ANATOMIC PATHOLOGY. Original Article

Transcription

1 ANATOMIC PATHOLOGY Original Article Intraductal Carcinoma Associated With Invasive Carcinoma of the Breast A Comparison of the Two Lesions With Implications for Intraductal Carcinoma Classification Systems NEAL S. GOLDSTEIN, MD, AND TRACY MURPHY, Intraductal carcinoma (DCIS) is a useful marker for predicting which women will develop a recurrent breast malignancy. The authors examined 5 consecutive, mammographically detected, Tl invasive carcinomas associated with DCIS to study the DCIS and compare it to its associated invasive carcinoma. Intraductal carcinoma nuclear grades were assigned to each duct on a scale of to. The percentage of DCIS ducts that were involved by each grade was quantitated into quartiles for cases with more than one DCIS nuclear grade. The predominant architectural pattern corresponding to each DCIS nuclear grade was recorded. Ninety-two percent of the 5 invasive carcinomas were of ductal type, 4% were tubular, and the remainder were various other subtypes. Nine percent of the DCIS cases were nuclear grade. The remaining % of cases were almost evenly distributed between mixed DCIS nuclear grades and (%), pure DCIS nuclear grade (4%), Intraductal carcinoma, or duct carcinoma in situ (DCIS) of the breast is a lesion that has increased in importance with the development of breast-sparing therapy, because it is a useful index for helping to predict which patients will develop a recurrent malignancy, either DCIS or invasive carcinoma. It is thought that DCIS is the progenitor of invasive carcinoma, yet only a few investigators have attended to the relation between types and grade of invasive carcinoma and different forms or grades of the associated DCIS. " Although authors have documented architectural heterogeneity within a given patient's DCIS, ' 4 " 7 little attention has been giving to the degree of nuclear heterogeneity that can be present in a patient's DCIS lesion. Addressing the degree of nuclear heterogeneity within From the Department of Anatomic Pathology, William Beaumont Hospital. Royal Oak, Michigan. Manuscript received November 4, 5; revision accepted February 8, 6. Address all correspondence to Dr. Goldstein: Department of Anatomic Pathology. William Beaumont Hospital, 6 West Mile Road, Royal Oak, MI 487. MD mixed DCIS nuclear grade to (5%), and pure DCIS nuclear grade (%). Two percent of cases were a mixture of DCIS nuclear grades and or,, and. All pure DCIS nuclear grade or mixed and were associated with well or moderately differentiated invasive carcinomas, whereas the majority (6%) of the pure DCIS nuclear grade cases were associated with poorly differentiated invasive carcinomas. There was no relation between the DCIS architectural pattern and the invasive carcinoma grade. In general, the DCIS nuclear grade correlates with the grade of the invasive carcinoma. Unlike DCIS architecture, nuclear grade heterogeneity within DCIS associated with invasive carcinoma is minimal. DCIS classification systems based on nuclear grade have merit because there is little variation in nuclear grade within a given patient's lesion. (Key words: Breast; Duct carcinoma in situ; Invasive carcinoma) Am J Clin Pathol 6; 6:-8. DCIS is relevant to the ongoing debate about DCIS classifications. 7 " We examined 5 consecutive, mammographically detected, small invasive carcinomas with DCIS accessioned over the -year period -4 to study the relation between DCIS nuclear grade and architecture and composite grade of the associated invasive carcinoma, and to document the degree of nuclear and architectural heterogeneity within DCIS. MATERIALS AND METHODS One hundred fifty consecutive, mammographically detected, pathologic stage-t invasive carcinomas accessioned by the Department of Anatomic Pathology at William Beaumont Hospital during to 4 were reviewed. were identified with the assistance of the hospital Tumor Registry. All slides from all the patients' breast-related procedures were examined, including biopsies, lumpectomies, and mastectomies. The invasive carcinomas were categorized by the criteria of Page and Anderson" and graded with the Ellston and Ellis criteria. Invasive carcinomas devoid of an intraductal carcinoma component were excluded. Downloaded from by guest on November 8

