388
Chapter 11
FIG. 11.72.
DCIS with vascular tumor emboli.
A,B:
After a
needle core biopsy revealed DCIS, this patient underwent
excisional biopsy. The specimen contained cribriform DCIS
shown here with carcinoma cells in an adjacent vascular
channel (
arrow
).
C:
Isolated cytokeratin (AE1/3) positive
cells were present in subcapsular sinuses of the SLNs.
assume a growth pattern that simulates DCIS (Fig. 11.73).
This occurrence is most easily appreciated in metastatic de-
posits at sites outside the breast such as the ALNs and less
frequently in visceral metastases. The phenomenon was de-
scribed by Cowen
302
in 1980 and in a later report, Cowen
and Bates
303
reported finding metastatic carcinoma with a
DCIS-like appearance in lymph nodes from 35 of 391 pa-
tients (9%) with axillary metastases. In two of these cases, no
intraductal component was found in the primary tumor, but
in the others the “pseudointraductal” carcinoma in metasta-
ses resembled DCIS in the primary lesion.
Barsky et al.
304
reported finding DCIS-like metastases in
ALNs from 21% of 200 cases. These foci were termed “re-
vertant” DCIS to reflect the hypothesis that this phenom-
enon is a manifestation of a condition in which metastatic
potential is inhibited or reversed by local factors. The au-
thors observed complete concordance between primary
and “revertant” DCIS with respect to architectural pat-
tern, nuclear size determined by digital image analysis, and
the expression of the prognostic markers p53, HER2, and
Ki67. “Revertant” DCIS featured circumferential basement
membranes demonstrated by immunoreactivity for laminin
and type IV collagen, but lacked myoepithelial cells. The
capacity of invasive carcinoma to assume an appearance
that resembles its
in situ
counterpart could complicate the
diagnosis of microinvasive carcinoma. Cowen and Bates
303
concluded that “since invasive breast carcinoma may mimic
intraductal growth, some cases of breast cancer diagnosed
histologically as DCIS may, in reality, be invasive.” This phe-
nomenon may be responsible for the rare patient found to
have axillary nodal metastases, especially as a result of SLN
mapping, when the breast appears to be the site of DCIS
with no demonstrable invasion.
The difficulties raised by the structural similarities of
in
situ
and invasive duct carcinoma are complicated by the re-
sults of studies that have demonstrated the presence of base-
ment membrane components around groups of invasive
carcinoma cells (Fig. 11.74). Arihiro et al.
305
found immuno-
reactivity for laminin at sites of invasive carcinoma in 54%
of 71 carcinomas. The presence of laminin was associated
with a greater degree of tubule formation. These findings
correlate with data obtained by Nadji et al.
306
as significantly
related to low histologic and nuclear grade.
In light of the foregoing discussion, it is evident that there
are instances in which the presence or absence of micro-
invasion can be difficult to determine with certainty, even
with the immunohistochemical reagents currently avail-
able. Some guidelines can be suggested based on experience:
1. Thepresenceofmyoepithelial cells at theperimeter of neo-
plastic glands is the most convincing evidence of DCIS,
especially if demonstrated with the p63 immunostain.
