Reprintedfrom
`
`Volume 20, Number 9
`
`May 1, 2002
`
`JOURNAL OF
`CLINICAL
`
`ONCOLOGY
`
`Assessment of Molecular Markers of Clinical Sensitivity to Single-
`Agent Taxane Therapy for Metastatic Breast Cancer
`
`ASCO_JCO
`
`Disclaimer
`The ideas and opinions exprtmed in the JonmaI of Clinical Oncology (JCO) do not necessarily reflect those of the American Society
`of Clinical Oncology (ASCO). The mention of any product, service, or therapy in any JCO article should not be construed as an
`endorsement of the products mentioned.
`
`It is the responsibility of the treating physician or other health care provider, relying on independent experience and knowledge of the
`patient, to determine drug dosages and the best treatment for the patient. Readers are advisod to check the appropriate medical
`literature and the product information currently provided by the manufacturer of each drug to be administered to verify the dosage,
`method, and duration of administration, or contraindications. Readers are also encouraged to contact the manufacturer with questions
`about the features or limitations of any products. ASCO assumes no responsibility for any injury or damage to persons or property
`arising out of or related to any use ofthe material contained in JCO orto any errors or omissions.
`
`Ascgo“
`
`American Society of Clinical Oncology
`
`Marking a world Qfdifl‘cr'crrcc in cancer (arr
`
`Celltrion v. Genentech
`
`IPR2017-01122
`
`Genentech Exhibit 2024
`
`

`

`
`
`
`
`
`
`PrintingorphotocopyingpermittedforuptothreecopiesCopyright2015ASCO-AllRightsRe
`
`
`
`
`
`
`
`Assessment of Molecular Markers of Clinical Sensitivity to
`Single-Agent Taxane Therapy for Metastatic Breast Cancer
`
`By Catherine Van Poznak, Lee Ton, Katherine S. Panogeos, Crispinito D. Arroyo, Clifford Hudis, Lorry Norton,
`and Andrew D. Seidman
`
`Purpose: The taxanes affect tubulin polymerization
`and interfere with mitotic transition. A checkpoint
`blockade at the 61-5 boundary would be expected to
`promote taxane-induced apoptotic cell death through a
`mechanism that may involve p27. Other proposed de-
`terminants of clinical taxane sensitivity]resistance in-
`clude p53, members of the epidermal growth factor
`receptor (EGFR) superfamily (e5, HER2, EGFR), and es-
`trogen receptors and progesterone receptors. These
`molecular markers and their correlation with clinical
`taxane sensitivity are investigated in this retrospective
`clinicopathologic study.
`Patients and Methods: We performed immunohisto-
`chemistry (IHC) for estrogen receptors, progesterone
`receptors, HER2, EGFR, p53, and p27 on 144 breast
`tumor specimens from patients treated for metastatic
`breast cancer on a series of clinical trials of single-agent
`taxane chemotherapy for correlation with clinical re-
`sponse (complete or partial response). Patient charac-
`teristics that could influence response (ie, perfor-
`mance status, extent of disease, and prior therapy)
`were also examined.
`
`In univariate analysis, Karnofsky perfor-
`Results:
`mance status 2 90% and no prior history of anthracy-
`cline therapy correlated with a good clinical response to
`single-agent taxane (P = .003 and P = .041 , respective-
`ly). None of the IHC variables tested were predictive of
`clinical response to taxane therapy, although p27 neg-
`ativity showed a trend toward significance (P = .075).
`Concordance between the polyclonal antibody with
`HercepTest (DAKO, Carpinteria, CA) and the monoclo-
`nal antibody CB-l 'I (BioGenex, San Ramon, CA) was
`noted (kappa = 0.943); however, neither univariate
`nor multivariate analysis demonstrated an associa-
`tion between HER2 status and response to taxane
`chemotherapy.
`Conclusion: The IHC biomarkers studied were not
`predictive of responso to single-agent taxane chemo-
`therapy in patients with metastatic breast cancer. Iden-
`tification of molecular correlates of taxane response
`remains an important goal.
`J Clin Oncol 20:2319-2326.
`Society of Clinical Oncology.
