on Sale.
2
To whom requests for reprints should be addressed, at The Sidney
Kimmel Comprehensive Cancer Center at Johns Hopkins, BuntingBlaustein Cancer Research Building, 1650 Orleans Street, Room 1M53,
Baltimore, MD 21231-1000. Phone: (443) 287-6489; Fax: (410) 6148160; E-mail: [email protected].
ation (immunohistochemistry, IHC) rates; and (c) expression (IHC) of estrogen receptor, HER2, bcl2, and p53.
Results: From April 1997 to June 2001, 29 women were
randomized. Twelve patients (42%) had a clinical complete
response (cCR), and 16 (55%) had a clinical partial response. Five women (17%) had a pathological complete
response, 7 (24%) had microscopic residual disease, and 17
(58%) had macroscopic residual disease. Higher baseline
apoptosis and proliferation were associated with an improved pathological response (P ⴝ 0.006 and 0.003, respectively). Among 14 evaluable patients, apoptosis increased in
women who had a cCR to the first agent but not in women
without a cCR. Estrogen receptor-positive patients had a
worse pathological response (P ⴝ 0.004).
Conclusions: The selected regimen is efficacious. It is
feasible to obtain serial core biopsies that are informative
for studies of apoptosis and IHC. This clinical design can
serve as a model for combining standard chemotherapy and
novel agents.
INTRODUCTION
Despite the widespread use of chemotherapy in cancer, the
molecular signatures of the activity of individual agents on
tumor cells during therapy are not well characterized. Although
breast cancer mortality has declined in recent years, many
women will not benefit from treatments available currently (1).
Traditional end points to assess treatment efficacy, such as time
to progression, disease-free survival, and overall survival require many months or years of follow up and large numbers of
patients. Thus, elucidating surrogate molecular or cellular markers of efficacy of traditional agents may provide earlier insights
into drug sensitivity and resistance.
Neoadjuvant or primary chemotherapy is generally administered to women with LABC.3 This treatment modality allows
for accurate tumor measurements, assessment of response to
therapy, and serial determinations of intratumoral characteristics
(2, 3). In addition to enhancing the likelihood of breast preservation, response to primary chemotherapy can be correlated
with long-term outcomes such as disease-free and overall survival (4 – 6).
3
The abbreviations used are: LABC, locally advanced breast cancer; A,
doxorubicin; T, paclitaxel; ER, estrogen receptor; cCR, clinical complete response; cPR, partial response; cSD, stable disease; pCR, pathological complete response; MiR, microscopic residual disease; MaR,
macroscopic residual disease; TUNEL, terminal deoxynucleotidyl transferase (Tdt)-mediated nick end labeling; IHC, immunohistochemistry;
C1D3, day 3 of cycle 1; C4D1, day 1 of cycle four.
Clinical Cancer Research 125
Surrogate end point biomarkers might be used to rapidly
predict response to neoadjuvant chemotherapy. Serial determinations of markers that represent downstream common events of
biochemical and physiological pathways may be useful in predicting sensitivity or resistance to specific therapies. Insight into
chemotherapy-induced biomarker changes might be used to
develop and study new therapeutic agents, administered either
alone or in combination with standard therapies.
Promising surrogate biomarkers of response to therapy
include induction of apoptosis, reduced proliferation, and
changes in markers related to the signal transduction pathways
perturbed by the agent. An increase in rate of apoptosis was
observed as early as 24 h after treatment in women who received
primary combination chemotherapy for breast cancer (7). In
patients with primary operable breast cancer who receive preoperative chemotherapy, improved response and survival were
associated with a decline in proliferation rate (Ki67) after treatment (8, 9).
Previous reports of biomarkers in preoperative therapy
have studied regimens of combination of several agents, including chemotherapy and hormone therapy (9). Doxorubicin
(Adriamycin) is considered the most active agent in breast
cancer (10). Recent studies have suggested that paclitaxel
(Taxol®) is equally effective in metastatic breast carcinoma
(11). Furthermore, paclitaxel is effective frequently in patients
treated previously with doxorubicin, suggesting a relative lack
of cross-resistance (12). Despite the widespread use of these
agents, the molecular markers of response or resistance to either
doxorubicin or paclitaxel have not been studied intensively in
clinical human breast cancer. Single agent sequential treatment
with dose-dense anthracyclines and taxanes has been extensively studied and is considered a promising approach (13, 14).
We initiated a prospective pilot clinical trial to: (a) assess the
feasibility of obtaining serial breast biopsies in patients receiving single agent sequential neoadjuvant doxorubicin followed
by paclitaxel or paclitaxel followed by doxorubicin for LABC;
(b) assess the feasibility of performing assays of apoptosis and
of markers that predict response on the specimens obtained; (c)
correlate the results of these assays with clinical and pathological response; and (d) compare the results of the assays in
patients treated with single agent doxorubicin versus single
agent paclitaxel.
PATIENTS AND METHODS
Eligibility. Women eligible for the study included those
who were 18 years or older, who were not pregnant or lactating,
with a histologically proven unresected clinical T3 or T4 infiltrating carcinoma of the breast. Women with prior hormonal or
cytotoxic therapy, or those with a known serious comorbid
medical or psychiatric conditions, were excluded. Eastern Cooperative Oncology Group performance status 0 –2, a baseline
left ventricular ejection fraction ⬎50%, and adequate blood
counts, and renal and liver functions were required.
