pregnancy-associated breast cancer,which can include cancers diagnosedup to 1e2 years after a delivery.

In this article, Abenhaim andcolleagues used the Healthcare Costand Utilization Project-NationwideInpatient Sample to evaluate deliveriesduring which breast cancer was alsoidentified over a 10-year period. Theyfound 6.5 cases per 100 000 births. Notsurprisingly, breast cancer diagnosedduring pregnancy was more prevalentin women older than 35 years who weretreated in urban hospitals and hadinsurance. There was a nonsignificantincrease in incidence per year duringthis 10-year period. No significantdifference in birth complications wasseen, except for an increase in induc-tion for those with cancer that was

likely secondary to wanting to proceedwith further local or systemic treatmentfor their cancer.

This analysis was an attempt to tryto better estimate the true incidence ofbreast cancer diagnosed during preg-nancy. It may continue to underesti-mate the incidence among women whoterminate the pregnancy, do not seekmedical care, or have limited resourcesduring a pregnancy and then presentlater with a more advanced breastcancer. It also doesn’t account forpregnancy-associated breast cancer,which is continuing to emerge as adistinct entity with potentially worseoutcomes. Overall, the authors pro-vided more evidence as to the extent ofthe intersection between pregnancyand breast cancer, and we need to

Breast D

continue to follow the trends as well asoutcomes for these women and theirchildren.

J. K. Litton, MD

References1. Smith LJ, Dalrymple JL,

Leiserowitz GS, Danielsen B,Gilbert WM. Obstetrical deliveriesassociated with maternal malignancyin California, 1992 through 1997. AmJ Obstet Gynecol. 2001;184:1504-1512.

2. Andersson TML, Johansson ALV,Hsieh C-C, Cnattingius S, Lambe M.Increasing incidence of pregnancy-associated breast cancer in Sweden.Obstet Gynecol. 2009;114:568-572.

TUMOR BIOLOGY

CHEK2*1100delC Heterozygosityin Women With Breast CancerAssociated With Early Death,Breast CancereSpecific Death,and Increased Risk of a SecondBreast Cancer

Weischer M, Nordestgaard BG, Pharoah P,et al (Univ of Copenhagen, Denmark; Univof Cambridge, UK; et al)

J Clin Oncol 30:4308-4316, 2012

Purpose.dWe tested the hypoth-eses that CHEK2*1100delC heterozy-gosity is associated with increased riskof early death, breast cancerespecificdeath, and risk of a second breast cancerin women with a first breast cancer.

Patients and Methods.dFrom 22studies participating in the Breast Can-

cer Association Consortium, 25,571whitewomenwith invasive breast cancerwere genotyped for CHEK2*1100delCand observed for up to 20 years (median,6.6 years). We examined risk of earlydeath and breast cancer-specific deathby estrogen receptor status and risk of asecond breast cancer after a first breastcancer in prospective studies.

Results.dCHEK2*1100delC het-erozygosity was found in 459 patients(1.8%). In women with estrogenreceptorepositive breast cancer, multi-factorially adjusted hazard ratios forheterozygotes versus noncarriers were1.43 (95% CI, 1.12 to 1.82; log-rankP¼ .004) for early death and 1.63(95% CI, 1.24 to 2.15; log-rankP < .001) for breast cancerespecificdeath. In all women, hazard ratio for asecond breast cancer was 2.77 (95%

CI, 2.00 to 3.83; log-rank P < .001) in-creasing to 3.52 (95% CI, 2.35 to 5.27;log-rank P < .001) in women withestrogen receptorepositive first breastcancer only.

Conclusion.dAmong women withestrogen receptor-positive breast can-cer, CHEK2*1100delC heterozygositywas associated with a 1.4-fold risk ofearly death, a 1.6-fold risk of breastcancer-specific death, and a 3.5-foldrisk of a second breast cancer. This isone of the few examples of a geneticfactor that influences long-term prog-nosis being documented in an extensiveseries of women with breast cancer.

In contrast to the high risk associ-ated with highly penetrant mutations inthe tumor suppressors BRCA1 andBRCA2, individuals heterozygous for

iseases: A Year Book� Quarterly 141Vol 24 No 2 2013

the CHEK2*1100delC variant in theCHEK2 cell cycle checkpoint genehave amodestly increased risk of breastcancer.1-3 CHEK2*1100delC is a raredysfunctional variant primarily carriedby individuals of Northern and EasternEuropean descent. Previously, it wasreported that CHEK2*1100delCheterozygotes with breast cancer had a2-fold increased risk of developingsecondary breast cancer and a lowerrecurrence-free survival rate.4

However, the retrospective nature ofthis cohort study and the relativelysmall number of carriers call for moreprecise estimates.

