Quantitation of HER2 expression or HER2: HER2 dimers and differential survival in a cohort of metastatic breast cancer patients carefully selected for trastuzumab treatment primarily

Quantitation of HER2 Expression or HER2:HER2 Dimersand Differential Survival in a Cohort of Metastatic Breast

Cancer Patients Carefully Selected for TrastuzumabTreatment Primarily by FISH

Christine Desmedt, PhD,* Jeff Sperinde, PhD,w Fanny Piette, PhD,zWeidong Huang, MD, PhD,wXueguang Jin, PhD,w Yuping Tan, PhD,w Virginie Durbecq, MD, PhD,*

Denis Larsimont, MD, PhD,* Rosa Giuliani, MD, PhD,* Colombe Chappey, PhD,wMarc Buyse, PhD,z John Winslow, PhD,w Martine Piccart, MD, PhD,*

Christos Sotiriou, MD, PhD,* Christos Petropoulos, PhD,w and Michael Bates, MDw

Abstract: The selection of patients with HER2-positive breast

cancer for treatment with trastuzumab is based on the

measurement of HER2 protein expression by immunohisto-

chemistry, or the presence of HER2 gene amplification by

fluorescence in situ hybridization (FISH). By using multivariate

analyses, we investigate the relationship between quantitative

measurements of HER2 expression or HER2:HER2 dimers and

objective response (Response Evaluation Criteria in Solid

Tumors), time to progression, and breast cancer survival after

trastuzumab treatment in a cohort of patients with metastatic

breast cancer who were primarily selected for treatment by

FISH. The VeraTag assay, a proximity-based assay designed to

quantitate protein expression and dimerization in formalin-

fixed, paraffin-embedded tissue specimens, was used to measure

HER2 protein expression and HER2:HER2 dimer levels. In a

Cox proportional hazards analysis, higher HER2 expression or

HER2:HER2 dimer levels were both correlated with longer

survival (P=0.0058 and P=0.016, respectively) after treatment

with trastuzumab in a population of patients that were either

FISH-positive (90%) or immunohistochemistry 3+ (10%).

Patients with higher levels of HER2 expression or HER2:HER2

dimers seemed to derive little benefit from the addition of

concomitant chemotherapy to trastuzumab, whereas those with

lower levels benefited significantly [interaction test P=0.43

(HER2 expression), P=0.27 (HER2:HER2 dimers)]. These

data suggest that more quantitative or functional measurements

of HER2 status may facilitate the development of more

personalized treatment strategies for patients with metastatic

breast cancer.

Key Words: HER2 expression, HER2 homodimers, metastatic

breast cancer, Cox multivariate analyses, trastuzumab treatment

(Diagn Mol Pathol 2009;00:00–00)

The HER-family of receptor tyrosine kinases (EGFR/HER1, ErbB2/HER2, ErbB3/HER3, and ErbB4/

HER4) plays a major role in the pathogenesis of manysolid tumors. In approximately 25% of breast cancers,overexpression of HER2 has been linked to both adverseprognosis and improved responsiveness to treatment withtrastuzumab (Herceptin, Genentech), a monoclonal anti-body directed at the extracellular domain of HER2.1–5

Trastuzumab offers significant disease-free and overallsurvival advantages in both the metastatic and adjuvantsettings in HER2 overexpressing patients.6–11 Currentmethodologies for the selection of HER2-positive patientsinclude immunohistochemistry (IHC) to detect HER2protein overexpression, and fluorescence in situ hybridi-zation (FISH) to detect HER2 gene amplification.12–18

Despite these methods, only a fraction of HER2-positivepatients with metastatic breast cancer (MBC) respond totrastuzumab, and a significant proportion of the respon-ders relapse within 1 year. These clinical observationssuggest the existence of both de novo and acquiredresistance to trastuzumab.

Here we describe the correlation between quantita-tive measurements of HER2 protein expression (H2T) orHER2:HER2 dimers (H2D) and breast cancer survival(BCS), time-to-progression (TTP), and Response Evalua-tion Criteria in Solid Tumors (RECIST) [RECIST criteriarefer to standardized methods of assessing the response ofsolid tumors to treatment. Predefined target lesions aremeasured radiographically (computed tomography scanor magnetic resonance imaging recommended) beforetherapy and then again at follow-up intervals that areCopyright r 2009 by Lippincott Williams & Wilkins

From the *Department of Medical Oncology, Jules Bordet Institute,Brussels; zIDDI, Louvain-la-Neuve, Belgium; and wMonogramBiosciences Inc, South San Francisco, CA.