Clinical Trial Design. After signing an informed consent approved by the Institutional Review Board at Georgetown
University Medical Center, eligible participants underwent a
baseline breast biopsy and were randomly assigned to three
cycles of A 90 mg/m2 followed by three cycles of T 250 mg/m2,
Fig. 1 Trial schema. A ⫽ doxorubicin (Adriamycin) 90 mg/m2 [45
mg/m2/day, days 1 and 2 (patients 21–29 were treated with A 60 mg/m2
on day 1, and T 175 mg/m2)]. T ⫽ paclitaxel (Taxol®) 250 mg/m2 over
3 h (patients 21–29 were treated with A 60 mg/m2 on day 1, and T 175
mg/m2). Each cycle was administered every 2 weeks, with filgrastim on
days 3– 8. , core biopsies were obtained before the first cycle (pre),
24 – 48 h after the first cycle (C1D3), and before the administration of
the second agent (before). , response evaluation includes Clinical
Response 1, and Clinical Response 2: clinical response to the first and
second agent, respectively. Pathological Response, assessment of residual disease on the surgical specimen.
or the reverse sequence (A3 T or T3 A; Fig. 1). Cycles were
administered 14 days apart, with granulocyte colony-stimulating
factor support. The purpose of randomly assigning patients to
the sequence was to allow for tumor sampling before and after
treatment with either single agent doxorubicin or paclitaxel. To
ensure equivalence between the two arms, the randomization
was stratified by menopausal status (pre/peri- or postmenopausal) and lesion stage (T3 or T4), and conducted in blocks of
4 patients as implemented in the RANLST module of the
STPLAN software package (15). Midway into accrual to this
trial, reports from several large prospective clinical trials suggested that higher doses of single agent chemotherapy were
clearly more toxic and might not be more beneficial than standard doses (16 –19). Therefore, for patients 21–29, the regimen
was changed to four cycles of A at 60 mg/m2 followed by four
cycles of T at 175 mg/m2, given 14 days apart, with granulocyte
colony-stimulating factor support, or the reverse.
After the chemotherapy, patients underwent a modified
radical mastectomy or a lumpectomy and axillary node dissection at the discretion of the treating breast surgeon. Postoperative chemotherapy was administered at the discretion of the
treating medical oncologist. Breast or chest wall irradiation was
administered at the conclusion of all of the chemotherapy. After
irradiation, all of the women whose tumors were ER and/or
progesterone receptor positive started tamoxifen therapy.
Response Criteria. Bidimensional tumor measurements
were obtained before each cycle and 2 weeks after the last cycle
of chemotherapy. Response to the regimen and to each agent
was determined using Unio Internationale Contra Cancrum criteria (20). cCR was defined as the disappearance of all of the
measurable clinical disease. cPR was defined as a reduction of
⬎50% in the product of the perpendicular dimensions of the
126 Neoadjuvant Doxorubicin and Paclitaxel in Breast Cancer
Table 1
Antigen
a
Antibodies and procedures used for IHC
Manufacturer
Scoring
Ki67
MAba
Antibody
Zymed, South San Francisco, CA
ER
MAb
Immunotech, Wesbrook, ME
HER2 (erbB-2)
MAb
Novocastra, Vector Laboratories,
Burlingame, CA
bcl-2
MAb
DAKO, Carpinteria, CA
p53
MAb (1801)
Novocastra, Vector Laboratories,
Burlingame, CA
Nuclear staining.
% positive
Nuclear staining.
ⱖ10% ⫽ positive
“DAKO system”
0, 1 negative
2, 3 positive
Cytoplasmic staining.
Intensity and % positive cells.
Score ⱖ6 ⫽ positive (32)
Nuclear staining.
ⱖ10% ⫽ positive
MAb, monoclonal antibody.
breast mass. cSD was defined as reduction of ⬍50% or an
increase in size up to 25%. A ⬎25% increase in size was
considered progressive disease.
It has been demonstrated that women with a pCR have a
significant improvement in prognosis (6). In addition, a cCR or
pathological MiR in the breast is also associated with improved
outcomes (5). Because we performed biological assays on the
breast tumor and not the lymph nodes, we correlated baseline
characteristics and change with response to the treatment in the
breast. All of the women underwent a definitive surgical procedure on completion of the regimen. The definitive surgical
specimen was carefully evaluated for the presence of residual
disease by the study pathologist. pCR was defined as absence of
invasive carcinoma in the breast (6). We correlated baseline and
change in biomarkers with pCR, or with the presence of MiR,
defined as ⬍1 cm, or MaR, defined as ⱖ1 cm or multiple foci
of invasive carcinoma throughout the specimen.
Breast Biopsy. Serial core biopsies were obtained exclusively for the purpose of this study for determination of potential
predictive surrogate markers of response. A core biopsy was
obtained using Bard Monopty disposable biopsy instrument
(Covington, GA). Two to four core biopsies were taken before
starting chemotherapy (1–7 days before starting therapy),
24 – 48 h after cycle one (C1D3), and on C4D1 or sooner if a
crossover between the agents had occurred. One to two corebiopsy specimens were fixed overnight in 10% buffered formalin and embedded in paraffin. Paraffin-embedded tissues were
sectioned into 4 ␮m-thick sections and mounted onto slides. An
additional core-biopsy specimen was rapidly frozen in liquid
nitrogen and immediately placed in a ⫺80°C freezer.
Apoptosis Assays. Apoptotic index was determined using the TUNEL method (TumorTACS In Situ Apoptosis Detection kit; Trevigen, Inc., Gaithersburg, MD). Slides were deparaffinized, hydrated, and endogenous peroxidase was removed.
Slides were incubated in labeling buffer for 5 min and then for
1 h at 37°C in a humid chamber with labeling buffer containing
TdT, deoxynucleoside triphosphate mix, and Mn2⫹. Slides were
then transferred into stop buffer for 5 min and washed in PBS.
Streptavidin-horseradish peroxidase was applied onto each sample for 10 min. Slides were washed in PBS twice, placed in
diaminobenzidine for 3 min, and counterstained with methyl
green. Slides were dehydrated and mounted. Apoptosis index
was calculated by counting and dividing the number of brown-
stained nuclei by the total number of cells seen by light microscopy field at ⫻400 magnification by a blinded investigator
(B. S.), and are expressed by percentage.
