Fax: (713) 745-4926
Positive surgical margins and ipsilateral breast tumor recurrence predict disease-specific survival after breast-conserving therapy
Version of Record online: 3 FEB 2003
Copyright © 2003 American Cancer Society
Volume 97, Issue 4, pages 926–933, 15 February 2003
How to Cite
Meric, F., Mirza, N. Q., Vlastos, G., Buchholz, T. A., Kuerer, H. M., Babiera, G. V., Singletary, S. E., Ross, M. I., Ames, F. C., Feig, B. W., Krishnamurthy, S., Perkins, G. H., McNeese, M. D., Strom, E. A., Valero, V. and Hunt, K. K. (2003), Positive surgical margins and ipsilateral breast tumor recurrence predict disease-specific survival after breast-conserving therapy. Cancer, 97: 926–933. doi: 10.1002/cncr.11222
- Issue online: 3 FEB 2003
- Version of Record online: 3 FEB 2003
- Manuscript Accepted: 12 DEC 2002
- Manuscript Received: 6 DEC 2002
- National Institutes of Health
- The University of Texas M. D. Anderson Cancer Center Physician-Scientist Program
- Department of Defense Career Development Award
- breast carcinoma;
- local recurrence;
- ipsilateral breast tumor recurrence;
- systemic recurrence;
- breast conservation;
- breast neoplasms;
The current study identified determinants of systemic recurrence and disease-specific survival (DSS) in patients with early-stage breast carcinoma treated with breast-conserving surgery and radiation therapy (breast-conserving therapy, or BCT).
The study population consisted of 1,043 consecutive women with Stages I or II breast carcinoma who underwent BCT between 1970 and 1994. Clinical and pathologic characteristics evaluated included age, tumor size, tumor grade, estrogen and progesterone receptor status, surgical margins, axillary lymph node involvement, and use of adjuvant therapy.
At a median follow-up time of 8.4 years, 127 patients (12%) had developed an ipsilateral breast tumor recurrence (IBTR), and 184 patients (18%) had developed a systemic recurrence. On multivariate logistic regression analysis, tumor size greater than 2 cm, positive lymph nodes, lack of adjuvant tamoxifen therapy, and positive margins (odds ratio [OR], 3.7; 95% confidence interval [CI], 1.1-12.3; P = 0.034) were predictors of systemic recurrence. When IBTR was added into the model, adjuvant therapy and surgical margins were not independent predictors; however, IBTR was an independent predictor of systemic recurrence (IBTR vs. no IBTR; OR, 6.2; 95% CI, 3.1-12.3; P < 0.001). The 10 year DSS rate after BCT was 87%. On multivariate Cox proportional hazards model analysis, the following factors were independent predictors of poor DSS: tumor size greater than 2 cm (vs. ≤ 2 cm; relative risk [RR], 2.3; 95% CI, 1.2-4.3; P = 0.010), negative progesterone receptor status (vs. positive; RR, 2.7; 95% CI, 1.4-5.1; P = 0.003), positive margins (vs. negative; RR, 3.9; 95% CI, 1.4-11.5; P = 0.011), and IBTR (vs. no IBTR; RR, 5.5; 95% CI, 2.8-11.0; P < 0.001).
Positive surgical margins and IBTR are predictors of systemic recurrence and disease-specific survival after BCT. Aggressive local therapy is necessary to ensure adequate surgical margins and to minimize IBTR. Cancer 2003;97:926–33. © 2003 American Cancer Society.
