This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Lagios, M. D. Articles by Silverstein, M. J. Search for Related Content PubMed PubMed Citation Articles by Lagios, M. D. Articles by Silverstein, M. J. Related Articles Related Reply

Prospective Study of Wide Excision Alone for Ductal Carcinoma In Situ of the Breast

The Breast Cancer Consultation Service, Tiburon, CA

David L. Page

Department of Pathology, Vanderbilt University, Nashville, TN

Melvin J. Silverstein

Keck School of Medicine, University of Southern California, Los Angeles, CA

To the Editor:

Clearly, the goal set by Wong et al to identify a subset of ductal carcinoma in situ (DCIS) patients who can be spared radiation treatment based on available clinical, mammographic, and pathologic criteria, is admirable. We question however, whether the methods employed by the investigators were appropriate and capable of identifying a subset that truly did not require radiation therapy. Both the University of Southern California Van Nuys (USC/Van Nuys) prospective study and the Eastern Cooperative Oncology Group (ECOG) registration trial require patients who are entered to meet strict criteria for grade, size, and margin determination. The ECOG trial is now maturing, but results from the USC/Van Nuys database can define subsets at low risk for local recurrence (ie, 5% at 12 years mean follow-up).¹

Identification of subsets requires highly correlated imaging and pathology, and complete sequential tissue processing for all surgical resections of DCIS. These are also requirements for entry in the ECOG trial. There is a growing consensus that such a resection should be processed sequentially in its entirety.^2-4 Sampling of such resections ensures neither the exclusion of invasive disease nor the determination of the extent of disease or margin status. Unfortunately, that critical part of the pathologic examination was not part of the study described by the investigators. How does this impact the potential results of the study?

Tumor Size
DCIS was estimated by the extent of mammographic microcalcification or by palpation. Although all patients entered were required to have a "clinical size" 25 mm, there was no measured pathologic size, except for those few cases in which the entirety of the DCIS was confined to a single slide. Ninety-six percent of the entire study group had size estimated by counting the number of blocks containing DCIS—a method that is highly dependent on the idiosyncrasies of sampling by an individual pathologist. The objective of having a "clinical size" 25 mm can be realized using neither the number of blocks, the extent of mammographic microcalcification, nor the area of palpable abnormality.

Margin
Margins were required to be 1 cm (10 mm), or to be re-excised until clear. However, since the resections that had a median 73-mL volume were only sampled by the protocol, only those margins that were identified could be measured. Sampling of resections for DCIS, which may not be marked by microcalcification, be visible grossly, and may involve fatty tissue, does not permit evaluation of the extent of disease, adequate margin assessment, or exclusion of microinvasion. Such resections should be sequentially sectioned and embedded in order to avoid the limitations of the type of pathologic examination that characterizes historical randomized trials of radiation for DCIS such as National Surgical Adjuvant Breast and Bowel Project B17 and European Organisation for Research and Treatment of Cancer 10853.

Eighty-three percent of such cases were nonetheless re-excised, but defining a clear margin by intraoperative shave re-excisions of the biopsy cavity is fraught with difficulty. Not infrequently, the entirety of the biopsy cavity is not included in the re-excision. Moreover, if the re-excision is sampled in the same fashion as the initial resection then margin status can remain unresolved.

Grade
Most recent classifications of DCIS use some combination of nuclear grade and necrosis to establish grades. Almost all use the highest nuclear grade identified. Some DCIS exhibit limited areas of high-grade disease, but are still classified as high grade in the USC/Van Nuys system, and in the earlier work from the Children's Hospital. Classifying DCIS on the basis of the predominant grade can potentially increase the risk of local recurrence for low and intermediate nuclear grade groups, and result in higher local recurrence rates. Indeed, this was noted by the investigators in this study. Necrosis was not part of the grading equation used by the investigators, despite the fact that many studies have shown it to be a significant prognostic indicator.

Definition of Recurrent Disease
In the current era, it is widely recognized and accepted that the risk to the irradiated breast outside of the quadrant containing the initial focus of DCIS is comparable with that in the contralateral breast. In evaluating the benefits of therapeutic intervention, it is essential to distinguish de novo events in the irradiated breast from true recurrences at the site of initial treatment. This is being documented in the ECOG trial. In the EORTC 10853 trial, most recurrences were at the site of treatment and shared a similar histology with the initial lesion. In this study 76% of the ipsilateral post-treatment events were classifiable as true recurrences; 23% occurred elsewhere in the treated breast. This establishes a true recurrence rate of 6.3% at 40 months. If an equal percentage of the invasive recurrences were also de novo events, then the three of 158 invasive true recurrences at 40 months would equal 1.8%; an uncorrected projection of recurrences to 5 years would be as follows: total in situ and invasive, 9.5%; invasive alone, 3.75%.

Projecting Recurrences at 5 Years
Projecting the local recurrence rate to 12% at 5 years does not match the slope of recurrence curves for well-excised DCIS treated by lumpectomy alone. For example, for high-grade DCIS, the mean interval to recurrence is 24 months, and most recurrences will have recurred by 5 years. To therefore project the likely recurrence rate in this group of patients followed for a median of 40 months may be problematic. Nonetheless, it is worth noting that the total recurrence rate for invasive disease was 2.5% (four of 158), and for the entirety of recurrences, including in situ disease, 8.2% (13 of 158).

Prospective studies of DCIS, which are now being conducted, may resolve such questions as the potential identification of subsets that could avoid radiation therapy with minimal risk. Some of us believe that this question has been answered, but we recognize that it must be corroborated by randomized trial data. The USC/Van Nuys approach has been validated by a number of investigators who have prospectively identified low-risk subsets. We certainly await the results of the ECOG registration trial, which addresses the same problem, and hopefully will provide a clearer answer to the question of a low-risk subset.

From a different perspective, the 158 patients entered onto this study were spared the costs and potential morbidities of radiation therapy. Four patients developed recurrent invasive carcinoma, but none developed metastases or died of disease. Overall, these are excellent results despite the inherent design flaws of the study. The nine patients with noninvasive recurrences can be treated with re-excision or mastectomy; the four with invasive recurrences can be treated by lumpectomy and irradiation, having preserved that option by avoiding initial radiation. No randomized trial has shown a survival benefit for those so treated, who do experience a small increase in cardiac mortality from the therapy itself.

Authors' Disclosures of Potential Conflicts of Interest

The authors indicated no potential conflicts of interest.

REFERENCES

1. Silverstein MJ: An argument against routine use of radiotherapy for ductal carcinoma in situ. Oncology 17:1511-1533, 2003[Medline]

2. Consensus Committee: Consensus conference on the classification of ductal carcinoma in situ. Cancer 80:1789-1802, 1997

3. Consensus Committee: Image-detected breast cancer: State of the art diagnosis and treatment. J Am Coll Surg 193:297-302, 2001[CrossRef][Medline]

4. Silverstein MJ, Lagios MD, Recht A, et al: Image-detected breast cancer: State of the art diagnosis and treatment. J Am Coll Surg 201:586-597, 2005[CrossRef][Medline]

Related Reply

• In Reply:
  Julia S. Wong, Susan C. Lester, and Barbara L. Smith
  JCO 2006 24: 3811-3812 [Full Text]

This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Lagios, M. D. Articles by Silverstein, M. J. Search for Related Content PubMed PubMed Citation Articles by Lagios, M. D. Articles by Silverstein, M. J. Related Articles Related Reply