IHC. Markers for proliferation (Ki67), expression of ER,
HER2/neu, bcl2, and p53 were evaluated by IHC. Slides were
deparaffinized and hydrated. Antigen retrieval consisted of 10
mM sodium citrate buffer (pH 6.0) at 95°C for 10 min. Endogenous peroxidase activity was blocked with peroxide block
(BioGenex, San Ramon, CA) and slides washed with PBS.
Nonspecific antibody binding was blocked with normal goat
serum (BioGenex) for 15 min at room temperature, and slides
were washed with PBS. Separate slides were immunostained
with commercially available antibodies (Table 1). Primary antibodies were diluted 1:20 –1:50 in primary antibody diluent
(BioGenex) and applied to the appropriate slide. After incubation at 37°C for 1–2 h, slides were washed in PBS, incubated
with secondary antibody (Link; BioGenex) for 20 min at room
temperature, washed in PBS, incubated with avidin-biotin complex (Label; BioGenex) for 20 min at room temperature, and
again washed with PBS. Slides were stained with diaminobenzidine, counterstained with hematoxylin and bluing solution
(Optimax Wash Buffer; BioGenex), dehydrated, and mounted.
IHC was scored for percentage of positive cells and/or relative
intensity by a blinded investigator (B. S.) using light microscopy. The antibodies, manufacturers, and scoring system are
listed in Table 1.
Statistical Methods. The primary end point, feasibility
of serial biopsies, was not addressed statistically. Additional end
points were addressed as follows. Patient baseline characteristics, the treatment regimen, and molecular markers were each
assessed for an association with clinical or pathologic response
using the Jonckheere-Terpstra test (21). Clinical and pathological responses to the regimen were reported as the proportion of
patients responding to therapy of the total number of patients
enrolled on the study, after the intent-to-treat principle. For
response to each specific agent, only those patients who were
evaluable for response to that agent are included in the response
rate. A patient was considered ineligible for response to the
second agent if she achieved a cCR to the first agent or if she did
not receive the second agent because of early withdrawal. All of
the patients were considered eligible for response to the first
agent.
Levels of apoptosis were compared by treatment response
Clinical Cancer Research 127
Table 2
Patient characteristics and response to treatment
A. Patient characteristics
Regimena
All
Characteristic
A3T
n (%)
All patients
29 (100)
Age
⬍50 years
22 (76)
ⱖ50 years
7 (24)
Menopausal status
Pre16 (55)
Peri4 (14)
Post9 (31)
Clinical stage at presentation
IIIA
14 (48)
IIIB
10 (35)
IV
5 (17)
Histologic subtype
c
Inf. ductal
25 (86)
Inf. lobular
2 (7)
Inf. duc and lob
1 (4)
Squamous cell
1 (4)
Baseline grade
Elston grade 5–6e
6 (21)
Elston grade 7–9
23 (79)
Clinical responseb
T3A
n (%)
n (%)
15 (52)
14 (48)
10 (67)
5 (33)
12 (86)
2 (14)
9 (60)
1 (7)
5 (33)
7 (50)
3 (21)
4 (29)
8 (53)
6 (40)
1 (7)
6 (43)
4 (29)
4 (29)
11 (73)
2 (13)
1 (7)
1 (7)
14 (100)
0 (0)
0 (0)
0 (0)
2 (13)
13 (87)
4 (29)
10 (71)
cCR
P
Pathological responseb
cPR
cSD
n (%)
n (%)
n (%)
12 (41)
16 (55)
1 (3)
10 (45)
2 (29)
11 (50)
5 (71)
1 (5)
0 (0)
0.39
pCR
P
MiR
MaR
n (%)
n (%)
n (%)
5 (17)
7 (24)
17 (59)
4 (18)
1 (14)
6 (27)
1 (14)
12 (55)
5 (71)
4 (25)
0 (0)
1 (11)
5 (31)
0 (0)
2 (22)
2 (14)
3 (30)
0 (0)
2 (14)
2 (20)
3 (60)
10 (71)
5 (50)
2 (40)
5 (20)
0 (0)
0 (0)
0 (0)
7 (28)
0 (0)
0 (0)
0 (0)
13 (52)
2 (100)
1 (100)
1 (100)
1 (17)
4 (17)
0 (0)
7 (30)
5 (83)
12 (52)
0.63
0.38
0.63
0.28
9 (56)
0 (0)
3 (33)
6 (38)
4 (100)
6 (67)
1 (6)
0 (0)
0 (0)
6 (43)
3 (30)
3 (60)
7 (50)
7 (70)
2 (40)
1 (7)
0 (0)
0 (0)
12 (48)
0 (0)
0 (0)
0 (0)
12 (48)
2 (100)
1 (100)
1 (100)
1 (4)
0 (0)
0 (0)
0 (0)
1 (17)
11 (48)
5 (83)
11 (48)
0 (0)
1 (4)
0.34
0.16
7 (44)
4 (100)
6 (67)
0.71
0.21
0.31
0.12
0.39
P
0.90
0.02d
0.34
B. Correlation of clinical and pathological response status by treatment
pCR
All
cCR
cPR
cSD
MiR
MaR
All
n (%)
A3T
n (%)
T3A
n (%)
All
n (%)
A3T
n (%)
T3A
n (%)
All
n (%)
A3T
n (%)
T3A
n (%)
All
n (%)
A3T
n (%)
T3A
n (%)
29 (100)
12 (42)
16 (55)
1 (3)
15 (52)
5 (17)
9 (31)
1 (3)
14 (48)
7 (24)
7 (24)
0 (0)
5 (17)
4 (14)
1 (3)
0 (0)
3 (10)
3 (10)
0 (0)
0 (0)
2 (7)
1 (3)
1 (3)
0 (0)
7 (24)
6 (21)
1 (3)
0 (0)
2 (7)
1 (3)
1 (3)
0 (0)
5 (17)
5 (17)
0 (0)
0 (0)
17 (59)
2 (7)
14 (48)
1 (3)
10 (34)
1 (3)
8 (28)
1 (3)
7 (24)
1 (3)
6 (21)
0 (0)
a
Percentages are based on the total within each regimen.