Supported by Monogram Biosciences.Disclaimers: Jeff Sperinde, Weidong Huang, Xueguang Jin, Yuping Tan,

Colombe Chappey, John Winslow, Christos Petropoulos, andMichael Bates are employees of and stockholders in MonogramBiosciences.

Reprints: Michael Bates, MD, Vice President, Clinical Research, Mono-gram Biosciences Inc, 345 Oyster Point Blvd., South San Francisco,CA 94080 (e-mail: [email protected]).

ORIGINAL ARTICLE

Diagn Mol Pathol � Volume 00, Number 0, ’’ 2009 1

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

protocol-specific, but usually at 6 to 8 weeks. Responsesare categorized as complete response (CR), partialresponse (PR), stable disease (SD), or progressive diseaseaccording to the specifications described by Therasse etal.19] responses after trastuzumab treatment in a cohort ofpatients with MBC who were carefully preselected foramplification of HER2 by FISH (90%) or overexpressionof HER2 by IHC (10%). Quantitative measurements aremade by the VeraTag assay (Monogram Biosciences) usingformalin-fixed, paraffin-embedded (FFPE) breast tumorspecimens. HER2:HER2 dimers are defined in this assay astwo HER2 receptors located within close proximity of oneanother. These could include HER2 molecules that interactalong a defined protein-protein interface (true HER2:-HER2 dimers), or HER2 monomers that are in closeproximity (<B100nm) as a result of pathologic over-expression. Whether such HER2 homodimers serve as amarker of trastuzumab susceptibility as a result of their rolein signal pathway activation or whether they simplyrepresent clustered HER2 molecules that may serve as abetter target for trastuzumab-mediated antibody-dependentcellular cytotoxicity is unclear at this time.

MATERIALS AND METHODS

The VeraTag Technology and FISHThe VeraTag technology is a proximity-based

method designed to accurately and reproducibly quanti-tate protein expression and protein-protein complexes,including cell surface dimers, in FFPE specimens.Detailed descriptions of the VeraTag assay and itsperformance characteristics are available elsewhere.20

The VeraTag AssayAll materials, cell culture procedures, fixation,

processing, paraffin-embedding, and microtomy for speci-men preparation were provided and performed asdescribed earlier. The H2T and H2D assays, includingcapillary electrophoresis, data analysis, and tumor con-tent normalization were carried out as detailed earlier.20

FISHThe detection of HER2/neu amplification was

conducted by an independent clinical laboratory (Pheno-Path Laboratories, Seattle, WA) by using the PathVysionassay (Abbott-Vysis Inc, Downers Grove, IL).

Description of the Clinical CohortSeventy-one patients with MBC were treated in the

trastuzumab Expanded Access Program or the IdentifiedPatient Program (IPP) sponsored by F. Hoffman-LaRoche in Belgium before May 2002. The mean age ofthe patients was 51 years and the mean follow-up timewas 20 months. The median TTP for the cohort was 8.1months. The median BCS of the cohort was undefined, as78.3% of the patients were still alive 18 months aftertreatment with trastuzumab. The patients were requiredto have MBC and most patients had prior experience with2 or more chemotherapy regimens. Patients from the IPPwere allowed to receive first line trastuzumab if it was

given with paclitaxel. The patients received trastuzumabas a single agent or in combination with chemotherapy,primarily paclitaxel (Table 1). On-site monitoring visitswere conducted to collect clinical data from the originalsource documents (medical records, case report forms, orimaging data). For the trastuzumab-only group, anindependent response review (IRR) committee evaluatedall data in a blinded fashion. The response was assessedradiographically, according to RECIST recommenda-tions. As there was high concordance between the IRRand the oncologists who coordinated the study, the IRRwas not required for the trastuzumab plus chemotherapygroup. The patients who were enrolled but never treatedand the patients who were receiving incomplete admin-istrations of trastuzumab (1 to 5wk) were excluded fromthe analysis, unless discontinuation was unequivocallyascribed to progressive disease. The patients withprevious or concomitant cancer other than breast wereexcluded, as were patients without archival samples oftheir invasive breast cancer. Central confirmation ofHER2 overexpression by FISH (N=64) or IHC (N=7)was mandatory. All study procedures were approved bythe Ethics Committee of the Jules Bordet Institute.

