Fax: (203) 785-4622
Management of suspicious or indeterminate calcifications and impact on local control†
Version of Record online: 13 APR 2005
Copyright © 2005 American Cancer Society
Volume 103, Issue 11, pages 2236–2240, 1 June 2005
How to Cite
Lally, B. E., Haffty, B. G., Moran, M. S., Colasanto, J. M. and Higgins, S. A. (2005), Management of suspicious or indeterminate calcifications and impact on local control. Cancer, 103: 2236–2240. doi: 10.1002/cncr.21044
Presented at the 26th Annual San Antonio Breast Cancer Symposium in San Antonio, Texas, December 3– 2003.
- Issue online: 18 MAY 2005
- Version of Record online: 13 APR 2005
- Manuscript Accepted: 18 JAN 2005
- Manuscript Revised: 8 DEC 2004
- Manuscript Received: 1 OCT 2004
- breast carcinoma;
- breast conservation therapy;
- suspicion or indeterminate calcifications;
- local excision;
- margin status;
The current study was undertaken to determine the potential advantage associated with complete removal of suspicious or indeterminate calcifications (SIC) before the initiation of irradiation as part of breast conservation therapy (BCT).
Of 2045 patients treated with BCT at Yale University School of Medicine (New Haven, Connecticut) and satellite facilities before 2002, 111 women, included 3 patients with bilateral disease, had a postexcision preirradiation mammogram (PREMAMMO) to evaluate residual SIC. Thus, 114 breasts were at risk for local disease recurrence. Seventy-five breasts at risk had no residual SIC and proceeded to undergo radiotherapy (XRT) without further surgery or mammography. Of the remaining 39 breasts at risk, only 3 underwent a PREMAMMO with documented removal of all calcifications (DRC). Thirty-six breasts at risk proceeded to XRT with either known SIC or with nondocumented removal of calcifications (NDRC) after another excision.
Of the 78 breasts at risk with DRC via PREMAMMO, there were 7 local failures (LF) and 1 distant failure. Of the 36 breasts with NDRC via PREMAMMO, there were 7 LF and 1 regional failure. Of the 34 breasts who underwent reexcision after detection of SIC by PREMAMMO, 20 (59%) were found to have residual disease.
Patients with DRC were found to have better local control than patients with NDRC. In addition, the presence of SIC on a PREMAMMO was associated with a high probability of detecting residual disease. Cancer 2005. © 2005 American Cancer Society.
Breast conservation therapy (BCT) is considered the standard treatment for in situ and early-stage invasive breast carcinoma. BCT is defined as excision of the primary breast tumor and adjacent normal breast tissue, with or without axillary lymph node dissection, followed by irradiation. The goals of BCT are 1) to eradicate, with moderate doses of radiotherapy (XRT), microscopic foci of cancer that may remain after limited surgery to remove the primary tumor; 2) to provide local control and survival rates equivalent to mastectomy; and 3) to maximize quality of life, which includes minimizing complications and achieving an acceptable cosmetic result.1
Margin status has been the principal pathologic factor used to evaluate excision adequacy and, ultimately, the risk of local failure (LF). Positive specimen margins have a well established and significant association with LF.2–10 To our knowledge, correlations between initial excision specimen margin status and the presence and amount of residual carcinoma in the adjacent breast parenchyma of reexcision specimens are unclear. Despite this inherent limitation, margin status has received the greatest attention as a predictor of residual invasive carcinoma or ductal carcinoma in situ (DCIS) in the adjacent breast parenchyma of reexcision specimens.11–15
The current guideline of the American College of Radiology recommends postexcision, preirradiation mammograms (PREMAMMO) to assess the completeness of resection of tumors with microcalcifications and thus eliminate the possibility of residual suspicious or indeterminate calcifications (SIC), particularly when an extensive intraductal component is present.1 To our knowledge, the utility and impact of PREMAMMO on local control has not been well defined. The current study was undertaken to determine the potential advantage associated with complete removal of SIC before the initiation of whole-breast irradiation.
MATERIALS AND METHODS
The study consists of a retrospective review of the medical records of all patients treated with BCT for breast carcinoma at the Yale University School of Medicine (New Haven, CT) and satellite facilities before 2002. Of the 2045 patients treated with BCT, 429 (21%) patients had SIC identified on the preoperative mammography. In the current study, the definition of SIC was determined from the original mammography report. No attempt was made to reinterpret the original findings. All patients underwent ≥ 1 local excision before the initiation of breast irradiation that typically consisted of a dose of 46–50 gray (Gy) to the intact breast and a boost to the tumor bed of 10–16 Gy. Regional lymph nodes were treated as clinically indicated. One hundred eleven women (ages 31–81 years) with breast carcinoma including 3 patients with bilateral primary tumors received PREMAMMO. Therefore, a total of 114 breasts were at risk for local disease recurrence and were included in the current study. Seventy-five breasts at risk had a PREMAMMO that showed no residual calcifications and they subsequently underwent XRT without further surgery or mammography. Thirty-nine breasts at risk had SIC via a PREMAMMO. Of these patients, 5 who were at risk did not undergo any further reexcision, whereas 34 at risk breasts underwent ≥ 1 reexcisions. In this latter group, 12 breasts at risk had an additional PREMAMMO to exclude residual SIC—9 breasts at risk had persistent SIC and proceeded to XRT without further reexcision and 3 breasts at risk began XRT with an additional PREMAMMO documenting no residual SIC. Therefore, the group with documented removal of calcifications (DRC) included 75 breasts at risk with a negative PREMAMMO, and 3 breasts at risk who had PREMAMMO with SIC, but subsequent reexcision resulted in an additional PREMAMMO documenting no SIC. The group for comparison includes patients at risk who began XRT and did not have documented removal of calcifications (NDRC). Specifically, the NDRC group comprised patients at risk who began XRT with either 1) known SIC on PREMAMMO (5 breasts at risk), 2) lack of an additional PREMAMMO indicating DRC after reexcision (22 breasts at risk), or 3) known SIC on an additional PREMAMMO after reexcision (9 breasts at risk). Figure 1 details the distribution of the patients.