Percentages are based on the total within each patient characteristic.
c
Inf, infiltrating; duc and lob, ductal and lobular.
d
Significant at the 0.05 level.
e
Includes 2 infiltrating lobular tumors.
b
level using a one-way ANOVA to assess whether apoptosis after
the first agent was associated with pathologic response (22).
Survival and follow-up estimates were calculated using the
methods of Kaplan and Meier (23).
RESULTS
Clinical Data. From April 1997 through June 2001, 29
women enrolled in the trial and completed the regimen. The
patient characteristics and response to treatment are summarized
in Tables 2, A and B. Twelve women (42%) had a cCR. Of
those, 7 patients had a cCR to the first drug (A3 T ⫽ 2;
T3 A ⫽ 5) and 5 to the second drug (A3 T ⫽ 3; T3 A ⫽ 2).
Four cCR were observed while administering A and 8 cCR
while administering T. Sixteen patients (55%) had a cPR to the
regimen (overall response rate 97%). One woman had a cSD.
Clinical response was similar in patients who received A3 T or
T3 A (Table 2B).
Five women had a pCR in the breast (17%). Of these 5
women, 3 had a pCR in the breast and the lymph nodes (10%).
Seven women (24%; A3 T ⫽ 2; T3 A ⫽ 5) had MiR in the
breast (⬍1 cm). Seventeen patients (59%; A3 T ⫽ 10; T3 A ⫽
7) had MaR in the breast (ⱖ1 cm or multiple foci). Pathological
response was similar in patients who received A3 T or T3 A
(Table 2B).
Of the 16 women with an initial T3 lesion, 7 (44%) underwent successful breast conserving surgery. After the surgical
procedure, 22 of 24 women with a stage III disease received
additional chemotherapy, usually single agent cyclophosphamide. Six of these women also received high-dose chemotherapy and peripheral stem cell rescue. Two women relapsed soon
after their surgery and were treated for metastatic disease. These
2 women did not receive irradiation. One additional woman
relapsed while receiving cyclophosphamide and did not receive
breast irradiation. A fourth woman did not receive chest wall
irradiation because of a previous irradiation to the same breast
after a lumpectomy for ductal carcinoma in situ. With a median
follow-up from time of surgery of 24 months (range, 10 –51
months), 13 of 24 (54%) women with stage III disease suffered
128 Neoadjuvant Doxorubicin and Paclitaxel in Breast Cancer
Table 3
Tissue availability for surrogate marker analysis
Number
Breast biopsy for study
Baseline
C1D3
C4D1
Definitive surgery
Malignant cells on biopsy
Baseline
C1D3
C4D1
Definitive surgery
Evaluable for apoptosis (TUNEL)
Baseline
C1D3
C4D1
Definitive surgery
Paired baseline and C1D3
samples available for TUNEL
Percent
29
25
12
29
100
86
41
100
20
17
8
24
69
59
28
83
26a
14
2
16
14
90
48
7
55
48b
a
Includes 6 specimens that were obtained from a diagnostic
biopsy.
b
Because of improper fixation, TUNEL was not evaluable in most
samples taken from patients 1–10. Paired samples were available in 3 of
the first 10 (30%) women and from 11 of the final 19 women (58%).
disease recurrence, and all 5 of the (100%) women with stage IV
disease suffered disease progression. Nine (38%) women with
stage III and 2 (40%) women with stage IV breast cancer died
of their disease. One additional patient with a stage IIIA disease
died without clinical evidence of recurrent disease.
Grade 3 and 4 toxicities were common for the first 20
patients and were mostly hematological. Of the first 20 women,
35% and 30% suffered grade 3/4 leukopenia and neutropenia,
respectively, and 25% suffered grade 3 neutropenic fevers.
Lymphopenia was reported in 3 women (15%). Of the first 20
patients, 15 did not receive erythropoietin, and 7 (44%) required
blood transfusions (median 2 units of packed RBCs, after three
to six cycles of chemotherapy), whereas 5 patients received
erythropoietin, and none required blood transfusions. Grade 3/4
nonhematological toxicities in the first 20 patients included
gastrointestinal (30%), fatigue (15%), bone pain/arthralgias
(15%), hand foot syndrome (10%), and peripheral neuropathy
(5%).
Of the 9 women who received the less intense regimen,
none suffered grade 3/4 leukopenia, neutropenia, nor required
erythropoietin or blood transfusions. Five women (56%) suffered lymphopenia. One woman suffered a grade 3 motor neuropathy from paclitaxel and was removed from the study. Other
chemotherapy-related grade 3/4 toxicities included infection
(22%) and arthralgias (11%). One of the women who suffered
infection was taken off study.
Correlative Tissues. Biopsies and tissues available for
analysis are summarized in Table 3. All 29 of the women
underwent a baseline breast biopsy. Twenty (69%) of the baseline samples that were obtained specifically for the study contained sufficient cancer cells for marker analysis. Whenever
sufficient tissue was not available at the time of the studyspecific baseline biopsy, the diagnostic paraffin block was retrieved, and sections were cut and used for baseline marker
analysis. Because of improper fixation in the beginning of the
Fig. 2 Pathological response to chemotherapy according to pretherapy
apoptosis and proliferation. A, response to A3 T or T3 A based on
baseline TUNEL staining (n ⫽ 26); B, response to A3 T or T3 A based
on baseline Ki67 (n ⫽ 27). ‰, A3 T arm; ⽧, T3 A arm; , median;
quartile range MiR, focus or ⬍1 cm; MaR, multifoci or ⱖ1 cm.
trial, several samples were not evaluable for baseline apoptosis
determination (Table 3). After we observed that ethanol fixation
produced false-positive TUNEL staining, all of the tissues were
fixed in formalin. Therefore, we were able to perform studies of
apoptosis and proliferation on 26 and 27 patients at baseline,
respectively. Twenty-five (86%) women had a biopsy on C1D3,
of which only 17 contained infiltrating carcinoma sufficient for
IHC staining, and only 14 samples were evaluable for apoptosis
determination. Thus, 14 women had paired samples before and
after the first cycle of chemotherapy evaluable for apoptosis.
