©

2003 Blackwell Publishing, Inc., 1075-122X/03/$15.00/0The Breast Journal, Volume 9, Number 4, 2003 307–311

Address correspondence and reprint requests to: Lester J. Layfield, MD,Department of Pathology, University of Utah, 50 North Medical Drive, Salt LakeCity, Utah 84132

Blackwell Publishing Ltd.Oxford, UKTBJThe Breast Journal1075-122X2003 Blackwell PublishingJuly/August 200394

Original Article

Four-Quadrant Sampling in Mastectomy Specimens

gupta et al

.

Utility of Four-Quadrant Random Sections in Mastectomy

Specimens

Dilip Gupta, MD,* Manju Nath, MD,

†

and Lester J. Layfield, MD

‡

*Magee Woman’s Hospital and

†

University of Pittsburgh, Pittsburgh, Pennsylvania, and

‡

University of Utah School of Medicine and ARUP Laboratories, Salt Lake City, Utah

�

Abstract:

Standard practice in surgical pathology dictatesthat random sections from the four quadrants of the breast betaken in mastectomy specimens. These sections are obtainedin addition to sampling of any grossly visible lesions within thebreast specimen. While tradition dictates the submission ofthese sections, we are unaware of any study supporting theirefficacy. We have investigated the utility and significance ofthese random sections in a series of 78 mastectomy speci-mens. This retrospective study identified mastectomy speci-mens from pathology files of Magee Woman’s Hospital, theUniversity of Pittsburgh, and the University of Utah School ofMedicine between 1997 and 2000. Clinical data (palpableversus nonpalpable), radiographic features (mammographicdiagnosis, presence of mass density and/or calcification), andpathologic features (size, histopathologic type, etc.) were stud-ied. The histologic sections of the cases were reviewed andthe random sections were specifically studied for pathologicfindings. Diagnosis and clinically significant features obtainedfrom examining these random sections, but not demonstrablein grossly selected sections, were tabulated. A total of 78 mas-tectomy specimens were analyzed. Diagnoses rendered wereinfiltrating ductal carcinoma (23), infiltrating ductal carcinomawith ductal carcinoma in situ (DCIS) (16), DCIS (25), infiltratinglobular carcinoma (4), biopsy cavity with no residual malig-nancy (4), infiltrating lobular carcinoma with lobular carcinomain situ (3), invasive ductal and lobular carcinoma (1), adenoidcystic carcinoma (1), and atypical ductal hyperplasia (1). Thenumber of random sections ranged from 2 to 17 (mean 9).Random sections provided additional information in 21 of 78mastectomies (27%). The multifocal/multicentric nature of the

lesion was diagnosed in 20 cases: DCIS (6), lobular carcinomain situ and invasive (2), invasive ductal carcinoma (6), invasiveand in situ ductal carcinoma (5), invasive lobular carcinoma (1),invasive ductal and lobular carcinoma (1). Additional findingsinclude lymphovascular invasion (2 cases), atypical ductalhyperplasia (1), DCIS at the operative margin (1), DCIS withinless than 1 mm of an operative margin (1), and atypical lobularhyperplasia (1). In the remaining 57 cases, random sections didnot provide any additional information. Histologic examinationof random sections from breast quadrants yielded importantinformation about the presence of multifocality, multicentricity,vascular invasion, and margin involvement by carcinoma in onlya minority of cases, many of which had a lobular morphology.

�

Key Words:

breast carcinoma,

mastectomy,

sampling

G

ross examination and submission of a sufficientnumber of tissue blocks to thoroughly, yet econom-

ically sample a specimen submitted for pathologic evalu-ation is an important component of tissue examination insurgical pathology. Studies recommending optimal num-bers of tissue blocks have been published for a variety ofspecimen types, including impalpable breast lesion speci-mens (1), needle localization breast biopsy specimens(2,3), and grossly benign breast biopsies (4). Traditionallyprotocols taught in pathology residency training pro-grams and published in textbooks of surgical pathology(5–7) have recommended that three sections be taken ofthe tumor, that sample blocks from all grossly or mam-mographically recognized lesions be taken, and that atleast one section randomly selected from each quadrantbe submitted. This recommendation is widely followed, but

308

•

gupta et al

.

has not, to our knowledge, been evaluated for efficiencyin a formal study.