The patient characteristics are presented in detail in Table 1. The median age between the the DRC and NDRC groups differed by 3 years. Compared with patients with DRC, the patients with NRDC did have a slightly higher percentage overall of having some form of infiltrating ducal pathology (51% vs. 41%) and either positive or unknown margin status (28% vs. 17%). More patients in the DRC group had known negative lymph node status (72% vs. 42%) than patients in the NDRC group. T classification and the use of adjuvant chemotherapy were evenly divided between the two groups. Finally, a higher percentage of patients in the NDRC group received adjuvant hormonal therapy (31% vs. 23%).
|Characteristics||DRC (%)||NDRC (%)|
|No. of patients at risk for local recurrence||78||36|
|Median age (yrs) (range)||54 (34–81)||51 (31–71)|
|Left||34 (44)||22 (61)|
|Right||44 (56)||14 (39)|
|Infiltrating ductal||11 (14)||6 (16)|
|Infiltrating ductal and DCIS||16 (21)||10 (28)|
|DCIS with focal invasion||5 (6)||2 (6)|
|DCIS||39 (50)||17 (47)|
|Infiltrating lobular||3 (4)||1 (3)|
|Tubular||4 (5)||0 (0)|
|Lymph node status|
|Positive||9 (12)||5 (14)|
|Negative||46 (59)||15 (42)|
|Unknown||23 (29)||16 (44)|
|Tis||39 (50)||17 (47)|
|T1||30 (38)||15 (42)|
|T2||9 (12)||4 (11)|
|Negative||40 (51)||20 (56)|
|<2 mm||25 (32)||6 (17)|
|Positive||4 (5)||5 (14)|
|Unknown||9 (12)||5 (14)|
|Yes||16 (21)||7 (19)|
|No||62 (79)||29 (81)|
|Adjuvant hormonal therapy|
|Yes||18 (23)||11 (31)|
|No||60 (77)||25 (69)|
|Local failures||7 (9)||7 (19)|
In the NDRC group, there were 7 LF and 1 regional failure (RF) for an absolute LF rate of 19% (after a median follow-up of 2.8 years). In the DRC group, there were 7 LF and 1 distant failure for an absolute LF rate of 9% (after a median follow-up of 3.2 years). Of the 34 patients who underwent reexcision after detection of SIC on PREMAMMO, 20 (59%) were found to have residual disease. Only five breasts at risk underwent reexcision because of both the SIC on PREMAMMO and the uncertainty of the margin status. Of these five breasts at risk, two breasts at risk, one breast at risk, and two breasts at risk, respectively, initiated XRT with unknown margins, positive margins, and negative margins. For the breasts that had residual disease on reexcision, the pre-needle localization margin status was 3 patients, 2 patients, 3 patients, and 12 patients for respective margin status of positive (3), within 2 mm (2), unknown (3), and negative (12).
Table 2 details the specific characteristics of the patients from our cohort who developed LF, RF, or distant failure after BCT. Only two of eight disease recurrences in each group were DCIS alone, because infiltrating ductal carcinoma was the predominant pathology. For the cohort of patients with NDRC who ultimately experienced a LF, only one of seven patients had a documented negative margin on the final pathology specimen. This compares to the DRC group in which four of seven patients had a documented negative margin. The majority of patients in both groups did not receive adjuvant hormonal therapy. The percentage of patients receiving chemotherapy was comparable.