2 GOLDSTEIN AND MURPHY Carcinoma FIG.. Nuclear grade duct carcinoma in situ (DCIS). Round uniform nuclei with occasional punctate nucleoli (hematoxylin and eosin, XI5). This photomicrograph demonstrates minimal nuclear variation. Although there is some heterogeneity in chromatin density between cells, none of the cells have course chromatin. Also, there is little variation in nuclear size and shape between cells. Despite the variation, all the cells are less atypical than grade DCIS. This photomicrograph was taken at X.8 lower magnification than Figures and to provide a larger field. Nuclear grades, on a scale of to, were assigned to each DCIS duct in each lesion. Grade was the lowest nuclear grade (Fig. ). The nuclei of grade DCIS were usually small with little variation in size. They had smooth nuclear membranes, uniform chromatin, and occasional punctate nucleoli. DCIS grade cells were monomorphic and were evenly spaced. Grade DCIS was composed of cells that were larger than grade DCIS cells (Fig. ). There was moderate variation of the FIG.. Nuclear grade duct carcinoma in situ (DCIS). Slightly larger nuclei than nuclear grade DCIS with moderate pleomorphism and occasional medium-sized nucleoli (hematoxylin and eosin, X). FIG.. Nuclear grade duct carcinoma in situ (DCIS). Highly atypical nuclei with extreme pleomorphism, course chromatin, and many large nucleoli (hematoxylin and eosin, XI). nuclear shape, and the nuclear membrane thickness varied slightly. The chromatin was less homogenous than in grade DCIS nuclei. Nucleoli were present in most cells, including macronucleoli. Grade DCIS cells had more size and shape variation than grade DCIS. Grade DCIS was composed of pleomorphic cells (Fig. ). There was marked variation in the size and shape of grade DCIS cells. Most nuclei were large, with nuclear membrane irregularities. Chromatin was very course and clumped, and usually there was at least one macronucleolus. The predominant nuclear grade in each duct was assigned when there was a mixture of nuclear grades within a duct. The percentage of the DCIS ducts composed of each nuclear grade was quantified into quartiles, (%, l%-5%, 6%-5%, 5%-75%, 76%-%, and %). The predominant architectural pattern and second most common architectural pattern associated with each DCIS nuclear grade were categorized as centrally necrotic, solid, cribriform, micropapillary, and clinging. For this study, centrally necrotic was used as a distinctive architectural pattern and could be formed by cells of any DCIS nuclear grade. Centrally necrotic architectural pattern was defined as a duct with a well-defined central zone of coagulative necrosis, ringed by tumor cells adherent to the duct wall. Occasional small glands within the peripheral ring of solid tumor cells were allowed within a centrally necrotic pattern and were not classified as cribriform pattern. Aside from being a mandatory constituent of the centrally necrotic architectural pattern, the presence of small to moderate amounts of necrosis within ducts was not evaluated. Downloaded from by guest on November 8 Vol. 6-No.

3 4 ANATOMIC PATHOLOGY Original Article TABLE. GRADE AND TYPE OF 5 PATHOLOGIC STAGE Tl INVASIVE CARCINOMAS Type No. of i of Total Invasive Carcinoma Grade Ductal, NOS Tubular Mucinous Lobular Mixed, tubular, and ductal Mixed, tubular, and lobular Total of all subtypes NOS = not otherwise specified. Values in parentheses are percentages. 8 7 The nonparametric Kendall Tau-B statistical analysis was used to study the relation between DCIS nuclear grade and the invasive carcinoma composite grade. The range of possible results for this test are to +. The closer the result to +, the stronger the relation between the two values being tested. RESULTS One hundred thirty-eight (%) of the 5 invasive carcinomas were ductal not otherwise specified (NOS) type; 7 (4%) were tubular; and the remainder were other subtypes (Table ). Fifty-one (4%) invasive carcinomas were grade, 58 (%) were grade, and 4 (7%) were grade. The mean size of the invasive carcinomas was. cm (range.-. cm). There were 4 cases (%) of pure nuclear grade DCIS (Table ). The remaining cases were almost evenly TABLE. DISTRIBUTION OF DCIS NUCLEAR GRADES DCIS Nuclear Grade Pure Mixed and 5-75% 6-5% -5% Pure Mixed and -5% 6-5% 5-75% No. of % of Total < < 4() 7() 5(4) 56 (4) 58() 4() 4(7) distributed between mixed nuclear grade and DCIS (8 cases, %), pure nuclear grade DCIS (5 cases, 4%), mixed nuclear grade and DCIS (7 cases, 5%), and pure nuclear grade DCIS ( cases, %). Three cases (%) were a mixture of nuclear grade and or nuclear grades,, and DCIS. A general association between the DCIS nuclear grade and invasive carcinoma grade was observed (Table ). Twelve of the 4 (86%) pure nuclear grade DCIS cases and 8 of the 8 (6%) mixed