UBC Faculty Research and Publications

Use of immunohistochemical markers can refine prognosis in triple negative breast cancer Tischkowitz, Marc; Brunet, Jean-Sébastien; Bégin, Louis R; Huntsman, David G; Cheang, Maggie C; Akslen, Lars A; Nielsen, Torsten O; Foulkes, William D Jul 24, 2007

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-12885_2007_Article_784.pdf [ 571.21kB ]
JSON: 52383-1.0220861.json
JSON-LD: 52383-1.0220861-ld.json
RDF/XML (Pretty): 52383-1.0220861-rdf.xml
RDF/JSON: 52383-1.0220861-rdf.json
Turtle: 52383-1.0220861-turtle.txt
N-Triples: 52383-1.0220861-rdf-ntriples.txt
Original Record: 52383-1.0220861-source.json
Full Text

Full Text

ralssBioMed CentBMC CancerOpen AcceResearch articleUse of immunohistochemical markers can refine prognosis in triple negative breast cancerMarc Tischkowitz*1,2, Jean-Sébastien Brunet1,3, Louis R Bégin4, David G Huntsman5, Maggie CU Cheang5, Lars A Akslen6, Torsten O Nielsen5 and William D Foulkes1,2Address: 1Program in Cancer Genetics, McGill University, Montréal, Québec, Canada, 2Cancer Prevention Centre, Segal Cancer Centre, Sir M.B. Davis-Jewish General Hospital, Montréal, Québec, Canada, 3Algorithme Pharma, Laval, Québec, Canada, 4Hôpital du Sacré-Coeur de Montréal, Québec, Canada, 5Genetic Pathology Evaluation Centre, BC Cancer Agency, UBC, Vancouver, BC, Canada and 6The Gade Institute, Section for Pathology, University of Bergen and Haukeland University Hospital, Bergen, NorwayEmail: Marc Tischkowitz* - marc.tischkowitz@mcgill.ca; Jean-Sébastien Brunet - jsbrunet@algopharm.com; Louis R Bégin - mdlrb@yahoo.ca; David G Huntsman - dhuntsma@bccancer.bc.ca; Maggie CU Cheang - chon@interchange.ubc.ca; Lars A Akslen - Lars.Akslen@gades.uib.no; Torsten O Nielsen - torsten@interchange.ubc.ca; William D Foulkes - william.foulkes@mcgill.ca* Corresponding author    AbstractBackground: Basal-like breast cancer has been extensively characterized on the basis of geneexpression profiles, but it is becoming increasingly common for these tumors to be defined on thebasis of immunohistochemical (IHC) staining patterns, particularly in retrospective studies wherematerial for expression profiling may not be available. The IHC pattern that best defines basal-liketumors is under investigation and various combinations of ER, PR, HER2-, CK5/6+ and EGFR+ havebeen tested.Methods: Using datasets from two different hospitals we describe how using differentcombinations of immunohistochemical patterns has different effects on estimating prognosis atdifferent time intervals after diagnosis. As our baseline, we used two IHC patterns ER-/PR-/HER2-("triple negative phenotype", TNP) and ER-/HER2-/CK5/6+ and/or EGFR+ ("core basal phenotype",CBP).Results: There was no overall difference in survival between the two hospital-based series, butthere was a difference between the TNP and non-TNP groups which was most marked at 3 years(76.8% vs 93.5%, p < .0001). This difference reduced with time, suggesting that long term survivors(beyond 10 years) in the TNP group may have comparable survival to non-TNP cases. A similardifference was seen if CBP was used instead of TNP. However when CK5/6 and/or EGFRexpressing tumors were analyzed without consideration of ER/PR status, the reduction in survivalincreased with time, becoming more pronounced at 10 years than at 3 years.Conclusion: Our findings suggests that CK5/6 and/or EGFR expressing tumor types have apersistently poorer prognosis over the longer term, an observation that may have importanttherapeutic implications as drugs that target the EGFR are currently being evaluated in breastPublished: 24 July 2007BMC Cancer 2007, 7:134 doi:10.1186/1471-2407-7-134Received: 10 March 2007Accepted: 24 July 2007This article is available from: http://www.biomedcentral.com/1471-2407/7/134© 2007 Tischkowitz et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.Page 1 of 11(page number not for citation purposes)cancer.BMC Cancer 2007, 7:134 http://www.biomedcentral.com/1471-2407/7/134BackgroundGene expression studies using DNA microarrays haveidentified several distinct breast cancer subtypes whichdiffer significantly in prognosis [1,2]. These subtypesinclude three main subtypes of estrogen receptor (ER)negative tumors – the basal-like (ER-/HER2-), the ER-/HER2+ subtype and the normal-like/unclassified subtype,and at least 2 types of ER+ tumors (luminal A and luminalB) [2]. Basal-like tumors typically show high expression ofgenes characteristic of the basal epithelial cells of the nor-mal mammary gland, including stratified epithelial cytok-eratins, such as cytokeratins 5, 14, 15 and 17 [3]. Basal-like breast cancers account for around 15% of all invasiveductal breast cancers of no special type [4] with a higherprevalence among African-American women [5]. Conven-tional histopathological as well as molecular studies ofbreast cancers with "basaloid" differentiation have shownthat basal-like tumors are often high grade [6], have areasof necrosis [7], may have a typical or an atypical medul-lary phenotype [8] and have a distinct pattern of geneticalterations [6], including frequent TP53 mutations [2]. Ahigh proportion of BRCA1 tumors exhibit the basal-likephenotype and germ-line BRCA1 mutations result