Initially, a biopsy was attempted on C4D1. Of the first 14
women, 11 biopsies were attempted at the C4D1 time point.
Whereas 7 samples contained cancer cells, only 2 contained a
sufficient number of malignant cells for proper marker analysis.
Thus, subsequent biopsies were obtained on C4D1 only when a
distinct mass was palpated. Finally, representative slides from
the definitive surgical specimen were evaluated by the study
pathologist (B. S.), and only 16 specimens (55%) contained
invasive carcinoma in the breast sufficient for marker analysis
(Table 3).
Apoptosis and Proliferation. Baseline apoptosis and
proliferation were evaluated in 26 and 27 samples respectively
(Figs. 2, A and B). Mean baseline TUNEL and ki67 staining
Clinical Cancer Research 129
Fig. 3 Baseline and change (C1D3) in apoptosis and response to the
first agent (A or T; n ⫽ 14). ‰, A3 T arm; ⽧ T3 A arm; *, subsequent
pCR in the breast and lymph nodes; #, subsequent 0.3 cm pathologic
residual disease in the breast and no lymph node metastases; &, subsequent cCR to the second agent, and pCR in the breast and lymph nodes
scores for all of the study participants were 0.34% and 33%,
respectively. For patients who achieved a pCR, mean baseline
values of apoptosis and proliferation were 0.52% and 46%,
respectively. Mean apoptotic or proliferative indices were 0.7%
and 60% for patients who had MiR in the breast, and 0.16% and
24% for those with MaR, respectively (P ⫽ 0.003 and 0.006 for
apoptosis and proliferation, respectively).
Apoptotic index on C1D3 was compared with baseline
index in 14 evaluable patients (Fig. 3). Because the second
sample was taken after the first cycle of chemotherapy, correlations were made only with the clinical response to the first
agent and not to the second agent. The 3 women who had a cCR
to the first agent had an increase in apoptosis (4-fold, 20-fold,
and 1 was not evaluable because of high degree of necrosis). In
contrast, in 11 women with a partial or no response to the first
agent, we did not observe a significant increase in apoptosis.
Mean baseline apoptosis for the 11 patients who did not have a
cCR to the first agent was 0.34% (range, 0 –1%), and at C1D3
0.45% (range, 0.1–1.9%). Although the sample size is small,
these results suggest that an increase in apoptosis may be
observed in women who had a cCR to the first agent, whereas
no significant change may be detected in women who had cPR
or cSD to that agent.
Predictive Markers of Response to Chemotherapy.
We correlated pathological response to A3 T or T3 A (pCR,
MiR, or MaR in the breast) with baseline status of ER, HER2,
bcl2, and p53. MiR was seen in 14% of women whose tumors
expressed ER compared with 33% pCR and 33% MiR in women
whose tumors were ER poor (P ⫽ 0.004; Fig. 4A). Overexpression of HER2 was associated with 30% pCR and 30% MiR
compared with 11% pCR and 21% MiR in HER2-negative
tumors (P ⫽ 0.169; Fig. 4B). No statistically significant difference in pathological response was observed based on bcl2
expression (P ⫽ 0.879; Fig. 4C) or p53 expression (P ⫽ 0.463;
Fig. 4D).
We then evaluated pretreatment characteristics and response to specific agent A or T. To determine response to each
agent, we calculated clinical response (cCR versus cPR/cSD) to
the first agent and to the second agent (Fig. 1, Clinical Response
1 and 2, respectively). Because the patients received both single
agents sequentially before undergoing surgery, correlation was
made between baseline tumor marker result and clinical response rather then pathological response. Furthermore, women
who had a cCR to the first chemotherapy agent had no measurable disease in the breast before starting the second agent and
were not evaluable for response to the second agent. For example, if a woman in regimen A3 T had a cCR to A, we concluded
that the cancer was sensitive to A but could not assess the
sensitivity to T. In contrast, if another woman who received
A3 T had cPR or cSD to A, we evaluated the clinical response
to A (tumor measurement before and after A) and to T (tumor
measurement before and after T). Seven women had cCR to the
first agent (2 to A and 5 to T) and, thus, were not evaluable for
response to the second agent. In addition, 2 women were taken
off study after receiving only one cycle of the second agent
(A3 T ⫽ 1 and T3 A ⫽ 1). Thus, only 23 women were
evaluable for clinical response to A, and 26 were evaluable for
clinical response to T (Table 4).
In ER-positive patients, the overall response to either A or
T was poorer than for ER-negative patients (ER-positive: 8%
cCR to A and 15% cCR to T; ER-negative: 36% cCR to A and
42% cCR to T; Fig. 5A). However, women whose tumors
overexpressed the HER2 oncogene were more likely to achieve
a cCR to A than to T (Fig. 5B; HER2-positive 56% cCR and
HER2-negative 0% cCR to A). Furthermore, HER2-positive
patients were slightly more likely to have cCR to T (HER2positive 33% cCR and HER2-negative 26% cCR to T), but this
difference was not as dramatic as that observed with A. Of note,
3 of the 6 women with tumors overexpressing HER2 who were
on regimen A3 T had a cCR to A and were, therefore, not
evaluable for response to T. Another woman received only one
of four scheduled cycles of T and was, thus, not evaluable for
response to T. We did not observe a difference in response to
individual agent A or T based on bcl2 or p53 expression (Fig. 5,
C and D, respectively). Finally, we did not observe significant
changes in expression of ER, HER2, bcl2, or p53 during or after
the chemotherapy (data not shown).