nuclear grade - cases were associated with grade invasive carcinomas. Of the 5 pure nuclear grade DCIS cases, 5 (4%) were associated with grade invasive carcinomas and (7%) were associated with grade invasive carcinomas. Eighteen (5%) of the 7 mixed nuclear grade and DCIS cases were associated with grade invasive carcinomas and 4 (6%) were associated with grade invasive carcinomas. Twenty of the (6%) of the pure nuclear grade DCIS cases were associated with grade invasive carcinomas. Nonparametric statistical analysis comparing the DCIS nuclear grade and the invasive carcinoma grade (Table ) produced a Kendall Tau-B value of.6. This confirms the observation of a moderately strong relation TABLE. DISTRIBUTION OF DCIS NUCLEAR GRADE AND INVASIVE CARCINOMA GRADE DCIS Nuclear Grade Invasive Carcinoma Grade No. of I i (%) Downloaded from by guest on November % 6-5% -5% Pure Mixed and and Mixed and DCIS = ductal carcinoma in situ. -5% 6-5% 5-75% Pure Mixed and Pure Mixed and Pure Mixed - and / Values are no. (%). DCIS = ductal carcinoma in situ. (86) 8(64) 5(4) 5() () (4) (6) (7) 8(5) (6) 7() 4(6) (6) 4() 8() 5() 7() () A.J.C.P.-September 6

4 GOLDSTEIN AND MURPHY 5 Intraductal Carcinoma TABLE 4. DISTRIBUTION OF DCIS ARCHITECTURAL PATTERNS Predominant Architectural Pattern DICS Nuclear Grade Cribriform Solid Micropapillary Centrally Necrotic Clinging No. of Pure Mixed and 4 4 Nuclear grades 8 No. of cases Pure Mixed and Nuclear grades No. of cases Pure grade DOS = ductal carcinoma in situ between the DCIS nuclear grade and the grade of the associated invasive carcinoma. All 4 cases of pure nuclear grade DCIS had a cribriform-predominant architectural pattern (Table 4). Of the 8 mixed nuclear grade and DCIS cases, cribriform was the predominant architectural pattern in (7%) of the grade and (75%) of the grade ducts. The predominant architectural pattern in almost all the other mixed nuclear grade and DCIS cases was solid. Eighteen of the 5 (5%) pure nuclear grade DCIS cases were predominantly cribriform, (4%) were solid, and 4(%) were predominantly centrally necrotic. Of the 7 mixed nuclear grade and DCIS cases, cribriform was the predominant architectural pattern in the grade ducts in 4 cases and of the grade ducts in 8 cases. Centrally necrotic pattern was the predominant architectural pattern in the nuclear grade ducts in of the mixed nuclear grade - cases. Centrally necrotic was also the predominant architectural pattern in 5 (46%) of the pure nuclear grade DCIS cases. Solid was the second most common architectural pattern in the pure nuclear grade DCIS cases. Second most common architectural patterns were evaluated for each DCIS nuclear grade. Similar to the predominant architectural pattern, there were no cases of pure grade DCIS or mixed nuclear grade and DCIS that had a centrally necrotic pattern. Other than this relation, there was marked heterogeneity of secondary architectural patterns in all the DCIS nuclear grade groups. Although some generalities could be observed, there was no statistically significant relation between the DCIS architectural pattern and the grade of the associated invasive carcinoma (Table 5). The majority of ducts composed of predominantly cribriform pattern DCIS were associated with grade invasive carcinomas. The predominantly solid pattern of DCIS was associated with all three grades of invasive carcinoma. All four cases of pure nuclear grade DCIS that formed predominantly centrally necrotic pattern were associated with grade invasive carcinomas. Thirteen of the pure nuclear grade, centrally necrotic pattern DCIS lesions were associated with grade invasive carcinomas, and the other two were associated with grade invasive carcinomas. DISCUSSION Intraductal breast carcinoma (DCIS) is a heterogeneous lesion, ' 4 " 7 in both its biologic behavior and morphology. What was initially recognized as a single entity has evolved over the last two decades into a rubric covering several entities. There are at least two subtypes of DCIS. The noncentrally necrotic, low nuclear grade subtype has low to minimal risk of recurrence after wide excision, whereas centrally necrotic, high nuclear grade (comedo) subtype has a high risk of recurrence despite TABLE 5. PREDOMINANT DCIS ARCHITECTURAL PATTERN AND INVASIVE CARCINOMA GRADE Predominant DCIS Architectural Pattern Cribriform Solid Micropapillary Clinging Centrally necrotic, pure grade Centrally necrotic, pure grade DCIS = ductal carcinoma in situ. 5 7 Invasive Carcinoma Grade / Downloaded from by guest on November 8 Vol. 6 «No.

