Fax: (011) 46 23 492 389.
Clinical relevance of the distribution of the lesions in 500 consecutive breast cancer cases documented in large-format histologic sections
Article first published online: 11 OCT 2007
Copyright © 2007 American Cancer Society
Volume 110, Issue 11, pages 2551–2560, 1 December 2007
How to Cite
Tot, T. (2007), Clinical relevance of the distribution of the lesions in 500 consecutive breast cancer cases documented in large-format histologic sections. Cancer, 110: 2551–2560. doi: 10.1002/cncr.23052
- Issue published online: 19 NOV 2007
- Article first published online: 11 OCT 2007
- Manuscript Accepted: 19 JUN 2007
- Manuscript Revised: 6 JUN 2007
- Manuscript Received: 26 MAR 2007
- breast cancer;
- diffuse growth;
- lymph node metastasis;
- large section histology;
- sick lobe theory
Breast carcinoma is a heterogeneous group of diseases deviating from each other not only in their clinical manifestations and outcome but also in their histologic appearance. The submacroscopic morphology of breast carcinomas, the distribution of the lesions, and the extent of the disease are seldom studied. Even more infrequently are these parameters included in surgical pathology reports. Conversely, the routine use of large-format histologic sections in workup of operated breast specimens provides better insight into the significance of these parameters. The aim of the study was to identify breast carcinoma growth patterns indicating increased metastatic potential of the tumor and a need for more aggressive therapy.
In all, 500 consecutive breast cancer cases, all of which were documented on large-format histologic sections, were retrospectively analyzed. The distribution of both in situ and invasive components of the tumors (unifocal/multifocal/diffuse) was defined, determined, and compared with the type of surgical intervention performed and the frequency of ipsilateral lymph node metastasis as endpoints. The extent of the disease, the size of the tumor, the presence or absence of lymphovascular invasion (LVI), and the proportion of invasive lobular carcinomas in the categories with different distributions were also analyzed.
Only 34% of the analyzed cases could be categorized as unifocal. This kind of tumor distribution was associated with lymph node metastasis in 28% of the cases, with LVI in 18%, with breast-conserving surgery in 67%, and with a proportion of 4% invasive lobular carcinomas. Tumors with a unifocal invasive component upgraded to multifocal or diffuse because of the distribution of the associated in situ component had similar characteristics. With their larger extent, tumors with a diffuse in situ component required mastectomy in 43% of cases. Multifocal distribution of the invasive component in the tumors was associated with higher frequency of LVI (42%) and lymph node metastases (48%), with a substantially lower number of cases undergoing breast-conserving surgery (33%) and with a higher proportion of lobular carcinomas (25%). If the multifocal invasive foci were associated with a diffuse in situ component, the proportion of invasive lobular carcinomas was only 5%. The extent of the lesions (defined as the area of breast tissue involved by in situ, invasive, and/or intravascular tumor foci) was ≥2 cm in >90% of multifocal cases and ≥4 cm in >70%. Diffusely growing invasive carcinomas were rare (only 20 cases), but were associated with lymph node metastasis in 60% of cases and resulted in mastectomy in 85% of the cases. Approximately two-thirds (65%) of these tumors belonged to invasive lobular carcinomas. The extent of diffusely growing invasive carcinomas was ≥4 cm in 75% of the cases. Although LVI was detected in only 10% of tumors with a diffusely growing invasive component, such tumors were found to have lymph node metastasis significantly more often (odds ratio of 2.33) and required mastectomy much more frequently (odds ratio of 2.58) compared with purely unifocal breast carcinomas.
These results indicate that the distribution of invasive and in situ tumor structures in breast carcinomas as defined in the current study, together with the extent of disease, are important morphologic parameters which determine the required surgical intervention and are related to biologic factors such as metastatic capacity. The method of large-section histology allows the examiner to properly document and demonstrate these important parameters, thus facilitating understanding of their clinical relevance. Cancer 2007. © 2007 American Cancer Society.
Classic whole-organ studies long ago demonstrated that the majority of breast carcinomas are multifocal and extensive.1–3 However, the most widely used methods in currently routine surgical pathology are insufficient to repeat these results in many cases. Despite the well-recognized importance of the distribution of cancer within the breast for optimizing breast-conserving surgery,4 there are conflicting results in the related literature regarding the significance of breast cancer multifocality, depending on the methodologic limitations and the various definitions used lacking a wide international consensus.1–3, 5–14 The diffuse growth pattern of some highly aggressive breast carcinomas10 remains mainly unrecognized.