`
`(0 2002 by American
`
`HE SELECTION OF chemotherapeutic agents and
`regimens for the treatment of breast cancer is gener-
`ally derived from large phase III clinical trials, where it is
`possible to estimate the comparative likelihood of benefit
`and toxicity for a specific regimen. The taxanes, paclitaxel
`and docetaxel, play important roles in treating breast can-
`cer' and large, randomized clinical trials of taxane-contain-
`ing chemotherapy have demonstrated a survival advantage
`over previous standard regimens both in the adjuvant setting
`and in the setting of metastatic disease.”5 However, these
`trials are not instructive as to the reason for response or
`resistance in an individual patient. The identification of
`patient-specific tumor characteristics that can improve the
`ability to predict response to therapy would help optimize
`treatment. This retrospective study was initiated to investi-
`gate clinical response in patients with metastatic breast
`cancer and the correlation with selected biomarkers (HERZ,
`ER, PR, p53, p27, and EGFR) as potential predictors of
`response to single-agent taxane chemotherapy.
`Although there are several molecular markers with prog-
`nostic utility in breast cancer, few have been evaluated as
`predictors of response to treatment. The estrogen receptor
`(ER) and progesterone receptor (PR) status is used to
`determine the likelihood of response to endocrine therapy
`and is a standard prognostic and predictive factor in breast
`
`cancers‘7 Few markers have shown similarly powerful
`predictive value for response to systemic therapy.
`Amplification and overexpression of certain oncogenes
`have been associated with an aggressive natural history.
`poor prognosis and,
`in some situations, chemoresistance,
`and in other situations, chemosensitivityfi'” For example,
`retrospective analysis of HER2 overexpression by Paik et
`2111" supports the hypothesis that anthracycline—containing
`adjuvant chemotherapy regimens confer benefit over non—
`anthracyclinc-containing regimens. Molecular markers may
`indicate specific cellular alterations that affect specific
`targeting mechanisms of a therapeutic treatment. We have
`speculated that human breast cancers overexpressing HER2
`
`Franz the Breast Cancer Medicine Service. the Pathology Depart-
`ment. and the Biostatistical Department, Memorial Sloan—Kettering
`Cancer Center. New York. NY.
`Submitted August 21 , 2001; accepted January 3/ , 2002.
`Supported by US. Army Medical Research and Materiel Command
`grant no. DAMDl 7~94rJ-4329.
`Address reprint requests 10 Catherine VII)? Poznak. MD. Memorial
`Sloan-Kettering Cancer Center, I 275 YonkAve. New York, NY 10021;
`email: vanpoznc@mskcc.org.
`© 2002 by American Society of Clinical Oncologz.
`0732-183M02/2009-2319/Ji20. 00
`
`Journal of Clinical Oncology, Vol 20, No 9 (May 1], 2002: pp 2319-2326
`DOI: 10.] 200/JCO,2002.08.125
`
`2319
`
`
`
`

`

`2320
`
`VAN POZNAK El" AL
`
`may be significantly more likely to respond to single-agent
`therapy with paclitaxel or docetaxel. Our prior assessment
`of tumor HER2 expression through monoclonal antibody
`(4D5) and the polyclonal antibody (pAb-l) demonstrated
`that 4D5 positivity was predictive of positive response to
`taxane monotherapy.17
`The taxanes are antimicrotubular agents that promote
`microtubular assembly from tubulin dimers and stabilize
`microtubules by preventing depolymeiization, thereby in-
`terfering with normal mitotic transition. Taxanes causes
`arrest of cell cycle progression in the mitotic phase of the
`cell cycle, with accumulation of cells in Gz-M interphase.18
`The ability of paclitaxel to activate MAP kinase and Raf-1
`kinase19 may explain how tubulin active agents (eg, tax-
`anes) could have enhanced cytotoxic effects in HER2
`overexpressing cells through downstream perturbation of
`the signal transduction cascade. Constitutively active epi—
`dermal growth factor receptor (EGFR) and HERZ in trans-
`fected cell cultures induce paclitaxel resistance and alter
`expression of beta-tubulin expression.20 EGFR overexpres-
`sion or altered expression has been reported for a variety of
`human tumors.