5 6 ANATOMIC PATHOLOGY Article radiation therapy and complete conservative excision. 44 " 8 Quantitative and qualitative components of DCIS contribute to its differences in recurrence risk. Authors have attempted to classify each component to predict the risk of residual tumor or recurrence after local excision (of the DCIS and/or associated invasive carcinoma) or to establish the risk of a coexistent invasive carcinoma. Quantitative components have included findings of Lagios and colleagues' that a DCIS lesion greater than.5 cm has a significant association with invasive carcinoma and multicentricity or that "extensive intraductal carcinoma" associated with an invasive carcinoma has a significant risk of having residual DCIS within the breast after local excision. " 5 Qualitative histologic parameters for DCIS have included nuclear size and grade, architecture, presence and pattern of necrosis, c-erb- protein, p5 protein, estrogen and progesterone receptor expression, and DNA analysis. 4 " 6 ' I4-6 " 4 Nuclear grade of the DCIS has been one of the most extensively studied qualitative histologic parameters.' " 7i467 ' 6 ' '- 5 " 8 However, the amount of DCIS nuclear heterogeneity that can exist within a lesion has received little attention. Our results show that, unlike DCIS architecture, which can be extremely heterogeneous, 5,6 nuclear heterogeneity within a patient's DCIS (associated with invasive carcinoma) is minimal. The clear majority of women in our study had either DCIS with a single nuclear grade or DCIS composed of two adjacent nuclear grades. Only of 5 cases, % of our cases, had a mixture of grade and grade DCIS. In contrast, we observed a continuous spectrum of DCIS nuclear grades from different women. Pure nuclear grade DCIS comprised the smallest group of lesions, % of the 5 cases. The remainder of the lesions were almost evenly distributed between nuclear grade and, pure nuclear grade, mixed nuclear grade and, and pure nuclear grade DCIS. Additionally, there was an almost even distribution of cases between the percentage of DCIS nuclear grades within the mixed-grade DCIS cases. These results require examination in light of how previous authors have approached DCIS nuclear grading. The classifications of DCIS nuclear grades have varied depending on the goal of the study. In descriptive studies of DCIS, such as the one by Lennington and colleagues, 6 a four-tiered grading scheme was used, and in cases with multiple nuclear grades, the nuclear grade of each pattern was averaged, producing a single final nuclear grade for each case. Moriya and associates' compared pure DCIS and DCIS associated with invasive carcinomas and classified the DCIS nuclear grade using a three-tiered system based on the highest nuclear grade. Patchefsky and coworkers 5 examined various features of DCIS in biopsies and related them to the findings seen in the mastectomy specimens. They also used a three-tiered nuclear grading system and classified the DCIS based on the highest nuclear grade. The greatest differences in how investigators have approached DCIS nuclear grading are found in studies that examined features of DCIS to predict failure with breastconserving therapy. Lagios and colleagues 4 classified DCIS nuclei into three grades and combined nuclear grading with architectural patterns to produce four DCIS subtypes, keeping each nuclear tier distinct. The NSABP Protocol B-7 had a similar goal. The authors initially classified DCIS nuclei using a three-tiered system and then, for statistical analysis, grouped nuclear grades and as "good" and as "poor'.' 7 DCIS lesions containing both "good" and "bad" nuclei were grouped as "poor" There are several extant DCIS classification systems, based primarily on the architectural pattern." A drawback to these classification systems is DCIS architectural heterogeneity. 