Conversely, the importance of the extent of the disease, corresponding to the entire area in the breast involved with malignant structures (ie, the tissue the surgeon should excise during a radical intervention), is well documented in the literature. Extensive lesions recur significantly more often that those of limited extent, both with regard to invasive11 and in situ tumors.12, 13
The routine use of large-format histologic sections in diagnostic breast pathology allows the observer to properly document and demonstrate the extent and distribution of the lesions in a case of malignant tumors and to compare these parameters with the surgical margins.15–17 Detailed radiologic/pathologic correlation in the workup of the specimen further enhances the sensitivity of this method.18, 19 These preparations are currently the best available source for analyzing the histologic distribution of the lesions and the extent of the disease in breast cancer cases.
The objective of the current study was to analyze a sufficiently large series of consecutive breast cancer cases with regard to the distribution of the lesions and the extent of the disease as documented in large histologic sections. The results were correlated with the histologic type of the invasive tumor, the type of surgical intervention performed, and to the status of the ipsilateral axillary lymph nodes, all with the aim of delineating the growth patterns of this heterogeneous disease indicating those with an unfavorable biologic potential and requiring more aggressive therapy.
MATERIALS AND METHODS
The study represents a retrospective analysis of 500 consecutive breast carcinoma cases diagnosed at the Department of Pathology and Clinical Cytology of the Central Hospital in Falun, Sweden, during the period between January 1, 2005 and March 15, 2007. Twelve of the patients were operated on for synchronous bilateral carcinomas; thus, the 500 carcinomas belonged to 494 patients, 3 of whom were men. Sixty-two of the patients had a history of previous breast carcinoma; in 46 cases, the cancer was contralateral.
All the cases underwent workup using the method of large-section histopathology, which has been a routine procedure in our laboratory since 1982. The method has been described in detail elsewhere.20, 21 Briefly, all the cases are discussed by a preoperative tumor board when the radiologic appearance is registered, including the radiologic extent and distribution. This information, together with the whole-specimen radiograph received with the surgical specimen, assists the pathologist during the workup. The sector-resection specimens are sliced into 3–4-mm thick tissue slices parallel with the pectoralis fascia. The slices are also radiographed. One to 5 of the most representative slices (measuring up to 9 cm × 8 cm) are selected for embedding into large paraffin blocks. Larger slices are bisected and embedded into separate blocks. Mastectomy specimens are sliced perpendicular to the pectoralis fascia to observe the surgical margin in a histologic plane.
All the archived large histologic sections belonging to this series (1 to 34 per case, most often 6 per case) were reviewed by the author. The type of the invasive component, the size of the tumor, the extent of the disease, and the distribution of the lesions were determined according to the diagnostic criteria cited below and registered, as well as the presence or absence of lymphovascular invasion (LVI). ‘Multicentricity,’ defined as the presence of malignant structures in different quadrants of the same breast, was not analyzed because it represents a clinical and/or radiologic parameter. The status of the axillary lymph nodes and the performed surgical intervention was taken from the files of our department. Mastectomy was 1 of the endpoints of the study, either performed as primary surgery or as a completing intervention within 6 months. The other endpoint was the presence of lymph node metastasis at the primary surgery or at the completion of the surgical intervention; only macrometastases were counted. The statistical analysis (chi-square test, odds ratio) was performed using commercially available software (MedCalc Statistics for Biomedical Research; MedCalc Software, Belgium) with a P value <.01 regarded as significant.
The invasive carcinomas were typed according to the recommendations of the World Health Organization classification of breast tumors.22 The lymph node metastases were classified as macrometastasis (measuring >2 mm in largest dimension), micrometastases (measuring >0.2 mm but not >2 mm), or isolated cancer cells/cell groups (measuring ≤0.2 mm), according to the 2002 TNM classification.23 Vascular invasion (defined as the presence of tumor emboli in endothel-lined spaces) was assessed in routinely stained large-format histology sections, in the peritumoral or intertumoral area. The size of the lesion in breast carcinoma was defined as the largest dimension of the largest invasive focus, measured in a large histologic section. The histologic size was compared with the radiologic measurements and with the macroscopically registered size; in the case of a discrepancy, the definite tumor size was determined using all this information on a postoperative tumor board.