inbreast cancers that are more likely to be basal-like innature [9-12].The breast cancer subtypes have been extensively charac-terized by gene expression analysis using DNA microar-rays and while this remains the gold-standard, it is notcurrently feasible for large-scale clinical applications orretrospective studies using formalin fixed, paraffin-embedded samples. In these situations the immunohisto-chemical staining profile (IHC) can be a useful surrogateof gene expression analysis. However the optimumimmunohistochemical profile of basal-like breast cancerremains unclear. The "triple negative phenotype", TNP(ER-, PR-, HER-) is increasingly used as a surrogate markerfor basal-like breast cancer as it has the advantage thatthese three stains are already used routinely in clinicalwork-up of breast cancers [13]. Although most basal-liketumors do not express ER, PR, and HER2, some may, andthe overlap between basal-like and TNP breast cancer isnot complete [14]. Moreover, the ER-, HER2-, CK5/6+and/or EGFR+ profile seems to correlate better with basal-like breast cancer gene expression profiles [3,15]. Wesought to clarify how the utilization of different markercombinations affects prognostic outcome.We have previously shown that tumors expressing theCK5/6 marker are associated with germline BRCA1 muta-tions based on data on unselected breast cancer cases froma single institution (Jewish General Hospital, JGH) [9].Here we present data on both the JGH series (n = 192) andrelations between IHC profiles and outcome in basal-likecancer.MethodsClinicopathological Review and IHCFor the JGH series, the study design is an ethnicallyrestricted single hospital-based retrospective cohort study,as described previously [9]. Of 309 consecutive cases ofAshkenazi Jewish women age 65 or less diagnosed with afirst primary, non-metastatic, invasive breast cancerbetween January 1, 1980 and November 1, 1995 at the SirMortimer B. Davis-Jewish General Hospital, Montreal,QC, 17 (5.5%) were excluded (because (a) we were una-ble to locate pathology blocks or (b) we found only carci-noma in situ was present on the available path blocks,leaving 292 cases. 192 cases had sufficient material to gen-erate a tissue microarray. Blocks were identified from eachof these women, and clinicopathological and follow-upinformation were obtained by chart review. All of thespecimens were reviewed by one pathologist (L. R. B.) forhistological type, nuclear/histological grade, and lymphnode status, and were stained for ER, PR and HER2 andCK5/6 IHC, as described previously [9]. The VGH studygroup comprised women with primary invasive breastcancer who underwent surgery for breast cancer between1974 and 1995 at Vancouver General Hospital. Thesewere consecutive cases, and the presence of invasive breastcarcinoma was the only selection criterion in this study.Outcome data were available for all of the patients, withmedian follow-up of 15.4 years (range, 6.3–26.6 years)and the assembly of archival tumor blocks into tissuemicroarrays, IHC and scoring were as described previously[16]. Epidermal growth factor receptor (EGFR) immunos-tains were also applied to both series, using methodsdescribed previously [15]. Information on the adjuvantuse of hormone therapy or chemotherapy was obtainedfrom the clinical record; these data were available on 440cases and 448 cases respectively out of the total of 456cases in the combined series. All cases had been collectedas part of studies which were subject to ethical approvalobtained from the local institutional ethical reviewboards. (McGill University/Jewish General Hospital andVancouver General Hospital).Statistical AnalysisClinical, pathological, and molecular data were collectedin a mutually blinded fashion. Patient characteristics werecompared using nonparametric Wilcoxon's test andFisher's exact test. Borderline statistical significance wasdefined as P-values between .05 and .10. Survival rateswere calculated from the date of primary surgery untildeath from breast cancer (breast cancer-specific survival).The median follow-up of those who did not die of breastPage 2 of 11(page number not for citation purposes)a second series of 264 breast cancer cases from the Van-couver General Hospital (VGH), focusing specifically cor-cancer was 11.13 years (n = 330; overall follow-up was8.84 years, n = 456). Ten-year survival curves were esti-BMC Cancer 2007, 7:134 http://www.biomedcentral.com/1471-2407/7/134mated using the Kaplan-Meier method, and significancewas assessed with the log-rank test.To estimate the relative risk (RR) of death from breast can-cer, three Cox proportional hazards models were built, allof which included the following measured prognostic fac-tors: Center, age of diagnosis, tumor size, axillary lymphnode status and histological grade. The first model wasbuilt to assess the importance of TNP and included termsfor TNP and [CK5/6 and/or EGFR] positive status. The sec-ond model was built to assess the importance of CBP andincluded terms for CBP and PR positive status. The thirdmodel was built to assess the importance of each compo-nent of the TNP and CBP criteria and included terms forER and HER2 negative status, PR and [CK5/6 and/orEGFR] positive status. In