`
`The cell cycle is regulated by a complex system of
`cyclins, cyclin—dependent kinases, and cyclin—dependent
`kinase inhibitors that are in turn governed by associated
`cyclins and by phosphorylation. p27 is a negative regulator
`implicated in Gl-phase arrest. Lower p27 levels in breast
`cancers assessed by immunohistochemistry (IHC) have
`been associated with shorter survival.2 “23 Increased p27 has
`been observed in a paclitaxel-resistant breast cancer cell
`line, compared with its sensitive parental
`(MDA 435)
`cells.24 Paclitaxel may be able to induce apoptosis through
`a non—p53-dependent pathway, possibly involving molecu—
`lar regulators of apoptosis such as bcl-Z, bax, and p27.25
`Inactivation of the tumor suppressor gene p53 has been
`implicated in the development and progression of a number
`of different cancers.26 Mutants of p53 are present in up to
`50% of invasive breast cancers, and loss of its function is
`associated with high proliferation index and poor clinical
`outcome. p53 is just one oncogene in a complex pathway
`that controls both proliferation and apoptosis. Other factors
`in this pathway include the bax and bcl-Z families. Bax
`transcription is upregulated by p53, which links the proapop-
`totic increase in p53 with the bcl-Z family of proteins.”33
`The identification of pretreatment tumor characteristics
`that are strong predictors of either drug sensitivity or
`resistance may minimize the proportion of patients who
`would not derive benefit and maximize the proportion that
`would benefit. Therefore, we attempted to assess immuno-
`histochemical phenotypes of human cancers that might be
`associated with poor or favorable clinical responsiveness to
`
`treatment with the taxanes, paclitaxel and
`Single-agent
`docetaxel, for metastatic breast cancer.
`
`PATIENTS AND METHODS
`
`Women with metastatic breast cancer treated at Memorial Sloan-
`Kettering Cancer Center (MSKCC) on nine successive clinical trials
`between 1991 to I998 with single-agent taxanes were identified. All
`patients within these nine trials were candidates for this present study;
`however, access to sufficient tumor block was limited to those patients
`who either had their primary surgery at MSKCC or who had their
`primary surgery elsewhere and had forwarded the tissue block to
`MSKCC. The specimens evaluated were selected by the availability of
`adequate quantity of tumor to perform multiple IHC studies. Patients
`were excluded if there was insufficient accessible material.
`
`Patients
`
`Patients treated with taxane chemotherapy on these nine clinical
`trials had histologically confirmed. bidimensionally measurable meta-
`static breast cancer, life expectancy more than 12 weeks. Kamofsky
`performance status (KPS) a 60%, and normal organ function (WBC
`count > 3,500 cells/mL, absolute neutrophil count > 2.000 cells/mL,
`platelet count > 100,000 cells/mL, creatinine < 2.0 mg/dL, bilirubin <
`1.5 mg/dL, and no significant cardiac disease or arrhythmia). Previous
`chemotherapy and/or hormonal
`therapy was allowed according to
`specific eligibility criteria for each individual protocol; no concomitant
`cytotoxic therapy, hormonal therapy, immunotherapy, or radiotherapy
`was permitted during protocol therapy with taxane. Exclusion criteria
`included clinically unstable brain metastases or carcinomatous menin—
`gitis, symptomatic lymphangitic pulmonary disease, and history of
`prior malignancies except completely excised in situ carcinoma of the
`cervix or nonmelanoma skin cancer. Patient characteristics that can
`influence response to chemotherapy were extracted from the medical
`record by chart review. These patient demographic factors are perfor—
`mance status. extent of disease, visceral dominant disease, extent of
`prior therapy, and prior anthracycline chemotherapy. All patients
`provided informed consent for protocols approved by the MSKCC
`Institutional Review Board.
`
`IHC Assays
`IHC studies were performed on paraffin blocks from tissues fixed in
`10% neutral buffered formalin. IHC analysis reported in this study was
`carried out in a single laboratory (MSKCC). All slides were reviewed
`by a pathologist (L.T.) without knowledge of the demographic or
`treatment response information. A hematoxylin and eosin—stained slide
`was prepared from each block and used for pathologic confirmation of
`the presence of invasive mammary carcinoma. Interpretation of the
`results was limited to the invasive portion of the tumor. Reagents used
`for IHC studies are listed in Table 1. Recognizing the possibility that
`different epitopes of the HERZ transmembrane protein could poten-
`tially have biologic significance regarding chemotherapy sensitivity,”
`two different antibodies were used: the polyclonal antibody HercepTest
`(DAKO, Carpinteria, CA) and the monoclonal antibody CBll (Bio—
`Genex, San Ramon, CA).
