Subsequent breast carcinoma risk after biopsy with atypia in a breast papilloma

25%

Subsequent Breast Carcinoma Risk after Biopsy with Atypia in a Breast Papilloma

David L. Page, M.D.'"

Kevin E. Salhany, M.O:

Roy A. Jensen, M.D."~

William D. Dupont, Ph.D.2

' Department of Pathology, Vanderbilt Univer- sity School of Medicine, Nashville, Tennessee.

Department of Preventive Medicine, Vanderbilt University School of Medicine, Nashville, Ten- nessee.

Department of Cell Biology, Vanderbilt Univer- sity School of Medicine, Nashville, Tennessee.

Department of Pathology and Laboratory Med- icine, University of Pennsylvania, Philadelphia, Pennsylvania.

Presented in part at the 1988 IAP Annual Meet- ing, Washington, DC.

Supported in part by NCI R01 CA31698 grant from the NIH.

Address for reprints: David L. Page, M.D., De- partment of Pathology, Vanderbilt University Medical School, Nashville, TN 37232.

Received December 29, 1995; revision received April 8. 1996; accepted April 15, 1996.

BACKGROUND. Risk of breast cancer after biopsy demonstrating a papilloma has long been variously interpreted on the basis of histologic pattern of multiplicity of papillomas. METHODS. A nested case control study was performed on women with surgical breast biopsies evidencing papillomas; cases who subsequently developed invasive carcinoma were compared with controls who did not. Presence of atypical hyperplasia (AH) within the papilloma as well as areas of AH in the surrounding parenchyma were evaluated in both cases and controls. The entire cohort (not tested) was separately evaluated for all variables except for atypia within papillomas. RESULTS. The relative risk of invasive carcinoma for women with papillomas containing AH was >4x that of papillomas without AH within or surrounding the papilloma. This risk may be greater with added atypical hyperplasia outside the papilloma and most strikingly, most of the subsequent invasive carcinomas developed in the same breast and probably near the site of the original papilloma. However, ordinary patterns of epithelial hyperplasia lacking specific features of AH within the papilloma do not add to the risk of subsequent carcinoma development over papillomas without hyperplasia. CONCLUSIONS. This study indicates that women having papillomas with AH have a similar or greater cancer risk than others with specifically defined patterns of atypical hyperplasia within the breast parenchyma (4-5x relative risk). Most im- portantly, this risk is largely local, in the region of the original papilloma. Cancer 1996; 78:258-66. 0 1996 American Cancer Society.

KEYWORDS: breast, papilloma, atypical hyperplasia, breast carcinoma, multiple, cancer risk, premalignant breast disease, proliferative breast disease.

ntraductal papillomas are relatively uncommon lesions in most I breast biopsies. However, they are relatively common in biopsies done on the uncommon indication of bloody nipple discharge. Papil- lomas may be found in either a central subareolar location, where they are often solitary, or in a more peripheral area of the breast, where they are usually smaller and multiple. The malignant or premalignant potential of these lesions has been debated, but the majority of authorities do not consider uncomplicated intraductal papillomas to be precursors of papillary carcinoma or invasive breast carcinoma (IBC) nor do they consider them to elevate the risk for subsequent IBC.'-' Other investigators, however, have concluded that these lesions may be precursors for papillary carcinoma or that they may elevate the risk of subsequent IBC.6-9 Haagensen et al.3*4 has stressed the importance of distinguishing solitary intraductal papillomas from multiple papillomas because he found that multiple papillomas were much more likely to be associated with concurrent or subsequent development of breast carcinoma, while solitary intraductal papillo-

0 1996 American Cancer Society

Cancer Risk of Breast Papillomas/Page et al. 259

mas were not. Studies of Carter’ and Others’O.” have also shown an association of inultiple intraductal papillomas with intraductal and invasive carcinoma of the breast.