all three models missing valueswere factored in by creating a dichotomized variable (X,IX) to identify whether or not the variable of interest wasmissing using the following method: X = 1 if positive, X =0 otherwise, IX = 1, if missing, IX = 0 otherwise. Thus asubject with missing values was (X, IX) = (0, 1), a positivemarker was (X, IX) = (1, 0) and a negative marker was (X,IX) = (0, 0). This allowed us to include all 456 subjects(compared to 327 subjects without adjustment for miss-ing).The survival model was reanalyzed separately for 244node-negative and 162 node-positive subjects. Survivaldata were analyzed and censored at both 3 years time andat 10 years time, and significance was assessed at the 5%level using two-sided tests. All three parsimonious modelswere built using the log-likelihood ratio test, employing abackward approach in which variables with the highestcontribution to the likelihood function were kept in themodel and where the parsimonious model was assumedwhen all P-values were below a 10% threshold. The parsi-monious model is thus built by removing the variableswith the least amount of influence to the likelihood func-tion; depending on the correlations between each covari-ate, some will be kept while others will lose statisticalsignificance and be removed. All three models start byincluding the same covariate, that is, center, age of diagno-sis, tumor size, axillary lymph node status and histologi-cal grade. Model 1 also includes the variable TNP and{CK5/6 or EGFR}. Model 2 also includes the variable CBPand PR. Model 3 also includes the variable ER, HER2, PRand {CK56 or EGFR}.As there was an upper age limit in the JGH series and noupper age limit for the VGH series, the analyses wererepeated without the VGH cases over 65 years. Since thefinal results did not essentially differ with and withoutolder cases, all subjects were kept in the statistical analysis.adjuvant chemotherapy or hormonal therapy on progno-sis.A Poisson regression model was built to examine the rela-tionship between the number of positive lymph nodesand tumor size in TNP+ and TNP- cases: ln(μ) = ln(Nexam)+ α + αTNP+ + β*Tsize + βTNP+ *Tsize where: μ = averagenumber of positive nodes, α = overall intercept, αTNP+ =extra intercept for TNP+ patients, β = overall slope, βTNP+= extra slope for TNP+ patients, and the natural logarithmof the number of nodes examined was used as an offset.ResultsAt 10 years time, there was no overall difference in sur-vival for all breast cancer types between the JGH and VGHcenters (p = .17) and TNP tumors made up 14% of casesin both series (JGH, 27 tumors, VGH, 36 tumors). In thecombined series, the median age at diagnosis was 9.4years younger in the TNP group (p = .0006) and themedian length of follow up in survivors was 6.75 years inthe TNP groups versus 9.09 years in the non-TNP group, p= .02. Exclusion of subjects who did not die of breast can-cer and were lost to follow up before 10 years did not alterthe statistical significance of the study results. There was asignificant overlap between the TNP and CBP groups with49/58 (84%) of TNP cases also falling in the CBP group.Comparison of clinical features in TNP and non-TNPcases in the combined series (Table 1) showed that TNPcases had an increased likelihood of a higher histologicalgrade (odds ratio (OR), for grade 3: 17.7 [95% confidenceinterval, C.I., 6.05–51.5], p < .0001), a larger tumor (ORfor tumor >2 cm: 1.85 [95% C.I. 1.04–3.32], p = .04) buthad a decreased likelihood of positive lymph nodes (OR= 0.44 [95% C.I. 0.23–0.84], p = .01). While there was aclear correlation between tumor size and the meannumber of positive lymph nodes in both the non-TNPand the TNP group, this correlation was less strong withthe TNP group (P = 0.01) and the interaction betweentumor size and TNP status on lymph node status was ofborderline significance (p = 0.10, Figure 1). Breast cancersurvival at 3 and 10 years correlated closely with histolog-ical grade, size and lymph node involvement (Table 2).The effect of TNP on prognosis was stronger at 3 yearsthan at 10 years, with TNP conferring a univariate RR of4.06 [95% C.I. 2.11–7.82], p = .0001 at 3 years (Table 3)compared to 1.71 [95% C.I. 1.05–2.78], p = .03 at 10years (Table 4). Although there is a degree of overlapbetween the confidence intervals at 3 years and at 10years, this is small and the fact that the TNP parameter isnot present in parsimonious model 1 at 10 years (Table 4)but is present in the same model at 3 years provides fur-ther evidence indicating that the differences are real. Asimilar pattern was seen with the CBP variable. Predicta-Page 3 of 11(page number not for citation purposes)Similar models were used to determine the influence of bly, TNP cases were less likely to receive hormone therapyand more likely to receive chemotherapy (Table 1). At 10BMC Cancer 2007, 7:134 http://www.biomedcentral.com/1471-2407/7/134years, survival was 63% in TNP cases treated with chemo-therapy versus 66% in the no treatment group; the corre-sponding figures for CBP were 68% and 62%. Thesedifferences were not significant and there was also no dif-ference in survival with adjuvant hormone therapy.In the combined JGH/VGH series, the difference in sur-vival