In this article, Weischer andcolleagues analyzed 22 studies from 12countries participating in the BreastCancer Association Consortium(BCAC), a large collection of studieswith prospective design that included25 571 women with early mortalityinformation, including 459CHEK2*1100delC heterozygotes. Theauthors found that CHEK2*1100delCcarriers were more likely to haveestrogen receptor (ER)epositive breastcancer compared to noncarriers andthat among ER-positive patients,CHEK2*1100delC heterozygoteshad a 1.4-fold higher risk of earlydeath, a 1.6-fold higher risk of breastcancerespecific death, and a 3.5-foldhigher risk of a second breast cancer.

The 2- to 4-fold increase in the riskof a second breast cancer reported inthese 2 studies is similar to theincreased risk of primary cancer asso-ciated with CHEK2*1100delC, sug-gesting that inherited susceptibilitylikely persists due to haploinsufficiencyof the CHEK2 gene. It is interestingthat, unlike the effect of the variant onbreast cancer risk, the associationbetween CHEK2*1100delC and earlyand breast cancerespecific deaths isfound only in ER-positive patients;however, the underlying biologicalmechanism of this association is

142 Breast Diseases: A Year Book� Quar

Vol 24 No 2 2013

unclear and additional confirmation inindependent studies may be necessary.

Using the same study populationfrom the BCAC, Fasching andcolleagues5 recently examined theprognostic relevance of 11 commonsingle-nucleotide polymorphisms iden-tified from genome-wide associationstudies of risk. The authors found thatonly one variant of the TOX3 gene wasmoderately associated with breastcancer survival, suggesting thatsurvival may be influenced by a distinctset of genetic variants.CHEK2 belongsto a class of medium- to high-risk breastcancer genes whose rare variants(minor allele frequency¼ 0.1% to0.5%) contribute to a 3- to 5-foldincreased risk of breast cancer. Thesame gene class also includes p53,PTEN, ATM, NBS1, RAD50, BRIP1,and PALB2. Rare variants in thesegenes have been linked to breast cancerprognosis (eg, survival, recurrence, andsecond cancers) in a variety of studies,although none of them, exceptCHEK2*1100delC, has been tested inlarge studies such as the BCAC study. Itis possible that rare variants of this geneclass contribute to the heritability ofbreast cancer survival and/or secondcancers; some of these variants areawaiting future discovery by further finemapping and targeted deep sequencing.

As mentioned by the authors, thereare several limitations to this study.Most of them are directly related tomissing key clinical data. For example,25% to 30% of the study subjects didnot have ER status information, and40% to 45% were missing lymph nodestatus. In addition, the treatment in-formation was completely absent. Al-though there was no direct evidence tosuggest that the missing clinical datawould alter the overall conclusions, itis possible that they might change theinterpretation of specific results. Forexample, if treatment information wereavailable, one might investigate

terly

whether radiation-based adjuvanttherapy could affect the associationbetween CHEK2*1100delC and sur-vival and second cancer occurrence.Since CHEK2 is activated by double-strand DNA breaks and is involved incell cycle control, DNA repair, andapoptosis, it is plausible that this raregenetic variant could function as apredictive factor for treatment out-come. Missing clinical information isnot an infrequent occurrence inconsortium-based epidemiologicstudies because of the logistic difficultyof collecting comprehensive clinicaldata from various study sites that areoften in different countries. Besidesgeographic and administrativebarriers, the clinical data from dispa-rate sites may be difficult to merge dueto differences in recording systems forclinical variables. Moreover, there maybe variations in treatment options andtreatment quality that could confoundthe clinical outcomes. Thus, despiteconcerted efforts to minimize missingdata, the collection and managementof clinical information still posea significant challenge in theconsortium-based approach tostudying cancer outcomes. Neverthe-less, the large sample size of this studyand the lengthy duration of prospectivefollow-up provide confidence that thedysfunctional CHEK2 variant inwomen with breast cancer plays aprognostic role.

X. Wu, MD, PhDH. Zhao, PhD

References1. Meijers-Heijboer H, van den

Ouweland A, Klijn J, et al; CHEK2-Breast Cancer Consortium.Low-penetrance susceptibility tobreast cancer due to CHEK2(*)1100delC in noncarriers of BRCA1or BRCA2 mutations. Nat Genet.2002;31:55-59.