In a previous study of proliferative breast disease (PD) we found the relative risk for subsequent IBC to be almost doubled in women with specific proliferative lesions, thus deriving the phrase “proliferative disease” because the histologic lesions are indicative of cellular proliferation and a raised risk of breast cancer is indicated by their presence. These included moderate and florid hyperplasia without atypia, sclerosing adenosis, and To determine the unique contribution of atypia within true papillary lesions with a branching fibrovascular framework to this elevated risk, we undertook a nested (within the larger cohort) case-control study of a portion of our patients with benign intraductal papillomas. Applying analogous criteria to proliferations of cells within the papilloma as in parenchymal”’“ we now report a signifi- cantly elevated risk of IBC for women with atypical hyperplasia within the papilloma (AH-PAP) and frequent association of AH-PAP with atypical hyperplasia (AH) in nonpapillary breast tissue. We also confirm the contribution of a slight breast carcinoma risk elevation for women with papillomas alone, apart from their association with other elements of proliferative disease.

MATERIALS AND METHODS Patient Selection and Study Design The method of patient selection has been previously described in detail.’’,l5 In brief, 10,366 breast biopsies originally diagnosed as benign from 3 Nashville hospi- tals between 1950 and 1968 were selected for histologic examination. From this group, follow-up was obtained on 5966 patients (89%) who met the appropriate criteria. This differs slightly from the original papers because follow-up is now extended to all women in the cohort.I5 The incidence of subsequent invasive breast cancer was compared with that of age-matched women from the Third National Cancer Survey in At- lanta who had the same length of time at risk.’”’‘ From this cohort of 5966 women with biopsies for benign breast disease, biopsies from 368 women were found on our histologic review to have benign intraductal papillomas as defined by Wellings et al.” as well as ourselves.” Elements of proliferative breast disease (PBD) were compared in the entire cohort.

Also, to evaluate atypia in papillomas (not done in original reviewI2) a nested case-control study was done. All 31 papilloma patients who subsequently developed invasive breast cancer were included in the case group. Both sniall and large papillomas were in-

cluded. A designation of “small” was given if the largest papilloma was 3 mm or smaller in greatest dimension. A random sample of 91 papilloma patients, stratified by lesion size (equalized small and large by percentage in each group), was selected from papilloma patients who did not develop invasive breast cancer and were used as the control group in the nested case-control study. Several clinical and histologic parameters were evaluated. A special effort was made to determine the information related to site of origin of original papillary lesions relative to the later developing carcinomas. Although this information was usually recorded in the original data retrieval, when incomplete with regard to the site within the breast we retrieved office records and operative reports relative to the diagnosis and treatment of the later carcinoma as well as the pathology report and slides.

We also evaluated the experience of all women in the entire cohort with large and small papillomas with and without concurrent elements of proliferative disease: hyperplasia without atypia andlor sclerosing adenosis.‘“

Histologic Evaluation For the nested study, histologic sections (numbering 1-5lpatient) from biopsies of the 31 patients and 91 controls were evaluated for epithelial proliferative lesions within the papillomas or elsewhere within the biopsy. Proliferative lesions within breast tissue and within papillomas were classified as ordinary epithelial hyperplasia or atypical hyperplasia of ductal or lobular type according to the criteria of Page and Rogers.I4 Biopsies with proliferative lesions were independently reviewed without knowledge of subsequent development of breast cancer and were classified in a similar manner. Agreement between two of the reviewers was over 95% in assigning atypia to areas within papilloma(s) (differing in 4 patients); differences were resolved with a double-headed microscope.

Diagnostic Criteria for Proliferative Lesions within Papillomas A papilloma was diagnosed when lesions within ducts had fronds of fibrovascular stroma covered by epithelium. This luminal epithelium is variable, but often cuboidal. Stalk attachment to the wall was sometimes present, but this may be inapparent in histologic sections available. Papillomas were considered to have epithelial proliferation if the epithelium covering the papillary fronds was piled up at least 4 or more cells thick, without any consideration to myoepithelial cells. When present, epithelial proliferative lesions within papillomas were classified as either ordinary epithelial hyperplasia (EH-PAP) or AH-PAP following criteria

260 CANCER July 15,1996 / Volume 78 / Number 2

FIGURE 1. Low power view in A is of a small papilloma with fibrovascular fronds evident. Note normal lobular units around the area. Note also that in one area there is a region of hyperplasia with proliferated cells extending beyond the supporting, fibrovascular fronds. Even at this low power the slit-like spaces are evident to indicate that this is a pattern of ordinary epithelial hyperplasia, although in the upper right corner there is a more uniform area with more sharply defined and rounded spaces. B has, at higher power, the atypical focus so recognized because of regularly rounded, intercellular spaces and uniform cells as seen in atypical hyperplastic lesions of ductal pattern.