The extent of the disease was defined as the area of the breast tissue involved by malignant structures including all the invasive, in situ, and intravascular tumor structures observed in large histologic sections. The extent was measured as an area in 2 dimensions.
The distribution of the invasive component and of the in situ component of the same lesion was determined separately. Invasive lesions were considered to be ‘unifocal’ if only 1 invasive focus could be observed in the large sections, the tumor focus containing or not containing an in situ component. ‘Multifocal’ invasive lesions were characterized by the presence of multiple well-delineated invasive tumor foci separated from each other by uninvolved breast tissue, regardless of the distance between the foci. Those tumors observed in the large sections to be dispersed over a large area much like a spider's web with no distinct tumor mass being evident were classified as ‘diffuse.’ In situ carcinomas were regarded as ‘unifocal’ if they appeared to involve a single terminal ductal lobular unit (TDLU) or several neighboring TDLUs without uninvolved breast tissue between; they were regarded as ‘multifocal’ if they involved several distant TDLUs with uninvolved breast tissue in between, and as ‘diffuse’ if they involved mainly the larger ducts. In breast cancer cases containing both in situ and invasive tumor structures, the distribution of these components was combined with each other. Diffuse distribution of either an in situ or invasive component qualified the lesion to be ‘diffuse.’ Tumors without evidence of diffuse growth were classified as ‘multifocal’ if either the in situ or the invasive component or both were multifocal. When assessing the distribution of the lesion, an attempt was made to summarize the findings in different large-section levels in every individual case trying to reconstruct the in vivo situation before surgery; detailed radiologic/pathologic correlation was also helpful. Typical cases of unifocal, multifocal, and diffuse tumors are illustrated in Figures 1–6.
The current series of 500 consecutive breast carcinoma cases contained 52 purely in situ carcinomas (10.4%) and 69 invasive carcinomas without a demonstrable in situ component (13.8%), but the vast majority of the tumors (379 cases; 75.8%) contained both in situ and invasive components. The distribution of the lesions, the size and type of the invasive component, and the extent of the disease could all be properly assessed in all but 12 cases (which either contained only a part of the lesion or the distribution was too complex to categorize). Accordingly, the distribution of the lesions, as demonstrated in Figure 7, was the following: 34% (170 of 500 cases) of the cases were unifocal, 36% (180 of 500 cases) were multifocal, 28% (138 of 500 cases) were diffuse, and the remaining 12 cases (2%) were considered to be unclassified.
Figure 8 demonstrates a more detailed classification of the cases according to the distribution of the lesions as well as clear differences in the incidence of the different categories. There were 16 possible combinations. The unifocal tumors were divided into 3 subgroups: those with a unifocal invasive component without demonstrable in situ component (52 cases), those with an in situ component exclusively within the area of the invasive structures (103 cases), and the purely in situ unifocal tumors (15 cases).
The multifocal tumors most often contained either a unifocal invasive focus associated with a multifocal in situ component (62 cases) or both multiple invasive and multiple in situ structures at the same time (84 cases). Lesions with a unifocal invasive focus associated with a distant unifocal in situ focus (3 cases) also belonged to the category of multifocal carcinomas as well as lesions containing only invasive but multiple foci (9 cases), lesions with multiple invasive foci, and a single in situ focus (8 cases) and the purely multifocal in situ carcinomas (14 cases).
The majority of the diffuse lesions were classified as ‘diffuse’ because of their in situ component involving the larger ducts and were associated with unifocal (40 cases) or multifocal (55 cases) invasive components. The 23 diffuse purely in situ lesions also belonged to the same category. Lesions classified as ‘diffuse’ because of their invasive component were rare (only 20 cases). The diffusely growing invasive breast carcinomas were associated with a diffuse in situ component in only 1 case; rather, they were either purely invasive (8 cases) or associated with multifocal in situ foci (10 cases). There was also a single case of a diffuse invasive carcinoma associated with a unifocal in situ component.