5,6 Recently, Holland and colleagues 4 published a DCIS classification system that uses nuclear grade as the primary classification criterion and DCIS architecture as a subclassification criterion. This classification system has support from prior studies,' 4 although, as pointed out by Fisher and colleagues, this classification may be more of a nosologic issue than a truly new classification. Another criticism of the Holland and colleagues' classification system noted by Lagios, is that it discounts nuclear size as an intricate component of the DCIS classification. Our results support the use of the Holland and colleagues' DCIS classification system, because we found that there is only a narrow spectrum of DCIS nuclear grades within a given patient's lesion, and, aside from the infrequent nuclear grade DCIS, there was an almost even distribution of DCIS nuclear grades within our study population. This provides evidence for consistency of Holland's primary grading criterion. Had there been a clustering of DCIS nuclear grades within the study population, or marked nuclear grade variation within a patients' DCIS, it would have cast doubts on the merits of this classification system. Ideally, there should only be one method of classifying DCIS. The ideal classification system would provide all the prognostically useful information that our clinical colleagues require. Also, it should be straightforward and simple enough that it can be applied consistently and uniformly by pathologists. If DCIS prognostic features were a static issue, the creation of an ideal DCIS classification system would be a relatively simple task. How- Downloaded from by guest on November 8 A.J.C.P.^ :r 6

6 GOLDSTEIN AND MURPHY 7 Intraductal Carcinoma ever, qualitative and qualitative DCIS prognostic parameters are in a state of flux. The significance of some parameters and the optimal methods of their evaluation are subjects of active debate. 7 " It must also be recognized that DCIS can occur as a pure lesion or be found in association with invasive carcinoma. It is unknown whether similar-appearing DCIS lesions act differently or have the same biologic or predictive significance when they are associated with invasive carcinomas versus when they occur in pure form. An ideal DCIS classification system must be adaptable to the changing qualitative and quantitative prognostic features as they evolve. Our study also examined the relationship between DCIS nuclear grade and grade of invasive carcinoma. We found, in general, a trend between the DCIS nuclear grade and the invasive carcinoma grade. Pure nuclear grade DCIS was associated only with grade invasive carcinomas. Mixed nuclear grade and DCIS was associated with grade or invasive carcinomas, and nuclear grades and/ or DCIS were associated predominantly with grade and invasive carcinomas. These findings corroborate those of other authors. - Lampejo and colleagues performed a similar study that included stage and invasive carcinomas. Only 8% of the tumors were pathologic T-stage (< cm). They classified DCIS using the Holland and colleagues' classification system. 4 They found a highly significant association between the grade of the infiltrating carcinoma and the DCIS grade. Grade DCIS was associated with welldifferentiated invasive carcinomas, grade DCIS was usually associated with moderately differentiated invasive carcinomas, and grade DCIS was associated with poorly differentiated invasive carcinomas. These results support the idea that well-differentiated DCIS gives rise to well-differentiated invasive carcinoma. Our results, which derived from the study of pathologic stage-tl invasive carcinomas (mean diameter of. cm), and those of Lampejo and colleagues, both suggest that the nuclear grade of DCIS associated with invasive carcinoma does not transform over the time that the invasive component is growing and enlarging. Seven percent of DCIS lesions in their series were grade. This is close to our figure of % of cases with pure nuclear grade DCIS. That both studies had almost the same percentage of grade DCIS cases, although almost of of their cases exceeded cm (mean.8 cm, range -8.5 cm), supports the concept that the nuclear grade of DCIS remains constant throughout much of the evolution of the DCIS lesion, even after the DCIS has eventuated into invasive carcinoma. We corroborate the findings of others that DCIS architectural patterns can be heterogeneous. ' ' 4 In general, low grade nuclear DCIS forms predominantly cribriform patterns, whereas DCIS with high-grade