CHICAGOResults from two published prospective randomized
clinical trials of women with ductal carcinoma in situ (DCIS) show
that radiotherapy after wide excision reduces local recurrence by 7%
to 15%, said Melvin J. Silverstein, MD, who does not believe
radiotherapy is suitable for all women with this form of breast
My reputation is that I dont want to give radiotherapy to
anybody. Thats not true; I give it to a lot of people. I just
dont want to give it to everybody, he said at the Second
Annual Lynn Sage Breast Cancer Symposium, sponsored by Northwestern
University Medical School.
Dr. Silverstein, professor of surgery, Keck School of Medicine,
University of Southern California, and director of the Harold E. and
Henrietta C. Lee Breast Center, Los Angeles, noted that radiotherapy
is expensive, inconvenient for patients, and confers only a slight
benefit in some subgroups of patients.
Radiotherapy also is a one-shot intervention. If you give
radiotherapy for DCIS and get an invasive recurrence, you cant
give it again. If you dont give it initially and get an
invasive recurrence, you can re-excise and irradiate, he said.
He noted that radiotherapy complicates the performance of
skin-sparing mastectomy for invasive recurrence. Doing a skin-sparing
mastectomy is more difficult because you have vascular
changes secondary to irradiation, he said.
Finally, the most compelling reason not to use radiotherapy for all
patients with DCIS, he said, is that the prospective randomized
trials have not shown a benefit in the single most important
endpoint, breast cancer mortality, regardless of treatment. Why
give an expensive, time-consuming, potentially harmful treatment when
there is no proven survival benefit? he asked.
Since 1995, Dr. Silverstein and his colleagues have been trying to
devise a reproducible scheme to select women who might be able to
avoid radiotherapy after excision of DCIS. After testing more than 30
histologic factors, they settled on tumor size, margin width, and
pathologic classification (based on nuclear grade and comedo-type
necrosis) as the basis for the Van Nuys Prognostic Index (VNPI).
In an observational study of 560 patients who underwent
breast-conserving procedures, there was little or no local recurrence
after wide excision in 163 women with DCIS who had small,
well-excised, low-grade tumors and therefore low VNPI scores of 3 or
4. The finding suggested that these women might be treated with
Although women with large, high-grade lesions with close margins and
higher VNPI scores achieved the greatest benefit from radiotherapy,
there still was a 50% local recurrence rate, suggesting that these
patients might be better managed by re-excision or mastectomy, he said.
Dr. Silverstein stated that complete tissue processing and 3D
reconstruction of the lesion were used to determine tumor size and
that this was difficult to reproduce from lab to lab. He, therefore,
decided that the distance between DCIS and the closest inked margin
was the most reproducible aspect of the VNPI and could be used by
itself as a surrogate marker. In one observational study, Dr.
Silverstein found a very small benefit from radiotherapy for 174
patients with margins of 10 mm or greater. In addition, there was no
difference in local recurrence rates between younger and older women
with margins of 10 mm or more.
Based on these data, Dr. Silverstein feels that the most important
cause of local failure is inadequate excision. He asked, Is
radiotherapy required for all cases of breast preservation? I
dont think so. The key to the future is to figure out which
lesions will become invasive and to treat them more aggressively.
Lawrence Wickerham, MD, associate director of the NSABP (National
Surgical Adjuvant Breast and Bowel Project), acknowledged that Dr.
Silversteins experience with single center studies has yielded
promising findings. Although we cant say it is wrong, it
is hard to say that it is absolutely correct. Observational studies,
if conducted meticulously, are appropriate for generating hypotheses.
However, they are fraught with methodological bias that can skew
results, Dr. Wickerham said.
The greatest potential bias arises from the use of historical
controls or control patients who were not treated concurrently.
To be effective, historical controls have to be the same as the
experimental groupthe same diagnosis, patient characteristics,
and evaluation, he said. That is extremely difficult to
do in this country, because medical care gets better, and patients
He cited improvements in mammography that have led to earlier breast
cancer diagnosis and pointed to bone marrow transplant for breast
cancer to show the problem of using historical controls. The
enthusiasm for bone marrow transplants was the result of comparisons
with historical controls. That enthusiasm has diminished as
randomized clinical trial results have become available, he said.
Dr. Wickerham stressed that the results of recent randomized clinical
trials of treatment for DCIS support radiotherapy as an adjunct to
excision. After 8 years of follow-up, NSABP B-17 revealed a 5%
significant difference in the rate of noninvasive disease for
patients who received radiotherapy. Studies of various segments of
the patient population were unable to find any group of women who
would not benefit from radiotherapy, regardless of margin or
The risk of ipsilateral breast tumor recurrence is incredibly
small in a group of 100 patients who have low-grade lesions that are
excised widely. But if you have one woman with this disease, her risk
is either zero or 100%. Based on the information I have today, I
would feel that offering her radiotherapy increases her chances of
being in that zero-risk group, Dr. Wickerham said.