●

mmcwcaeodcobaatcK©

Im

Cr4Actu

spmboctitorqDcda

A

Increased Breast Calcifications in Women With ESRD on Dialysis:Implications for Breast Cancer Screening

Mario Castellanos, MD, Seema Varma, MD, Kathleen Ahern, PhD, RN, Sue-Jane Grosso, MD,Shalom Buchbinder, MD, Denis D’Angelo, MD, Carolyn Raia, MD,

Morton Kleiner, MD, and Suzanne Elsayegh, MD

Background: Different appearances of breast calcification on mammography can differentiate benign fromalignant disease. An increased incidence of breast calcifications in dialysis patients is established, but data fororphological characteristics in renal patients, incidence of benign and malignant calcifications, and clinical

onsequences are limited. In this study, we compare mammograms and the workup of abnormal calcifications ofomen on hemodialysis with that of women with normal renal function. Methods: This is a retrospectivease-control study; 45 women on hemodialysis had their screening mammograms reviewed. A control group of 86ge-matched women with normal renal function was randomly obtained for comparison. Mammograms werexamined and the recommended workup was traced. Breast calcification morphological characteristics, incidencef benign versus malignant calcifications, callback rate, and biopsy recommendation rates were compared toetermine whether breast calcifications in renal patients led to excessive workups. Results: Overall, breastalcifications in the renal group were statistically significantly increased compared with controls, mostly becausef several benign-appearing morphological characteristics. No statistically significant difference was presentetween the 2 groups with respect to callback rates. However, incidences of malignancy-associated calcificationnd hence biopsy recommendation rate were slightly greater for the renal group. Conclusion: Renal patients haven increase in breast calcification, mostly caused by several benign calcifications. The callback rate is no greaterhan that in the general population. However, there is a slightly greater incidence of malignancy-associatedalcifications; hence, once called back, they have a greater probability of being recommended for biopsy. Am Jidney Dis 48:301-306.2006 by the National Kidney Foundation, Inc.

NDEX WORDS: Breast calcification; end-stage renal disease (ESRD); chronic kidney disease (CKD); screening

ammography.

iaiuc

wwWb

IV

4

o

DDAs

URRENTLY IN THE United States, 431,284patients are being treated for end-stage

enal disease (ESRD). Females account for about3%, with the majority older than 45 years.1,2

ccording to current recommendations for breastancer screening, most of these patients are inhe age range in which patients are encouraged tondergo screening mammography.Breast calcification is an important finding on

creening mammography; the appearance andattern may be specific, reflecting a benign,alignant, or indeterminate pattern. Patterns of

reast calcification in renal dialysis patients aref particular interest because there is an in-reased incidence of breast calcification in pa-ients with renal disease.3-10 There were attemptsn the past to describe patterns of breast calcifica-ions in renal patients, but these included only 2ld small studies5,11 and a few scattered caseeports,4,9,12 significantly limited by the pooruality of mammogram resolution of that period.ata for morphological characteristics of breast

alcification associated with renal disease, inci-ence of benign versus malignant calcifications,

nd their clinical consequences clearly are lim-

merican Journal of Kidney Diseases, Vol 48, No 2 (August), 2006

ted, although thousands of women with ESRDre eligible for mammogram screening. This datas important to aid in management and avoidnnecessary costly erroneous workups and psy-hological stress for the patient.

In this study, we compare mammograms ofomen with ESRD on long-term dialysis therapyith those of women with normal renal function.e describe morphological characteristics of

reast calcifications in each group and compare

From the Departments of Medicine and Radiology, Statensland University Hospital; and Regional Radiology, Stincent’s Medical Center of Richmond, Staten Island, NY.Received October 21, 2005; accepted in revised form May

, 2006.Originally published online as doi:10.1053/j.ajkd.2006.05.001

n June 30, 2006.Support: None. Potential conflicts of interest: None.Address reprint requests to Mario Castellanos, MD,

irector of Medical Women’s Health Division, Associateirector of Research, Department of Medicine, 401 Seaviewve, Staten Island, NY 10305. E-mail: mario_castellanos@iuh.edu

© 2006 by the National Kidney Foundation, Inc.0272-6386/06/4802-0015$32.00/0

doi:10.1053/j.ajkd.2006.05.001

: pp 301-306 301

cncr

S

Afcshwiwwichgaba

M

gumear

aIcou

sacs

uncccwrr

F

stsgaptnttona

S

eu

uC

CASTELLANOS ET AL302

allback rates, incidences of benign and malig-ant calcifications, diagnostic workups for suspi-ious calcifications, and biopsy recommendationates for each group.

