Improving local control with breast-conserving therapy
A 27-year single-institution experience
Article first published online: 23 MAY 2005
Copyright © 2005 American Cancer Society
Volume 104, Issue 1, pages 20–29, 1 July 2005
How to Cite
Cabioglu, N., Hunt, K. K., Buchholz, T. A., Mirza, N., Singletary, S. E., Kuerer, H. M., Babiera, G. V., Ames, F. C., Sahin, A. A. and Meric-Bernstam, F. (2005), Improving local control with breast-conserving therapy. Cancer, 104: 20–29. doi: 10.1002/cncr.21121
- Issue published online: 17 JUN 2005
- Article first published online: 23 MAY 2005
- Manuscript Accepted: 18 FEB 2005
- Manuscript Revised: 21 JAN 2005
- Manuscript Received: 30 NOV 2004
- National Institutes of Health. Grant Number: 1K08-CA-91895-01
- breast-conserving surgery;
- local recurrence;
- hormonal therapy;
- invasive breast carcinoma;
The risk of ipsilateral breast tumor recurrence (IBTR) after breast-conserving therapy (BCT) is associated with treatment and tumor-related variables, such as surgical margin status and the use of systemic therapy, and these variables have changed over time. Correspondingly, the authors of the current study hypothesized that the contemporary multidisciplinary management of breast carcinoma would lead to an improvement in IBTR rates after BCT.
Between 1970 and 1996, 1355 patients with pathologic Stage I-II invasive breast carcinoma underwent BCT (breast-conserving surgery and adjuvant radiation therapy) at The University of Texas M. D. Anderson Cancer Center. Contemporary methods of analyzing surgical margins were in routine use by 1994. To analyze the effect of this variable and others, patient and tumor characteristics and IBTR rates in patients treated during 1994–1996 were compared with those in patients treated from 1970 to 1993.
Characteristics were similar in patients treated during 1994–1996 (n = 381) and those treated before 1994 (n = 974) except for patients aged >50 years (63.3% vs. 51.7%, P < 0.001), and patients who had a family history of breast carcinoma (37.9% vs. 30.8%, P = 0.017). Patients treated after 1994 were less likely to have positive or unknown margins (2.9 % vs. 24.1 %, P = 0.0001), more likely to receive chemotherapy (40.5% vs. 26%, P < 0.001), and more likely to receive hormonal therapy (33.3% vs. 19.4%, P < 0.001), but less likely to receive radiation boosts to the primary tumor bed (59.8% vs. 89%, P < 0.001). The 5-year cumulative IBTR rate was significantly lower among patients treated in 1994–1996 than among patients treated before 1994 (1.3% vs. 5.7%, P = 0.001) largely because of the drop in IBTR rates among patients aged ≤ 50 years (1.4 % vs. 9.1 %, P = 0.0001). On multivariate analysis, age > 50 (hazards ratio [HR] = 0.401; P = 0.0001), presence of negative surgical margins (HR = 0.574; P = 0.017), and use of adjuvant hormonal therapy (HR = 0.402; P = 0.05) were independent predictors of improved 5-year IBTR-free survival. On subgroup analysis, use of chemotherapy was associated with increased IBTR-free survival among women aged ≤ 50 years (HR = 0.383; P = 0.001). Although 5-year cumulative IBTR rates were lower among women aged > 50 years than among younger women before 1994 (2.6 % vs. 9.1%, P < 0.0001), no such difference was found in the group treated in 1994–1996 (1.2 % for age > 50 yrs vs. 1.4 % for ≤ 50 yrs, P = 0.999).
The IBTR rate after BCT appears to be declining, especially among patients < 50 years of age. However, long-term follow-up is necessary to confirm this finding. This finding may reflect changes in surgical approaches and pathologic evaluation as well as an increased use of systemic therapy. The current low incidence of IBTR with multidisciplinary management of breast carcinoma may result in more patients choosing BCT over mastectomy. Cancer 2005. © 2005 American Cancer Society.