Preselection for HER2 Gene Amplificationand Objective Response

Eighty percent of the cohort received treatment inthe adjuvant setting and 58% received chemotherapyconcomitantly with trastuzumab, whereas the remainderwere treated with trastuzumab alone. It is worth notingthat greater than 90% of the patients were selected forinclusion on the basis of FISH positivity (Table 1). Therewas mandatory confirmation of HER2 status and centra-lized collection of clinical data. As such, the stringency ofselection for HER2 overexpression or gene amplificationwas high.

Statistical MethodsMeasurements of H2T and H2D were tested for

correlation with BCS, defined as the time from the startof Herceptin-containing treatment to cancer-associateddeath or the end of the follow-up period. Deaths notattributed to breast cancer were considered censored atthe time of death. TTP was defined as the time from thestart of Herceptin-containing treatment to progression(RECIST) or censor [end of follow-up (n=4) or changein treatment (n=23)]. Patients were classified as eitherhigh or low for each of the 2 variables, H2T and H2D,using the median of the distribution. High and low groupswere compared for time-to-event end points using theKaplan-Meier (KM) analysis. Cox proportional hazardsmodels were fitted to determine the most significantcorrelates with outcome. Potential prognostic factorsincluded HER2 expression, HER2:HER2 dimers, HER2gene copy number as measured by centromere-correctedFISH (FISH/CEP17), treatment group (trastuzumab-only or trastuzumab plus chemotherapy), estrogen andprogesterone receptor status, exposure to adjuvanthormonal therapy or adjuvant chemotherapy, time from

Desmedt et al Diagn Mol Pathol � Volume 00, Number 0, ’’ 2009

2 r 2009 Lippincott Williams & Wilkins


metastasis to treatment, time from surgery to treatment,presence of metastases in the liver, lung, skin, bone,lymph nodes, brain, or other viscera, tumor size, numberof involved lymph nodes, and total number of sites wheremetastases were found. Metastatic sites occurring in lessthan 10% of the total cohort were not considered asindividual variables, but were included in the calculationof the total number of sites. A log transformation wasapplied to H2T and H2D to reduce the asymmetry ofthe distribution of these variables. A forward selectionprocedure was applied to enter successively the moststatistically significant covariates until none of them wassignificant at the 0.05 level. The Cox models were used toclassify patients in risk groups based on a multivariatescore (equal to the linear combination of the prognosticfactors retained as statistically significant in the model).Cox analyses were run separately for the entire cohort(N=71) and for the subset selected by FISH (N=64) toaccount for any variability introduced by the minoritypopulation selected by IHC. KM curves were estimatedseparately for patients at high and low risk by using themedian of the distribution of the multivariate score asthe cut-point. All statistical analyses were performedat Monogram Biosciences (Sperinde, Chappey) and alsoindependently at IDDI (Piette, Buyse).

RESULTS

Correlation of HER2 Expression by VeraTagWith IHC and FISH Copy Number

Data generated at Monogram by using 170 FFPEbreast tumor controls show that the VeraTag assaymeasures subtle differences in HER2 expression that aredistributed over a 3-log dynamic range compared with thesemiquantitative readout of IHC assessed on a 0 to 3+scale, or by using the more quantitative histoscore(H score), where staining intensity is corrected for percentageof tumor involvement (Shi et al, Fig. 6). By using 19 FFPEmetastatic breast tumor control samples obtained from acommercial tissue vendor (MT Group, Santa Monica, CA)and not linked to clinical outcome data, we assessed HER2gene copy number by FISH by using an independentclinical laboratory (PhenoPath Laboratories, Seattle, WA)and correlated the results with H2T. VeraTag measures ofH2T showed a general correlation with total and FISH/CEP17 copy number. However, the data suggest thattumors with similar numbers of HER2 gene copies mayexpress significantly different amounts of HER2 protein(Figs. 1A, B). Similarly, a positive correlation of HER2expression by the VeraTag assay with FISH/CEP17 copynumber in the Bordet cohort was also observed, as shown

TABLE 1. Clinical Characteristics of the Bordet Cohort

Characteristic Jules Bordet cohortGeographic origin BelgiumType of cohort Expanded Access Program/

Identified Patient ProgramCentralized HER2 assessment YesCentralized clinical data collection YesTotal no. patients 71Mean age (y) 51Mean months of follow-up (range) 20 (2.6-44.9)Median time to progression (mo) 8.1 Tumor size No. patients (%)Median overall survival Undefined (78.3% alive

at 18mo)r2 cm2 24 (33.8)