METHODS

tudy PopulationWe performed a retrospective case-control study in 2001.

fter institutional review board approval, medical recordsrom 3 dialysis units from a university-affiliated medicalenter were reviewed. All women who were on hemodialy-is therapy were considered. From this cohort, patients whoad a previous mammogram performed at the medical centerere assigned to the study group. Women attending the

nstitution’s medical clinics who had normal renal functionere randomly selected to form the control group. Patientsith normal renal function were obtained by randomly select-

ng patients with normal creatinine clearance and serumalcium levels. Patients were screened to ensure that eachad a mammogram at the institution. Patients in the controlroup were age matched to the renal group to create a similarge distribution to the study group because the incidence ofreast calcification is influenced by age.5 Figure 1 shows thege distribution of renal patients.

ammographyMammograms of patients from both the renal and control

roups were reviewed. Mammograms were obtained bysing a standard 4-view examination with Lorad MIII mam-ography units (Trex Medical, Danbury, CT). Three experi-

nced mammographers reviewed the films (S.-J.G., D.D.,nd C.R.). They were blinded to patient renal function andeexamined the mammograms independently (Fig 2).

Calcifications were described, categorized, and reportedccording to the American College of Radiology Breastmaging Reporting and Data Systems (BIRADS)13 (Table 1).Aomponent of the BIRADS is that it gives standard terminol-gy for describing findings seen on mammography that aresed to provide consistency among radiologists. Using this

ystem, breast calcifications can be grouped as being associ-ted with benign (9 calcification patterns), malignant (3alcification patterns), and indeterminate breast pathologicaltates (2 calcification patterns).

The final diagnosis of mammograms was reported bysing the BIRADS final assessment categories. Negative/ormal studies are placed into category 1; benign lesions,ategory 2; probably benign lesions, category 3; suspi-ious lesions, category 4; and highly suspicious lesions,ategory 5. Incomplete screening that requires additionalorkup is placed in category 0. BIRADS 1, 2, and 3 were

egarded as negative studies and BIRADS 4 and 5 wereegarded as positive studies for malignancy.

ollow-UpPatients were called back for abnormal findings on initial

creening mammogram, such as a mass, asymmetric densi-ies, and suspicious breast calcifications. The callback rate, atandard mammography measure, was examined in bothroups. The callback rate is the number of cases called backfter initial screening mammography of the total number ofatients screened. Additional workup included magnifica-ion views of the breast with or without breast ultrasound, aseeded. Patients who had a final BIRADS description afterhe additional workup consistent with suspicious calcifica-ions for malignancy were recommended to undergo a bi-psy. Biopsy recommendation rates were calculated as totalumber of patients recommended for biopsy in each groupgainst the total number of patients screened.

tatistical AnalysesNonparametric statistics were used. Chi-square and Fisher

xact tests were performed as appropriate, with Fisher exactsed in all instances in which cell sizes were less than 5.

RESULTS

Forty-five renal patients from the 3 dialysisnits had mammograms at the medical center.ontrols consisted of 86 age-matched patients.

Fig 1. Age distribution ofrenal patients.

hvt

ctT

Tcc(ccsTwtmc

clmld

i

M

B

I

BREAST CALCIFICATIONS IN RENAL DISEASE 303

From the renal group, 38 of 45 patients (84.4%)ad calcifications on their mammograms. Con-ersely, 51 of 86 controls (59.3%) had calcifica-ions on their mammograms. The incidence of

Fig 2. Mammogram show-ng vascular calcifications.

