Outcome after Invasive Local Recurrence in Patients with Ductal Carcinoma in Situ of the Breast

Silverstein, Melvin J et al.
Abramson Cancer Center of the University of Pennsylvania
Last Modified: November 1, 2001

Reviewers: Kenneth Blank, MD and John Han-Chih Chang, MD
Source: Journal of Clinical Oncology, Volume 16 (Number 4): pages 1367-1373.
The Van Nuys Breast Center, California


Ductal carcinoma in situ (DCIS) is a non-invasive cancer of the breast. It is a highly curable disease with mastectomy or lumpectomy (localized excision) with or without radiation therapy (RT).

Bernard Fisher, MD, documented on the National Surgical Adjuvant Breast and Bowel Project (NSABP) trial B - 17 in the February 1998 Journal of Clinical Oncology that RT (5000 cGy or rads to the whole breast) in addition to lumpectomy decreased local recurrence rates. Ipsilateral (same sided) DCIS recurrence rate at 8 years decreased from 13% to 8% with RT added, and the incidence of invasive ipsilateral breast cancer at 8 years declined from 13% to 4% -- all of which were statistically significant and with a median follow up of 90 months.

Melvin J Silverstein, MD, in a prior June 1996 Cancer article, attempted to find a subset of patients that may benefit from RT following lumpectomy. He reviewed retrospectively data from DCIS patients treated at the Van Nuys Breast Center and the Children's Hospital in San Francisco. He created the Van Nuys Prognostic Index shown here:




Tumor Size

? 15 mm

16 - 40 mm

> 40 mm


? 10 mm

1 - 9 mm

< 1 mm


Intermediate to low grade without evidence of comedo necrosis

Intermediate to low grade with evidence of comedo necrosis

High grade

The sum total of all 3 categories gives the prognostic score. Patients with a score of 3 - 4 had no statistical improvement in local recurrence free survival rates at 8 years (100% lumpectomy + RT versus 97% lumpectomy alone. While those with a score of 5, 6 or 7 had an improvement from 68% to 85% with the addition of RT (p-value of 0.017). Patients with scores of 8 or 9 had recurrence rates of greater than 60%, but were significantly benefited with the addition of RT. Despite the improvement in the 8 9 score patients, high rates of recurrence would suggest more of a role for mastectomy in those poor prognostic patients.

Lawrence Solin, MD, reported in a March 1996 Journal of Clinical Oncology article on nearly 270 DCIS patients treated with breast conservation therapy (lumpectomy + RT). Those that demonstrated comedo necrosis and nuclear grade 3 pathology had NO significant local failure rate difference at 10 years (18%) compared to those without one or both of those pathologic findings (15%). The 15 year overall survival was 87%, cause-specific survival and freedom from distant relapse rate was 96% and an actuarial local failure rate of 19% at 15 years. The median time to local failure was 5.2 years with half being invasive carcinomas. The contralateral breast cancer incidence was 9% at 15years.

Materials and Methods

Seven hundred and seven patients were reviewed. They were treated from 1972 to 1997 at the Breast Center in Van Nuys and the Children's Hospital in San Francisco. The treatment philosophy was described as "evolving." For the most part, the Children's Hospital treated patients on a pilot study of excision alone for patients with a non-palpable lesion seen only on mammogram, less than 25 mm in size, clear margins of excision by 1 mm and no microcalcifications on post operative mammogram. Other treatment of DCIS patients was not described. The patients at Van Nuys were treated based on patient preference with recommendations made to them based on tumor size: greater than 4 cm lesions were to have mastectomies, while those with smaller lesions were given a choice to have breast conservation surgery with or without RT or mastectomy. The great majority of patient data came from the Van Nuys center. The RT given was initially to the whole breast for 4000 - 5000 cGy or rads followed by a boost to the tumor bed for another 1000 - 2000 cGy with Iridium 192 radiation seeds or external electron or photon beam.