`Tissue sections of 4-p.m thickness were prepared, mounted on
`polylysine-coated slides, deparaffinized, and rehydrated with distilled
`water. With the exception of CB1 I (no treatment, Ventana automated
`stainer [Ventana Medical Systems, Inc, Tucson, AZ]), the following
`epitope retrieval was used for the remaining antibodies: microwave on
`high energy for 5 minutes twice in 0.01 mol/L citrate buffer (pH 6.0)
`
`
`
`
`
`PrintingorphotocopyingpermittedforuptothreecopiesCopyright2015ASCOaAllRightsRe
`
`
`
`
`
`
`
`
`
`

`

`
`
`
`
`PrintingorphotocopyingpermittedforuptothreecopiesCopyright2015ASCO-AllRightsRe
`
`
`
`
`
`
`
`MOLECULAR MARKERS OF TAXANE SENSITIVITY
`
`Table I. Reagents Used for the IHC Studies
`
`
` Antibody Clone Dilution Manufacturer and Location
`
`
`ER
`PR
`EGFR
`p53
`
`p27
`
`ID5
`1A6
`3lG7
`PABI 801
`
`DCS72
`
`l:5
`1:50
`1:5
`I : I ,000
`
`I :500
`
`Beckman Coulter, Fullerton, CA
`Novocastra, Burlingame, CA
`Zymed, San Francisco, CA
`CalbiochemvNovabiochem,
`La Jolla, CA
`Calbiochem-Novabiochem,
`La Jolla, CA
`HercepTest; DAKO, Carpinteria,
`CA
`I:IOMonoclonal CBI IHER2 BioGenex, San Ramon, CA
`
`HER2
`
`Polyclonal A0485
`
`I:I
`
`
`
`
`
`
`
`for ER, PR, p53, and p27; 01% pepsin at 37°C for 20 minutes for
`EGFR; and boiling at 95° to 99°C for 40 minutes in diluted epitope
`retrieval solution (DAKO) for HercepTest. ER and PR were scored as
`positive when at least 10% of the carcinoma cell nuclei were immu-
`noreactive. For p27 and p53, any percentage of positive nuclei was
`noted. The intensity of membrane staining of HERZ for both HercepT-
`est and CB” was evaluated according the criteria set forth by the
`DAKO HercepTest: score 0 = no or up to 10% membrane staining;
`score 1+ = partial and/or faint membrane staining in more than 10%
`of tumor cells; score 2+ = weak to moderate complete membrane
`staining in more than 10% of tumor cells; and score 3+ = strong
`complete membrane staining in more than 10% of tumor cells. EGFR
`was scored as either negative or positive on the basis of the presence of
`any membrane staining in the tumor cells.
`
`Clinical Response Criteria
`A complete response (CR) was defined as disappearance of all
`measurable and assessable disease, signs, symptoms, and biochemical
`changes related to the tumor for more than 4 weeks, during which time
`no new lesion(s) appeared. A partial response was defined as at least a
`50% reduction in the sum of the products of the perpendicular
`diameters of all measurable lesions lasting more than 4 weeks, with no
`new lesion(s) and no enlargement of existing lesions. Minor response
`was defined as a reduction of25% to 49% in the sum of the products of the
`perpendicular diameters of all measurable lesions lasting more than 4
`weeks, with no new lesions and no enlargement of existing lesions. Stable
`disease was defined as less than a 50% reduction and less than a 25%
`increase in the sum of the products of two perpendicular diameters of all
`measurable lesions and lack of appearance of new lesions. Progression of
`disease was defined as either an increase in the product of the two
`perpendicular diameters of any measurable lesion by more than 25% over
`the size at entry or the appearance of new areas of malignant disease All
`tumor responses were confirmed and radiographic responses reviewed by
`a designated reference radiologist. For four of these trials, additional
`external, independent review of response occurred.
