382
Chapter 11
If the sample provided for histologic diagnosis is an ex-
cisional biopsy limited to the region of the index lesion, the
material is not suitable for determining whether the patient
has multicentric carcinoma. According to some definitions,
more than a quadrantectomy is necessary to detect multi-
centricity. In practice, the distinction between multifocal-
ity and multicentricity is difficult to make with certainty in
slides prepared by random sampling of breast specimens for
diagnostic purposes. Serial sequential sectioning of smaller
excisional biopsy specimens facilitates mapping of the extent
of disease, and this method should be utilized whenever fea-
sible. Advanced techniques such as stereoscopic dissection
and subgross microscopic–radiologic correlative studies are
more reliable methods for identifying true multicentricity
and multifocality, but they are too costly and time consum-
ing to be practical procedures for routine diagnostic work.
Given the limited resources available in most pathology lab-
oratories, it may be unrealistic to expect diagnostic reports
on breast biopsies to routinely distinguish multicentricity
from multifocality.
It is likely that multifocal and multicentric DCIS will be
increasingly detected with growing use of MRI. In a study
of 285 patients with newly diagnosed DCIS, MRI exami-
nation performed for the evaluation of extent of disease
showed separate foci of invasive carcinoma “elsewhere” in
the breast, that is, either multicentric or contralateral, in 16
(5.6%) patients.
279
A meta-analysis of 19 studies (
n
=
2,610)
demonstrated that MRI detected additional foci of carci-
noma (invasive and/or
in situ
) in 16% of invasive and/or
in
situ
carcinoma cases that had not identified by traditional
evaluation.
280
In summary,
multicentric
DCIS refers to its presence in
distinctly different regions of the breast, usually in two or
more quadrants, whereas
multifocal
DCIS is confined to one
quadrant and is generally regarded as that which is spatially
closer, with at least some normal breast tissue separating the
neoplastic foci. The adjectives, multicentric and multifocal
ought not to be used interchangeably owing to the divergent
clinical implications of each term.
281
MICROINVASION
Genomic and phenotypic similarities, among other clini-
cal and pathologic forms of evidence, support the conclu-
sion that DCIS is a nonobligate precursor of invasive ductal
carcinoma. Microinvasive carcinoma (abbreviated as T1
mic
in the TNM American Joint Committee on Cancer [AJCC]-
Union Internationale Contre le Cancer [UICC] staging
system) is the earliest manifestation and stage of invasive
mammary ductal carcinoma. The upper limit of the extent
of invasion for an invasive carcinoma to qualify for this des-
ignation is 0.1 cm (i.e., 1 mm).
Ultrastructural studies have detected foci of discontinuity
in the basement membranes of ducts with DCIS,
282,283
and
similar observations have been reported in tissues studied by
IHC.
284
Breaks in the basement membrane were more com-
mon when DCIS was of high nuclear grade with or without
single carcinomatous focus typically limited to one region
or quadrant. The latter condition is commonly referred
to as
multifocality
. One commonly employed criterion for
establishing the presence of multifocality depends on the
number of histologic sections that show DCIS. For example,
Fisher et al. stated that “ductal carcinoma
in situ
in only
one section of two or more obtained from different blocks
was considered to be unifocal. Its presence in sections from
two or more different blocks was considered multifocal.”
On the basis of this definition, 329 (60.8%) of 541 evalu-
able specimens of DCIS were classified as multifocal in data
from National Surgical Adjuvant Breast and Bowel Project
(NSABP) Protocol B17.
265
Silverstein et al.
269
considered
DCIS to be multifocal when “. . . separate foci of DCIS [duct
carcinoma
in situ
] more than 2 cm from primary site . . .”
were found in a mastectomy specimen. On the basis of this
criterion, multifocality was present in 41 (41%) of 98 breasts
examined after mastectomy. Multicentricity, defined as car-
cinoma outside the index quadrant, was present in 15% of
these breasts. Multifocality and multicentricity were not
significantly related to the histologic category of DCIS when
stratified as comedo and noncomedo type.
269
Hardman
et al.
273
found multicentric carcinoma in 27% of mastec-
tomy specimens from patients with carcinoma of the com-
edo type. Multicentricity was reported in 33%
18
and 37%
274
of mastectomies performed for diverse types of DCIS. A
proportion of “multifocal” DCIS, albeit a minor one, may
be a product of artifact, because of the inherent problem of
two-dimensional viewing of a three-dimensional arborizing
disease process.
In an effort to circumvent these technical issues, some
investigators have set anatomic limits on the distribution
of carcinoma to somewhat arbitrarily distinguish between
unicentric and multicentric diseases. For example, carci-
noma has been deemed to be multicentric if it is detected
in more than one quadrant or if it is 5 cm from the index
lesion.
93,274–276
Silverstein et al.
277
classified as multicentric if
two foci were separated by more than 2.0 cm.
Schwartz et al.
93,278
studied the frequency of multicentric-
ity in the breasts of patients with DCIS who underwent mas-
tectomy. Multicentricity was defined as “the presence [of]
invasive ductal or lobular carcinoma, microinvasive ductal
carcinoma, or DCIS in an area or quadrant outside the bi-
opsy site . . . . If the lesion was centrally located, the cancer
was considered multicentric only if additional foci of carci-
noma were found outside a perimeter of 5 cm from the edge
of the nipple and areola.”
93
Multicentricity was found in 18
(36%) of 50
93
and 4 (36%) of 11
278
breasts. Multicentricity
was more often present in lesions detected because of nipple
discharge (71%) or Paget disease (50%) than in those found
by mammography (38%).
93
When classified on the basis of
the predominant growth pattern, micropapillary DCIS was
more often multicentric (86%) than papillary (33%) or com-
edo (42%) DCIS. No multicentricity was encountered in five
cribriform and seven solid DCIS. Bellamy et al.
86
also found
multicentricity to be present significantly more often in pa-
tients with micropapillary DCIS than in those with other
patterns of DCIS.