FIGURE 2. (A) High power view of an area within a papilloma showing development of sharply defined spaces between a fairly evenly placed population of similar cells. This was about the full extent of this change and indicated atypical hyperplasia (AH) in this papilloma. (6) Low power view of an area seen at high power in A. 6 shows that this is an encysted area with a papillary lesion inside. Note uniform population of proliferated cells with an even pattern of spaces between the cells defining AH, but confined to a small area (<3 mm).

similar to those for ordinary epithelial hyperplasia arid atypical hyperplasia elsewhere in the breast.I4 EH-PAP was diagnosed if epithelial cells were piled up three or more cells thick, often forming irregular slit-like spaces (Fig. 1) and lacking cytologic uniformity. AH- PAP was diagnosed if histologic patterns and cytologic features of hyperplastic epithelium approached but did not fulfill standard diagnostic criteria for ductal pattern carcinoma in situ (Figs. 2-4). We attempted to analogize as much as possible with the atypical hyperplastic lesions of ductal pattern (ADH) diagnosed in

nonpapilloma 10cations.l~ This approach specifically used rules of extent (or size) hoping to analogize the intrapapilloma location to the intralobular unit location of most ADH lesions. Specifically, any area of uniform histology and cytology consistent with noncom- edo ductal carcinoma in situ (DCIS) extending >3 mm in greatest dimension was considered to be a minor DCIS lesion; whereas the same features in -3 mm in greatest dimension were regarded as AH. This is analogous to the approach of Tavassoli and N ~ r r i s . ' ~ During the time of case accrual (1950- 1968) Nashville

Cancer Risk of Breast PapillornadPage et al. 261

FIGURE 3. These two photographs are from the histologically most atypical of all the cases. (A) Low power view of most of the atypical area which is continuous with papilloma at left. Note there is a cribriform focus seen at low power and some micropapillae similar to micropapillary ductal carcinoma in situ seen in both the low power (A) and high power (B) pictures. (6) High power photograph of a micropapillary atypia in an adjacent area to that seen as lower power in A.

dence intervals (CI) for RR were derived using the a- symptotic variance-covariance matrix of the coeffi- cient estimates from the regression

In Table 3 the risk of breast cancer among patients relative to the risk among females from the Third Na- tional Cancer Survey in Atlanta" was derived from standard morbidity ratios using the approach of Bres- low and Day."

FIGURE 4. This demonstrates an area suggestive of atypicality but notice that the more hyperchromatic nuclei present adjacent to the lumen are gradually attained and that there is regional variation in nuclear features and amount of cytoplasm. Therefore, this population of cells is not uniform as required in atypia, and has not developed sufficient cytologic uniformity to qualify as atypical hyperplastic lesions of ductal pattern.

pathologists usually diagnosed larger papillomas with such atypical foci as 'papillary carcinoma' so that many females with such lesions had mastectomy. Such patients were excluded from follow-up.

Statistical analysis Relative risks (RR) in Table 1 were estimated from odds ratios (OR) derived from logistic regression analyses. These risks were adjusted for age at biopsy by including age as a covariable in the regression model. Confi-

RESULTS Histologic Findings Review of biopsies from 122 patients with benign papillomas (31 patients and 91 controls) revealed proliferative lesions within the epithelial component of the papillomas from 47 patients. Epithelial proliferative lesions within the papillomas were usually EH-PAP (39 patients). Foci of AH-PAP were present in 8 patients. These foci were always of the ADH type; no foci of atypical lobular hyperplasia (ALH) were found within the papillomas. These foci of ADH within the papillomas usually occupied only a small portion of the papilloma (<25%) (Figs. 1-31, but in the most severe lesions, involved nearly half of the epithelium associated with the smaller papillomas. The patterns of ADH within atypical papillomas most frequently had features mimicking the cribriform pattern of DCIS (Figs. 1, 2, 4), but occasionally had a micropapillary pattern (Fig. 3). Papillary epithelium typical of benign papillomas could easily be found adjacent to foci of AH-PAP in all biopsies including the most severe case (Fig. 3). None of the papillomas meet the criteria for noninvasive (in situ) papillary carcinoma21,22 which would en- tail more extensive atypical changes. Papillomas fre-