`
`Statistical Analysis
`response, minor
`Taxane response was identified as CR, partial
`response, stable disease, progression of disease, or inassessable. Sta-
`tistical analysis was performed on the CR plus partial response subset
`versus all other tumor response categories combined Clinical variables
`were dichotomized as follows: KPS (60% to 80% v 90% to 100%),
`number of prior treatment regimens (S two v 2 three regimens), prior
`anthracycline exposure (yes v no), number of organ systems with
`metastases (5 two v 2 three systems), and visceral dominant disease
`
`2321
`
`
`Table 2. Taxane Monotherapy Regimens (N = 188)
`No, of
`Regimen
`Patients
`
`
`23
`Pacliiaxel I00 mg/m2 I hour weekly
`62
`Paclitaxel 135-175 mg/m2 3 hours every 3 weeks
`69
`Poclitaxel 175-250 mg/m2 24 hours every 3 weeks
`9
`Paclitaxel I40 mg/m2 96 hours every 3 weeks
`
`Docetoxel 100 mg/m2 1 hour every 3 weeks 25
`NOTE. Patients were on one of nine clinical
`trials. The clinical
`trials
`outnumber the treatment regimens For two reasons. The trials varied by the
`entry criteria such thato regimen at paclitaxel I 75 mg/m2 over 3 hours in one
`clinical trial may have been used in patients with refractory metastatic disease,
`and in a dilierent clinical trial, the some regimen may have been used as a
`first-line chemotherapy.2 The above table demonstrates ranges of paclitoxel
`used. The dose of drug used in one study may have been paclitaxel I75
`mg/rn2 over 24 hours, whereas another trial used paclitaxel 250 mg/m2 over
`24 hours
`
`(yes v no). Associations between taxane sensitivity and the clinical
`characteristics or molecular markers were assessed univariately by X2
`analysis. The IHC markers that showed a trend toward statistical
`significance were analyzed by logistic regression to adjust for clinical
`characteristics known to be associated with taxane sensitivity.
`Both HERZ IHC (HercepTest and CB-l 1) results were considered
`negative if scored 0 or 1 and positive if scored 2 0r 3. The kappa
`statistic was used to assess concordance between (HercepTest and
`CB-l 1). Additional HERZ statistical analysis was perfomied using 0 to 2+
`staining as a negative result and 3+ as positive. A cut~point analysis was
`performed by the maximum X2 with P value adjustment method to
`determine positive values for p27 and p53. This analysis revealed p27-
`negative as 0% staining and any staining as p27-positive and p53-negative
`as less than 10% staining and 10% staining as positive.
`
`RESULTS
`
`The database consists of patients treated at MSKCC on
`nine different clinical trials (1991 to 1998) using single-
`agent taxane (paclitaxel, eight trials; docetaxel, one trial)
`chemotherapy for metastatic breast cancer, The treatment
`regimens used are listed in Table 2. Dose adjustments were
`directed by toxicities and protocol. The single-agent taxane
`protocol database contains 340 patients. Not all patients had
`accessible and/or sufficient tissue for investigational immu—
`nohistochemical studies. Many patients had been referred to
`MSKCC and had only stained pathology slides submitted
`for confirmation of disease before protocol enrollment. The
`tumor blocks of 188 patients were available, and the
`medical records of these 188 patients were reviewed for
`clinical data. The clinical characteristics of these patients
`are listed in Table 3.
`
`Of the 188 patients whose paraffin-embedded tumor
`block was accessible, 144 had adequate quantity of material
`with evidence of invasive breast cancer for planned IHC
`studies. Of the patients with available tissue, 61 of 14-4
`(42%) experienced a CR or partial response with single-
`
`
`
`
`
`

`

`2322
`
`103
`85
`
`155
`31
`30
`77
`48
`19
`9
`2
`1
`
`68
`118
`
`125
`63
`
`111
`76
`
`8
`67
`28
`44
`32
`9
`
`75
`113
`
`Table 3. Demographic Variables“w
`
`Frequency
`Total No.
`Clinical Characteristic
`No. oi Patients
`%
`—.—.—.___—________
`KPS
`< 90%
`2 90%
`No. oi prior treatment regimens
`s 2
`2 3
`0
`1
`2
`3
`4
`5
`2 6
`Prior doxorubicin
`Yes
`No
`No. oF organ systems involved
`5 2
`2 3
`Visceral dominant
`Yes
`No
`Taxane response
`CR
`PR
`MR
`SD
`POD
`INEV
`Taxane response
`CR, PR
`Other
`Toxane response in patients
`with sutticient tumor block
`ior immunohistochemistry
`
`61 42CR, PR 144
`
`VAN POZNAK ET AL
`
`
`
`Table 4. Immunohistochemistrya
`
`Frequency
`No. of
`Total No.
`96
`lHC Assay‘
`Patients
`EE—
`ER
`
`Negative
`Positive
`
`PR
`
`95
`49
`
`66.0
`34.0
`
`144
`
`144
`
`141
`
`139
`
`141
`
`141
`
`1 18
`26
`
`85
`20
`15
`21
`
`80
`26
`13
`20
`
`92
`49
`
`76
`65
`
`81.9
`18. 1
`
`60.3
`14.2
`10.6
`14.9
`
`57.6
`18.7
`9.3
`14.4
`
`65.2
`34.8
`
`53.9
`46.1
`
`Negative
`Positive
`HER2 HercepTest score
`0
`1
`2
`3
`HER2 CB—l 1 score
`0
`1
`2
`3
`p53
`Negative
`Positive
`p27
`Negative
`Positive
`EGFR
`139
`85.6
`1 19
`Negative
`
`
`20Positive 1 4. 4M
`NOTE. Not all tumor blocks contained sutiicient materials for all immuno-
`histochemical assays of interest; therefore] n is less than 144 for some assays.