Table 1. Abbreviations of BreastCalcification Categories

Pattern Abbreviation

alignant associationInterrupted, fine, linear, or branching ICPleomorphic heterogeneous HCAmorphous AC

enign associationSkin SCVascular VCLarge rod-like LGEggshell ECSpherical or lucent center OCSuture SC1Dystrophic DCPunctate (�0.5 mm) PCMilk of calcium MC

ndeterminateCoarse CC

cRound (0.5-1 mm) RC

alcifications in the renal group versus the con-rol group was statistically significant (P � 0.003;able 2).The frequency of each category is listed in

able 3. Most of the increase in incidence ofalcifications in the renal group appears to beaused by several benign calcification patternsFig 3; Table 3). Comparing categories of benignalcifications, vascular calcification was the mostommon pattern seen in both groups, with notatistical difference between groups (P � 0.46).he benign pattern that differed between groupsas parenchymal spherical or lucent calcifica-

ion pattern. Renal patients were significantlyore likely to have this calcification pattern than

ontrols (P � 0.0004; Table 3).Comparing only incidence of calcifications

ommonly considered to be associated with ma-ignancy (BIRADS 4 and 5), amorphous, pleo-orphic heterogeneous, and interrupted, fine,

inear, or branching, no statistically significantifference was identified between the renal and

ontrol groups for the incidence of each indi-

vgad(t

4gs8bfT

fritcTdrcmsww

NNCF

( tly diffe

M

B

B

vn

CASTELLANOS ET AL304

idual calcification (Table 3). However, as aroup, the final diagnosis of malignancy-associ-ted calcifications provoking a biopsy recommen-ation was significantly greater in the renal group4 of 45 patients) versus controls (1 of 86 pa-ients; P � 0.047; Tables 2 and 4).

On initial screening mammography, there werecases with suspicious calcifications in the renalroup of 45 patients screened and 3 cases withuspicious calcifications in the control group of6 patients screened (Table 2). Therefore, call-ack rates for abnormal calcifications were 8.8%or the renal group versus 3.48% for controls.his was not statistically significant (P � 0.23).

Table 2. Results o

o. of patientso. of cases with calcificationsallback rate for suspicious calcificationsinal BIRADS description consistent with malignancy-assocalcification and cases sent for biopsy

NOTE. Values expressed as number (percent).*As a group, malignancy-associated calcifications were

1.16%), and biopsy recommendation rates were significan

Table 3. Frequencies of Morphological Characteand Co

Morphological Characteristics ofCalcification Abbreviation

alignantInterrupted, fine, linear, or branching ICPleomorphic heterogeneous HCAmorphous AC

enignSkin SCVascular VCLarge rod-like LGEggshell ECSpherical or lucent center OCSuture SC1Dystrophic DCPunctate (�0.5 mm) PCMilk of calcium MC

enign or malignantCoarse CCRound (0.5-1 mm) RC

NOTE. As a group, malignancy-associated calcificationersus controls (1 of 86 patients; P � 0.047). Vascular calco statistical difference between groups.

*Categories are statistically significant.

Upon further workup of patients called backor suspicious calcifications, all 4 cases in theenal group, but only 1 case of the 3 callback casesn the control group, had a final BIRADS descrip-ion consistent with a malignancy-associated cal-ification pattern, and biopsy was recommended.herefore, the incidence of final radiologicaliagnosis of malignant-associated calcifications,eflecting the biopsy recommendation rate, wasalculated as the total number of patients recom-ended for biopsy of the total number of patients

creened in each group. In the renal group, thisas 8.8%, and in controls, 1.16% (P � 0.047,hich is statistically significant; Tables 2 and 4).

mogram Review

Renal Patients Controls P

45 8638 (84.4) 51 (59.3) 0.003

4/45 (8.8) 3/86 (3.48) 0.23

4/45 (8.8) 1/86 (1.16) 0.047*

cantly greater in the renal group (8.8%) versus controlsrent (P � 0.047).

of Calcifications on Mammograms of the Renalroups

ncy of Calcificationtern in the Renal

Group (%)