Large, randomized trials have demonstrated that survival rates after breast-conserving therapy (BCT) for invasive breast carcinoma (i.e., breast-conserving surgery [BCS] followed by radiation therapy) are similar to those after mastectomy.1, 2 Increased breast cancer awareness and use of mammographic screening have facilitated the early diagnosis of breast carcinoma, thereby increasing the numbers of patients eligible for BCT. Concurrent successes with use of preoperative chemotherapy to control locally advanced disease or to shrink large operable tumors have further expanded the eligibility criteria for BCT to include selected patients with advanced disease who have shown a favorable response to chemotherapy.3–8
Studies with long-term follow-up have shown that the risk of ipsilateral breast tumor recurrence (IBTR) after BCT is 9% to 15% (0.5–1% per year).1, 2 This risk has been linked to patient- and tumor-related variables and was higher with positive surgical margins but lower with the use of systemic therapies such as chemotherapy and hormonal therapy.9–12 Three major changes in the management of breast carcinoma took place at our institution during the past decade that could influence the outcome in BCT. First was the transition to using detailed pathologic methods for analyzing surgical margins, which began in the early 1990s and had become routine by 1994. Second was the increase in use of chemotherapy in both adjuvant and preoperative settings. Third was the increase in use of tamoxifen for patients with hormone-receptor–positive tumors in both premenopausal and postmenopausal women. We hypothesized that these changes and others in the multidisciplinary management of breast carcinoma would result in lower rates of IBTR after BCT, and we reviewed our 27-year institutional experience with BCT to test this hypothesis.
MATERIALS AND METHODS
Between January 1970 and December 1996, 1355 patients with pathologic Stage I–II invasive breast carcinoma were identified through a database review as having undergone BCT at The University of Texas M. D. Anderson Cancer Center. Charts were reviewed to extract information on clinical and pathologic variables, including use of systemic therapy (hormones or chemotherapy) and local therapy (surgery and radiotherapy), disease recurrence, and survival. Excluded from the analysis were patients with ductal carcinoma in situ (DCIS) with or without microinvasion, patients who did not receive adjuvant radiotherapy, and patients who presented with systemic metastases at the initial diagnosis. Patients whose tumors decreased in size with the use of preoperative chemotherapy, thus allowing breast-conserving surgery, (n = 121) were included. Pathologic stage was used to select the study population to avoid interobserver variability in assignment of clinical stage. Tumor and lymph node staging were performed according to the 6th edition of the AJCC Cancer Staging Manual.13
Patients were evaluated by a multidisciplinary team comprising a surgical oncologist, a radiation oncologist, and a medical oncologist to plan definitive treatment. Generally, patients with multicentric disease or those for whom radiation therapy was contraindicated were considered ineligible for BCT. Selected patients who received preoperative chemotherapy underwent BCT according to our institutional practice as described by Chen et al.6 and as noted below.
Patients who were initially treated with BCS underwent wide local excision (grossly normal tissue margins of 1 cm) for the aim of complete tumor removal. In patients who underwent preoperative chemotherapy, the residual primary tumor was excised if a partial response to chemotherapy was obtained; if a complete clinical response was obtained, the site of the initial primary tumor, as determined by pretreatment placement of metallic markers, was excised. In all cases, excision of the primary tumor was macroscopically complete. If surgical margins were positive, reexcision was performed in most cases. If margin status was unknown or close, reexcision was performed at the discretion of the treating surgeon. Patients treated during the study period underwent Level I and II axillary lymph node dissection (ALND), with or without sentinel lymph node biopsy, at the discretion of the surgeon.
In all cases in which the primary tumor was excised at an outside facility, pathology slides were reviewed by pathologists at M. D. Anderson Cancer Center for histologic analysis and evaluation of surgical margins. Contemporary methods of analyzing surgical margins were initiated in late 1980s and performed routinely after 1994. Segmental mastectomy specimens were evaluated intraoperatively and postoperatively in a standard fashion as described previously by Schnitt14 and Sahin.15. Briefly, a multicolor inking system was used to mark the superior, inferior, lateral, medial, anterior, and posterior surfaces of the resection specimen. The specimen was then sectioned into 3–5 mm sections perpendicular to the longest axis through the tumor mass. The sections were then grossly examined for proximity of the tumor mass to the margins. If the mass appeared to approach or touch a margin, a frozen section analysis was performed to examine the questionable margin, and/or an immediate reexcision was performed at the same surgery. In nonpalpable lesions, the whole specimen was additionally radiographed to identify the targeted mammographic lesion and then the inked specimen that was sectioned into 3- to 5-mm slices was again radiographed and reviewed to determine the extent of any radiographic abnormalities and their proximity to margins. If the surgical margin was identified as positive or close intraoperatively, additional margins were excised to obtain adequate margins. Final margin status was determined by examination of permanent paraffin-embedded sections. Surgical margins were defined as positive if cancerous cells were present within < 1 mm of the inked specimen margin.