The median follow up for all 707 patients was 69 months. Seventy-four (10.5%) locally recurred; 39 (5.5%) had a DCIS recurrence, while 35 (4.9%) had an invasive recurrence. The author documented the estimated the local recurrence probability at 8 years at 12.5%, while the estimated local invasive recurrence probability was 6.5%. The breast cancer-specific mortality was 1.4% and the overall mortality was estimated to be 7.6% at 8 years. Table 1 has the data on the 8-year rates along with some of the patient characteristics: (Highlights of chart shown here):


All breast conservation patients

Excision + RT

Excision only

Patient numbers





Local recurrence probability at 8 yr

0.5% (2)

18.8% (72)

16.6% (36)

21.1% (36)

Local invasive recurrence probability at 8yr

0.5% (2)

9.3% (33)

8.8% (18)

11.5% (14)

Distant recurrence probability at 8yr

0.5% (1)

2.3% (6)

3.4% (5)

1.0% (1)

Breast cancer specific mortality at 8yr





Overall mortality probability at 8yr





Median follow up

67 months

70 months

91 months

55 months

Number of actual patients in parenthesis

All the differences stated above were non-significant statistically. The follow up is not comparable enough to make judgments on 8-year estimates.

For the 74 patients that recurred locally, the 8 year probability of developing distant recurrence was 20.8% and the breast cancer-specific mortality was 8.8% at 8 years. They describe that NO patients with DCIS recurrence had distant recurrences or died of their disease. Therefore, one only needs to look at the 35 that had an invasive recurrence, for which the distant metastatic recurrence rate was 27.1% and cancer specific mortality rate was 14.4% at 8 years. Figure 1 is a time to recurrence graph of excision only versus excision + RT; the significance in the difference seen was not stated. The median time to any type of local recurrence was 23 months for the excision only group versus 56 months for those that got RT along with excision. Table 2 lists the median time to recurrences and recurrent tumor characteristics - note the increased size and the decreased ability to clinically detect (palpate) those recurrences in those treated with RT post excision. Figure 2 documents the number of local recurrences as a function of time with the median time to local DCIS being 22 months versus 58 months for invasive local recurrences. Figure 3 documents the stage of local recurrence at the time of diagnosis. These data are not complete since most patients had a lymph node dissection at the time of initial excision for the DCIS lesions. Thus, lymph node status was not available on the majority (30 of 35) of those with invasive recurrences.


It is a contention of the author that an increased size at the time of presentation of local invasive cancer recurrence was due to radiation fibrosis causing a decreased ability of a MD to palpate an early small recurrence. He also proposed that this was also the reason that mammograms have a difficult time detecting a locally recurrent lesion. Thus, breasts treated with RT lead to later presentation of the recurrent tumor with more aggressive biological features. Dr. Silverstein also states that this may not be an issue now, because most of his patients were treated with sub-optimal RT technique: treating 4 times per week with 200cGy fractions and utilizing brachytherapy boost (12 of 18 patients with local invasive recurrence and 4 of 5 with distant disease) instead of photons or electrons. This is mostly speculation. While his data are minimally support his statements, the numbers are too small and insignificant on which to base his contentions. For example, the follow up on the brachytherapy (Iridium 192 implant) patients (whom he states have a worse outcome) is twice as long as on the electron and photon patients (120 months versus 60 months, respectively). Thus, more follow up leads to detecting more adverse events. Few if any other reported literature would support his many claims.

Dr. Silverstein touts his mastectomy data as far superior than most others, but he explains that this may be due to the fact that there is a significant amount of exclusion of patients found at the time of mastectomy to have foci of invasive cancer within the breast. The amount may be very significant since the pathologist that processes the specimens for the Van Nuys Breast Center patients is very aggressive at sectioning and mounting the whole organ. This would find many other foci of invasive cancer that would not be usually found in mastectomies performed at community and most university hospitals (only representative slices are taken). This is a way to eliminate those with invasive cancer in the mastectomy patients from the data set, while those with possibly invasive lesions in the breast conservation group will never be discovered and contribute to the minimal and non-significant difference in outcome.

What would have been helpful in this paper is a description on what salvage were performed and correlate that with the outcome, since this was a paper supposedly dedicated to the outcome of patients with invasive recurrences. A few good items were brought to light: 1) He is an advocate of placement of surgical clips in the tumor bed (not a generalized practice nationwide) to assist the radiation oncologist in his treatment planning 2) He admits that for most patients mastectomy is over-treatment 3) He admits that cosmetically, properly done breast conservation therapy yields a far better result than any surgical breast reconstruction.


Don’t Look Back with Cancer
by Bob Riter
April 5, 2013

CancerLand Bookshelf: The Summer of Her Baldness
by Alysa Cummings
December 30, 2010

Want up to date Cancer Treatment News? Subscribe to our eNewsletter. Subscribe Now