262 CANCER July 15, 1996 / Volume 78 / Number 2

TABLE 1 Age-Adjusted Relative Risk of IBC in Women with Intraductal Papillomas

95% Confidence Numerator of relative risk No:' Denominator of relative risk No." Relative risk interval P value

AH-PA!' AH-PAP no AH AH-PAP t AH EH-PAP no AH

618 nonAH-P.4F 251114 9.6 1.62-56.91 0.013 0.036 213 n o n A H - P , P n o AH 221105 7.5 0.131 415 n o n A t W 4 P no AH 221105 15.8

7 / 3 5 nonEH-PAP no AH 17/73 0.78 0.32- 1.90 0.576

1.14-48.71 0.44-57.13

I B C inyasive breast cancer; AH-PAP: papilloma with atypical hyperplasia within papilloma; nonAl+PAP: papilloma without atypical hyperplasia within papilloma: AH-PAP t AH papilloma with arypia within and in breast parenchyma outside papilloma: no AH no atypical hyerplasia in surrounding parenchyma; EH-PAP papilloma with moderate or florid degrees of nonahFical hyperplasia (noAH-PAP]; nonEH-PA!? no epithelial hyperplasia within the papilloma. 'These fractions indicate the number of women developing invasive carcinoma divided by the total number of women in the group: (women with later canceriall women in group).

TABLE 2 Follow-Up Data for Patients with AH-PAP Who Subsequently Developed IBC (n = 6)

Presence of survival Patient Location of Bx Size of AH outside Interval Histologic S i e o f LN after no. Age with AH-PAP papilloma the papilloma to IBC Location of IBC type of IBC IBC status IBC Dx

1 45 1. 1 cm ADH 4 ITS L." invasive cribriform 2 cm tieg A&W,

2 7 3 R. LOQ lsubareolar) > I c m ADH 5 yrs R. LOQ' (subareolar) facally invasive 2.5 cm neg A&W,

3 70 R. UIQ (subareolar) 1.3 cm absent Z y r s L.LOQ NST <1 c m 3 pos Dead,

4 37 L. UOQ 0.1 c m ALH 13 yrs 1. UOQ (subareolar) NST 2.5 c m n e g A&W,

5 65 R. LOQ (subareolar) 1.1 cni absent 5 ys R. O Q (peripheral) NST <2 cm n e g Dead, 5 Yr

6 46 R. OQ [subareolar) 0.1 cn i ADH 6 yrs R. NST 3 cm n e g Dead, 5 Yr

15 yr

papillary 9yr

3yr

(subareolar)

AH-PAP atypical hyperplasia within the papilloma; IBC: invasive breast carcinoma: LOU: lower outer quadrant; LIQ: lower inner quadrant; UOQ: upper outer quadrant; UlQ: upper inner quadrant; ADH atypical ductal hyperplasia; A L H atypical lobular hyperplasia: NST invasive breast carcinoma of no special type (ductal); N A not available; A&W alive and well. " Specific documentation of occurrence within previous papilloma biopsy site.

quently had foci of apocrine metaplasia, sclerosis, or hemorrhage, and occasionally had areas of infarction and calcification. Epithelial proliferative lesions were also found outside the papilloma in 65 of 122 patients studied, with atypical hyperplasia present in 14, ADH in 10, and ALH in 4. However, ADH or ALH was found in the surrounding breast tissue in 5 of 8 patients with AH-PAP (4 ADH, 1 ALH).

Assessment of Relative Risk in Nested Design The RR estimates for patients in the nested study eval- uating atypia within the papillomas is summarized in Table 1. Six of 31 patients who developed invasive breast cancer had ADH within their papillomas coin- pared with 2 of 91 controls who did not develop invasive carcinoma (RR = 9.6; P = 0.013). AH-PAP was associated with ADH or ALH in the surrounding breast

tissue of 4 of the 6 patients who developed IBC. There were 9 patients and 8 controls with AH elsewhere in their biopsies. After elimination of these patients from the analysis (AH-PAP without AH vs. nonAH-PAP without AH) (Table 11, we found an RR of 7.5 for development of breast cancer among patients with AH-PAP. However, this difference rests on only 2 patients. When patients with AH-PAP had AH elsewhere in the biopsy; the relative risk doubled from 7.5 to 15.8. Overall (ab- solute risk), 53% (9117) of women with AH in parenchyma or AH-PAP developed invasive carcinoma in the follow-up period which was 17 years for the 91 controls.