between the TNP and non-TNP groups (Figure 2)was most marked at 3 years with an absolute reduction of16.7% in the TNP group (76.8% versus 93.5%, p < .0001).Although the absolute reduction in survival of 9.2% at 10years in the TNP group was still significant (p = .03), thedifference appeared to be reducing with time, suggestingthat long term survivors in the TNP group may have acomparable survival to non-TNP cases. When using CBPinstead of TNP a similar overall survival pattern emergedwith a significant difference at 3 years (77.4% versus93.4%, p =< .0001) that also became less marked at 10years.However, when tumors negative for CK5/6 and EGFRexpression were compared to tumors that expressed eitherCK5/6 or EGFR (Figure 3) the absolute survival differencewas notably greater at 10 years (17.1%, p = .0007), thanthat at 3 years (7.8%, p = .02). This was reflected in themultivariate parsimonious models (Table 3 and 4) whichshowed that at 3 years time, both TNP and CBP parame-ters not only remained in their respective parsimoniousmodels but both also worked well in predicting outcome(models 1 and 2). As both models share ER negative andHER2 negative status, these appear to be the main drivingfactors predicting early outcome. Indeed, when all com-ponents of TNP and CBP are analyzed separately (model3), only ER negative and HER2 negative status remainedin the parsimonious model while positive PR status and[CK5/6+ and/or EGFR+] status fell out of the model.In contrast, the data at 10 years indicates that ER negativeand HER negative status diminishes in influence withincreasing time, with [CK5/6+ and/or EGFR+] statusbecoming the main driving factor. Therefore CBP (whichincorporates CK5/6 and/or EGFR+) may be a better modelat 10 years.DiscussionThe data presented here show that different immunohis-tochemical marker combinations may influence progno-sis at different points in time. This is in agreement with therecent findings of Anderson et al. who using the SEERdatabase found that at 17 months, ER- hazard ratespeaked at 7.5% per year then declined, whereas ER+ haz-ard rates were comparatively constant at 1.5–2% per year;falling ER- and constant ER+ hazard rates crossed at 7years after which time prognosis was better for ER+ cases[17].TNP had a marginally greater effect on prognosis in lymphnode negative patients compared to lymph node positivepatients (Figures 4 and 5). The univariate relative risk forbreast cancer death at 3 years in the TNP group versus theTable 1: Age at diagnosis, tumor characteristics and treatment given in the TNP and non-TNP groups.Clinical Feature No. of cases Non-TNP No. of cases TNP Odds Ratio [95% CI] P-value*Center VGH 228 36JGH 165 27 1.04 [0.61; 1.77] 0.89Age at diagnosis 25–49 years 116 3250–95 years 277 31 0.41 [0.24; 0.70] 0.001Histological grade Grade 1 120 4Grade 2 193 15 2.33 [0.76; 7.19] 0.15Grade 3 68 40 17.65 [6.05; 51.45] <0.0001Missing 12 4Tumor size 0–2 cm 161 182.01–15 cm 217 45 1.85 [1.04; 3.32] 0.04Missing 15 0Lymph node status Negative 205 41Positive 149 13 0.44 [0.23; 0.84] 0.01Missing 39 9Hormonal Treatment None 219 49Yes 161 11 0.31 [0.15; 0.61] 0.0003Missing 13 3Chemotherapy None 279 35Yes 108 26 1.92 [1.10; 3.34) 0.02Missing 6 2Page 4 of 11(page number not for citation purposes)* P-values based on Fisher's exact test.BMC Cancer 2007, 7:134 http://www.biomedcentral.com/1471-2407/7/134non-TNP group was 5.40 and 3.48 among lymph nodenegative patients and lymph node positive patientsrespectively, giving a magnitude of reduction in survival at3 years of 0.68 in lymph node positive patients comparedto lymph node negative patients. Whether lymph nodestatus has less prognostic value in basal-like breast cancersremains a contentious issue. For example in their recentstudy of African-American women, Carey et al did not seea poor survival in lymph node negative basal-like breastcancer, but it was very poor in lymph node positive cases[5] while other groups have found that node-negativebasal-like breast cancer also carries a poor prognosis [18].In this study the limited correlation between tumor sizeand mean number of lymph nodes in the TNP group andthe modest difference of the effect of TNP between lymphnode negative and lymph node positive groups suggestthat lymph node involvement is a less reliable predictor ofprognosis in the TNP group.Data from the first cohort studied (JGH) indicate thatbreast cancers with the TNP have a worse prognosis atleast in the first three years after diagnosis but this differ-ence in prognosis may diminish with time from diagno-sis. These data were validated in a second largeTable 3: Cox proportional model for survival until death from breast cancer at 3 years.Univariate Full Multivariate ParsimoniousModel Variable RR (95% CI) P RR (95% CI) P RR (95% CI) PModel 1 CK5/6+ and/or EGFR+ No 1. 1.Yes 2.24 [1.14; 4.39] 0.02 1.20 [0.50; 2.85] 0.68 not presentTNP No 1. 1. 1.Yes 4.06 [2.11; 7.82] 0.0001 3.42 [1.37; 8.49] 0.008 4.40 [2.23; 8.69] 0.0001Model 2 PR No 1. 1.Yes 0.43 [0.23; 0.81] 0.009 0.70 [0.34; 1.45] 0.34 not presentCBP No 1. 1. 1.Yes 3.73 [1.88; 7.39] 0.0002 3.13 [1.26; 7.77] 0.01 4.29 [2.12; 8.68] 0.0001Model 3 ER- HER2- No 1. 1. 1.Yes 4.39 [2.34; 8.23] 0.0001 4.82 [1.95; 11.9] 0.0006 5.38 [2.77; 10.5] 0.0001PR No 1. 