>2 cm2 and r5 cm2 23 (32.4)Parameter No. patients (%) >5 cm2 7 (9.9)HER2 status by IHC 7 (9.9) Missing 17 (23.9)HER 2 status by FISH 64 (90.1)

Prior adjuvant therapyHormone receptor status Adjuvant hormonal therapy 29 (40.9)ER+ PR+ 11 (15.5) Adjuvant chemotherapy 47 (66.2)ER+ PR� 11 (15.5) Adjuvant hormonal

therapy only10 (14.1)

ER� PR+ 6 (8.5) Adjuvant chemotherapy only 28 (39.4)ER� PR� 36 (50.7) Adjuvant hormonal+adjuvant

chemotherapy19 (26.8)

ER unknown, PR� 2 (2.8) None 14 (19.7)ER unknown, PR unknown 5 (7.1)

No. metastatic sitesNodal status 1 or 2 55 (77.5)Negative 17 (23.9) 3 or 4 16 (22.5)1 to 3 positive nodes 16 (22.5)4 to 10 positive nodes 13 (18.3) TreatmentZ10 positive nodes 19 (26.8) Trastuzumab+chemotherapy 41 (57.8)Status missing 6 (8.5) Trastuzumab only 30 (42.3)

ER indicates estrogen receptor; PR, progesterone receptor.

Diagn Mol Pathol � Volume 00, Number 0, ’’ 2009Quantitative HER2 Expression and

Trastuzumab Treatment

r 2009 Lippincott Williams & Wilkins 3


in Figure 2. Although the correlation with FISH/CEP17is weaker in the Bordet cohort than in the MT dataset, thesample sizes are small, and the differences may simplyreflect the effects of random variation. Larger datasets areneeded to gain a better understanding of these relation-ships.

Higher Levels of HER2 Expression or HER2Homodimers Correlate With Better ClinicalOutcomes on Trastuzumab

Several different types of analyses were used toexplore the relationship between quantitative levels ofHER2 expression or HER2:HER2 dimers and improvedoutcomes after trastuzumab treatment. In an analysisof response to trastuzumab by quartiles of H2T or H2D,the relationship between the VeraTag measurements andclinical benefit (CR+PR+SD>6mo) as assessed byRECIST criteria, although not statistically significant inthis small cohort, hint that higher H2T or higher H2Dmay portend better responses to trastuzumab (test fortrend P=0.37 and 0.12, respectively) (Fig. 3).

Multivariate Cox proportional hazards regressionanalyses were performed (Table 2) by using BCS and TTPas end points. In the complete dataset including all 71patients, 3 prognostic factors were included in the finalmodel after forward selection (Table 2, model A): (1)number of metastatic sites (P=0.00019), (2) HER2expression by VeraTag (P=0.0058), and (3) treatmentwith trastuzumab alone (P=0.036). Increasing numbersof metastases and treatment with trastuzumab alone wereassociated with increased hazard ratios for death, whereasincreasing expression of HER2 was associated with areduced risk of death.

H2T and H2D were strongly correlated in thisdataset (r=0.73, P<0.0001). The substitution of H2Dfor H2T showed a statistically significant correlationwith survival (model B). Model C evaluated only the 64patients who were selected for inclusion by FISH, andfound that H2T was still correlated significantly with BCSin this analysis.

Cox analyses using TTP as the end point yieldedseveral models that could explain the data. Modelsobtained by the forward selection procedure includedthe variable lung metastases and failed to identify H2T(P=0.17) or H2D (P=0.12) as significant prognosticvariables. However, when the 3 prognostic factorsidentified in model A were forced in a model fitted onthe TTP data (models D, E), H2T or H2D was correlatedwith outcome. Model F evaluated only the 64 patientswho were selected for inclusion by FISH. Of particularnote, 21 patients (30%) were censored early on accountof a change in therapy in the absence of progression,perhaps reducing the power of this analysis.