`‘ER, < 10% = negative; PR, < 10% = negative; p53/ < 10% = negative;
`EGFR, 0% = negative; p27, 0% = negative.
`
`54.8
`45.2
`
`83.3
`16.7
`16.1
`41.4
`25.8
`10.2
`4.8
`1.1
`0.5
`
`36.6
`63.4
`
`66.5
`33.5
`
`59.4
`40.6
`
`4.3
`35.6
`14.9
`23.4
`17.0
`4.8
`
`39.9
`60.1
`
`188
`
`186
`
`186
`
`188
`
`187
`
`188
`
`188
`
`
`
`
`
`NOTE. The medical records of two patients were limited in their information
`regarding prior chemotherapy (at outside institutions) and the dominant site of
`tumor burden was unclear for one patient.
`Abbreviations: PR, partial response; MR, minor response; SD, stable dis-
`ease; POD, progression oi disease; lNEV, inevaluable.
`
`agent taxane therapy (listed in Table 3 as a subset of the 188
`patients). One hundred four tumor blocks assessed were
`from the primary breast cancers; 40 tumor blocks were from
`metastatic sites. The results of the IHC assays are listed in
`Table 4. Most patients had hormone receptor—negative
`disease, with ER-negativc disease in 66% and PR-negative
`disease in 81.9%. Consistent with prior reports, the HER2
`assays revealed that approximately one quarter of patients
`had tumors overexpressing HER2 2+/3+ by IHC. The
`results of both the polyclonal HercepTest and the monoclo-
`nal assay CB-ll are listed in Table 4. Concordance between
`
`the HercepTest and CB—ll was high (kappa = .943; 95%
`confidence interval, 0.877 to 1.0).
`Univariate analysis, X2 analysis, and multivariate analysis
`performed to assess both the clinical and IHC results for
`
`predictive value of response to taxane therapy revealed that
`only better KPS and lack of prior exposure to anthracycline
`chemotherapy were predictive of response to single—agent
`taxane chemotherapy. Univariate analysis is shown in Table
`5 and multivariate analysis in Table 6. None of the biomar-
`kers assessed by IHC (ER, PR. EGFR, HERZ, p27, and p53)
`showed a statistically significant association with clinical
`response to taxane therapy. Univariate and multivariate
`analysis of HER2 score 0 to 2+ v 3+, and analyses using
`cutoff 0 to 1+ v 2 to 3+, showed no association with
`clinical
`taxane response by either HercepTest or (2811.
`Multivariate analysis of p27, KPS, and prior exposure to
`doxorubicin by logistic regression demonstrated persistence
`of a trend for p27 negativity and clinical taxane response.
`Our analysis used the cut point of any p27 staining as
`positive. The assessment of p27 by quartile categories 0% to
`
`oD
`
`:
`ID..
`a:
`.9
`D:
`
`4;.
`
`OUma
`
`ONa
`
`:1.0t—
`.
`.E
`m
`':
`.
`
`>a
`
`oO m
`
`.2a.
`aU
`:L
`.C.s.
`oa.
`o~
`3t.
`oa.
`‘oG}so
`_
`{11
`a.
`at
`.E>a.
`0U
`o._.
`
`oa
`
`2E.
`
`o2a
`
`.
`
`5mC
`
`.5C
`':
`a.
`
`
`
`

`

`
`
`
`
`PrintingorphotocopyingpermittedforuptothreecopiesCopyright2015ASCO-AllRightsRe
`
`
`
`
`
`
`
`MOLECULAR MARKERS OF TAXANE SENSITIVITY
`
`2323
`
`
`
`Table 5. x2 Univariate Analysis'
`Taxane Response
` Variable P
`
`KPS 2 90%
`.003
`No prior anihracycline
`.04]
`p27
`.075
`p53
`.508
`HercepTest
`(3-1 V 2-3
`0-2 v 3
`EGFR
`ER
`Nonvisceral dominant disease
`PR
`S 2 organ syslems involved
`g 2 chemotherapy prior regimens
`CBvl l
`.l92
`O-l V 23
`
`0-2 V 3 .866
`
`.5l I
`.999
`.858
`.931
`.094
`.995
`.105
`.I60
`
`'Concorclonce between the DAKO HercepTest and CB-l l: kappa = .943
`(95% confidence interval, 0.87 to 1.0).