The distribution of the 448 invasive carcinomas in the current series by size, extent, proportion of performed mastectomy, LVI, lymph node status, and proportion of invasive carcinomas with lobular morphology is demonstrated in Table 1. Categories were separated into clearly unifocal cases, unifocal invasive carcinomas upgraded to multifocal or diffuse according to the in situ component, cases that were multifocal or diffuse because of the growth pattern of the invasive component, and unclassified cases.
|Tumor size, mm*||Extent ≥2 cm||Extent ≥4 cm||Mastectomy†||LVI||Lymph node metastasis||Invasive lobular component||No.|
|U‡||16.3 (94,941)||41% (65/158)||1% (2/158)||33% (49+3/158)||18% (28/158)||28% (44/158)||4% (6/158)||158|
|U invasive MF in situ||14.0 (73,696)||87% (54/62)||52% (32/62)||35% (15+7/62)||19% (12/62)||11% (7/62)||10% (6/62)||62|
|U invasive D in situ||14.0 (95,018)||95% (38/40)||68% (27/40)||43% (8+9/40)||28% (11/40)||30% (12/40)||3% (1/40)||40|
|MF invasive||20.2 (113,253)||92% (93/101)||74% (75/101)||66% (51+16/101)||42% (42/101)||48% (48/101)||25% (25/101)||101|
|MF invasive D in situ||16.5 (95,371)||100% (55/55)||84% (46/55)||76% (29+13/55)||53% (29/55)||42% (23/55)||5% (3/55)||55|
|D invasive||NA||100% (20/20)||75% (15/20)||85% (14+3/20)||10% (2/20)||60% (12/20)||65% (13/20)||20|
|Unclassified||NA||NA||NA||17% (2/12)||NA||33% (4/12)||8% (1/12)||12|
|All invasive tumors||16.7 (98,998)||75% (325/436)||45% (197/436)||49% (219/448)||28% (124/448)||33% (150/448)||12% (55/448)||448|
The category of unifocal invasive tumors (without an in situ component or with an in situ component within but not outside the invasive focus) was characterized by a 33% mastectomy rate (with the remainder of the tumors being treated with breast-conserving surgery), 18% rate of LVI, 28% rate of positive lymph nodes (only macrometastases included), and a proportion of only 4% lobular carcinomas. The results were similar in the categories when unifocal invasive carcinomas were associated with a multifocal or diffuse in situ component, although the proportion of cases treated with mastectomy was somewhat higher. Carcinomas with a multifocal invasive component with or without an in situ component had a significantly higher proportion of cases resulting in mastectomy (66% and 76%, respectively) and a substantially increased proportion of cases with LVI (42% and 53%, respectively) and cases with positive lymph nodes (48% and 42%, respectively) compared with unifocal tumors. The cases with a diffuse growth pattern of the invasive component required mastectomy in as many as 85% of cases and were associated with lymph node metastasis in 60%. However, LVI was only found in 10% (2 of 20 cases) of the cases in this subgroup.
Statistical analysis did not reveal any significant difference in tumor size between the categories of invasive carcinomas, as demonstrated in Table 1. The average tumor size (arithmetic mean) for the 448 invasive carcinomas in this series was 16.7 mm (range, 1–80 mm; P = .8266 by the chi-square test); the median tumor size was 15 mm (95% confidence interval, 14–16 mm; P = .7816). However, there were significant differences in the extent of the disease both in categories with an extent ≥2 cm and those with an extent ≥4 cm. In unifocal carcinomas the extent was most often equal to the size of the lesion, although the peritumoral appearance of the in situ component or peritumoral vascular invasion could increase the extent of these lesions. These tumors were better characterized by their size in this study than by their extent. Thus, the extent of these lesions (which was ≥4 cm in only 2 cases [1%] and ≥2 cm in 65 cases [41%]) was clearly deviating from the extent in multifocal and diffuse cases. Conversely, no statistically significant differences were found when the extent of the categories of multifocal and diffuse tumors was tested (P = .8622 for an extent ≥2.0 cm, P = .0924 for an extent ≥4.0 cm). In diffuse cases, which usually lack a well-delineated tumor body, the size of the lesions was found to correspond to the extent of the disease in multifocal cases rather than to the size of unifocal tumors. Their size was hardly measurable and was usually equal or similar to their extent.
The statistical analysis revealed highly significant differences in the proportion of mastectomies performed as well as in the proportion of cases with LVI and lymph node-positive cases between the categories with different tumor distribution, as demonstrated in Table 1. Tumors with a diffusely growing invasive component were found to have lymph node metastasis significantly more often (odds ratio of 2.33) and required mastectomy much more frequently (odds ratio of 2.58) compared with purely unifocal breast carcinomas.