Most patients were treated with an anthracycline-containing regimen. Between 1994 and the end of the study period, a few patients were treated with paclitaxel only, delivered every 3 weeks. Adjuvant chemotherapy was offered to patients with histologic evidence of lymph node involvement and to some patients with lymph-node–negative disease. The indications for using preoperative chemotherapy were based on clinical protocols designed to compare the efficacy of two chemotherapy regimens or on physician preference for patients with operable large breast carcinoma to achieve tumor shrinkage to facilitate BCT.
Tamoxifen was considered for patients with estrogen or progesterone-receptor–positive tumors after chemotherapy or, for patients who did not receive chemotherapy, after surgery. Most of the patients to whom tamoxifen was offered were postmenopausal. However, tamoxifen was more frequently offered to premenopausal women who had estrogen or progesterone-receptor–positive tumors in the latter treatment period.
All patients underwent adjuvant external-beam radiation therapy to the ipsilateral breast with medial and lateral tangential fields. The total dose was 45–50 Gy, delivered in 25 fractions over a 5-week period. Radiation was delivered with tangential fields of cobalt 60 or 6-MeV photons or, rarely, for patients with large breasts, 18-MeV photons. A 10–20 Gy boost was delivered to the tumor bed with reduced fields of electrons or an interstitial implant at the discretion of the radiation oncologist in all patients who had received preoperative chemotherapy and in some additional patients according to the proximity of the surgical margin or other risk factors for IBTR. Radiation to the regional lymph nodes was also delivered at the discretion of the radiation oncologist.
Most patients were followed with history and physical examination at least every 6 months for the first 5 years and then, if they remained free of disease, every year thereafter. Mammograms were obtained 6 months following the completion of radiation therapy and then annually thereafter. Suspicious lesions were evaluated by additional breast imaging and biopsy. Any new carcinoma detected in the ipsilateral breast was considered an IBTR. Patients determined to have an IBTR were also evaluated with additional diagnostic imaging studies to evaluate for evidence of distant metastasis.
Descriptive statistics were used to assess frequency distributions among the groups. Because contemporary methods of margin analysis were in routine use starting in 1994, patient and tumor characteristics and treatment variables and outcome were compared for patients treated from 1994 to 1996 and for those treated before 1994. The endpoints for analysis were cumulative rates of IBTR and IBTR-free survival. All events were measured from the date of histologic diagnosis of the initial diagnostic biopsy specimen. IBTR-free survival time was defined as the interval between diagnosis of the primary tumor and diagnosis of the IBTR or the last follow-up.
The SPSS 10.1 software package (SPSS Inc., Chicago, IL) was used for statistical analyses. Categorical variables were compared between groups with the chi-square test, and continuous variables were compared with the Mann–Whitney U-test. Survival analyses were calculated by the Kaplan–Meier method, with comparisons among groups performed with two-sided log-rank tests. All tests were two-tailed, with P values of 0.05 or less considered significant. A Cox proportional hazards model was used to identify variables that were independently associated with IBTR. Variables that were significant in univariate analyses in Kaplan–Meier survival analyses were put into the multivariate Cox regression model to identify the independent factors affecting IBTR-free survival.
Patient and Tumor Characteristics
Patient and tumor characteristics are shown in Tables 1 and 2. The median age of the patients at diagnosis was 51.6 years (range, 22–88 yrs). Demographic features were similar for patients treated during 1994–1996 (n = 381) and those treated before 1994 (n = 974) except for proportions of patients who were older than 50 years (63.3% vs. 51.7%, P < 0.001) and who had a family history of breast carcinoma (37.9% vs. 30.8%, P = 0.017).