While it is generally assumed that submission of theserandom quadrant sections will on occasion yield diagnos-tically or prognostically important new information, thefrequency at which such new information is obtained isunknown. Because of cost containment issues, we ques-tioned the necessity of these random tissue blocks, andundertook the present study to determine if the yield ofdiagnostically or prognostically important findings wassufficient to justify their continued submission. We stud-ied a series of 78 consecutive mastectomies drawn fromthree different institutions to determine the frequency ofdiagnostically or prognostically important new findingsin random quadrant sections. Herein we report the find-ings of that study.

MATERIALS AND METHODS

Following approval of the study by the institutionalreview boards at the University of Utah, Magee Women’sHospital, and the University of Pittsburgh, we searchedthe files of the pathology departments at each institutionfor all mastectomies performed between January 1997and December 2000. Only cases where all of the slideswere available for review and random quadrant sectionshad been taken were selected for the study. Cases alsoincluded in the study were required to have a grossly iden-tifiable tumor or biopsy cavity in the mastectomy speci-men. Sixty-five cases were obtained from the pathologyfiles of Magee Women’s Hospital and the University ofPittsburgh, and 28 were obtained from the University ofUtah. The clinical and radiographic reports were reviewedfor evidence of preoperatively recognized multicentricityor metastatic disease.

The slides from each case were reviewed microscopi-cally by the authors. The sections of gross tumor andgrossly observed abnormalities were reviewed and allfindings were documented on a protocol sheet. Followingthis, the random quadrant sections, varying in numberfrom 4 to 17 (mean 9) were reviewed for pathologic find-ings. These were separately documented. The two sets ofhistopathologic findings were then compared and anynew findings present only in the random sections wererecorded on the protocol sheet. These were then judged asto whether they represented significant new findings. Inthis study, multifocality was defined as discrete foci ofcarcinoma occurring within an area not more than 5 cmin diameter and multicentricity was defined as separatefoci of tumor more than 5 cm apart.

RESULTS

Seventy-eight mastectomy specimens fulfilled thecriteria for availability of slides, the presence of randomquadrant sections, and no clinically documented evidenceof multifocality, multicentricity, or intramammary metas-tasis. The number of random quadrant sections takenranged from 4 to 17 (mean 9). Table 1 shows the histologictypes of carcinoma found within the mastectomy specimensand/or the original biopsies. In four cases, no residualcarcinoma was detected within the mastectomy, but abiopsy cavity was present.

Table 2 shows the histologic types of carcinoma whererandom sections disclosed new information. Randomsections provided additional information in 21 of 78 cases(27%). The multifocal or multicentric nature of the lesionwas diagnosed in 21 cases. Table 3 shows the distributionof histologic types of carcinoma where multifocality ormulticentricity was documented in the random sectionsas well as the nature of the multifocal or multicentric lesion.Table 4 documents the location of the tumor in caseswhere random sections disclosed new findings. In 19 of 21cases, the new findings occurred in a quadrant differentfrom the quadrant containing the biopsied neoplasm orwere more than 5 cm from the biopsy site or clinicallyrecognized tumor.

Table 1. Histologic Types of Carcinoma Included inStudy

Histologic type Number

Ductal carcinoma in situ 25Invasive ductal carcinoma 27Invasive and in situ ductal carcinoma 16Invasive lobular carcinoma 4Invasive and in situ lobular carcinoma 3Adenoid cystic carcinoma 1Atypical ductal hyperplasia 1Invasive ductal and lobular carcinoma in situ 1

Table 2. Distribution of Histologic Types ofCarcinoma Present in the 21 Cases Where RandomSection Disclosed New Findings

Histologic type Number (%)

Ductal carcinoma in situ 6 (28)Invasive ductal carcinoma 6 (28)Invasive and in situ ductal carcinoma 5 (24)Invasive and in situ lobular carcinoma 2 (10)Invasive lobular carcinoma 1 (5)Invasive ductal and lobular carcinoma in situ 1 (5)

Four-Quadrant Sampling in Mastectomy Specimens •

309

In addition to the documentation of multifocality ormulticentricity, random sections also demonstratedotherwise undocumented foci of lymphovascular invasion(two cases), an additional focus of atypical ductal hyper-plasia, a single case of ductal carcinoma in situ (DCIS)at an operative margin, and a focus of DCIS within 1 mm ofan operative margin. Finally, in one case a random sectiondisclosed a new focus of atypical lobular hyperplasia.