Treatment of early-stage breast carcinoma with breast-conserving surgery (lumpectomy or segmental mastectomy) plus radiation therapy is well accepted as an alternative to mastectomy for most patients with Stages I and II breast carcinoma. Twenty year follow-up of prospective randomized trials confirms that there is no significant difference in distant disease-free survival or disease-specific survival between patients treated with mastectomy and those treated with breast-conserving therapy (BCT).1, 2 However, 4–20% of patients with early-stage breast carcinoma have an ipsilateral breast tumor recurrence (IBTR) within 10 years of BCT.1–9 The risk of IBTR in patients undergoing BCT is significantly higher than the risk of local recurrence after mastectomy.1
The impact of IBTR on the risk of systemic recurrence and survival remains controversial. Although IBTR has been associated with an increased risk of systemic recurrence,8, 10–13 it has been proposed that local relapse is not the cause of but simply an indicator of poor prognosis.14 Thus, it has been argued that local management has little, if any, effect on survival.15
Since the early 1970s, The University of Texas M. D. Anderson Cancer Center has used BCT to treat selected patients. In the current study, we reviewed the records of 1043 consecutive patients with Stages I and II breast carcinoma treated with BCT and followed those patients long term. We aimed to identify factors predictive of systemic recurrence and disease-specific survival and to determine the influence of local control on survival.
PATIENTS AND METHODS
Between May 1970 and December 1994, 1153 women with Stages I or II breast carcinoma were treated with BCT at M. D. Anderson Cancer Center. The medical records of these patients were reviewed for information on clinical and pathologic variables, disease recurrence, and survival. Patients were excluded if they had a history of a prior invasive breast carcinoma (n = 40) or another primary malignancy except basal cell carcinoma of the skin or carcinoma in situ of the cervix (n = 30) or if they had received preoperative chemotherapy (n = 40).
Before surgery patients were evaluated by a surgical oncologist, and treatment options (mastectomy vs. BCT) were discussed. Patients with multicentric disease, large tumor size relative to breast size, or contraindications to radiation therapy were considered ineligible for BCT.
Patients were treated with breast-conserving surgery with the aim of complete tumor removal with grossly normal tissue margins of 1 cm. All patients had macroscopically complete excision of the primary tumor. If surgical margins were positive (i.e., microscopically involved with tumor), a re-excision was performed in most patients, unless the margin was at the pectoral fascia. If the status of the surgical margins was unknown, a re-excision was usually performed, at the discretion of the surgeon. During this period of study, patients were treated with Levels I and II axillary lymph node dissection (ALND) at the discretion of the surgeon.
The ipsilateral breast was treated to a total dose of 45–50 Gy, delivered in 25 fractions over a 5 week period. Radiation was delivered using medial and lateral tangential fields of cobalt 60 or 6-MeV or, in rare patients with large breasts, 18-MeV photons. A 10–20 Gy boost was delivered to the tumor bed using reduced fields of electrons or an interstitial implant.
Doxorubicin-based adjuvant chemotherapy was typically used for patients with histologically proven lymph node involvement and selected patients with lymph node-negative disease. Tamoxifen was only considered for patients aged 50 years or more with estrogen receptor–positive tumors at the completion of chemotherapy or following surgery in patients who did not receive chemotherapy.
Information on tumor histology and surgical margins was obtained from the original pathology reports. For all cases in which the primary surgical procedure took place at an outside facility, pathology slides were reviewed by the pathologists at the M. D. Anderson Cancer Center.
Most patients returned for follow-up examinations at least every six months for the first five years and then annually if they remained disease free. A careful history and physical examination was performed at each visit. Mammograms were obtained six months after irradiation and then annually. Suspicious lesions were further studied with biopsy. Any new breast carcinoma involving the ipsilateral breast was considered an IBTR. In patients with IBTR, diagnostic imaging studies were routinely performed to determine distant relapse and whether systemic recurrence had occurred. In patients without IBTR, diagnostic imaging studies were performed when clinically indicated.
Patients who stopped returning for scheduled follow-up were contacted by letter or telephone call annually, either by the clinicians responsible for initial treatment or by the M. D. Anderson Department of Epidemiology.
Descriptive statistics were used to assess frequency distributions among the groups. Categorical variables and continuous variables were compared between groups using the chi-squared test. Disease-specific survival was defined as the interval between diagnosis and last follow-up. Patients who were alive or had died of a cause other than breast carcinoma were censored for analysis of disease-specific survival. Disease-specific survival probabilities were calculated by the product-limit method of Kaplan and Meier, and survival rates were compared using log-rank tests. A P value of 0.05 or less was considered to be statistically significant. The SPSS 10.1 software package (SPSS Inc., Chicago, IL) was used for statistical analysis.