DISCUSSION
In this pilot prospective trial, we studied baseline and
changes in surrogate end point biomarkers after single agent
sequential administration of doxorubicin or paclitaxel as neoadjuvant chemotherapy for LABC. High rates of apoptosis and
proliferation at baseline were associated with improved pathological response to both regimens (A3 T or T3 A). Increased
rate of apoptosis 24 – 48 h after treatment with doxorubicin or
paclitaxel was associated with improved clinical response.
Our results are similar to other reports in the literature. As
reported previously, we observed that adequate training and
experience are required to obtain adequate tissue for analysis
(24). Nonetheless, our data are consistent with reports that early
changes in apoptosis correlate with subsequent response to the
regimen (7). In addition, prospective evaluation of response
to single agent therapy may permit us to evaluate molecular
signatures of sensitivity or resistance to individual agents. On
the basis of these characteristics, several hypotheses can be
generated.
One hypothesis is that breast carcinomas with very low
baseline apoptosis may have a poor response to chemotherapy.
130 Neoadjuvant Doxorubicin and Paclitaxel in Breast Cancer
Fig. 4 Baseline status of predictive markers and pathologic response to the regimen (A3 T or T3 A). A, ER; B, HER2; C, bcl2; D, p53. 䡺,
percentage of pCR; u, percentage of MiR (focus or ⬍1 cm); f, percentage of MaR (multifoci or ⱖ1 cm); (), number in response group/total number
in group.
Table 4
a
b
c
Evaluable patients for correlation of baseline markers and response to chemotherapy
Group
n (%)
Clinical responsea
Pathological responseb
Patients in A3T arm
Evaluable for response to A
cCR to A
Stopped T early to do AEc
Evaluable for response to T
cCR to T
Patients in T3A arm
Evaluable for response to T
cCR to T
Stopped A early to do AE
Evaluable for response to A
cCR to A
Patients in A3T or T3A
Evaluable for response to A
cCR to A
Evaluable for response to T
cCR to T
15 (52)
15 (52)
2 (10)
1 (3)
12 (38)
3 (10)
14 (48)
14 (48)
5 (17)
1 (3)
8 (28)
2 (7)
29 (100)
23 (79)
4 (17)
26 (86)
8 (28)
14 (93)
13 (87)
5 (33)
2 (100)
9 (75)
14 (100)
13 (93)
2 (67)
7 (50)
4 (80)
6 (75)
28 (97)
19 (83)
2 (100)
12 (41)
4 (100)
22 (85)
6 (75)
A patient is considered to have a clinical response if she had either cCR or cPR.
A patient is considered to have a pathological response if she had either pCR or MiR.
AE, adverse event.
The sample size in this study was too small to conduct a
multivariate analysis, but it appeared that tumors with low
apoptotic ratios were ER-positive. Other investigators have assessed the effects of chemoendocrine therapy on proliferation
and did not observe a difference in response rates for ER-rich or
ER-poor tumors (8, 25). Our results are consistent with retrospective evaluations that suggested that hormone receptornegative tumors are more chemosensitive than those that lack
Clinical Cancer Research 131
Fig. 5 Baseline status of predictive markers and clinical response to single agent A or T. A, ER; B, HER2; C, bcl2; D, p53. 䡺, percentage of cCR;
f, percentage of cPR/cSD; (), number in response group/total number in group.
hormone receptors in both the metastatic and neoadjuvant settings (6, 26, 27). It is now clear that patients with ER-positive
tumors benefit substantially from endocrine manipulations (28).
Such observations have led to hypotheses that chemotherapy
may add little benefit to adequate hormone therapy in receptorrich patients, although this hypothesis requires confirmation in
prospective randomized clinical trials.
A second hypothesis that can be generated is that higher
baseline apoptotic rates are associated with better response to
chemotherapy. Therefore, if available, pretreatment with agents
that specifically induce apoptosis might result in high response
to chemotherapy. We suggest that this trial design is a good
clinical model to test this theory.
These data also suggest that it may be possible to determine
as early as 24 – 48 h after administration of chemotherapy
whether a woman is likely to respond to a specific agent or not.
Such information might help to make early decisions regarding
a change in treatment. Moreover, the single agent nature of this
trial design permits assessment of changes in surrogate markers
(such as increased apoptosis) in a more straightforward fashion
than if the chemotherapy is given in combination. Larger studies
are required to assess whether anthracyclines and taxanes induce
different changes in apoptosis or proliferation. Such results may
provide biological insights into the mechanisms of action of
both standard and novel antineoplastic treatments.
The novel approach in this study can also answer questions
regarding the role of other markers and response to individual
therapies. For example, it has been suggested that overexpression or amplification of HER2 may be associated with an
improved response to doxorubicin and possibly paclitaxel (29,
30). Our data are also consistent with this hypothesis.
Other markers that are closely related to cell cycle and cell
death include bcl2 and p53. Although we did not observe a
significant correlation between expression of bcl2 and response
to the chemotherapy, we have not evaluated posttranslational
modification of the protein nor ratio with other bcl2 family
members. It is also possible that other methods of detecting an
alteration in a gene or a gene product may be more sensitive to
predict for response to specific therapy. One study suggested
that wild-type p53 was associated with an improved response to
anthracyclins, whereas a mutated p53 was associated with an
improved response to paclitaxel (31).
In summary, we present a clinical design incorporating
single agent sequential chemotherapy in LABC that can be used
in future investigation not only to correlate surrogate end point
biomarkers with response to single agent therapy. The model
can also be used to incorporate novel agents with standard
treatments. Changes in apoptosis and proliferation can be used
to determine the efficacy of the combination.
132 Neoadjuvant Doxorubicin and Paclitaxel in Breast Cancer
ACKNOWLEDGMENTS
We thank Drs. Marc Lippman, Susan Honig, and Michael Hawkins
for helpful discussions. We thank Drs. Said Baidas, Robert Buras,
Phillip Cohen, Patricia Conrad-Rizzo, and Robert Warren for enrolling
their patients in this study. We thank Dr. Sofia Merajver for a critical
review of the manuscript.