|Characteristics||Disease recurrence||Age (yrs)||Pathology||Lymph node status||T classification||Margin status||Adjuvant chemotherapy||Adjuvant hormonal therapy|
|1||Breast only||60||Infiltrating ductal||Negative||T1||Unknown||No||No|
|4||Breast only||53||Infiltrating ductal||Negative||T2||Positive||No||No|
|5||Breast only||34||Infiltrating ductal and DCIS||Negative||T2||Negative||Yes||No|
|6||Breast only||47||Infiltrating ductal||Positive||T1||Positive||Yes||No|
|7||Breast only||70||Infiltrating ductal||Negative||T2||Negative||No||Yes|
|8||Distant failure||47||Infiltrating ductal and DCIS||Negative||T2||Within 2 mm||Yes||No|
|1||Breast only||55||Infiltrating ductal and DCIS||Unknown||T2||Negative||No||No|
|3||Breast only||64||Infiltrating ductal and DCIS||Unknown||T1||Within 2 mm||No||No|
|4||Breast only||63||DCIS||Unknown||T0||Within 2 mm||No||No|
|5||Breast only||31||Infiltrating ductal||Unknown||T1||Unknown||Yes||No|
|6||Breast only||58||Infiltrating ductal||Positive||T2||Negative||Yes||No|
|7||Breast only||44||DCIS with focal invasion||Negative||T2||Negative||Yes||No|
|8||Local lymph nodes||38||Infiltrating ductal||Positive||T2||Unknown||Yes||Yes|
To optimize the results for women undergoing BCT, all evidence of macroscopic tumor should be removed. As evident by the impact of margin status, failure to do so has negative consequences. The utilization of PREMAMMO to define the adequacy of tumor excision is controversial. However, the presence of microcalcifications in the tumor bed is suspicious for residual disease. Teixidor et al.16 reported the results of a study of 120 consecutive patients who had PREMAMMO before XRT. Ten patients had residual calcifications, resulting in biopsy-proven residual disease for six patients.
Gluck et al.17 reported the results of 43 women who required a reexcision secondary to margin status. All patients underwent mammography before the second surgery. They reported 29 patients with SIC on mammography, which included 20 patients who had residual disease. Waddell et al.18 reported similar results in 67 patients with DCIS who underwent PREMAMMO. Sixteen patients were found to have SIC. Of the 12 patients who underwent further surgery, 9 were found to have residual disease. In six patients, the margin status was unknown or positive. Finally, Aref et al.19 reported the results of 90 patients with breast carcinoma with adequate margins who underwent PREMAMMO. Sixteen were found to have SIC. These results lead to reexcision in 12 patients, 8 of whom were found to have residual disease. All of these results are consistent with our results, which found SIC on PREMAMMO to be correlated with residual disease in greater than one-half of the patients.
In women with indeterminate or malignant calcifications on regular mammography, microfocus (0.1-mm focal spot) magnification views in orthogonal projections are useful.20 However, in patients who have undergone a surgical procedure, the formation of hematomas and scarring may limit the sensitivity and specificity of PREMAMMO for SIC. Furthermore, there may be interobserver variability in the interpretation of these films. Several radiologists were involved in the original interpretation of mammograms for the patients in the current study. We chose to perform the current study from a practical clinical situation, without attempting to reinterpret and bias original findings. We acknowledge this potential weakness of variability, but view it as a reality of the actual practice pattern.
At our institution, the current policy is to obtain a minimum of one PREMAMMO for patients for whom the original lesion is associated with SIC. If residual SIC suspicious or indeterminate calcifications are identified, the patients are strongly encouraged to undergo reexcision before the initiation of XRT. SIC are typically removed with the guidance of a needle localization procedure, and the majority of patients in this series underwent needle localization to remove these residual calcifications. We believe that this selects for the optimal local control in patients undergoing BCT.
To our knowledge, the current study is the only published report to date that evaluates quantitatively the utility of PREMAMMO in terms of LF in a cohort of patients with contemporary BCT management. Although the sample of patients analyzed is small, we found that NDRC on PREMAMMO was associated with a higher LF rate. Also, based on our evaluation of pathologic specimens from patients who underwent reexcision after the detection of SIC on a PREMAMMO, it appears that residual SIC are associated with a high incidence of residual disease. Therefore, if the margin status is positive and SIC are identified on PREMAMMO, we believe that reexcision of SIC may decrease LF. If the margin status is negative, the role of multiple reexcisions for persistent SIC on additional PREMAMMO is of questionable value. Additional surgery may not be necessary because the remaining calcifications may not be associated with residual disease. Moreover, performing multiple reexcisions may compromise the cosmetic outcome. A major goal of BCT is to maximize the patient's quality of life (i.e., to minimize complications and achieve an acceptable cosmetic result.
We recommend that the findings of the PREMAMMO should be used in conjunction with margin status and solid clinical judgment in the management of patients with persistent SIC. Further studies focused on investigating this subset of patients are needed to elucidate the optimal use of imaging techniques and surgical intervention necessary to optimize local control with acceptable cosmetic results.
- 1Conservative surgery and radiation in the treatment of stage I and II carcinoma of the breast. American College of Radiology. ACR appropriateness criteria. Radiology. 2000; 215(Suppl): 1193–1205., , , et al.
- 3Patients with early stage invasive cancer with close or positive margins treated with conservative surgery and radiation have an increased risk of breast recurrence that is delayed by adjuvant systemic therapy. Int J Radiat Oncol Biol Phys. 1999; 44: 1005–1015., , , et al.
- 20Appropriate imaging work-up of breast microcalcifications. American College of Radiology. ACR appropriateness criteria. Radiology. 2000; 215(Suppl): 973–980., , , et al.