Clinicopathologic Correlation Follow-up data for the 6 patients with AH-PAP who subsequently developed IBC is listed in Table 2. Of the

Cancer Risk of Breast Papillomas/Page et al. 263

TABLE 3 Relative Risk of Cancer in Women with Papillomasa

No. other With other Patients No AH with AH No FH with FH No AH or FH PD PDWA

Micropapilloma 3.35b 4.40 3.42b 3.77 3.54b 1.90 4.42 95% CI (1.94-5.77) (1.10-17.58) (1.98-5.88) (0.94-15.06) (2.01-6.23) (0.60-5.74) (2.38-8.21) No. Dev. M a t risk 131154 2/14 13146 2122 121134 3/59 10195

Large papilloma 2.30' 13.1b 3.Wb 3.20 2.10 2.27 2.42

No. Dev. CAiat risk 91130 4/16 111125 2/21 71112 71100 2/30 95% CI (1.20-4.42) (4.93-35) (1.70-5.53) (0.80-12.81) (1-4.40) (1.08-4.76) (0.61-9.68)

No AH: no atypical hyperplasia in parenchyma surrounding papilloma; FH: history of breast carcinoma in at least one first degree relative; P D proliferative disease, AH, or any sclerosing adenosis or hyperplasia associated with cancer risk elevation; PDWA: PD without any atypical ductal or atypical lobular hyperplasia; No. Dev. CA. number of women developing invasive carcinoma in this group: at risk number of women in this group at risk for developing cancer; CI: confidence interval. "Risks are derived relative to women from the Third National Cancer Sumey in Atlanta and are adjusted for age at biopsy and length of follow.up.

' l w o of [he 9 women in this group who developed invasive cancer had AH in the original papilloma. If they are removed, risk falls to about 1.8. P value < 0.001.

6 patients with AH-PAP who later developed IBC, the IBC occurred in the same breast as the papilloma in 5; and at least 2 of the IBCs were documented to have occurred in the papilloma biopsy site. The IBC in an- other patient occurred in the same quadrant and general location, but we could not document if it occurred at the papilloma biopsy site. In at least one patient with AH-PAP and IBC in the same breast, the IBC did not occur at the papilloma biopsy site. Inadequate historical information regarding the precise location of the IBC was available for the other patient who developed IBC in the same breast as the AH-PAP. The IBC in the sixth patient with AH-PAP occurred in the contralateral breast. Interestingly, the papilloma from this sixth patient was the one which came closest to meeting the criteria for noninvasive papillary carcinoma (Fig. 3 ) . Four of these patients developed invasive mammary carcinoma of no special type, one had invasive cribriform carcinoma, and the sixth developed focally invasive papillary carcinoma. It is note- worthy that the invasive papillary carcinoma occurred in the same site as the previous AH-PAP with an interval of 5 years between lesions. All six of these carcinomas developed from 2 to 13 years following the entry biopsy. Three of these six atypical papillomas were large subareolar masses (1 cm), two were microscopic papillomas, and the localization of the last of the six within the breast could not be determined. Of the 25 patients who developed IBC but did not have atypical papillomas, the IBC occurred in the ipsilateral breast in 6, the contralateral breast in 6, and was bilateral in 2 patients. Ipsilateral or contralateral location of the IBC could not be determined in the other 11 patients because of bilateral entry biopsies or inadequate documentation in clinical or surgical pathology records of

the breasts involved. The median interval from the papilloma biopsy to IBC in these patients was 8 years (range: 1-24 yrs), reflecting the experience of the entire cohort.