KM analyses were performed to examine therelationship between the quantitative levels of H2T orH2D and BCS. In the absence of data suggesting anoptimum cutoff, ‘‘high’’ was defined as greater than orequal to the median value of the distribution and ‘‘low’’

as less than the median of the distribution. The left panelof Figure 4A shows the univariate KM plot for H2T.Figure 4C shows the KM plot for the multivariate scorebased on Model A. Correspondingly, the left panel of

FIGURE 1. A and B, VeraTag correlated with FISH. Correlationof HER2 expression by VeraTag with HER2 amplification bytotal FISH copy number (A), and centromere 17 corrected FISHcopy number (B) in 19 FFPE breast cancer controls from theMT Group. FFPE indicates formalin-fixed, paraffin-embedded;FISH, fluorescence in situ hybridization.




Figure 4B shows the univariate KM plot for H2D, andFigure 4D shows the KM plot of the multivariate scorebased on Model B.

In the subgroup of patients treated with trastuzu-mab alone (N=30), patients with high H2T or H2Dsurvived longer than those in the low group (HR=0.27,P=0.035 for H2T, HR=0.29, P=0.049 for H2D)(Figs. 4E, F).

Quantitative Levels of HER2 Expression orHER2:HER2 Dimers and Response toTrastuzumab Alone or in CombinationWith Chemotherapy

Finally, as shown in Figures 4A and B, treatmentresponse seems to differ between patient groups with‘‘high’’ H2T or H2D and those with ‘‘low’’ H2T or H2Das discriminated by VeraTag. Those in the high groupseemed to gain little from the addition of chemotherapyto trastuzumab (see right panels), whereas those inthe low group benefited significantly from concomitantchemotherapy and trastuzumab. The interaction test wasnot statistically significant. There was no correlation

FIGURE 2. VeraTag versus FISH in Bordet. Correlation withcentromere 17-corrected FISH copy number in the Bordetcohort (N = 64). FISH indicates fluorescence in situ hybridization.

FIGURE 3. Clinical benefit by RECIST. Correlation of VeraTagmeasures of HER2 expression and HER2:HER2 dimers withclinical benefit by RECIST responses in the Bordet cohort(clinical benefit = CR+PR+SD>6 mo). Quartile 1 = lowest 25%of HER2 expression/dimer distribution, quartile 4 = highest25% of HER2 expression/dimer distribution. RECIST indicatesResponse Evaluation Criteria in Solid Tumors.

TABLE 2. Multivariate Cox Proportional Hazards Analyses

Model Variable PHazard Ratio (95%

Confidence Interval)

Cox Proportional Hazards Models: prognostic factors for breast cancersurvivalA No. metastatic sites 0.00019 2.4 (1.5-3.9)N=71 HER2 expression (log10) 0.0058 0.24 (0.087-0.66)

Trastuzumab-onlytreatment arm

0.036 2.8 (1.07-7.3)

B No. metastatic sites 0.00024 24 (1.5-3.8)N=71 HER2:HER2 dimers

(log10)0.016 0.42 (0.21-0.85)


0.021 3.2 (1.2-8.6)

C No. metastatic sites 0.00048 2.5 (1.5-4.1)N=64* HER2 expression (log10) 0.017 0.24 (0.075-0.77)

Cox Proportional Hazards Models: prognostic factors for time toprogressionD No. metastatic sites 0.0031 1.7 (1.2-2.4)N=71 HER2 expression (log10) 0.031 0.51 (0.28-0.94)


0.0017 3.0 (1.5-5.9)

E No. metastatic sites 0.0053 1.63 (1.2-2.3)N=71 HER2:HER2 dimers

(log10)0.039 0.61 (0.33-0.98)


0.0015 3.1 (1.5-6.2)

F No. metastatic sites 0.0075 1.6 (1.1-2.3)N=64* HER2 expression (log10) 0.067 0.52 (0.26-1.05)


0.0074 2.6 (1.3-5.4)

All measures of HER2 expression and HER2:HER2 dimers were made usingthe VeraTag technology.

Cox proportional hazards analyses were performed to identify prognosticfactors for clinical outcome. Using either BCS or TTP as the end point, modelswere generated to explain the observed data. The final models for each end point(BCS and TTP) are shown.

*Models C and F only include patients who are positive by fluorescence in situhybridization.

BCS indicates breast cancer survival; TTP, time to progression.





between these groupings and estrogen receptor status(data not shown).

DISCUSSIONCurrent methodologies employed in the selection of

patients with MBC for treatment with trastuzumab aresemiquantitative (IHC and FISH) or are indirect mea-sures (FISH) of the expression of the target of trastuzu-mab, HER2. Only a subset of patients with MBC whoare selected to receive trastuzumab based on these assaysrespond, and a significant percentage of the respondersrelapse within 1 year.2,10,11,21 In this paper, we describethe initial correlations of HER2 expression and HER2:HER2 dimers, as measured by the VeraTag assay, withclinical outcomes in a well-described cohort of patientswith HER2-amplified or overexpressed MBC treated withtrastuzumab.