REFERENCES
1. Peto, R., Boreham, J., Clarke, M., Davies, C., and Beral, V. UK and
USA breast cancer deaths down 25% in year 2000 at ages 20 – 69 years.
Lancet, 355: 1822, 2000.
2. Fisher, B., and Mamounas, E. Preoperative chemotherapy: A model
for studying the biology and therapy of primary breast cancer. J. Clin.
Oncol., 13: 537–540, 1995.
3. Wolff, A. C., and Davidson, N. E. Primary systemic therapy in
operable breast cancer. J. Clin. Oncol., 18: 1558 –1569, 2000.
4. Fisher, B., Bryant, J., Wolmark, N., Mamounas, E., Brown, A.,
Fisher, E. R., Wickerham, D. L., Begovic, M., DeCillis, A., Robidoux,
A., Margolese, R. G., Cruz, A. B., Jr., Hoehn, J. L., Lees, A. W.,
Dimitrov, N. V., and Bear, H. D. Effect of preoperative chemotherapy
on the outcome of women with operable breast cancer. J. Clin. Oncol.,
16: 2672–2685, 1998.
5. Feldman, L. D., Hortobagyi, G. N., Buzdar, A. U., Ames, F. C., and
Blumenschein, G. R. Pathological assessment of response to induction
chemotherapy in breast cancer. Cancer Res., 46: 2578 –2581, 1986.
6. Kuerer, H. M., Newman, L. A., Smith, T. L., Ames, F. C., Hunt,
K. K., Dhingra, K., Theriault, R. L., Singh, G., Binkley, S. M., Sneige,
N., Buchholz, T. A., Ross, M. I., McNeese, M. D., Buzdar, A. U.,
Hortobagyi, G. N., and Singletary, S. E. Clinical course of breast cancer
patients with complete pathologic primary tumor and axillary lymph
node response to doxorubicin-based neoadjuvant chemotherapy. J. Clin.
Oncol., 17: 460 – 469, 1999.
7. Ellis, P. A., Smith, I. E., McCarthy, K., Detre, S., Salter, J., and
Dowsett, M. Preoperative chemotherapy induces apoptosis in early
breast cancer. Lancet, 349: 849, 1997.
8. Chang, J., Powles, T. J., Allred, D. C., Ashley, S. E., Clark, G. M.,
Makris, A., Assersohn, L., Gregory, R. K., Osborne, C. K., and Dowsett,
M. Biologic markers as predictors of clinical outcome from systemic
therapy for primary operable breast cancer. J. Clin. Oncol., 17: 3058 –
3063, 1999.
9. Dowsett, M., Archer, C., Assersohn, L., Gregory, R. K., Ellis, P. A.,
Salter, J., Chang, J., Mainwaring, P., Boeddinghaus, I., Johnston, S. R.,
Powles, T. J., and Smith, I. E. Clinical studies of apoptosis and proliferation in breast cancer. Endocr. Relat. Cancer, 6: 25–28, 1999.
10. Ellis, M., Hayes, D. F., and Lippman, M. E. Treatment of metastatic
disease. In: J. Harris, M. Lippman, M. Morrow, and C. K. Osborne
(eds.), Diseases of the Breast, Ed. 2, pp. 749 –798. Philadelphia:
Lippincott-Raven, 2000.
11. Sledge, G., Neuberg, D., Ingle, J., Martino, S., and Wood, W. Phase
III trial of doxorubicin vs. paclitaxel vs. doxorubicin ⫹ paclitaxel as first
line therapy for metastatic breast cancer: An Intergroup trial. Proc. Am.
Soc. Clin. Oncol. Annu. Meet., 16: 1a, 1997.
12. Seidman, A. D., Reichman, B. S., Crown, J. P., Yao, T. J., Currie,
V., Hakes, T. B., Hudis, C. A., Gilewski, T. A., Baselga, J., Forsythe, P.,
and et al. Paclitaxel as second and subsequent therapy for metastatic
breast cancer: activity independent of prior anthracycline response.
J. Clin. Oncol., 13: 1152–1159, 1995.
13. Hudis, C., Fornier, M., Riccio, L., Lebwohl, D., Crown, J.,
Gilewski, T., Surbone, A., Currie, V., Seidman, A., Reichman, B.,
Moynahan, M., Raptis, G., Sklarin, N., Theodoulou, M., Weiselberg, L.,
Salvaggio, R., Panageas, K. S., Yao, T. J., and Norton, L. 5-year results
of dose-intensive sequential adjuvant chemotherapy for women with
high-risk node-positive breast cancer: a phase II study. J. Clin. Oncol.,
17: 1118 –1126, 1999.
14. Fountzilas, G., Papadimitriou, C., Dafni, U., Bafaloukos, D., Skarlos, D., Moulopoulos, L. A., Razis, E., Kalofonos, H. P., Aravantinos,
G., Briassoulis, E., Papakostas, P., Abela, K., Gogas, E., Kosmidis, P.,
Pavlidis, N., and Dimopoulos, M. A. Dose-dense sequential chemotherapy with epirubicin and paclitaxel versus the combination, as first-line
chemotherapy, in advanced breast cancer: a randomized study conducted by the Hellenic Cooperative Oncology Group. J. Clin. Oncol.,
19: 2232–2239, 2001.
15. Brown, W. B., Lovato, J., and Rundell, B. RANLST software.
Houston, TX: The University of Texas, M. D. Anderson Cancer Center,
Department of Biomathematics ([email protected]), 1992.
16. Fisher, B., Anderson, S., Wickerham, D., DeCillis, A., Dimitrov,
N., Mamounas, E., Wolmark, N., Pugh, R., Atkins, J., Meyers, F.,
Abramson, N., Wolter, J., Bornstein, R., Levy, L., Romond, E., Caggian,
V., Grimaldi, M., Jochimsen, P., and Deckers, P. Increased intensification and total dose of cyclohosphamide in a doxorubicin-cyclophosphamide regimen for the treatment of primary breast cancer: Findings from
NSABP B-22. J. Clin. Oncol., 15: 1858 –1869, 1997.