Patient and Treatment Characteristics
Median follow-up of the study population was 8.4 years (range 0.25–28.9 years). The study population comprised 1043 patients. The median age at diagnosis was 50 years (range, 21–88 years). The median pathologic tumor size at diagnosis was 1.5 cm (range, 0.1–7.0 cm). Four hundred thirteen patients (40%) underwent re-excision of the tumor site at our institution. Final excision margins were positive in 35 patients (3%), negative in 774 patients (74%), and unknown in 234 patients (22%).
ALND was performed in 874 patients (84%). In the remaining patients, ALND was omitted because of comorbid conditions or older patient age (median age was 50 years in the ALND group vs. 53 years in the no-ALND group; P = 0.013). Two hundred thirty-two (27%) of the patients who underwent ALND had histologically positive lymph nodes.
All patients received radiation therapy. Seven hundred forty-one (71%) received a boost to the primary tumor site. Two hundred five patients (20%) received adjuvant hormonal therapy, and two hundred forty-four (23%) received adjuvant chemotherapy.
Systemic Recurrence after Breast-Conserving Therapy
The median follow-up time for the study population was 8.4 years, (range, 0.25–28.9 years). During follow-up, 127 patients (12%) developed an IBTR, and 184 patients (18%) developed a systemic recurrence. The median time from beginning of local treatment to systemic recurrence was 44 months (range, 1.7–287.9 months). The 5-year and 10-year systemic recurrence-free survival rates were 89% and 83%, respectively (Fig. 1).
In a univariate analysis (Table 1), the following factors were predictive of systemic recurrence: young age (50 years or younger), tumor size greater than 2 cm positive lymph nodes, omission of ALND, receipt of a radiation therapy boost, lack of tamoxifen therapy, and IBTR.
|Variable||No. of patientsa||No. (%) of patients with systemic recurrence||Pb||5 yr DSS rate, %||P|
|≤ 50 years||535||111 (21)||0.007||87||0.030|
|≥ 50 years||508||73 (14)||93|
|≤ 2 cm||726||108 (15)||< 0.001||92||0.030|
|≥ 2 cm||290||74 (26)|
|Estrogen receptor status|
|Progesterone receptor status|
|Lymph node status|
|Negative||641||80 (12)||< 0.001||93||< 0.001|
|Lumpectomy + ALND||874||142 (16)||90|
|Radiation therapy boost|
|No||916||125 (14)||< 0.001||92||0.001|
Multivariate analysis was carried out in patients who had undergone ALND. On multivariate logistic regression analysis, predictors of systemic recurrence were tumor size greater than 2 cm (vs. ≤ 2 cm; odds ratio [OR], 2.4; 95% confidence interval [CI], 1.3–4.1; P = 0.002), positive lymph nodes (vs. negative; OR, 2.1; 95% CI, 1.2–3.8; P = 0.011), lack of adjuvant tamoxifen therapy (vs. receipt of tamoxifen therapy; OR, 2.0; 95% CI, 1.1–3.9; P = 0.032), and positive margins (vs. negative; OR, 3.7; 95% CI, 1.1–12.3; P = 0.034). When IBTR was added into the multivariate model, adjuvant therapy and surgical margins were not independent predictors; however, IBTR was an independent predictor of systemic recurrence (IBTR vs. no IBTR; OR, 6.2; 95% CI, 3.1–12.3; P < 0.001).
Of the 127 patients with IBTR, 68 (54%) had an IBTR as their only event, 3 (2%) had an IBTR after a systemic recurrence, 22 (17%) had a synchronous systemic recurrence (i.e., within three months before or after diagnosis of the IBTR), and 34 (27%) had an IBTR before a systemic recurrence; thus, 102 patients had an isolated IBTR as their first event. Patients who had an IBTR as their first event had a significantly lower 10 year systemic recurrence-free survival rate than patients who had no IBTR (76% vs. 86%, P = 0.001, Fig. 2).
