398
Chapter 11
guiding further patient management.” In this series, margin
width was the strongest predictor of successful conservation
treatment. When compared with patients who had a clear
margin of at least 10 mm, the RR for breast recurrence was
2.5 if the margin was 1 to 9 mm and 22 for a margin of less
than 1 mm. Warnberg
402
reported no statistically significant
differences in relapse-free survival between patients with
DCIS stratified into the three VNPI prognostic groups.
The VNPI should be validated in a prospective random-
ized trial before acceptance as a basis for clinical practice.
This is especially important because of significant concerns
about the database from which it was derived. A major is-
sue is the lack of a consistent treatment program illustrated
by the following quotation that described the study group
403
:
Until 1988 all patients with DCIS who elected breast con-
servation were advised to add breast irradiation to their
treatment. Most patients accepted this recommendation; a
few refused and were treated with careful clinical follow-
up without irradiation. Beginning in 1989, the physicians
within The Breast Center were no longer convinced of the
overall value of radiation therapy for DCIS, and all breast
conservation patients with uninvolved biopsy margins
(clear by 1 mm or more) were offered the option of care-
ful clinical follow-up without radiation therapy. Many pa-
tients accepted this option; some refused and were treated
with breast irradiation. Outside patients with DCIS re-
ferred to our radiation oncologists for radiation therapy
continued to be treated with radiation therapy in accord
with the wishes of their referring physicians.
Other uncontrolled variables included differing radiation
schedules and inconsistent boost treatment.
403
Lesion size was one of the original three variables included
in the VNPI. As discussed elsewhere in this chapter, there is no
reliable or generally acceptedmethod for measuring the size or
extent of DCIS, especially with the precision that is required for
the VNPI scoring system. In lesions limited to a single tissue
block, it may be possible to distinguish between foci smaller
and larger than 15 mm, but the distinction between 15 and 40
mm and larger than 40 mm is likely to be very unreliable. De-
termining size when DCIS is distributed in more than one tis-
sue block from a single biopsy specimen or if it is in more than
one biopsy specimen is very imprecise. The methods for de-
termining lesion extent by counting 4× fields of involvement
The VNPI was developed to stratify patients with DCIS,
to distinguish between women who are most likely to be
treated successfully by breast conservation and those who
might be candidates for mastectomy because of a relatively
high risk of breast recurrence.
399
The original VNPI was a
numerical score of 3 to 9 based on the assessment of three
variables: size of DCIS, distance between DCIS and margin,
and a pathologic classification based on necrosis and nuclear
grade. Each variable was divided into three categories, which
were ranked (scored) from most to least favorable as 1 to 3
(Table 11.4). The original VNPI was derived from the sum of
scores for individual variables.
Follow-up of patients withDCIS grouped into three VNPI
categories (scores 3,4; scores 5,6,7; and scores 8,9) showed
significant differences in recurrence-free survival, with the
most favorable outcome associated with the lowest scores.
Patients were stratified within the VNPI groups according to
whether they received radiotherapy in addition to excision.
Radiated patients in the VNPI 3,4 group did not differ sig-
nificantly from those who were not radiated, but radiation
appeared to be beneficial in the intermediate VNPI group.
Recurrences were “unacceptably” frequent in the VNPI 8,9
group, even when radiotherapy was administered.
399
On the
basis of these observations, it was suggested that women
with DCIS classified as VNPI 3,4 could be treated by exci-
sion alone, that excision with radiotherapy be employed for
the VNPI 5 to 7 group, and that mastectomy should be rec-
ommended if the VNPI is 8 or 9.
Retrospective studies have not confirmed the VNPI as a
prognostic guide for local control of DCIS. For example, de
Mascarel et al.
400
found a significant difference in local re-
currence between women in the low- and intermediate-risk
VNPI groups. In univariate analysis, the local recurrence rate
increased with the size of the DCIS, with decrease in distance
to the margin, with higher histologic grade, and with the
percentage of paraffin blocks involved by DCIS. When these
variables were considered in multivariate analysis, the per-
centage of paraffin blocks with DCIS was the only significant
predictor of local recurrence. Boland et al.
401
retrospectively
studied 237 patients with DCIS and confirmed that margin
width and grade were significant risk factors for breast re-
currence. However, when stratified by VNPI, 78% of patients
were in the moderate risk group, a result that led the authors
to conclude that “the VNPI lacked discriminatory power for
Table 11.4
The University of Southern California/Van Nuys Prognostic Index (USC/VNPI) Scoring System
Score
1
2
3
Size (mm)
≤
15
6–40
≥
41
Margins (mm)
≥
10
1–9
<
1
Pathologic classification
Nonhigh grade
Nonhigh grade
High grade
(
−
) necrosis
(
−
) necrosis
(
±
) necrosis
Age (yr)
≥
61
40–60
≤
39
Modified from Silverstein MJ, Lagios MD. Choosing treatment for patients with ductal carcinoma in situ: fine tuning the University of Southern
California/Van Nuys Prognostic Index.
J Natl Cancer Inst Monogr
2010;2010(41):193–196.