Cancer risk of papillomas compared with general population When all of the women in our cohort undergoing surgical biopsy in Nashville between 1950 and 1968 were analyzed for cancer risk and compared with the general population, we found women with papillomas had an elevated cancer risk not attributable to positive family history of cancer or atypical hyperplasia outside of the papillomas (Table 3 ) . The groups were analyzed with the end in mind to ascertain the adequacy of the presence of micropapilloma or large papilloma alone to indicate increased risk of breast cancer apart from other lesions of proliferative breast disease known to produce elevated risk. With regard to interaction with atypical hyperplasia, only small papillomas when present without atypical hyperplasia within or outside the papilloma appear to have a greater than doubling of cancer risk. This risk was highly significant, approached almost 3.5, and was maintained even when women with a positive family history were removed. However, it is also evident that much of the risk indicated by micropapillomas does depend on the concurrent presence of other forms of proliferative disease, predominantly usual, moderate, and florid hyperplasia because only 3 of the 59 women without concurrent proliferative disease of some kind and micropapillomas developed invasive breast cancer. This was without statistical significance. It is when present with other forms of proliferative disease (sclerosing adenosis and usual patterns of hyperplasia), that risk was


highly significant and possibly higher than that of other forms of proliferative disease alone. It would seem that large papillomas (most of our cases were solitary) have somewhat less indication of increased risk, particularly if found without other diagnoses indicating proliferative disease. When two of the nine women developing invasive breast carcinoma at a later time who had AH-PAP are removed from consideration, the risk of women with large papilloma without atypical hyperplasia in the parenchyma falls to about 1 . 8 ~ (Table 3 ) .

DISCUSSION The major aim of this study was to ascertain the importance of AH-PAP. This was done with a nested study looking at all of the women with papillomas having developed IBC compared with a control group (Table 1). Table 2 details the specifics of the six women who developed IBC and had atypia present within their papilloma(s). This atypia was defined as closely as possible to being analogous to the diagnosis of atypical ductal hyperplasia made in the parenchyma, where it is usually confined to a single lobular unit. Note that four of these women also had atypical hyperplasia outside the papilloma in the parenchyma, arid only two, both with large papillomas, had only AH- PAP itself. The occurrence of most later cancers in the area of the original papilloma indicates that this risk is likely preferentially directed at the local area.

The finding that papillomas 5 3 mm in greatest dimension may have greater implication for canc:er risk than larger ones may be related to the evidence that they are often multiple; while larger papillomas are usually ~ o l i t a r y . ' ~ ~ ~ ~ " ' ~ Haagensen et al.3.4,'4 have championed this view, but analysis of this series is complicated by the presence of frequent patterns of histologic atypia in his cases.I8 While this study design cannot evaluate the effect(s) of multiplicity, these findings indicate that two added histologic features are of importance: presence of AH in or around a papilloma and concurrent presence of proliferative disease without atypia in association with small papillomas.

In the early part of this century papillomas were treated by mastectomy, a posture replaced in the sec- ond quarter of this century by a belief that papillomas were benign and could be cured by local excision. Papers indicating that there might be some papillomas more bothersome than others were few and it has been widely believed that there are lesions to be regarded as carcinoma which have a papillary growth pattern with an abrupt separation from benign Saphir and Parker" did indicate that proliferative ac- tivity within papillomas might be of some concern

with regard to the development of malignancy, and about 10 years later Haagensen et al.'" indicated that multiplicity of papillomas was an indicator of the likelihood of evolution into malignancy. While it is highly likely that many of the cases with multiple papillomas have areas of atypia and indeed carcinoma in sit^,^ this relation is not settled at the present time. Papotti et aL2' did point to the fact that patterns seen in non- comedo DCIS may be present within papillomas and were concerned that evolution to invasive disease might well be likely in such lesions. Ohuchi et al.1',2fi also talked about such patterns in papillomas, further characterized the growth patterns of these lesions, and emphasized their continuity along ducts. They noted that solitary papillomas are confined to large ducts, whereas women with multiple papillomas regularly had involvement of proliferating cells in adjacent lobular units. Multiplicity of papillomas has also been held to indicate an elevated likelihood of later breast cancer development. The current study did not evaluate multiplicity, indeed that situation is probably quite uncommon with Haagensen collecting only about 50 cases in his study extending over many years.4 It would seem that there is some correlation between multiplicity of papillomas and the likelihood of atypical epithelial proliferation within them. Indeed, whether the cancer risk is due to the multiplicity of papillomas or the atypical hyperplasias commonly present within them may never be resolved because of the difficulty of assembling a sufficiently large series of patients with definition of disease pattern and extent. Such a study will be made even more difficult by the fact that with the frequent over diagnosis of foci of AH-PAP as carcinoma, most of these lesions will be treated by ade- quate excision at time of diagnosis.