2. CHEK2 Breast Cancer Case-ControlConsortium. CHEK2*1100delC andsusceptibility to breast cancer:a collaborative analysis involving10,860 breast cancer cases and 9,065controls from 10 studies. Am J HumGenet. 2004;74:1175-1182.

3. Weischer M, Bojesen SE, Ellervik C,Tybjaerg-HansenA,NordestgaardBG.

CHEK*1100delC genotyping forclinical assessment of breast cancerrisk: meta-analyses of 26,000 patientcases and 27,000 controls. J ClinOncol. 2008;26:542-548.

4. Schmidt MK, Tollenaar RA, deKemp SR, et al. Breast cancersurvival and tumor characteristics inpremenopausal women carrying the

Breast Di

CHEK2*1100delC germlinemutation. J Clin Oncol. 2007;25:64-69.

5. Fasching PA, Pharoah PD, Cox A,et al. The role of genetic breast cancersusceptibility variants as prognosticfactors. Hum Mol Genet. 2012;21:3926-3929.

Clinically Used Breast CancerMarkers Such As EstrogenReceptor, ProgesteroneReceptor, and Human EpidermalGrowth Factor Receptor 2 AreUnstable Throughout TumorProgression

Lindström LS, Karlsson E, Wilking UM,et al (Karolinska Institutet and KarolinskaUniversity Hospital, Solna, Sweden; et al)

J Clin Oncol 30:2601-2608, 2012

Purpose.dTo investigate whetherhormonal receptors and human epi-dermal growth factor receptor 2(HER2) change throughout tumorprogression, because this may alterpatient management.

Patients and Methods.dThe studycohort included female patients withbreast cancer in the Stockholm healthcare region who relapsed from January1, 1997, to December 31, 2007. Eitherbiochemical or immunohistochem-ical (IHC)/immunocytochemical (ICC)methods were used to determine es-trogen receptor (ER), progesteronereceptor (PR), and HER2 status, whichwas then confirmed by fluorescent insitu hybridization for IHC/ICC 2+ and3+ status.

Results.dER (459 patients), PR(430 patients), and HER2 (104 patients)from both primary tumor and relapsewere assessed, revealing a change in

32.4% (McNemar’s test P < .001),40.7% (P < .001), and 14.5% (P¼ .44)of patients, respectively. Assessment ofER (119 patients), PR (116 patients),and HER2 (32 patients) with multiple(from two to six) consecutive relapsesshowed an alteration in 33.6%, 32.0%,and 15.7% of patients, respectively.A statistically significant differentialoverall survival related to intraindivid-ual ER and PR status in primary tumorand relapse (log-rank P < .001) wasnoted. In addition, women with ER-positive primary tumors that changedto ER-negative tumors had a significant48% increased risk of death (hazardratio, 1.48; 95% CI, 1.08 to 2.05)compared with women with stable ER-positive tumors.

Conclusion.dPatients with breastcancer experience altered hormonereceptor and HER2 status throughouttumor progression, possibly influencedby adjuvant therapies, which signifi-cantly influences survival. Hence,marker investigations at relapse maypotentially improve patient manage-ment and survival.

Breast cancer treatment is deter-mined by the tumor subtype; knowl-edge of the ER, PR, and HER2 statusis fundamental for treatment selectionamong not only patients with early-stage breast cancer but also for thosewith metastatic disease.1 The lack ofstability in the expression of the

hormone receptors and HER2 hasbeen described by several groups,2-6

but the frequency of discordance, itsrelationship to changes in treatment,its prognostic implications, andwhether this change is influenced byadjuvant therapies remain areas ofgreat interest.

In this study, Lindstrom andcolleagues assessed intraindividualER, PR, and HER2 status in a largecohort of breast cancer patients bycontrasting primary tumor and relapsemarker status and by comparingintraindividual marker status inconsecutive relapses. The authorsidentified a total of 459, 430, and 104patients with paired determination ofER, PR, and HER2, respectively.A very high frequency of discordancewas observed. Among patients withpositive ER, PR, and HER2 status inthe primary tumor, a change to nega-tive status was observed in 24.6%,33.0%, and 8.7% of relapses, respec-tively. Of less magnitude but similarinterest, the ER, PR, and HER2 statuschanged from negative in the primarytumor to positive in the relapse in7.8%, 7.7%, and 5.8% of patients,respectively. The absolute discordancerate was 32.4%, 40.7%, and 14.5% forER, PR, and HER2, respectively.Among the patients with multiplerelapses, the authors evaluated thediscrepancy among relapses and foundthat the discrepancy rate was 33.6%,

seases: A Year Book� Quarterly 143Vol 24 No 2 2013