1.Yes 0.43 [0.23; 0.81] 0.009 0.82 [0.38; 1.76] 0.61 not presentCK5/6+ and/or EGFR+ No 1. 1.Yes 2.24 [1.14; 4.39] 0.02 0.99 [0.42; 2.34] 0.98 not presentAll statistical modeling adjust for variables with missing values (see methods). The full multivariate model also includes terms for centre, age of diagnosis, histological grade, tumor size and lymph node. Parsimonious model was assumed when all P-values were below 10%. All three Table 2: Univariate Cox proportional hazards model for survival until death from breast cancer at 3 years and 10 years.3 years 10 yearsVariable Definition RR (95% CI) P RR (95% CI) PCenter VGH 1. 1.JGH 0.67 [0.35; 1.31] .25 0.76 [0.51; 1.13] .17Age of diagnosis < 50 years 1. 1.≥ 50 years 1.10 [0.55; 2.16] .79 0.97 [0.66; 1.44] .89Histological grade*, § 1 1. 1.2 2.62 [0.88; 7.80] .08 1.63 [0.97; 2.73] .063 5.47 [1.84; 16.2] .002 2.37 [1.36; 4.11] .002Tumor Size§ < 2 cm 1. 1.≥ 2 cm 3.22 [1.42; 7.32] .005 2.82 [1.78; 4.46] .0001Lymph nodes§ No 1. 1.Yes 2.78 [1.37; 5.65] .005 2.21 [1.46; 3.32] .0002* Test of overall histological grade. P-value = 0.004 (at 3 years) and 0.009 (at 10 years).§ Adjusting for missing values.The total number of subjects is 456. The total number of events is 39 at 3 years and 111 at 10 years.Page 5 of 11(page number not for citation purposes)parsimonious models also included terms for tumor size and lymph node status. Parsimonious model-2 also included a term for Histological grade while parsimonious model-3 also included a term for the centre. The total number of subjects is 456. The total number of events is 39 at 3 years.BMC Cancer 2007, 7:134 http://www.biomedcentral.com/1471-2407/7/134Page 6 of 11(page number not for citation purposes)Poisson regression curve examining the relationship between tumor size, lymph node status and TNP groupFigure 1Poisson regression curve examining the relationship between tumor size, lymph node status and TNP group. The number of positive lymph nodes showed a closer correlation with tumor size in the non-TNP group compared to the TNP group.ER-, PR-, HER2- versus others30Others (n=334) ER-, PR-, HER2- (n=52)25201015Number of positivelymphnodes505 6 7 8 9 10 110 1 2 3 4Tumor size (cm)Results:Etumor size =  0.3077    P-value  <.0001 ln(tot_node)  Offset=  0.1857    P-value  =.1085interaction slopeE= -1.6688  P-value  =.0109ER-,PR-,HER2-D= -2.9452  P-value  <.0001overallDPrediction is made at the overall median total node examined (i.e. 12).Table 4: Cox proportional model for survival until death from breast cancer at 10 years time.Univariate Full Multivariate ParsimoniousModel Variable RR (95% CI) P RR (95% CI) P RR (95% CI) PModel 1 CK5/6+ and/or EGFR+ No 1. 1. 1.Yes 2.00 [1.33; 3.02] 0.0009 1.95 [1.18; 3.21] 0.009 1.92 [1.27; 2.89] 0.002TNP No 1. 1.Yes 1.71 [1.05; 2.78] 0.03 0.96 [0.51; 1.79] 0.89 not presentModel 2 PR No 1. 1. 1.Yes 0.59 [0.40; 0.85] 0.005 0.64 [0.42; 0.98] 0.04 0.62 [0.42; 0.92] 0.02CBP No 1. 1.Yes 1.72 [1.06; 2.80] 0.03 1.18 [0.64; 2.17] 0.60 not presentModel 3 ER- HER2- No 1. 1. 1.Yes 1.96 [1.30; 2.97] 0.002 1.41 [0.80; 2.49] 0.24 1.64 [0.99; 2.70] 0.05PR No 1. 1.Yes 0.59 [0.40; 0.85] 0.005 0.77 [0.49; 1.20] 0.24 not presentCK5/6+ and/or EGFR+ No 1. 1. 1.Yes 2.00 [1.33; 3.02] 0.0009 1.54 [0.92; 2.58] 0.10 1.69 [1.04; 2.74] 0.03All statistical modeling adjust for variables with missing values (see methods). The full multivariate model also includes terms for centre, age of diagnosis, histological grade, tumor size and lymph node. Parsimonious model was assumed when all P-values were below 10%. All three parsimonious models also included terms for tumor size and lymph node status. Parsimonious model-2 also included a term for Histological grade while parsimonious model-3 also included a term for the centre. The total number of subjects is 456. The total number of events is 111 at 10 years.BMC Cancer 2007, 7:134 http://www.biomedcentral.com/1471-2407/7/134independent data set (VGH). Both data sets showed verysimilar overall survival curves suggesting that they are gen-erally comparable. Although both data sets were identi-fied retrospectively, this is counterbalanced by the factthat they originate from different centers and the degree ofconsistency between the two data sets which strengthensthe overall findings. However, given the small numbers ofcases analyzed, any impact of treatment on survival strat-ified by TNP status would have had to be very large tobecome apparent and further larger, preferably rand-omized, studies are required to assess this in greater detail.The data presented here add to the growing body of evi-dence that basal-like breast tumors have a worse progno-sis [13,15,18,19] and respond less well to chemotherapyat relapse [20], although there remains a significantdegree of heterogeneity within this group [6,21-24]. Someof these studies have used microarray-based gene expres-sion studies to identify the basal-like group and, whilethese studies are likely to be more accurate in delineatingpractice. The advantage of this study is that it uses markersthat are readily available in most pathology departmentsand is therefore directly translatable into routine clinicalmanagement, and can be applied