Frequency of CalcificationPattern in the Control

Group (%) P

7.8 0 0.072.06 1.9 0.10 0 0

2.6 7.8 0.3855 45 0.467.8 13.7 0.5

10.5 11.7 0.150 13.7 0.0004*0 0 00 0 05.2 13.7 0.130 0 0

21 17.6 0.5115.7 21.5 0.37

significantly greater in the renal group (4 of 45 patients)n was the most common pattern seen in both groups, with

f Mam

ciated

signifi

risticsntrol G

FrequePat

s wereificatio

fHibtfirwdusp

otttatlrmotibRm

ragFcdtcfafioTrtuctcbsdtmcPtBr

amtctg

BREAST CALCIFICATIONS IN RENAL DISEASE 305

DISCUSSION

Thousands of patients with ESRD are eligibleor breast cancer screening with mammography.owever, there is a concern that mammographic

nterpretation in this population may be difficultecause of the increased incidence of breastissue calcification. Metastatic soft-tissue calci-cation as a result of secondary hyperparathy-oidism in patients with chronic renal disease isell described.5,7,8,10 In 1987, Sommer et al5

escribed breast calcifications in 68% of patientsndergoing dialysis. Since then, there have beeneveral reports describing breast calcifications inatients with renal disease.3,4,6,9

There is concern that the increased incidencef breast calcifications in patients on dialysisherapy may pose a diagnostic dilemma. Despitehis possibility, only 2 old studies5,11 attemptedo describe and categorize morphological char-cteristics of breast calcifications in dialysis pa-ients. These studies were small and significantlyimited by the unavailability of current high-esolution mammography techniques. Further-ore, investigators of these studies5,11 and previ-

us case reports4,9,12 did not have uniformerminology available for categorizing and report-ng breast calcifications because they were doneefore institution of the American College ofadiology BIRADS. Sommer et al5 compared

Fig 3. Frequencies of benignnd malignant calcifications onammograms of the renal and con-

rol groups. The frequency of spheri-al or lucent (OC) calcification pat-ern was more common in the renalroup, P � 0.0004.

ammograms of 15 patients with compensated r

enal insufficiency, 22 on hemodialysis therapynd 14 who had renal transplants, with a controlroup of 100 who had normal kidney function.requency, size, structure, and location of cal-ific lesions were assessed. Patients with renalisease had significantly more calcifications thanhose with normal kidney function. Parenchymalalcifications (61% versus 27%) were the mostrequent, followed by vascular (45% versus 8%)nd ductal (29% versus 9%). Parenchymal calci-cations were categorized as small, round, orval; scattered; and not associated with a mass.his appears to correspond to punctate and

ound calcifications in the BIRADS groups. Duc-al calcifications, large rod-like in current terms,sually were found behind the areola and alsoould be seen in other areas of the breast in theypical orientation toward the nipple. No calcifi-ation looked like malignant calcification of thereast, according to the investigators. Evans et al11

tudied mammograms of 16 women on renalialysis therapy and compared them with a con-rol group of 32 women attending for routineammographic screening. They categorized cal-

ifications as vascular, parenchymal, and ductal.arenchymal calcifications were grouped as punc-

ate, teacup, or pearl shaped. Using the currentIRADS lexicon, these terms probably would

epresent dystrophic, milk of calcium, and round,

espectively. They categorized ductal calcifica-

tBcptcogmtcamtdco

c(tcbrTaredag(prerttticfrhoasi18

aaicfsaswsmsig

22

p(

KR

i

BR

a1

ScA

I

c

JruA

c3

t

p

CASTELLANOS ET AL306

ions as granular, casting, or linear, which, inIRADS terms, describes large rod-like calcifi-ations. They found a significant increase inarenchymal, followed by vascular, calcifica-ions in women on renal dialysis therapy. Ductalalcifications were no more prevalent in patientsn dialysis therapy compared with the controlroup, and in no case did calcifications simulatealignancy. These 2 studies correctly concluded

hat there was an increased incidence of breastalcifications on mammograms of renal patients,nd most of these calcifications did not simulatealignancy with the available imaging resolu-