Disease-Specific Survival after Breast-Conserving Therapy
At last follow-up, 746 patients (72%) were alive with no evidence of disease, and 39 patients (4%) were alive with disease. The 5- and 10-year disease-specific survival rates were 94% and 87%, respectively (Fig. 3).
In a univariate analysis (Table 1), the following factors were predictive of disease-specific survival: tumor size smaller than 2 cm, negative lymph node status, axillary surgery, use of adjuvant tamoxifen, and no IBTR. When patients with unknown characteristics were excluded, positive estrogen receptor status, positive progesterone receptor status, and negative margin status were also predictive of disease-specific survival.
On multivariate analysis using the Cox proportional hazards model (Table 2), four factors were independent predictors of poor disease-specific survival: tumor size greater than 2 cm (vs. ≥ 2 cm), negative progesterone receptor status (vs. positive), positive surgical margins (vs. negative), and IBTR (vs. no IBTR). The 10 year disease-specific survival rate for patients with positive surgical margins was 77%, compared to 90% for patients with negative surgical margins (P = 0.0449, Fig. 4). The 10 year disease-specific survival rate for patients who had IBTR as a first event was 82%, compared to 90% for patients who did not develop an IBTR (P = 0.0001, Fig. 5).
|Variable||Relative risk||95% CI||P|
|Tumor size > 2 cm||2.3||1.2–4.3||0.010|
|Negative progesterone receptor status||2.7||1.4–5.1||0.003|
BCT is widely used as an alternative to mastectomy in patients with early-stage breast carcinoma. In the current study, we reviewed the records of 1043 consecutive patients with Stages I and II breast carcinoma treated with BCT and followed those patients long term. Our goals were to identify factors predictive of systemic recurrence and disease-specific survival and to determine the impact of local control on survival. We found that positive surgical margins and IBTR were predictors of systemic recurrence and disease-specific survival.
BCT and mastectomy lead to similar rates of distant disease-free survival and disease-specific survival in patients with early stage breast carcinoma.3, 16–22 The IBTR rate at eight years after BCT has previously been reported to be 5–20%,23–25 similar to the 12% rate observed in the current study. Although IBTR has been associated with an increased risk of systemic recurrence in several studies,8, 10–13 it has been proposed that the development of IBTR only indicates an increased risk for and is not a cause of distant metastases.14
The importance of local treatment in terms of outcomes after BCT is controversial. It has been proposed that breast carcinoma is a systemic disease at the onset.26 The similar survival rates for BCT and mastectomy in spite of the higher local recurrence rate with BCT suggest that the prognosis of breast carcinoma is linked to the presence or absence of occult distant metastases at presentation and not to the extent of local surgery.1 Recently, della Rovere and Benson wrote that the belief that histologically clear margins of excision are necessary to diminish local recurrence may be incorrect. If local recurrence does not impact survival, they wrote, why risk overtreating the patient by trying to prevent that recurrence?15
In the current study, positive surgical margins and IBTR were associated with an increased risk of systemic recurrence and decreased disease-specific survival. The current results regarding positive surgical margins agree with those previously reported by three other groups.27–29 Schnitt et al.27 and Voogd et al.28 reported that microscopic involvement of the surgical margin was an independent predictor of systemic recurrence. Voogd et al. also reported that the 10 year actuarial risk of systemic recurrence after BCT was 60% for patients with positive margins, compared with 29% for those with negative margins.28 DiBiase et al. recently reported significantly lower 10 year disease-free survival and overall survival rates for women with positive surgical margins compared with women with negative margins (disease-free survival 71% vs. 82%; overall survival 79% vs. 90%).29 Taken together, these results suggest that local control indeed plays an important role in outcomes after BCT.