The mean size of the carcinomas found in the 52 caseswithout additional findings was 1.9 cm (range 0.3–5.0cm). The size of the neoplasm in the 21 cases with randomsections showing additional findings was 3.7 cm (range0.3–12.0 cm). Thus larger tumors were more likely tocontain findings within the random sections. In two cases,removed from further analysis, the carcinoma was notgrossly visualized and random quadrant sections revealedthe tumor to be considerably larger than estimated fromsections selected on the basis of minimal gross abnormal-ities. Both of these carcinomas were lobular carcinomas.

DISCUSSION

The selection of at least one random tissue block fromeach quadrant of a mastectomy specimen has been thetraditional method of sampling and has been recommendedin textbooks of surgical pathology (5–7). While widelyaccepted as the protocol of choice, there appears to beno significant body of published work documenting thefrequency with which these random quadrant sections

add diagnostically or prognostically useful new informa-tion. In this age of cost containment, an evaluation of thisapproach seems timely.

Our study of 78 mastectomy specimens derived fromthree institutions revealed a relatively high incidence(27%) of clinically important findings obtainable onlyfrom these random quadrant sections. Most frequentlythis new information was the documentation of apparentmulticentricity or multifocality unrecognized in thesections selected on the basis of grossly visible abnorma-lities. This information may be important in evaluating theextent of carcinoma present. Page (8) has emphasized thatmulticentricity and multifocality may not be as clinicallyimportant as once believed. Of potentially greater impor-tance was the occasional documentation of grossly unsus-pected positive or close operative margins. Unsuspectedmultifocality/multicentricity was documented in 27% ofcases, and an unsuspected positive or close operativemargin was found in two cases (3%). Such findingscould have significant clinical importance in helping theclinician decide whether postoperative chemotherapy isrequired. The relationship between tumor size and unex-pected findings in the random quadrants is not surprising,as larger carcinomas were more likely to be associatedwith random sections containing carcinoma.

Of equal interest was the relationship between the his-topathologic type of carcinoma and the occurrence of newfindings in the random quadrant sections. While invasiveductal carcinomas were more numerous in our review, ahigher percentage of lobular carcinomas was associatedwith new findings in the random sections. This suggeststhat lobular carcinomas are less easily recognized grossly.

The findings of clinically and grossly unsuspectedmulticentric disease is of particular interest. In our series,multicentric disease (including grossly recognized lesions)was found in approximately one-third of cases. This multi-centricity was not grossly apparent and was found only in therandomly selected sections in 27% of cases. Our overallrate of multicentricity fell within the range documentedin the literature (9–75%) (9–24). Many of the studies

Table 3. Distribution of Multifocality and Multicentricity Documented in Random Sections Among HistologicTypes of Carcinoma Present

Ductal carcinoma in situ Multifocal, 4; multicentric, 5; multifocal and multicentric, 1Invasive ductal carcinoma Multifocal, 3Invasive and in situ ductal carcinoma Multifocal and multicentric, 1; multifocal, 1Invasive ductal carcinoma and lobular carcinoma in situ Multicentric, 2Invasive lobular carcinoma Multicentric, 2Lobular carcinoma in situ Multifocal, 2Total 21

Table 4. Location of the Tumor in Cases WhereRandom Sections Disclosed New Findings

Upper outer 4Upper inner 3Lower outer 3Lower inner 3Outer quadrants 1Upper quadrants 1Inner quadrants 1Central 1All four quadrants 1

310

•

gupta et al

.

documenting multicentricity based their findings onselection of blocks from grossly suspicious areas, mam-mographic findings, and random sections (15,17,21,22).A few studies used large tissue sections (9), whileothers relied on multiple sections from each quadrant(10,12,14,23). As would be suspected, the frequency ofmulticentricity varies greatly among these findings, butaverage around 40% (25).