nuclei forms centrally necrotic or solid patterns. Centrally necrotic architectural pattern was not seen in cases with DCIS nuclear grade, All architectural patterns were seen in the cases with higher DCIS nuclear grades. Four cases of pure nuclear grade, predominantly centrally necrotic DCIS, were associated with grade invasive carcinomas, whereas of 5 cases of pure nuclear grade, predominantly centrally necrotic DCIS were associated with grade invasive carcinomas. Although this number of cases is too small to derive any significant conclusions, it supports the hypothesis that nuclear grade, rather than DCIS architecture, is the most important predictor of the grade of an associated invasive carcinoma. In summary, we found that there is a great variation of nuclear grade within a large group of women with DCIS; however, there is minimal DCIS nuclear grade variation within each individual patient's lesion. Regardless of the nuclear uniformity within a DCIS lesion, there could be marked architectural variation of the DCIS. In general, the grade of the DCIS is associated with the grade of the invasive carcinoma. Acknowledgments. The authors thank Drs. J. Watts. J. Neill, and R. Goldstein for their editorial comments. REFERENCES. Moriya T, Silverberg SG. Intraductal carcinoma (ductal carcinoma in situ) of the breast: A comparison of pure noninvasive tumors with those including different proportions of infiltrating carcinoma. Cancer 4;74: Lampejo OT, Barnes DM, Smith P, Millis RR. Evaluation of infiltrating ductal carcinomas with a DCIS component: Correlation of the histologic type of the in situ component with grade of the infiltrating component. Sent Diag Pathol 4; :5-.. Enjoji M. Toyoshima S. Ductal carcinoma of the breast: An analysis of proportions of intraductal and invasive carcinoma. Path ResPract ;87: Lagios MD, Margilin FR, Westdahl PR, Rose MR. Mammographically detected duct carcinoma in situ: Frequency of local recurrence following tylectomy and prognostic effect of nuclear grade in local recurrence. Cancer 8;6: Patchefsky AS, Schwartz GF, Finkelstein SD, et al. Heterogeneity of intraductal carcinoma of the breast. Cancer 8:6: Lennington WJ. Jensen RA, Dalton LW, Page DL. Ductal carcinoma in situ of the breast: Heterogeneity of individual lesions. Cancer 4:7: Fisher ER, Costantino J, Fisher B, et al. Pathologic findings from the National Surgical Adjuvant Breast Project (NSABP) Protocol B-7: Intraductal carcinoma (ductal carcinoma in situ). Cancer 5;75:-. 8. Page DL. Lagios MD. Pathologic analysis of the National Surgical Adjuvant Breast Project (NSABP) B-7 trial: Unanswered questions remaining unanswered considering current concepts of ductal carcinoma in situ. Cancer 5;75:-.. Fisher ER, Costantino J, Fisher B, Palekar AS, Mamounas E. Response-Blunting the counterpoint. Cancer 5;75:-7. Downloaded from by guest on November 8 Vol. 6-No.

7 8 ANATOMIC PATHOLOGY Original Article. Lagios MD. Ductal carcinoma in situ: Controversies in diagnosis, biology, and treatment. The Breast Journal 5; : Page DL, Anderson TJ. Diagnostic Histopathology of the Breast. Edinburgh: Churchill Livingstone, 87.. Elston CW, Ellis IO. Pathological prognostic factors in breast cancer. I. The value of histological grade in breast cancer: Experience from a large study with long-term follow-up. Histopathology ;:4-4.. Fechner RE. One century of mammary carcinoma in situ: What have we learned? Am J Clin Pathol ; : Solin U, Yeh IT, Kurtz J, et al. Ductal carcinoma in situ (intraductal carcinoma) of the breast treated with breast-conserving surgery and definitive irradiation: Correlation of pathologic parameters with outcome of treatment. Cancer ; 7: Poller DN, Silverstein MJ, Galea M, et al. Ductal carcinoma in situ of the breast: A proposal for a new simplified histological classification association between cellular proliferation and c- erbb- protein expression. Mod Pathol 4;7: Ottesen GL, Graversen HP, Blichert-Toft M, Zedeler K, Andersen JA. Ductal carcinoma in situ of the female breast: Short-term results of a prospective nationwide study. Am J Surg Pathol ;6: Bellamy COC, McDonald C, Salter DM, Chetty U, Anderson TJ. Noninvasive ductal carcinoma of the breast: The