ion. However, they did not compare the inci-ence of malignant calcifications as a group,linical consequences of increased calcifications,r callback and biopsy recommendation rates.In our study, breast calcifications were more

ommon in patients on long-term dialysis therapy84.4%) compared with age-matched control pa-ients (59%), and the increased incidence wasaused mostly by calcifications associated withenign morphological characteristics catego-ized according to the current standard BIRADS.his is reflected by the similar callback ratesnd is consistent with results of the 2 previouslyeported studies and case reports.5,7,8,10 How-ver, compared with the control group, the inci-ence of malignancy-associated calcificationss a group was slightly greater in the renalroup, as was the biopsy recommendation rateP � 0.047). Therefore, it appears that renalatients, once called back, were more likely to beecommended for biopsy compared with the gen-ral population. However, an increase in cancerates in renal patients remains uncertain at thisime because the study is not designed to clarifyhis question. Bias related to interval length be-ween mammogram screenings may exist, lead-ng to group differences. Therefore, what can beoncluded from this study is that the greaterrequency of breast calcifications leads to a greaterate of breast calcifications of indeterminate be-avior warranting biopsies. The callback rate atur institution for the entire year of 2005 wasbout 10%. The renal group in the study had aimilar rate of 8.8%. The biopsy rate at thenstitution is 1.5%, similar to that of controls at.16%. However, this is much lower than the

.8% biopsy rate for the renal group in the study. A

Additional studies with a larger sample sizere recommended to confirm this new findingnd determine the positive predictive value ofndeterminate calcifications and determine can-er rates in the hemodialysis population. There-ore, at this time, every woman on dialysis therapyhould be referred for screening mammography,nd the physician taking care of these patientshould be aware of the slightly increased risk fororkup and possibly resultant psychological

tress after screening in dialysis patients. Mam-ography is an important screening tool and

hould be offered to all dialysis patients accord-ng to standard health care guidelines of theeneral population.

REFERENCES

1. US Renal Data System: Excerpts from the USRDS004 Annual Data Report. Am J Kidney Dis 45:S1-S280,005 (suppl 1)2. US Renal Data System: Annual Report. Incidence and

revalence of ESRD. Am J Kidney Dis 34:S40-S50, 1999suppl 1)

3. Sarah Rovno HD, Feig SA, Hughes JS, Hurford MT,arasick D, Filippone E: Breast imaging case of the day.adiographics 18:1599-1604, 19984. Cooper RA, Berman S: Extensive breast calcification

n renal failure. J Thorac Imaging 3:81-82, 19885. Sommer G, Kopsa H, Zazgornik J, Salomonowitz E:

reast calcifications in renal hyperparathyroidism. AJR Am Joentgenol 148:855-857, 19876. McDougal BA, Lukert BP: Resolution of breast pain

nd calcification with renal transplantation. Arch Intern Med37:375-377, 19777. Conger JD, Hammond WS, Alfrey AC, Contiguglia

R, Stanford RE, Huffer WE: Pulmonary calcification inhronic dialysis patients: Clinical and pathologic studies.nn Intern Med 83:330-336, 19758. Parfitt AM: Soft tissue calcification in uremia. Arch

ntern Med 124:544-556, 19699. Han SY, Witten DM: Diffuse calcification of the breast in

hronic renal failure. AJR Am J Roentgenol 129:341-342, 197710. Massry SG, Coburn JW, Popovtzer MM, Shinaberger

H, Maxwell MH, Kleeman CR: Secondary hyperparathy-oidism in chronic renal failure: The clinical spectrum inremia, during hemodialysis, and after renal transplantation.rch Intern Med 124:431-441, 196911. Evans AJ, Cohen MEL, Cohen GF: Patterns of breast

alcification in patients on renal dialysis. Clin Radiol 45:343-44, 199212. Evans SE, Whitehouse GH: Extensive calcification in

he breast in chronic renal failure. Br J Radiol 64:757-759, 199113. American College of Radiology: Breast Imaging Re-

orting and Data Systems (BIRADS) (ed 4). Reston, VA,

merican College of Radiology, 2003