These data suggest several interesting interpreta-tions. First, in a cohort of patients considered to beHER2-amplified or HER2-overexpressed on the basis ofFISH or IHC, respectively (90% selected by FISH, 10%by IHC with mandatory confirmation of HER2 statusand independent review), quantitative measures of HER2expression or HER2:HER2 dimers, as provided byVeraTag, describe a continuum of response to trastuzu-mab. The Cox proportional hazards analyses of BCSsuggest that higher HER2 expression or HER2:HER2dimer levels may predict better outcomes after trastuzu-mab treatment than lower levels.

MBCs that express very high levels of HER2 orHER2:HER2 dimers may be more susceptible to trastu-zumab because the signaling cascades that drive theirproliferation and survival may be more dependent onassociations that are targets of the drug, whereasthose tumors that express lower levels of HER2 or havelower levels of HER2:HER2 dimers may derive their

FIGURE 4. A to F, Kaplan-Meier analyses. Univariate KM plots of BCS and HER2 expression (A) or HER2:HER2 dimers (B). Patientswere treated with trastuzumab (H only) or trastuzumab+chemotherapy (H+CT). C and D, Breast cancer survival by multivariatescore based on Cox models A and B. E and F, Univariate KM plots in the subpopulation of patients who were treated withtrastuzumab alone.




proliferative potential from a more heterogeneous set ofprotein interactions, not all of which are specificallyantagonized by trastuzumab. Tumors that are moreheterogeneous, from a signaling perspective, may bebetter equipped to develop resistance to trastuzumabthrough alternative signaling mechanisms than tumorsthat are more homogeneously dependent on signalingthrough protein complexes that are targeted by trastuzu-mab. Prime candidates that have been hypothesized toexplain in vivo reductions in susceptibility to trastuzumabinclude HER2:HER3 dimers, HER1:HER2 dimers, andp95/HER2 (the constitutively activated, cleaved intra-cellular domain of HER2), and signaling mediated bynon–EGFR-family receptor tyrosine kinases such asIGF-1R.22–33 The identification of such complexes inbreast tumors that exhibit clinical resistance to trastuzu-mab offers the possibility of designing combinationtherapy trials including trastuzumab and targeted inhibi-tors of HER1 (eg, erlotinib, gefitinib, lapatinib), HER3(pertuzumab), or p95/HER2 (lapatinib) for patientswhose tumors depend on heterogeneous signaling eventsfor their proliferation and survival. VeraTag assaysdesigned to make these measurements are currently indevelopment.

Second, the ability of the VeraTag assay to measureprotein dimerization using FFPE specimens is a novelcapability, and suggests a way to test the hypothesis thatcombination therapy regimens selected on the basisof ‘‘functional’’ pathway activation may be more clinicallyuseful than measuring the presence or absence of aparticular gene or gene product. In this study, weobserved a high correlation between HER2 expressionlevels and HER2:HER2 dimer levels, and this may reflectthe biology of HER2 self-association. It remains to bedetermined which of these markers is a better predictor ofsusceptibility to trastuzumab, but it seems likely that theability to measure HER2 heterodimers may have greaterclinical relevance.

Third, the observation that patients whose tumorsexpress high levels of HER2 or HER2:HER2 dimers seemto benefit little from the addition of chemotherapy totrastuzumab whereas their counterparts in the lower halfof the distribution seem to benefit significantly fromcombination therapy is potentially important. These datamust be interpreted cautiously, as patients were notprospectively randomized to treatment with trastuzumabalone, but rather were selected by their treating oncolo-gists in the course of their individualized clinical manage-ment. Although provocative, these data must beconsidered as hypothesis-generating, and require furtherinvestigation in a prospective trial to be confirmed.