This paper was planned to gain the information available about papillomas from a large cohort study of patients undergoing breast biopsies in Nashville during the 19-year period between 1950 and 1968. The study design was a case-control study nested within the large cohort in which we gathered all 31 patients with papillomas on initial review who developed IBC and compared them with 3x as many controls who had not developed carcinoma. Thus, the study design does not allow us the ability to report RR figures for AH-PAP relative to the general population. However, the overall RR figures from the entire cohort presented in Table 3 are not confounded by the presence of atypia within small papillomas ( 5 3 mm). This is because all of these patients had atypia in the parenchyma as well as within the small papillomas. Thus we can remove all patients with small papillomas containing atypia from the analysis with those cases si- multaneously with the removal of patients with atypia

Cancer Risk of Breast PapillomaslPage et al. 265

in parenchyma outside the papilloma (Table 3). Two of the females with large papillomas containing atypia developed invasive cancer arid had atypia only in the papillomas. Although the relative likelihood of later cancer development in patients with papillomas and atypia as contrasted with women without such patterns may not be discerned directly, removing these 2 women from the 9 who had large papillomas without atypia in the parenchyma would produce an approxi- mate change from 2.2 to 1.8 in overall risk. This implies a doubling of risk for large papillomas without atypia within the papilloma or remainder of the biopsy speci- men.

For the current study, to determine what histologic factors within papillomas might be responsible for elevating breast cancer risk in women with papillomas we undertook a nested case control study of papilloma patients. We found a greatly increased risk of breast cancer for women with papillomas when atypical hyperplasia was present within the papilloma (RR = 9.61, compared with papillomas 1ackingAH. The 95% CI (1.6-56.9) indicates that the risk of IBC for AH-PAP is at least 1 . 6 ~ that for papillomas without AH. The cancer risk of patients with papillomas which lack AH either within the papilloma or elsewhere closely ap- proximates the risk for proliferative disease without atypia in general. These findings suggest that much of the increased risk of IBC found in women with papillomas is in part related to AH either within or outside of the papilloma, but that papillomas without AH also indicate an elevated breast cancer risk, although of only a slight magnitude. The insightful review of Car- ter' indicated that papillomas might have some ability to predict risk of later carcinoma and also suggested that greater extensiveness of hyperplastic changes might increase this predictiveness. This current study presents suggestive evidence that AH in any papilloma should raise concerns about the risk of local invasive cancer development. Also, papillomas of any kind indicate an elevation of breast cancer risk about double that of the general population. This may be slightly greater when the various forms of proliferative disease without atypia (florid or moderate hyperplasia, sclerosing adenosis) coexist in the breast tissue outside the papillomas.

We conclude that it is clinically significant to have AH within any papilloma and that this should be regarded with concern and certainly close mammographic follow-up. This close surveillance should probably direct special vigilance to the area within the breast of the papilloma found to have AH. We suggest that the practical, surgical implications of this study are just that, implications awaiting further confirma- tion of the local risk of cancer after local excision of

papillomas with AH. We do not think that wider local re-excision is regularly indicated, but that careful mammographic surveillance (as with specifically atypical ductal or lobular hyperplasia) is indicated with special attention to the local region and quadrant of the papilloma with AH.

REFERENCES 1.

2.

3.

4.

5.

6.

7.

8.

9.

10

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

Azzopardi JG. Problems in Breast Pathology. London: W.B. Saunders, 1979: 150-239. Hart D. Intracystic papilloniatous tumors of the breast, benign and malignant. Arch Siirg 1927; 14:793-9. Haagensen CD, Bodian C, Haagensen DE Breast Carcinoma, Risk and Detection. Philadelphia: W.B. Saunders 1981; 146- 9. Haagensen CD. Diseases of the breast. Philadelphia: Saun- ders, 1986. Kraus FT, Neubecker RD. The differential diagnosis of papillary tumours of the breast. Cancer 1962; 15:444-55. Buhl-Jorgensen SE, Fischermann K, Johansen H, Peterson B. Cancer risk in intraduct papilloma and papillomatosis. Szrrg Gyrrecol Obstet 1968; 127:1307-12. Carter D. Intraductal papillary tumors of the breast: a study of 78 cases. Cancer 1977;39:1689-92. McDivitt RW, Holleb AI, Foote FWJ. Prior breast disease in patients treated for papillary carcinoma. Arch Path01