to archival specimensfor which long-term follow-up information is alreadyavailable.A large number of new markers are emerging which aimto further delineate the basal phenotype [23,25]. How-ever, ER, PR, HER2, EGFR and CK5/6 are already routinelyavailable in most centers. ER, PR, and HER2 in particularare used to guide treatment decisions in breast cancer [26]and it is therefore important to know exactly how expres-sion of these markers affects prognosis. Haffty et al. haverecently published prognosis data on a series of 482patients, 117 of which had a TNP [13]. The median followup time was 7.9 years and TNP was an independent pre-dictor of disease-specific survival (hazard ratio = 1.79;95% CI 1.03–3.22). Another recent study showed that inthe neoadjuvant setting, patients with ER negative andSurvival until breast cancer death by TNP statusFigure 2Survival until breast cancer death by TNP status. Survival at 3 years time was 76.8% in the TNP cases versus 93.5% among non-TNP cases (p < .0001); Survival at 10 years time were, respectively 65.0% and 74.2% (p = .03).01020304050607080901000.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0years after breast cancer occurrence% survivalOthersER-, PR-, HER2-n at risk                 Others: 393                 388                377               354                330               303                282                258                228               197                172ER-, PR-, HER2- :  63                   58                 53                  45                  42                 38                  33                  30                  26                 26                  22p=<.0001 at 3 yearsp= .03 at 10 yearsPage 7 of 11(page number not for citation purposes)the basal phenotype than standard immunohistochemi-cal methods, they are not yet routinely used in clinicalHER2 negative breast cancer have higher sensitivity toanthracycline-based chemotherapy than the luminal sub-BMC Cancer 2007, 7:134 http://www.biomedcentral.com/1471-2407/7/134type, and have higher rates of pathologic completeresponse [27]. Several new drugs that target the EGFR inbreast cancer are currently being evaluated [28] and theobservations presented here suggest that the effects ofthese drugs may become more apparent over the longerterm, beyond the time over which a typical drug trialwould extend. This would have important implicationsfor trial design and interpretation of results. Finally, anincreasing number of immunohistochemical markers arebeing utilized in the identification of basal-like andBRCA1-related breast cancers [18,29-31], and further val-idation of these additional markers will be required if theyare also to be used as a guide to clinical prognosis andtherapeutic choices [14].ConclusionOur data confirms that TNP is useful as a prognosticmarker, but also suggests that the effect of TNP on survivalreduces over time and that focusing on CK5/6 and/orEGFR expression may provide a better marker for longterm prognosis (beyond three years).Competing interestsThe author(s) declare that they have no competing inter-ests.Authors' contributionsMT and WDF conceived and designed the study anddrafted the manuscript, JSB undertook the statistical anal-yses, MCUC, LRB, DGH, LAA and TON were involved inSurvival until breast cancer death by Ck5/6 and EGFR statusFigure 3Survival until breast cancer death by Ck5/6 and EGFR status. Survival at 3 years time was 85.6% among CK5/6 and/or EGFR positive cases versus 93.4% among cases that were negative for both CK5/6 and EGFR (p = .02); Survival at 10 years time was 61.4% and 78.5% (p = .0007) respectively.p=.02 at 3 yearsp=.0007 at 10 years01020304050607080901000.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0years after breast cancer occurrence% survivalCK5/6 - and EGFR -CK5/6+ or EGFR+n at risk                   Others:  244                243                239               226                215               199                186                173               158                148               130CK5/6+ or EGFR+ :  131                131                131               130                128               128                128                127               127                127               125Page 8 of 11(page number not for citation purposes)data acquisition and interpretation, and critically revisingthe manuscript.BMC Cancer 2007, 7:134 http://www.biomedcentral.com/1471-2407/7/134Page 9 of 11(page number not for citation purposes)Relationship between TNP, survival and lymph node status – lymph node positive subgroupFigure 4Relationship between TNP, survival and lymph node status – lymph node positive subgroup. In TNP cases, the magnitude of the decrease in survival was approximately 1.5-fold greater at 3 years in the lymph node positive subgroup (Fig-ure 4) compared to the lymph node negative subgroup (Figure 5).01020304050607080901000.