The clinical impact of these data is uncertain at thistime, and more data from larger, well-controlled clinicaltrials are needed to confirm these observations. However,these findings are novel and potentially important inseveral respects. Although it would be incorrect toconclude that patients in the low HER2 expression/dimer group are not benefiting from trastuzumab, thedata suggest that not all patients selected for trastuzumab

therapy by the best currently available methods have thesame probability of response and survival. This isconsistent with what is already known about clinicaloutcome after trastuzumab therapy in patients withMBC: not all patients fare equally well on trastuzumab.The VeraTag technology may offer the opportunity todiscriminate between groups of patients with differentprobabilities of clinical response, and to begin to makebetter decisions regarding their management, as well as toidentify patients as appropriate candidates for trastuzu-mab who may be missed by currently available assays.The knowledge that particular patients have a higherprobability of early failure may identify them as bettercandidates for combination therapy trials than theircounterparts who may be unlikely to show demonstrablebenefit from the addition of other targeted inhibitors to atrastuzumab-based regimen. In addition, the identifica-tion of a subgroup of patients for whom concomitantchemotherapy adds little benefit, and who could bespared the unnecessary risk of toxicity, would be highlydesirable. Finally, the ability to use measurements ofdimerization events between HER-family receptor tyro-sine kinases to characterize reduced susceptibility totrastuzumab, and to identify functionally active proteintargets in individual patients, offers hope that the rationaldesign of combination therapy trials for patients withbreast cancer may be on the horizon.

REFERENCES1. Burstein HJ. The distinctive nature of HER2-positive breast cancers.

N Engl J Med. 2005;353:1652–1654.2. Cobleigh MA, Vogel CL, Tripathy D, et al. Multinational study of

the efficacy and safety of humanized anti-HER2 monoclonalantibody in women who have HER2 overexpressing metastaticbreast cancer that has progressed after chemotherapy for metastaticdisease. J Clin Oncol. 1999;17:2639–2648.

3. Eppenberger-Castori S, Kueng W, Benz C, et al. Prognostic andpredictive significance of ErbB-2 breast tumor levels measured byenzyme immunoassay. J Clin On-col. 2001;19:645–656.

4. Slamon DJ, Clark GM, Wong SG, et al. Human breast cancer:correlation of relapse and survival with amplification of the HER-2/neu oncogene. Science (New York, NY). 1987;235:177–182.

5. Slamon DJ, Godolphin W, Jones LA, et al. Studies of the HER-2/neu protooncogene in human breast and ovarian cancer. Science(New York, NY). 1989;244:707–712.

6. Baselga J, Perez EA, Pienkowski T, et al. Adjuvant trastuzumab:a milestone in the treatment of HER-2-positive early breast cancer.The Oncologist. 2006;11(suppl 1):4–12.

7. Hortobagyi GN. Trastuzumab in the treatment of breast cancer.N Engl J Med. 2005;353:1734–1736.

8. Piccart-Gebhart MJ, Procter M, Leyland-Jones B, et al. Trastuzu-mab after adjuvant chemotherapy in HER2-positive breast cancer.N Engl J Med. 2005;353:1659–1672.

9. Romond EH, Perez EA, Bryant J, et al. Trastuzumab plus adjuvantchemotherapy for operable HER2-positive breast cancer. N EnglJ Med. 2005;353:1673–1684.

10. Slamon DJ, Leyland-Jones B, Shak S, et al. Use of chemotherapyplus a mono clonal antibody against HER2 for metastatic breastcancer that overexpresses HER2. N Engl J Med. 2001;344:783–792.

11. Vogel CL, Cobleigh MA, Tripathy D, et al. Efficacy and safety oftrastuzumab as a single agent in first-line treatment of HER2-overexpressing metastatic breast cancer. J Clin Oncol. 2002;20:719–726.

12. Dressler LG, Berry DA, Broadwater G, et al. Comparison of HER2status by fluorescence in situ hybridization and immunohistochemistry





to predict benefit from dose escalation of adjuvant doxorubicin-basedtherapy in node-positive breast cancer patients. J Clin Oncol. 2005;23:4287–4297.

13. Pauletti G, Dandekar S, Rong H, et al. Assessment of methods fortissue-based detection of the HER-2/neu alteration in human breastcancer: a direct comparison of fluorescence in situ hybridization andimmunohistochemistry. J Clin Oncol. 2000;18:3651–3664.

14. Perez EA, Suman VJ, Davidson NE, et al. HER2 testing by local,central, and reference laboratories in specimens from the NorthCentral Cancer Treatment Group N9831 intergroup adjuvant trial.J Clin Oncol. 2006;24:3032–3038.

15. Press MF, Sauter G, Bernstein L, et al. Diagnostic evaluation ofHER-2 as a molecular target: an assessment of accuracy andreproducibility of laboratory testing in large, prospective, random-ized clinical trials. Clin Cancer Res. 2005;11:6598–6607.