`
`25%, 26% to 50%, 51% to 75%, and 76% to 100% yielded
`no correlation with response of taxane chemotherapy (data
`not shown). Analysis by individual protocol for a relation—
`ship between the individual biomarkers and response to
`taxane regimen did not reveal a correlation (data not
`shown). Although serial biopsy assessment was not per—
`formed, the results of IHC interpatient results from primary
`breast cancer tissue versus tissue from metastatic sites
`revealed a significant difference only for PR status, where
`more primary tissues were found to be PR-positive
`(P = .041).
`Interestingly, ER and PR status were not
`predictive of taxane response in either univariate or multi-
`variate analysis.
`
`DISCUSSION
`
`We performed a retrospective clinicopathologic correla—
`tive science study to evaluate potential molecular correlates
`of clinical responsiveness to single—agent taxane chemother-
`apy for metastatic breast cancer. Because the database was
`created while investigating the optimal dosing and schedule
`of taxane therapy, there is heterogeneity in the regimens
`within nine different clinical
`therapeutic protocols. The
`
`Table 6. Multivariate Analysis: Logistic Regression'
`Variable
`Adjusted Odds Ratio
`95% Confidence Interval
`P
`r94
`0953.95
`.068
`2.52
`I .24~5.IO
`.01 0
`1.91
`0.92-3.95
`.081
`
`p27
`KPS
`Prior doxorubicin
`
`’Association oi p27 slatus and clinical taxane response adiusted Ior KPS
`and prior doxorubicin therapy.
`
`clinical factors of KPS 2 90% (P = .003) and no prior
`antliracycline exposure (P = .041) were predictive of a
`positive response to taxane therapy by both univariate and
`multivariate analyses. None of the biomarkers assessed by
`IHC (ER, PR, EGFR, HER2, p53, and p27) showed a
`statistically significant ability to act as predictors of clinical
`response to taxane therapy, although p27 negativity showed
`a trend toward significance by both univariate and multi-
`variate analyses when adjusted for KPS and prior anthracy—
`cline exposure. Because demographic data correlated with
`clinical
`response to taxane monotlrerapy for metastatic
`breast cancer, we felt
`it appropriate to include the full
`clinical data sample (n = 188) and not
`to limit our
`presentation to include only those patients whose tissue was
`available (n = 144). These observations await independent,
`prospective confirmation and remain hypothesis generating.
`Although well characterized for their prognostic and
`predictive power for hormonal therapies, ER and PR status
`was not predictive of response to single-agent
`taxane
`chemotherapy in either univariate or multivariate analysis
`within our study population. Of note,
`the majority of
`patients within our study were ER-negative and/or PR—
`negative (66% and 82%, respectively). The predominance
`of hormone receptor—negative patients within our study may
`reflect the patient referral pattern to the phase II clinical
`trials of single—agent taxane chemotherapy for metastatic
`breast cancer at MSKCC during the years 1991 to 1998.
`This finding could be important in light of the fact that no
`benefit was observed for
`the addition of paclitaxel
`to
`doxorubicin and cyclophosphamide in the ER-positive sub-
`set of patients treated in Cancer and Leukemia Group B
`study 9344.34
`Approximately one quarter of tumor samples tested
`positive for overexpression of HER2 (2 to 3+) as would be
`expected. Our results are noteworthy for the lack of corre-
`lation between HER2 status as assessed by either HercepT~
`est or CB-ll and response to single-agent taxane therapy.