In the current study, LVI was found in 28% (124 of 448 cases) of invasive carcinomas, in 16% (45 of 286 cases) of cases without any lymph node involvement, and in 53% (79 of 150 cases) of cases with lymph node metastases. Tumors with a unifocal invasive component were found to have LVI in 20% of cases (51 of 260 cases), tumors with a multifocal invasive component were found to have LVI in 46% of cases (71 of 156 cases), and tumors with a diffuse invasive component were found to have LVI in 10% of cases (2 of 20 cases). LVI was present in 41% (11 of 27 cases) of cases in multifocal tumors without an in situ component, in 42% (31 of 74 cases) of cases in multifocal invasive tumors with a multifocal in situ component, and in 53% (29 of 55 cases) of cases in multifocal invasive cancers with a diffuse in situ component.
Invasive lobular carcinomas corresponded to 12% (55 of 448 cases) of the cases in the current series. The analysis demonstrated that this type of invasive carcinoma is much more frequent among the multifocal tumors when associated with a multifocal in situ component, but not in cases in which multifocal invasive foci are associated with a diffuse in situ component. As many as 65% (13 of 20 cases) of the diffusely growing invasive carcinomas were invasive lobular carcinomas, with the remainder being ductal invasive carcinomas (6 cases) and inflammatory breast carcinoma (1 case).
The proportion of ‘multifocal’ invasive cases varies in different series of breast carcinomas depending on the histologic method used, applied diagnostic criteria, and case selection. Values as low as 11%7 and as high as 75%1 have been reported. Using, for example, a criterion of at least ≥2 cm distance between the separated individual foci to term the breast cancer ‘multifocal’ results in including fewer distant multiple lesions into the ‘unifocal’ group, covering some multifocal cases with larger ‘size’ of the tumor. Such studies are usually based on a traditional morphologic workup focused on macroscopically observed dominant masses and are characterized by a low proportion of multifocal cases and a lack of significant differences between ‘unifocal’ and ‘multifocal’ cancers.8, 9 With such an approach, the lesion in Figure 2 would have been classified as ‘unifocal.’ Defining multifocality using distance between the individual foci appears to impede the ability to obtain comparable and significant results, first because the biologic importance of a distance is hard to test, but also because it is impossible to achieve consensus regarding the required amount of millimeters. In the current study, we tried to escape these methodologic limitations by studying a nonselected population (consecutive cases), by using the large-section technique, and by applying qualitative rather than quantitative diagnostic criteria. Accordingly, our findings are not directly comparable to the results of the above-cited and some other published series of breast carcinomas because their findings match the characteristics of the category of unifocal tumors in the current series rather than the results of the entire study. This statement is also valid with regard to the proportion of lymph node-positive cases and the percentage of cases treated with breast-conserving surgery.24
Conversely, our results are concordant with the findings in studies using a large-section technique or radiology-guided whole-organ examination. Using a technique that was very similar to that in the current study, Gallager and Martin1 reported “extensive neoplastic transformation of the ductal epithelium and multiplicity of the invasive sites” in approximately three-quarters of the 38 mastectomy specimens studied. Holland et al.3 also reported that 63% of the breast carcinomas in Tis and T1-2 stadium had tumor foci around the “referent mass”; 20% within a 2-cm distance and 43% even more distantly (with 27% being in situ foci). Luttges et al.25 worked up 166 cases with whole-organ sectioning and found that 76 cases (46%) were multifocan and, in addition, 36 cases (22%) were multicentric. Although these studies did not delineate the diffuse growth pattern (in the study of Holland et al., it was an exclusion criterion), they verified a low proportion of purely unifocal cases (25%, 37%, and 32%, respectively), fully comparable to the rate of 34% noted in the current study.