Whether the development of IBTR is the cause of distant metastasis cannot be definitively resolved by current data. Biologically, it would be difficult to argue that cancer cells in the primary tumor can metastasize and affect survival, while cancer cells in the recurrent tumor cannot. More likely, IBTR is associated with poor survival through both mechanisms: IBTRs are indicators of more aggressive primary tumors, which have an increased likelihood of systemic spread, and the recurrent tumor can potentially also lead to secondary tumor dissemination and, therefore, added mortality. In fact, using a controversial mathematical model, Koscielny et al. found that nearly all of the excess metastases observed in patients with local recurrence had been initiated after the initial treatment.30, 31 The relationship between IBTR and distant metastasis remains to be definitively characterized. In the meantime, since IBTR may have an adverse effect on outcome, aggressive local therapy needs to be pursued to minimize the risk of IBTR.
Achieving widely negative surgical margins is one way to decrease the risk of IBTR; pathologic margin status is one of the most important factors associated with IBTR.24 The current results suggest that a margin-positive resection is an independent predictor of poor disease-specific survival. Thus, the residual breast carcinoma left at the resection margins may also be a source of systemic spread and, ultimately, disease-specific mortality. Of interest, IBTR rates after BCT have been found to vary significantly between medical centers.4 This emphasizes the importance of a skilled multidisciplinary team of surgeons, pathologists, radiologists, medical oncologists and radiation oncologists in the treatment of early-stage breast carcinoma.
A second approach to decreasing the risk of IBTR is the use of breast irradiation. Breast irradiation has been found to uniformly decrease the risk of IBTR after BCT,32–34 and this risk is further reduced with tumor bed boosts.35 Higher doses of radiation, however, are unable to provide adequate local control in patients with positive margins.29 Therefore, breast irradiation should not be used in place of adequate surgery but rather should be given to supplement definitive surgery to further decrease the risk of IBTR.
If IBTR leads to distant metastasis, one would expect radiation therapy to be associated with an increase in survival. Most studies that have found a decrease in local recurrence with irradiation after breast-conserving surgery have not found an improvement in systemic disease-free survival or overall survival.32, 33, 36 However, a Scottish trial found a trend toward fewer distant metastases in irradiated patients,34 and a Milan trial reported a survival advantage in a subgroup of node-positive patients.25 A difference in systemic-disease free survival and disease-specific survival may be observed with a larger sample size and longer follow-up. The 20 year follow-up of the National Surgical Adjuvant Breast and Bowel Project B-06 trial indicated that radiation therapy in addition to breast-conserving surgery was associated with a marginally significant decrease in deaths due to breast carcinoma (hazard ratio, 0.82; 95% CI, 0.68–0.99; P = 0.04).2 This survival advantage was partially offset by an increase in deaths from other causes.Until recently, a similar controversy surrounded the survival benefit of postmastectomy irradiation. Postmastectomy radiation therapy has now been shown to produce not only a two-thirds reduction in local recurrence but also an increase in overall survival in node-positive breast carcinoma.37, 38
Another way to minimize IBTR may be careful patient selection. Patients who have IBTR after BCT can often undergo salvage mastectomy, and patients who undergo salvage mastectomy have ultimate locoregional control rates similar to those of patients who have undergone mastectomy as initial treatment.23, 32 However, since IBTR is associated with an increased risk of systemic recurrence and decreased survival, it would be preferable to be able to predict before initial surgery which patients would develop a local recurrence with BCT but not with a mastectomy. Voogd et al. addressed this question using data pooled from the two randomized trials conducted by the European Organization for Research and Treatment of Cancer and the Danish Breast Cancer Cooperative Group.28 In patients aged 35 years and younger, the risk of local recurrence was significantly higher after BCT than after mastectomy; the 10 year actuarial rate of local recurrence was 35% after BCT and 7% after mastectomy. Voogd et al. also found that patients with an extensive intraductal component were more likely to have a local recurrence after BCT than after mastectomy.28 However, other studies have suggested that an extensive intraductal component is not prognostic of an increased risk of IBTR if adequate margins can be obtained.24, 27, 39, 40 Further study is needed to better identify clinical and biologic parameters predictive of IBTR after BCT.