The relatively high rate of clinically and grossly unsus-pected multicentricity in our study and that documentedin the literature (9–24) supports the utilization of radia-tion in the entire breast following a lumpectomy in amajority of patients. Page (8) has de-emphasized theimportance of multicentric and multifocal disease by cor-rectly pointing out the relatively low rate of recurrenceafter lumpectomy despite the relatively high reportedincidence of multifocality and multicentricity in breastcancer. Radiation therapy to the residual breast tissue mayexplain the low rate of recurrence.

Other grossly unsuspected findings in the randomquadrant sections were a single case of an unsuspectedpositive margin, a second case of tumor closely approxi-mating an operative margin (<1 mm), and in two cases,documentation of lymphovascular or perineural invasion.The findings of a positive margin would appear to havegreater immediate clinical significance than the documen-tation of multicentricity. The discovery of unexpectedpositive margins or “close” margins might indicate theneed for further surgery following mastectomy, or alter-nately, modifications in the dose or distribution of radio-therapy given postoperatively. As in cases of lumpectomywith positive margins, the surgeon might elect to reexcisethe area of involvement.

In two cases not included in the present study,increased size of the carcinoma was detected by randomsections. These cases were excluded from the final analysisbecause no grossly identifiable tumor mass or biopsycavity was seen in the mastectomy specimen.

The demonstration of increased size of the primarytumor by random section is important for prognosticpurposes as well as the selection of postoperative therapy.Size is a well-known prognostic indicator, with carcinomasless than 2 cm having a more favorable outcome thanthose larger than 2 cm (25). In two cases the size of the car-cinoma was dramatically increased by the demonstrationof diffuse confluent multiquadrant involvement by lobu-lar carcinoma. The tissue blocks selected on the basis ofgross abnormalities in these two cases markedly underes-timated the size of the tumor. The random quadrant sec-tions revealed its presence in areas that grossly appeared

to be benign fibroadipose tissue. Because of the submis-sion of random sections, the size of the main tumor masswas dramatically increased.

It is possible that some of the findings documented inour random sections would have been detected by seniorsurgical pathologists, but they were not identified byresidents and pathology assistants. Approximately half ofour cases were grossly examined by first-year residents,while the remainder were grossly examined by pathologyassistants and/or surgical pathology faculty. Wiley andKeh (26) detected 44 major diagnostic discrepancies dueto errors in gross dissection and sampling in a series of 520mastectomies. In their study, slightly more than half ofthese errors were committed by first-year residents andresulted from incomplete inspection of the specimen orfailure to recognize a lesion. The residents in our programs“grossing in” specimens were predominantly first-yearresidents, suggesting that some of the random quadrantsections might have been recognized as abnormal by moreexperienced observers.

Gross review of a subpopulation of the glass slidesof the random sections disclosed a bland appearance offibrofatty tissue without obvious abnormalities withthe unaided eye. Only microscopic evaluation detectedthe neoplastic change. Thus, in our study, the randomsections would appear to be obtained from grosslyunremarkable tissue.

While random quadrant sections increase the cost ofhistopathologic evaluation of breast mastectomy speci-mens, the submission of four random quadrant sectionsappears to be in line with the cost of block submission asrecommended by authors studying needle localizationbreast biopsies and biopsies of grossly benign breast tissue(1–4). Our study indicates that in a significant percentageof mastectomies, random sections derived from each ofthe four quadrants disclose clinically important new find-ings. On this basis we would recommend the continueduse of the traditional random quadrant bloc submissionswhen decisions concerning adjuvant radiation therapyfollowing mastectomy are based on the extent and size ofcarcinomas present in the mastectomy specimen.

REFERENCES

1. Mower GA, Leong AS. Gross handling of impalpablebreast lesion specimens.

Pathology

1992;24:261–63.2. Owings DV, Hann L, Schnitt SJ. How thoroughly should

needle localization breast biopsies be sampled for microscopicexamination? A prospective mammographic/pathologic corre-lation study.

Am J Surg Pathol

1990;14:578–83.