The results of the current study proved that the extent of the disease in breast carcinomas, defined as the area of the breast tissue involved by in situ, invasive, and/or intravascular malignant structures, influences the success of breast-conserving surgery in invasive cases. Faverly at al11 demonstrated that breast carcinomas of limited extent (defined as tumors with a distance between the invasive foci of <1 cm) are proper candidates for breast-conservingsurgery. These cases corresponded to approximately half of their study material. Using the largest dimension of the involved area for determining the extent of the disease and not the distance between the individual foci makes the results of the current study hardly comparable to those of Faverly et al.11, but the current study does demonstrate the same thing: more extensive lesions require mastectomy more often. Defining the extent by the dimensions of the involved area also proved to be a sensitive approach in our previous study regarding in situ breast carcinomas, showing a nearly 10-fold difference in the local recurrence rate between extensive (area > 2 cm) and nonextensive ductal in situ breast carcinomas.13
Our hypothesis, the theory of the sick lobe,13 defines breast carcinoma as a lobar disease originating in a single breast lobe in the majority of cases. In situ carcinoma may appear in the ducts and the lobules of the sick lobe simultaneously in large areas or with time differences involving small parts of it. The invasive component of the tumor may, naturally, infiltrate the tissue beyond the borders of the sick lobe, especially in advanced cases, but the multiple tumor foci are also lobar in their origin. Sector resection, the method of choice of breast-conserving surgery in Scandinavia, represents an attempt to excise the entire sick lobe together with the malignant structures. The complex lobar anatomy of the breast, with considerable variations of size and shape of the lobes and with their complex interrelation,26 together with the lack of methods of intravital visualization of the sick lobe, make the achievement of this objective very difficult.
The biologic significance of tumor multifocality remains unclear. In 1982, Egan2 reported that women with a single-site carcinoma had substantially better survival than those with multiple sites (2.5% vs 15% mortality per year). More recent studies could not confirm these results. For example, Pedersen et al.8, although finding a positive correlation between multifocality and the number of positive lymph nodes (similar to our findings), could not verify a significant effect of multifocality on overall survival on multivariate statistical analysis. In contrast to their findings, some recent studies could not show any association of tumor multifocality and lymph node status.9 These studies also pointed to the importance of tumor size in multifocal cancers, which appears to be an independent prognostic factor. The results of the current study also demonstrated some differences in tumor size between unifocal and multifocal lesions, but was focused on testing the importance of tumor distribution and extent.
Breast carcinomas with multiple invasive foci appear to have a different biologic potential from the unifocal tumors, possibly as an expression of increased tumor volume. Ignoring the multiplicity of the invasive foci may underestimate the real tumor burden; conversely, measuring together multiple foci close to each other may lead to a false judgment of the tumor as a cohesive mass. The criteria for multifocality regarding the invasive component of the tumors as defined and applied in the current study appear to have biologic relevance because the incidence of lymph node macrometastasis were significantly different between unifocal (28%), multifocal (48%), and diffuse (60%) invasive carcinomas. These findings indicate a greater propensity of multifocal tumors to disseminate compared with unifocal ones and are in concordance with the results of Coombs and Boyages,7 who also reported a statistically significant difference in the proportion of lymph node-positive cases between unifocal and multifocal breast carcinomas (37.5% and 52.1%, respectively) as well as with the results of some other related studies.5, 6
LVI was clearly found to be associated with lymph node metastasis in the current study. These findings as well as the proportion of cases with LVI were comparable to the results in many other publications.27–29 We demonstrated that LVI is also associated with multifocality of the invasive foci. The presence or absence of a multifocal in situ component in the cases with multiple invasive tumors appeared to have no influence on the proportion of cases with LVI (41% vs 42%). However, diffuse growth of an in situ component appears to have this capacity. Multifocal invasive tumors have a larger surface area than unifocal tumors with similar tumor volume; thus, the increased tumor/vascular interface may be the explanation for the more frequent appearance of LVI and lymph node metastasis in multifocal cases. Vascular invasion may also cause ‘satellite’ tumor foci if the tumor cells invade around the involved vascular space. Thus, LVI appears to be both the cause and consequence of tumor multifocality. It is interesting to note that the diffuse invasive tumors, although associated with lymph node metastasis in 60% of cases, had LVI in only 10% of the cases. In diffuse invasive tumors, especially lobular carcinomas, tumor cells may already be present in the so-called ‘prelymphatic’ spaces,30 which are not endothelized, but connected to lymph vessels, a phenomenon representing a possible explanation of the high proportion of lymph node-positive cases observed in this subgroup.