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0years after breast cancer occurrence% survivalOthersER-, PR-, HER2-n at riskOthers:                  149                146               140               128                116               106               100                  87                  76                62                 56ER-, PR-, HER2-    13                  13                 11                 11                    7                   7                   7                    7                    7                  4                   3(lymph node positive)p=.02 at 3 yearsp=.20 at 10 yearsBMC Cancer 2007, 7:134 http://www.biomedcentral.com/1471-2407/7/134AcknowledgementsMT is funded by the Jewish General Hospital Weekend to End Breast Can-cer, Rethink Breast Cancer Canada, and the Canadian Foundation for Inno-vation. The Genetic Pathology Evaluation Centre is supported by an unrestricted educational grant from Sanofi-Aventis. DGH and TON are scholars of the Michael Smith Foundation for Health Research. WDF is funded by CBCRA.References1. Perou CM, Sorlie T, Eisen MB, van de Rijn M, Jeffrey SS, Rees CA, Pol-lack JR, Ross DT, Johnsen H, Akslen LA, Fluge O, Pergamenschikov A,Williams C, Zhu SX, Lonning PE, Borresen-Dale AL, Brown PO, Bot-stein D: Molecular portraits of human breast tumours.  Nature2000, 406(6797):747-752.2. Sorlie T, Perou CM, Tibshirani R, Aas T, Geisler S, Johnsen H, HastieT, Eisen MB, van de Rijn M, Jeffrey SS, Thorsen T, Quist H, Matese JC,Brown PO, Botstein D, Eystein Lonning P, Borresen-Dale AL: Geneexpression patterns of breast carcinomas distinguish tumorsubclasses with clinical implications.  Proc Natl Acad Sci U S A2001, 98(19):10869-10874.3. Yehiely F, Moyano JV, Evans JR, Nielsen TO, Cryns VL: Decon-structing the molecular portrait of basal-like breast cancer.Trends Mol Med 2006, 12(11):537-544.4. van de Rijn M, Perou CM, Tibshirani R, Haas P, Kallioniemi O,cytokeratins 17 and 5 identifies a group of breast carcinomaswith poor clinical outcome.  Am J Pathol 2002, 161(6):1991-1996.5. Carey LA, Perou CM, Livasy CA, Dressler LG, Cowan D, Conway K,Karaca G, Troester MA, Tse CK, Edmiston S, Deming SL, Geradts J,Cheang MC, Nielsen TO, Moorman PG, Earp HS, Millikan RC: Race,breast cancer subtypes, and survival in the Carolina BreastCancer Study.  Jama 2006, 295(21):2492-2502.6. Jones C, Nonni AV, Fulford L, Merrett S, Chaggar R, Eusebi V, LakhaniSR: CGH analysis of ductal carcinoma of the breast withbasaloid/myoepithelial cell differentiation.  Br J Cancer 2001,85(3):422-427.7. Tsuda H, Takarabe T, Hasegawa F, Fukutomi T, Hirohashi S: Large,central acellular zones indicating myoepithelial tumor differ-entiation in high-grade invasive ductal carcinomas as mark-ers of predisposition to lung and brain metastases.  Am J SurgPathol 2000, 24(2):197-202.8. Tot T: The cytokeratin profile of medullary carcinoma of thebreast.  Histopathology 2000, 37(2):175-181.9. Foulkes WD, Brunet JS, Stefansson IM, Straume O, Chappuis PO,Begin LR, Hamel N, Goffin JR, Wong N, Trudel M, Kapusta L, PorterP, Akslen LA: The prognostic implication of the basal-like (cyc-lin E high/p27 low/p53+/glomeruloid-microvascular-prolifer-ation+) phenotype of BRCA1-related breast cancer.  CancerRes 2004, 64(3):830-835.10. Sorlie T, Tibshirani R, Parker J, Hastie T, Marron JS, Nobel A, Deng S,Johnsen H, Pesich R, Geisler S, Demeter J, Perou CM, Lonning PE,Brown PO, Borresen-Dale AL, Botstein D: Repeated observationof breast tumor subtypes in independent gene expressionRelationship between TNP, survival and lymph node status – lymph node negative subgroupFigure 5Relationship between TNP, survival and lymph node status – lymph node negative subgroup. In TNP cases, the magnitude of the decrease in survival was approximately 1.5-fold greater at 3 years in the lymph node positive subgroup (Fig-ure 4) compared to the lymph node negative subgroup (Figure 5).01020304050607080901000.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0years after breast cancer occurrence% survivalOthersER-, PR-, HER2-P -value = .08   (@10 yrs)               = .001 (@  3 yrs)n at riskOthers:                  205                204               201               191                183               174                159               148                134               123               107ER-, PR-, HER2-  :  41                  41                 37                 34                  31                 31                  27                 24                  24                 20                 18(lymph node negative)p=.001 at 3 yearsp=.08 at 10 yearsPage 10 of 11(page number not for citation purposes)Kononen J, Torhorst J, Sauter G, Zuber M, Kochli OR, Mross F, Die-terich H, Seitz R, Ross D, Botstein D, Brown P: Expression of data sets.  Proc Natl Acad Sci U S A 2003, 100(14):8418-8423.Publish with BioMed Central   and  every scientist can read your work free of charge"BioMed Central will be the most significant development for disseminating the results of biomedical research in our lifetime."Sir Paul Nurse, Cancer Research UKYour research papers will be:available free of charge to the entire biomedical communitypeer reviewed and published immediately upon acceptancecited in PubMed and archived on PubMed Central BMC Cancer 2007, 7:134 http://www.biomedcentral.com/1471-2407/7/13411. Foulkes WD, Stefansson IM, Chappuis PO, Begin LR, Goffin JR, WongN, Trudel M, Akslen LA: Germline BRCA1 mutations and abasal epithelial phenotype in breast cancer.  