Four-Quadrant Sampling in Mastectomy Specimens •

311

3. Alberhasky MT. Mammographic and gross pathogenicanalysis of breast needle localization specimens. Use of a tissueanalysis device.

Am J Clin Pathol

1989;92:452–57.4. Schmitt SJ, Wang HH. Histologic sampling of grossly

benign breast biopsies. How much is enough?

Am J Surg Pathol

1989;13:505–12.5. Rosai J.

Ackerman’s Surgical Pathology

, 8th ed. St.Louis: Mosby, 1996:2640–42.

6. Lowe DG, Jeffrey IM.

Surgical Pathology Techniques

.Philadelphia: BC Decker, 1990:67.

7. Schmidt WA.

Principles and Techniques of SurgicalPathology

. Menlo Park, CA: Addison-Wesley, 1983:371–79.8. Page D. Breast disease in the 1990s. Patchwork quilt and

growth industry.

Am J Clin Pathol

1993;99:238–43.9. Qualheim RE, Gall EA. Breast carcinoma with multiple

sites of origin.

Cancer

1957;10:460–68.10. Tellem M, Prive L, Meranze DR. Four-quadrant study

of breast removed for carcinoma.

Cancer

1962;15:10–17.11. Shah JP, Rosen PP, Robins GF. Pitfalls of local excision

in the treatment of carcinoma of the breast.

Surg Gynecol Obstet

1973;136:721–25.12. Fisher ER, Gregorio R, Redmond C,

et al.

Pathologicfindings from the national surgical adjuvant breast project(protocol no. 4). I. Observations concerning the multicentricityof mammary cancer.

Cancer

1975;35:247–54.13. Morgenstern L, Maufman PA, Friedman NB. The cases

against tylecotomy for carcinoma of the breast. The factor ofmulticentricity.

Am J Surg

1975;130:251–58.14. Rosen PP, Fraccia AA, Urban JA,

et al.

“Residual”mammary carcinoma following simulated partial mastectomy.

Cancer

1975;35:739–47.15. Lagios MD. Multicentricity of breast carcinoma demon-

strated by routine correlated serial subgross and radiographicexamination.

Cancer

1977;40:1726–34.

16. Patchefsky AS, Shaber GS, Schwartz GF,

et al.

Thepathology of breast cancer detected by mass population screen-ing.

Cancer

1977;40:1659–70.17. Lagios MD, Rose MR, Margolion FR. Tubular carci-

noma of the breast. Association with multicentricity, bilaterality,and family history of mammary carcinoma.

Am J Clin Pathol

1980;73:25–30.18. Schwartz GF, Patchefsky AS, Feig SA,

et al.

Clinicallyoccult breast cancer. Multicentricity and implications fortreatment.

Ann Surg

1980;191:8–12.19. Schwartz GF, Patchefsky AS, Feig SA,

et al.

Multicentricityof non-palpable breast cancer.

Cancer

1980;45:2913–16.20. Westman-Naeser S, Bengtsson E, Eriksson O,

et al.

Multifocal breast carcinoma.

Am J Surg

1981;142:255–57.21. Egan RL. Multicentric breast carcinomas: clinical-

radiographic-pathologic whole organ studies and 10-yearsurvival.

Cancer

1982;49:1123–30.22. Lagios MD, Westdahl PR, Margolin FR,

et al.

Ductcarcinoma in situ. Relationship of extent of non-invasive diseaseto the frequency of occult invasion, multicentricity, lymphnode metastases, and short-term failures.

Cancer

1982;50:1309–14.

23. Lesser ML, Rosen PP, Kinne DW. Multicentricity andbilaterality in invasive breast carcinoma.

Surgery

1982;91:234–40.

24. Sarnelli R, Squartini F. Multicentricity in breast cancer:a submicroscopic study. In: Sommer SC, Rosen PP, Fechner RE,eds.

Pathology Annual

. Norwalk, CT: Appleton-Century-Crofts, 1986:143–58.

25. Tavassoli F.

Pathology of the Breast

. Norwalk, CT:Appleton and Lange, 1992:36.

26. Wiley EL, Keh P. Diagnostic discrepancies in breastspecimens subjected to gross reexamination.

Am J Surg Pathol

1999;23:876–79.