To our knowledge, there are no generally accepted criteria of multifocality regarding in situ breast carcinomas. Fisher et al.31 defined multifocality of in situ carcinoma as the presence of in situ tumor in ≥2 different tissue blocks from the same case (shown in 60.8% of their cases). Silverstein et al.32 considered in situ carcinomas to be multifocal when separate in situ foci were found > 2 cm from the primary site (found in 41% of cases). Faverly et al.33 demonstrated discontinuity of in situ tumor cell proliferation in 3-dimensional thick-section images assessed with a stereoscopic microscope. They defined multifocal in situ carcinoma as those with discontinuity of <4 cm (if more, the tumor was categorized as multicentric). They also showed that multifocality is related to the grade of the in situ tumor and that high-grade in situ ductal carcinomas have continuous growth most often. The classification used in the current study is similar to that of Andersen et al.,34 who divided cases of ductal carcinoma in situ into microfocal, tumor-forming, diffuse, and mixed. Trying to characterize the development of in situ tumor as a process disturbing normal lobularization and arborization in the breast, we rather connect the distribution of the lesions in in situ carcinomas to the anatomic structures of ducts and lobule (TDLUs) than to simply categorize histologic images. If disturbing the lobularization in situ carcinoma is localized to the TDLUs, either next to or distant from each other, this may give rise to unifocal or multifocal in situ lesions. When involving the larger ducts and not (only) the TDLUs, in situ carcinoma forms a diffuse network of tubes, which are hard to delineate and measure.13, 14 This type of tumoral growth frequently characterizes high-grade ductal cancer in situ, which has been identified as a negative morphologic prognostic parameter with survival as endpoint.35–37
The current study appears to be unique in its detailed analysis of the growth pattern and demonstration of the heterogeneity of tumoral growth in breast carcinomas, with 16 possible different combinations of the distribution of in situ and invasive tumor components. The current study results clearly indicate the need to recognize diffuse tumoral growth of both the in situ and invasive components of the tumor, a category that is missing in the related literature. The majority of the diffuse cases in the current series were categorized as such because of their in situ component. The diffuse growth of the invasive component appears to be even more relevant. We have previously described “invasive lobular carcinoma of diffuse type” and reported the unfavorable prognosis in this breast cancer subtype.10 More infrequently, invasive ductal carcinomas may also have the same growth pattern, as demonstrated in the current study. The diffuse growth pattern appears to indicate a lack of cohesion between the tumor structures and, together with the absence of a stromal reaction, is a hallmark of very aggressive invasive breast carcinomas, requiring mastectomy in nearly every case and giving rise to lymph node metastasis in greater than half of the cases. These results indicate that a category of diffuse tumors characterized by extent rather than by size of the lesions should be included in staging systems such as the TNM classification.
In conclusion, the results of the current study indicate that the distribution of in situ and invasive tumor structures in breast carcinomas and their extent are significant morphologic parameters determining the type of required surgery and are related to biologic factors and prognosis. There is a pressing need for an international consensus in defining tumor multifocality and recognizing the diffuse growth pattern as powerful negative prognostic parameters. The surgical intervention should provide a sufficiently large contiguous piece of tissue for proper assessment of the distribution of the lesion and extent of the disease in every breast cancer case. The method of large-section histology facilitates the proper histopathologic workup of the specimen, allowing the examiner to adequately document and demonstrate these important morphologic parameters.
- 14Radio-pathologic correlation of ductal carcinoma in situ of the breast using 2- and 3-dimensional large histologic sections. Semin Breast Dis. 2005; 8: 144–151., .
- 17Large section (macrosection) histologic slides. In: SilversteinMJ, editor. Ductal Carcinoma in Situ of the Breast.2nd ed. Philadelphia: Lippincott, Williams & Wilkins; 2002: 249–254., , .
- 19Enhancement of mammographic-pathologic correlation utilizing large format histology for malignant breast diseases. Semin Breast Dis. 2005; 8: 152–162., .
- 20Practical Breast Pathology. Stuttgart: Thieme; 2002: 116–123., , .
- 21Breast Cancer, The Art and Science of Early Detection with Mammography. Stuttgart: Thieme; 2005: 405–438., , .
- 22TavassoliFA,DevileeP, editors. World Health Organization Classification of Tumors. Pathology and Genetics of Tumors of the Breast and Female Genital Organs. Lyon, France: IARC Press; 2003.
- 23GreeneFL,PageDL,FlemingID, et al., editors. AJCC Cancer Staging Manual.6th ed. New York: Springer; 2002.
- 31Pathologic findings from the National Surgical Adjuvant Breast Project (NSABP) Protocol B-17. Intraductal carcinoma (ductal carcinoma in situ). The National Surgical Adjuvant Breast and Bowel Project Collaborating Investigators. Cancer. 1995; 75: 1310–1319., , , , , .
- 37Breast Cancer: Early Detection with Mammography. Casting Type Calcifications: Sign of a Subtype with Deceptive Features. Stuttgart: Thieme; 2007., , .