16. Press MF, Slamon DJ, Flom KJ, et al. Evaluation of HER-2/neugene amplification and overexpression: comparison of frequentlyused assay methods in a molecularly characterized cohort of breastcancer specimens. J Clin Oncol. 2002;20:3095–3105.

17. Ross JS, Fletcher JA, Linette GP, et al. The Her-2/neu gene andprotein in breast cancer 2003: biomarker and target of therapy.Oncologist. 2003;8:307–325.

18. Wolff AC, Hammond ME, Schwartz JN, et al. American Societyof Clinical Oncology/College of American Pathologists guidelinerecommendations for human epidermal growth factor receptor 2testing in breast cancer. J Clin Oncol. 2007;25:118–145.

19. Therasse S, Arbuck SG, Eisenhauer E, et al. New guidelines toevaluate the response to treatment in solid tumors. J Natl CancerInst. 2000;92:205.

20. Shi Y, Huang W, Tan Y, et al. A novel proximity assay for thedetection of proteins and protein complexes: quantitation of HER1and HER2 total protein expression and homodimerization informalin-fixed, paraffin-embedded cell lines and breast cancer tissue.Diagn Mol Pathol. 2009;18:11–21.

21. Pegram MD, Lipton A, Hayes DF, et al. Phase II study of receptor-enhanced chemosensitivity using recombinant humanized anti-p185HER2/neu monoclonal antibody plus cisplatin in patients withHER2/neu-overexpressing metastatic breast cancer refractory tochemotherapy treatment. J Clin Oncol. 1998;16:2659–2671.

22. Albanell J, Baselga J. Unraveling resistance to trastuzumab(Herceptin): insulin-like growth factor-I receptor, a new suspect.J Natl Cancer Inst. 2001;93:1830–1832.

23. DiGiovanna MP, Stern DF, Edgerton SM, et al. Relationship ofepidermal growth factor receptor expression to ErbB-2 signalingactivity and prognosis in breast cancer patients. J Clin Oncol. 2005;23:1152–1160.

24. Hendriks BS, Opresko LK, Wiley HS, et al. Coregulation ofepidermal growth factor receptor/human epidermal growth factorreceptor 2 (HER2) levels and locations: quantitative analysis ofHER2 overexpression effects. Cancer Res. 2003;63:1130–1137.

25. Holbro T, Beerli RR, Maurer F, et al. The ErbB2/ErbB3heterodimer functions as an oncogenic unit: ErbB2 requires ErbB3to drive breast tumor cell proliferation. Proc Natl Acad Sci USA.2003;100:8933–8938.

26. Nahta R, Yu D, Hung MC, et al. Mechanisms of disease:understanding resistance to HER2-targeted therapy in humanbreast cancer. Nat Clin Pract. 2006;3:269–280.

27. Nahta R, Yuan LX, Zhang B, et al. Insulin-like growth factor-Ireceptor/human epidermal growth factor receptor 2 heterodimeriza-tion contributes to trastuzumab resistance of breast cancer cells.Cancer Res. 2005;65:11118–11128.

28. Pandolfi PP. Breast cancer—loss of PTEN predicts resistance totreatment. N Engl J Med. 2004;351:2337–2338.

29. Pinkas-Kramarski R, Soussan L, Waterman H, et al. Diversificationof Neu differentiation factor and epidermal growth factor signalingby combinatorial receptor interactions. EMBO J. 1996;15:2452–2467.

30. Sergina NV, Rausch M, Wang D, et al. Escape from HER-familytyrosine kinase inhibitor therapy by the kinase-inactive HER3.Nature. 2007;445:437–441.

31. Slamon DJ. The FUTURE of ErbB-1 and ErbB-2 pathwayinhibition in breast cancer: targeting multiple receptors. Oncologist.2004;9(suppl 3):1–3.

32. Yarden Y, Sliwkowski MX. Untangling the ErbB signallingnetwork. Nat Rev. 2001;2:127–137.

33. Saez R, Molina MA, Ramsey EE, et al. p95HER-2 predicts worseoutcome in patients with HER-2-positive breast cancer. Clin CancerRes. 2006;12:424–431.




Documents

Quantitation of HER2 expression or HER2: HER2 dimers and differential survival in a cohort of metastatic breast cancer patients carefully selected for trastuzumab treatment primarily