17. Fisher, B., Anderson, S., DeCillis, A., Dimitrov, N., Atkins, J. N.,
Fehrenbacher, L., Henry, P. H., Romond, E. H., Lanier, K. S., Davila,
E., Kardinal, C. G., Laufman, L., Pierce, H. I., Abramson, N., Keller,
A. M., Hamm, J. T., Wickerham, D. L., Begovic, M., Tan-Chiu, E.,
Tian, W., and Wolmark, N. Further evaluation of intensified and increased total dose of cyclophosphamide for the treatment of primary
breast cancer: findings from National Surgical Adjuvant Breast and
Bowel Project B-25. J. Clin. Oncol., 17: 3374 –3388, 1999.
18. Henderson, I. C., Berry, D., Demetri, G., Cirrincione, C., Goldstein,
L., Martino, S., Ingle, J., Cooper, M., Canellos, G., Borden, E., Fleming,
G., Holland, J., Graziano, S., Carpenter, J., Muss, H., and Norton, L.
Improved disease-free and overall survival from the addition of sequential paclitaxel but not from the escalation of doxorubicin dose level in
the adjuvant chemotherapy of patients with node-positive primary breast
cancer. Proc. Am. Soc. Clin. Oncol. Annu. Meet., 17: 101a, 1998.
19. Winer, E., Berry, D., Duggan, D., Henderson, I. C., Cirrincione, C.,
Cooper, R., and Norton, L. Failure of higher dose paclitaxel to improve
outcome in patients with metastatic breast cancer. Proc. Am. Soc. Clin.
Oncol. Annu. Meet., 17: 101a, 1998.
20. Monfardini, S., Brunner, K., and Crowther, D. Evaluation of the cancer
patient and the response to treatment. In: UICC-Manual of Adult and
Pediatric Medical Oncology, pp. 22–38. Berlin, Germany: Springer, 1987.
21. Hollander, M., and Wolfe, D. A. Nonparametric Statistical Methods. New York: John Wiley & Sons, 1973.
22. Scheffe, H. The Analysis of Variance, pp. 55– 87. New York: John
Wiley & Sons, Inc., 1959.
23. Kaplan, E., and Meier, P. Nonparametric estimation from incomplete observations. J. Am. Stat. Assoc., 53: 457– 481, 1958.
24. Symmans, W. F., Volm, M. D., Shapiro, R. L., Perkins, A. B., Kim,
A. Y., Demaria, S., Yee, H. T., McMullen, H., Oratz, R., Klein, P.,
Formenti, S. C., and Muggia, F. Paclitaxel-induced apoptosis and mitotic arrest assessed by serial fine-needle aspiration: implications for
early prediction of breast cancer response to neoadjuvant treatment.
Clin. Cancer Res., 6: 4610 – 4617, 2000.
25. Makris, A., Powles, T. J., Allred, D. C., Ashley, S. E., Trott, P. A.,
Ormerod, M. G., Titley, J. C., and Dowsett, M. Quantitative changes in
cytological molecular markers during primary medical treatment of
breast cancer: a pilot study. Breast Cancer Res. Treat., 53: 51–59, 1999.
26. Lippman, M. E., Allegra, J. C., Thompson, E. B., Simon, R.,
Barlock, A., Green, L., Huff, K. K., Do, H. M. T., Aitken, S. C., and
Warren, R. The relation between estrogen receptors and response rate to
cytotoxic chemotherapy in metastatic breast cancer. N. Engl. J. Med.,
298: 1223–1228, 1978.
27. Colleoni, M., Minchella, I., Mazzarol, G., Nole, F., Peruzzotti, G.,
Rocca, A., Viale, G., Orlando, L., Ferretti, G., Curigliano, G., Veronesi,
P., Intra, M., and Goldhirsch, A. Response to primary chemotherapy in
breast cancer patients with tumors not expressing estrogen and progesterone receptors. Ann. Oncol., 11: 1057–1059, 2000.
Clinical Cancer Research 133
28. Early Breast Cancer Trialist’s Collaborative Group. Tamoxifen for
early breast cancer: an overview of the randomised trials. Lancet, 351:
1451–1467, 1998.
29. Yamauchi, H., Stearns, V., and Hayes, D. F. When is a tumor
marker ready for prime time? A case study of c-erbB-2 as a predictive
factor in breast cancer. J. Clin. Oncol., 19: 2334 –2356, 2001.
30. Konecny, G., Thomssen, M., Pegram, M., Luck, H., Untch, M., Pauletti,
G., Dandekar, U., Ramos, L., Kuhn, W., Eidtmann, H., DuBois, A., Olbricht,
S., Steinfeld, D., Moebus, V., Minckwitz, G., and Slamon, D. J. HEr-2/neu gene
amplification and response to paclitaxel in patients with metatatic breast cancer.
Proc. Am. Soc. Clin. Oncol. Annu. Meet., 20: 23a, 2001.
31. Kandioler-Eckersberger, D., Ludwig, C., Rudas, M., Kappel, S.,
Janschek, E., Wenzel, C., Schlagbauer-Wadl, H., Mittlbock, M., Gnant,
M., Steger, G., and Jakesz, R. TP53 mutation and p53 overexpression
for prediction of response to neoadjuvant treatment in breast cancer
patients. Clin. Cancer Res., 6: 50 –56, 2000.
32. Elledge, R. M., Green, S., Howes, L., Clark, G. M., Berardo, M.,
Allred, D. C., Pugh, R., Ciocca, D., Ravdin, P., O’Sullivan, J.,
Rivkin, S., Martino, S., and Osborne, C. K. bcl-2, p53, and response
to tamoxifen in estrogen receptor-positive metastatic breast cancer: a
Southwest Oncology Group study. J. Clin. Oncol., 15: 1916 –1922,
1997.