`These findings are partly in contrast to our earlier analysis.'7
`In this earlier analysis of fewer cases, HER2 status as
`assessed by the monoclonal antibody 4D5 was predictive of
`positive response to taxane monotherapy, whereas HER2
`assessment with the polyclonal antibody pAb-l, was not. In
`the present study, we demonstrated strong concordance
`between the two IHC methods of evaluating HER2 status
`(HercepTest and CB—l l), which was not observed between
`4D5 and pAb-l. Evaluating HER2 status by IHC assesses
`protein overexpression, whereas an alternative means of
`studying HER2 status,
`fluorescent
`in situ hybridization
`(FISH), measures gene amplification. The molecular anal—
`ysis of amplification offers a more quantitative approach,
`and there appears to be a close correlation between ampli-
`
`
`
`

`

`2324
`
`VAN POZNAK ET AL
`
`fication reflected in increased copy number and the ability to
`detect HER2 immunohistochemically.35'37 A retrospective
`analysis of patients with metastatic breast cancer random—
`ized to either paclitaxel and epirubicin or cyclophosphamide
`and epirubicin demonstrated that the patients in the pacli-
`taxel and epirubicin arm whose tumors had HER2 gene
`amplification by FISH experienced a statistically significant
`improvement in response rate and progression-free survival
`and a borderline significant improvement in overall surviv-
`al.3g Although FISH may become the preferred method of
`defining HERZ status, a functional assay (eg, PNZA, an
`antibody that is more specific for the activated, phosphor-
`ylated HERZ receptor) may also be relevant for prediction
`of therapeutic benefit.39
`In virtually all reports to date, loss of p27 in epithelial
`cancers has been shown to correlate significantly with
`high-grade, poorly differentiated tumors showing signifi-
`cantly lower p27 protein expression than their well-differ-
`entiated counterparts. A potential taxane-induced mecha-
`nism of p53—independent apoptosis by induction of p27 has
`been reported.40 In this study, both univariate and multivar-
`iate analyses demonstrated a trend toward statistical signif-
`icance with p27 negativity and response to single-agent
`taxane chemotherapy. Although most studieszl‘23‘41 have
`scored cases as 0% to 25%, 25% to 50%, 50% to 75%, and
`more than 75% or using the cutoff of high (> 50% of cells
`staining) versus low stainingfz‘43 our statistical analysis has
`shown no effect in separating these groups when assessing
`for clinical response to taxane therapy.
`The observation of prolonged disease-free survival for
`patients with HERZ overexpressing primary tumors receiv-
`ing more dose-intense anthracycline—based chemotherapy
`suggests that the presence of a marker of more malignant
`phenotype might allow identification of a subset of patients
`who would optimally benefit from a specific therapeutic
`strategy.44 The assessment of two or more coexpressed
`genes may further increase the predictive power of biologic
`markers. The clinical trial Cancer and Leukemia Group B
`8541 examined three different dose-intensity regimens of
`adjuvant cyclophosphamide, doxorubicin, and fluorouracil
`in women with stage II breast cancer. Retrospective analysis
`of HERZ and p53 expression in a subset of these patients
`demonstrated that HER2 positivity and p53 negativity
`showed an improved survival with high-dose cyclophosph-
`amide, doxorubicin, and fluorouracil.45
`Approximately one quarter of the tissue samples that
`underwent biopsy in our study were obtained from a
`metastatic site. As serial biopsies (comparing primary with
`metastatic biopsy specimens within the same patient) were
`not part of this study, we cannot draw conclusions on the
`effects of the site of biopsy on our results. However, other
`
`investigators have performed IHC assays for bcI-Z and p53
`on serial biopsy specimens in breast cancer patients and
`have demonstrated consistent results within individual pa-
`tients.46 Assessment of both primary and metastatic lesions
`for I-IER2 and p53 in breast cancer patients demonstrated
`consistent results between the primary tumor and the met-
`astatic lesion for both markersf‘7
`
`locally
`Serial biopsy in the neoadjuvant setting for
`advanced breast cancer may offer a more optimal scenario
`for correlative studies of molecular predictors of chemosen-
`sitivity. Marcus et al48 reported analysis on 19 breast cancer
`patients with locally advanced breast cancer who had
`sequential tumor assessment for HER2, p53, cyclin D1, and
`p27. The biopsy specimens were obtained at diagnosis, after
`two cycles of doxorubicin and docetaxel, and at mastectomy
`after six cycles of doxorubicin and docetaxel. Although
`pretreatment tumor specimens strongly expressing HER2 or
`p53 predicted response to doxorubicin and docetaxel in this
`small study, cyclin D1, bc1-2, and p27 did not. Although this
`small study suggests markers possibly predictive of re-
`sponse to combination chemotherapy, extrapolation to ei-
`ther single—agent doxorubicin or taxane chemotherapy is not
`possible. HERZ, p53, and bcl-2 have been investigated as
`predictors of response to doxorubicin and paclitaxel as
`single—agent therapy for metastatic breast cancer.” How-
`ever,
`these markers did not demonstrate a correlation
`between expression of these markers and clinical response
`to either agent