J Natl Cancer Inst2003, 95(19):1482-1485.12. Lakhani SR, Reis-Filho JS, Fulford L, Penault-Llorca F, van der Vijver M,Parry S, Bishop T, Benitez J, Rivas C, Bignon YJ, Chang-Claude J,Hamann U, Cornelisse CJ, Devilee P, Beckmann MW, Nestle-Kram-ling C, Daly PA, Haites N, Varley J, Lalloo F, Evans G, Maugard C, Mei-jers-Heijboer H, Klijn JG, Olah E, Gusterson BA, Pilotti S, Radice P,Scherneck S, Sobol H, Jacquemier J, Wagner T, Peto J, Stratton MR,McGuffog L, Easton DF: Prediction of BRCA1 status in patientswith breast cancer using estrogen receptor and basal pheno-type.  Clin Cancer Res 2005, 11(14):5175-5180.13. Haffty BG, Yang Q, Reiss M, Kearney T, Higgins SA, Weidhaas J, Har-ris L, Hait W, Toppmeyer D: Locoregional relapse and distantmetastasis in conservatively managed triple negative early-stage breast cancer.  J Clin Oncol 2006, 24(36):5652-5657.14. Cleator S, Heller W, Coombes RC: Triple-negative breast can-cer: therapeutic options.  Lancet Oncol 2007, 8(3):235-244.15. Nielsen TO, Hsu FD, Jensen K, Cheang M, Karaca G, Hu Z, Hernan-dez-Boussard T, Livasy C, Cowan D, Dressler L, Akslen LA, Ragaz J,Gown AM, Gilks CB, van de Rijn M, Perou CM: Immunohisto-chemical and clinical characterization of the basal-like sub-type of invasive breast carcinoma.  Clin Cancer Res 2004,10(16):5367-5374.16. Makretsov NA, Huntsman DG, Nielsen TO, Yorida E, Peacock M,Cheang MC, Dunn SE, Hayes M, van de Rijn M, Bajdik C, Gilks CB:Hierarchical clustering analysis of tissue microarray immu-nostaining data identifies prognostically significant groups ofbreast carcinoma.  Clin Cancer Res 2004, 10(18 Pt 1):6143-6151.17. Anderson WF, Chen BE, Jatoi I, Rosenberg PS: Effects of estrogenreceptor expression and histopathology on annual hazardrates of death from breast cancer.  Breast Cancer Res Treat 2006,100(1):121-126.18. Rodriguez-Pinilla SM, Sarrio D, Honrado E, Hardisson D, Calero F,Benitez J, Palacios J: Prognostic significance of basal-like pheno-type and fascin expression in node-negative invasive breastcarcinomas.  Clin Cancer Res 2006, 12(5):1533-1539.19. Minn AJ, Gupta GP, Siegel PM, Bos PD, Shu W, Giri DD, Viale A,Olshen AB, Gerald WL, Massague J: Genes that mediate breastcancer metastasis to lung.  Nature 2005, 436(7050):518-524.20. Banerjee S, Reis-Filho JS, Ashley S, Steele D, Ashworth A, Lakhani SR,Smith IE: Basal-like breast carcinomas: clinical outcome andresponse to chemotherapy.  J Clin Pathol 2006, 59(7):729-735.21. Sotiriou C, Neo SY, McShane LM, Korn EL, Long PM, Jazaeri A, Mar-tiat P, Fox SB, Harris AL, Liu ET: Breast cancer classification andprognosis based on gene expression profiles from a popula-tion-based study.  Proc Natl Acad Sci U S A 2003,100(18):10393-10398.22. Rakha EA, Putti TC, Abd El-Rehim DM, Paish C, Green AR, Powe DG,Lee AH, Robertson JF, Ellis IO: Morphological and immunophe-notypic analysis of breast carcinomas with basal and myoep-ithelial differentiation.  J Pathol 2006, 208(4):495-506.23. Laakso M, Tanner M, Nilsson J, Wiklund T, Erikstein B, Kellokumpu-Lehtinen P, Malmstrom P, Wilking N, Bergh J, Isola J: Basoluminalcarcinoma: a new biologically and prognostically distinctentity between basal and luminal breast cancer.  Clin CancerRes 2006, 12(14 Pt 1):4185-4191.24. Fulford LG, Reis-Filho JS, Ryder K, Jones C, Gillett CE, Hanby A, Eas-ton D, Lakhani SR: Basal-like grade III invasive ductal carci-noma of the breast: patterns of metastasis and long-termsurvival.  Breast Cancer Res 2007, 9(1):R4.25. Gusterson BA, Ross DT, Heath VJ, Stein T: Basal cytokeratins andtheir relationship to the cellular origin and functional classi-fication of breast cancer.  Breast Cancer Res 2005, 7(4):143-148.26. Hennessy BT, Gonzalez-Angulo AM, Hortobagyi GN: Individualiza-tion of neoadjuvant therapy for breast cancer according tomolecular tumor characteristics.  Nat Clin Pract Oncol 2005,2(12):598-599.27. Carey LA, Dees EC, Sawyer L, Gatti L, Moore DT, Collichio F, OllilaDW, Sartor CI, Graham ML, Perou CM: The triple negative par-adox: primary tumor chemosensitivity of breast cancer sub-types.  Clin Cancer Res 2007, 13(8):2329-2334.28. Wakeling AE: Inhibitors of growth factor signalling.  Endocr Relat29. Arnes JB, Brunet JS, Stefansson I, Begin LR, Wong N, Chappuis PO,Akslen LA, Foulkes WD: Placental cadherin and the basal epi-thelial phenotype of BRCA1-related breast cancer.  Clin Can-cer Res 2005, 11(11):4003-4011.30. Moyano JV, Evans JR, Chen F, Lu M, Werner ME, Yehiely F, Diaz LK,Turbin D, Karaca G, Wiley E, Nielsen TO, Perou CM, Cryns VL:AlphaB-crystallin is a novel oncoprotein that predicts poorclinical outcome in breast cancer.  J Clin Invest 2006,116(1):261-270.31. Rakha EA, El-Sayed ME, Green AR, Paish EC, Lee AH, Ellis IO: Breastcarcinoma with basal differentiation: a proposal for pathol-ogy definition based on basal cytokeratin expression.  Histopa-thology 2007, 50(4):434-438.Pre-publication historyThe pre-publication history for this paper can be accessedhere:http://www.biomedcentral.com/1471-2407/7/134/prepubyours — you keep the copyrightSubmit your manuscript here:http://www.biomedcentral.com/info/publishing_adv.aspBioMedcentralPage 11 of 11(page number not for citation purposes)Cancer 2005, 12 Suppl 1:S183-7.


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items