Moore SW. Pearce J, Ring E. Intraductal papilloma of the breast. Surg Gynecol Obstet 1961; 112:153-64. Murad TM. Contesso G, Mouriesse H. Papillary tumors of large lactiferous ducts. Cancer 1981; 48(122): 122-33. Ohuchi N , Rikiya A, Kasai M. Possible cancerous change of intraductal papillomas of the breast: a 3D reconstruction study of 25 cases. Cuncer 1984;54(605):605-11. Dupont WD, Page DL. Risk factors for breast cancer in women with proliferative breast disease. N Engl / Med

Page DL, Dupont WD, Rogers LW, Rados MS. Atypical hyperplastic lesions of the female breast. A long-follow-up study. Cancer 1985; 55(2698):2698-708. Page DL, Rogers LW. Combined histologic and cytologic criteria for the diagnosis of mammary atypical ductal hyperplasia. Hurn Puthol 1992;23:1095-7. Page DL. Dupont Wn, Jensen RA. Papillary apocrine change of the breast associated with atypical hyperplasia and risk of breast cancer. Cancer Epidemiol, Biornarkers Prev 1996. In press. Cutler SJ, Young JLJ. Third National Cancer Survey: incidence data DHEW. Bethesda (MD): National Cancer Insti- tute; 1975. Publication No.: NIH-NHLI-69-2185-4. Wellings SR, Jensen HM, Marcum RG. An atlas of subgross pathology of the human breast with special reference to possible precancerous lesions. / Narl Cancer Inst 1975; 55:231-73. Page DL, Anderson TJ. Papilloma and related lesions. Diag- nostic Histopathology of the Breast. Edinburgh: Churchill Livingston, 1987; 104-119. Tavassoli FA, Norris HI. A comparison of the results of long- term followup for atypical intraductal hyperplasia and intraductal hyperplasia of the breast. Cancer 1990;65:518-529. Breslow NE. Day NE. Statistical Methods in Cancer Re- search: Volume 11-The design and analysis of cohort studies. Lyon, France: IARC Scientific Publication, 1987.

1968;85:117-24.

1985; 3 12: 146-5 1.


21. Page DL, Anderson TJ, Rogers LW. Carcinoma in situ. In: Page DL, Anderson TJ, editors. Diagnostic Histopathology of the Breast. Edinburgh: Churchill Livingstone, 1987; 157- 92.

22. Carter D, Orr SL, Merino MI. lntracystic papillary carcinoma of the breast. After mastectomy, radiotherapy or excisional biopsy alone. Cancer 1983;52:14-9.

23. Dawson AE, Mulford DK. Benign versus malignant papillary neoplasms of the breast: diagnostic clues in fine needle aspi- ration cytology. Acta Cytol 1994;38(1):23-8.

24. Haagensen CD, Stout AP, Phillips JS. The papillary neoplasms of the breast. Ann Surg 1951; 133:18-36.

25. Lefkowitz M, Lefkowitz W, Wargotz ES. Intraductal (intracystic) papillary carcinoma of the breast and its variants: a clinico- pathological study of 77 cases. Hum Pathol 1994;25:802-9.

26. Ohuchi N, Abe R, Takahashi T, Tezuka F. Origin and exten- sion of intraductal papillomas of the breast: a three-dimen- sional reconstruction study. Breast Cancer Res Treat 1984;

27. Saphir 0, Parker ML. Intracystic papilloma of the breast. Am

28. Papotti M, Gugliotta P, Ghiringhello B, Bussolati G. Associa- tion of breast carcinoma and multiple intraductal papillomas: an histological and immunohistochemical investiga- tion. Histopathology 1984; 8:963-75.

4(117):117-28.

J Path01 1940; 16~189-98.

Subsequent breast carcinoma risk after biopsy with atypia in a breast papilloma

Documents

Transcript of Subsequent breast carcinoma risk after biopsy with atypia in a breast papilloma