relevance of histologic categorization. Hum Pathol ;4: Page DL, Dupont WD, Rogers LW, Jensen RA, Schuyler PA. Continued local recurrence of carcinoma 5-5 years after a diagnosis of low grade ductal carcinoma in situ of the breast treated by biopsy only. Cancer 5;76:7-.. Lagios MD, Westdahl PR, Margolin FR, Rose MR. Duct carcinoma in situ: Relationship of extent of noninvasive disease to the frequency of occult invasion, multicentricity, lymph node metastases, and short-term treatment failures. Cancer 8; 5: -4.. Boyages J, Recht A, Connolly J, et al. Early breast cancer: Predictors of breast recurrence for patients treated with conservative surgery and radiation therapy. Radiother Oncol ; :-4.. Jacquemier J, Kurtz J, Amalric R, et al. An assessment of extensive intraductal component as a risk factor for local recurrence after breast-conserving therapy. Br J Cancer ;6: Fourquet A, Campana F, Zafrani B, et al. Prognostic factors of breast recurrence in the conservative management of early breast cancer: A 5-year follow-up. Int J Radial Oncol Biol Phys 8;7: Borger J, Kemperman H, Hart A, et al. Risk factors in breast-conservation therapy. J Clin Oncol 4; : Schnitt SJ. Connolly J, Khettry U, et al. Pathologic findings on reexcision of the primary site in breast cancer patients considered for treatment by primary radiation therapy. Cancer 87; 5: Holland R, Connolly J, Gelman R, et al. The presence of an extensive intraductal component following a limited excision correlates with prominent residual disease in the remainder of the breast. J Clin Oncol ; 8: Leal CB, Schmitt FC, Bento MJ, Maia NC, Lopes CS. Ductal carcinoma in situ of the breast: Histologic categorization and its relationship to ploidy and immunohistochemical expression of hormone receptors, p5, and c-erb-b- protein. Cancer 5;75:-. 7. Wilber DC, Barrows GH. Estrogen and progesterone receptor and c-erbb- oncoprotein analysis in pure in situ breast carcinoma: An immunohistochemical study. Mod Pathol ;6: Bur ME, Zimarowski MJ, Schnitt SJ, Baker S, Lew R. Estrogen receptor immunohistochemistry in carcinoma in situ of the breast. Cancer ; 6: Lodato RF, Maguire HC, Greene MI, Weiner DB, LiVolsi VA. Immunohistochemical evaluation ofc-erbb- oncogene expression in ductal carcinoma in situ and atypical ductal hyperplasia of the breast. Mod Pathol ;: Schimmelpenning H, Eriksson ET, Pallis L, et al. Immunohistochemical c-erbb- proto-oncogene expression and nuclear DNA content in human mammary carcinoma in situ. Am J Clin Pathol ;7(suppl ):S48-S5.. Killeen JL, Namiki H. DNA analysis of ductal carcinoma in situ of the breast. Cancer ;68: Ramachandra S, Machin L, Ashley S, Monaghan P, Gusterson BA. Immunohistochemical distribution ofc-erbb- in in situ breast carcinoma: A detailed morphological analysis. J Pathol ;6:7-4.. Zifrani B, Leroyer A, Fourquet A, et al. Mammographically-detected ductal in situ carcinoma of the breast analyzed with a new classification. A study of 7 cases: Correlation with estrogen and progesterone receptors, p5, and c-erbb- proteins, and proliferative activity. Sem Diag Pathol 4; : De Potter CR, Schelfhout A, Verbeeck P, et al. Neu overexpression correlates with extent of disease in large cell ductal carcinoma in situ of the breast. Hum Pathol 5;6: Fisher ER, Siderits R. Value of cytometric analysis for distinction of intraductal carcinoma of the breast. Breast Cancer Res Treat ;: Eusebi V, Feudale E, Foschini MP, et al. Long-term follow-up of in situ carcinoma of the breast. Sem Diag Pathol 4; : Silverstein MJ, Poller DN, Waisman JR, et al. Prognostic classification of breast ductal carcinoma in situ. Lancet 5:45: Sneige N, McNeese MD, Atkinson EN, et al. Ductal carcinoma in situ treated with lumpectomy and irradiation: Histopathological analysis of 4 specimens with emphasis on risk factors and long-term results. Hum Pathol 5;6: Rosen PP, Oberman HA. Intraepithelial (preinvasive or in situ) carcinoma. In: Atlas of Tumor Pathology, ed. Washington, DC: Armed Forces Institute of Pathology Press,, pp Holland R, Peterse J, Millis RR, et al. Ductal carcinoma in situ: A proposal for a new classification. Sem Diag Pathol 4; : Downloaded from by guest on November 8 A.J.C.P. September 6