Rosen's Breast Pathology, 4e

381

Ductal Carcinoma In Situ

for consultation, Lennington et al. 115 recorded size from the “outside surgical pathology report” for “extensive lesions” but for “smaller, less extensive lesions, size was measured di- rectly from the glass slides.” Lesions 2 mm or less in diame- ter were excluded by definition, because they were classified as atypical hyperplasia. The authors stated that they “. . . rec- ognize the lack of precision, but believe the measurements are usually within 3–5 mm of true extent. There is probably a greater error in larger lesions.” This range of error in mea- suring the size of DCIS is a further impediment to using le- sion size as a criterion for distinguishing hyperplasia from DCIS. In the face of these substantial limitations, the authors reported the following distribution of “precise” mean sizes in relation to histologic subtype: all pure DCIS exclusive of micropapillary, 8.5 mm; mixed noncomedo histologic types, 13.1 mm; noncomedo with necrosis, 11.6 mm; comedo, 16.2 mm; and micropapillary, 19.1 mm. 115 The terms “extensive” intraductal component or “exten- sive” intraductal carcinoma should not be used as synonyms for widespread DCIS, as these terms have a different clinico- pathologic connotation. “Extensive” DCIS ought to be used only when DCIS is associated with invasive carcinoma, and the DCIS therein, constituting more than 25% of the tumor mass, extends beyond the dominant tumor mass into adja- cent breast parenchyma. In sum, sequential sectioning of surgical specimens (opti- mally conducted with radiologic correlation) is essential for the precise assessment of the extent of DCIS. The consistent practice of sampling in this manner will also optimize ob- servational research on issues relating to size and margins. MULTICENTRICITY AND MULTIFOCALITY There is no uniformly accepted definition of multicentric- ity in DCIS. The concept of multicentricity was advanced by Cheatle and Cutler, 10 Cheatle, 270 and Charteris 271 as a result of observations made on whole-organ sections of mastectomy specimens. Carcinoma was considered to be multicentric when there were foci that were separate from the clinically detected tumor. Lagios et al. 272 defined mul- ticentricity as “ the presence of separate independent foci of carcinoma within the breast—separate from the lesion which is clinically or mammographically evident, that is, the reference tumor.” Multicentric DCIS is generally defined as that present in multiple quadrants. 273–276 The use of this definition for the term precludes its use in routine lumpectomy specimens, unless the term is used for DCIS separated by 5 cm or some other arbitrary dimension. The rate of multicentricity of DCIS has traditionally been stated to be approximately 25%, with a range from 0% to 75% or so, such a wide range that reflects differences in the definition of the term. For practi- cal purposes, multicentricity refers to foci of carcinoma in distinctly different regions of the breast, usually in two or more quadrants. Multicentricity should be distinguished from in- traepithelial extension within ducts and lobules of a

increased with the nuclear grade of DCIS on mammogra- phy, and it increased as the mammographic breast density decreased. 264 Although preoperative imaging has limitations in determining the extent of DCIS, specimen slice radiogra- phy can improve these estimates. 265 It is sometimes possible to determine the size of a lesion microscopically if it is limited to a single group of contiguous ducts (unifocal), especially when the area is confined to the histologic section froma single paraffin block that contains tis- sue from an excisional biopsy. Reporting the size of the small lesions in this category is confounded by the fact that some authors exclude abnormalities smaller than 2 mm from the diagnosis of DCIS regardless of histologic appearance. 115,266 A significant proportion of DCIS are multifocal and not confined to a single coherent palpable lesion or to a micro- scopic focus that will fit in the confines of a single paraffin block. It is noteworthy that the dimensions of a DCIS limited to one standard paraffin block would not ordinarily exceed 2.0 to 2.5 cm or approximately 1 inch. Lagios et al. 267 studied mastectomy specimens from patients with DCIS by a serial subgross method and reported that the frequency of multi- centricity and occult invasion was substantially greater for lesions larger than 2.5 cm. It is on the basis of these studies that 2.5 cm came to be viewed as an important size criterion in the selection of patients for breast-conserving therapy. La- gios 268 expressed skepticism about the quantitation of DCIS, observing that “quantitation of DCIS will remain a problem since the association of the extent of DCIS, and the extent of microcalcifications, the only preoperative measure available at present, is quite variable.” Presently, there is no consensus on a method for deter- mining the extent of DCIS on the basis of the proportion of slides showing the lesion. This approach is highly depen- dent on the completeness of sampling and biopsy size, both of which determine the denominator. This issue would be partially addressed if the denominator were biopsy weight in grams with the numerator representing the number of slides showing DCIS. This calculation would be most useful in situations where all tissue has been processed for histologic examination. Until a standardized method has been vali- dated and widely adopted, the number of sections involved in sequentially taken samples (or at a minimum the propor- tion of slides with DCIS) will remain a crude measure of the extent of DCIS. Many patients have more than one diagnostic proce- dure performed (e.g., needle core biopsy, followed by exci- sion and reexcision of various margins) with DCIS in more than one specimen. It is not practical to reassemble the foci of DCIS from two or more specimens to obtain a single measurement. For the foregoing reasons, a 1997 consensus report on the classification of DCIS left the issue of size largely unan- swered and was unable to address this question. 151 A classifi- cation system for DCIS that includes lesion size is currently impractical. Nonetheless, data purportedly describing the “size” of DCIS are reported. The mean size of 227 lesions in one series was 2.1 cm, ranging from 1.5 cm in cribriform to 2.5 cm in comedo DCIS. 269 In a study of cases referred