|Variable||Treatment period||P valuea|
|All patients (n = 1355) no. (%)||1970-1993 (n = 974) no. (%)||1994-1996 (n = 381) no. (%)|
|Age (yrs)||< 0.001|
|≤ 50 yrs||610 (45)||470 (48.3)||140 (36.7)|
|> 50 yrs||745 (55)||504 (51.7)||241 (63.3)|
|White||1121 (82.7)||818 (84)||303 (79.5)|
|Black||91 (6.7)||61 (6.3)||30 (7.9)|
|Hispanic||121 (8.9)||82 (8.4)||39 (10.2)|
|Other||22 (1.6)||13 (1.3)||9 (2.4)|
|Family history of breast cancer||0.017|
|Yes||393 (33)||255 (30.8)||138 (37.9)|
|No||798 (67)||572 (69.2)||226 (62.1)|
|Initial clinical disease stage||0.280|
|I||726 (62.3)||492 (60.8)||234 (65.7)|
|II||396 (34)||286 (35.4)||110 (30.9)|
|III||43 (3.7)||31 (3.8)||12 (3.4)|
|Initial clinical T size||0.316|
|T1||544 (60.9)||413 (61.6)||131 (58.5)|
|T2||325 (36.4)||236 (35.2)||89 (39.7)|
|T3||25 (2.8)||21 (3.1)||4 (1.8)|
|Radiation boost||< 0.001|
|Yes||817 (82.9)||695 (89)||122 (59.8)|
|No||168 (17.1)||86 (11)||82 (40.2)|
|Yes||407 (30.1)||253 (26)||154 (40.5)|
|No||945 (69.9)||719 (74)||226 (59.5)|
|Hormonal therapy||< 001|
|Yes||314 (23.3)||189 (19.4)||125 (33.3)|
|No||1034 (76.7)||784 (80.6)||250 (66.7)|
|All patients no. (%)||1970-1993 no. (%)||1994-1996 no. (%)||P valuea|
|Pathologic tumor size||0.972|
|≤ 2 cm (T1)||1002 (73.9)||720 (73.9)||282 (74)|
|>2 cm (T2)||353 (26.1)||254 (26.1)||99 (26)|
|Pathologic nodal status||0.504|
|Negative||935 (77.6)||645 (77.1)||290 (78.8)|
|Positive (1–3)||270 (22.4)||192 (22.9)||78 (21.2)|
|I||719 (59.5)||484 (57.9)||235 (63.2)|
|II||489 (40.5)||352 (42.1)||137 (36.8)|
|Invasive ductal ± other||1088 (95.1)||928 (95.3)||361 (94.8)|
|Invasive lobular ± other||66 (4.9)||46 (4.7)||20 (5.2)|
|Well & intermediate||657 (66.2)||427 (66.5)||230 (65.5)|
|Poorly differentiated||336 (33.8)||215 (33.5)||121 (34.5)|
|Estrogen receptor status||0.754|
|Negative||358 (35.9)||239 (36.2)||119 (35.2)|
|Positive||640 (64.1)||421 (63.8)||219 (64.8)|
|Progesterone receptor status||0.179|
|Negative||352 (40.5)||207 (38.7)||145 (43.3)|
|Positive||518 (59.5)||328 (61.3)||190 (56.7)|
|Final margin status||0.0001|
|Negative||1109 (81.8)||739 (75.9)||370 (97.1)|
|Positive||39 (2.9)||30 (3.1)||9 (2.4)|
|Unknown||207 (15.3)||205 (21)||2 (0.5)|
Initial clinical stage, clinical or pathologic tumor size, and pathologic nodal status were distributed similarly in both groups. No differences were found in any other tumor characteristic, including nuclear grade and hormone-receptor positivity, between the two groups (Table 2).
As would be expected from the increased emphasis on pathologic evaluation of surgical margins since the early 1990s, the number of patients whose final surgical margin status was positive or unknown after BCT dropped dramatically after 1994 (2.9% for 1994–1996 vs. 24.1% before 1994, P= 0.0001) (Table 1). Patients treated during the later period were more likely to receive chemotherapy (40.5% from 1994–1996 vs. 26% from before 1994, P < 0.001), and more likely to receive hormonal therapy (33.3% vs. 19.4%, P < 0.001), but less likely to receive radiation boosts to the primary tumor bed (59.8% vs. 89%, P < 0.001) than those treated before 1994 (Table 1). Use of hormonal therapy increased in 1994–1996 regardless of patient age (13.1% vs. 7.7 for age ≤50 yrs, P = 0.048; and 45% vs. 30.4% for age > 50 yrs, P < 0.001).
At a median follow-up time of 7.4 years, the cumulative 5-year IBTR rate was 4.5 % (61 of 1355 patients) (Table 3). Notably, the cumulative 5-year IBTR rates declined throughout the 27-year period: 7.1% for patients treated during 1970–1984; 6.4% for those treated 1985–1989; 4.2% for 1990–1993; and 1.3% for patients treated during 1994–1996. Patients treated during 1994–1996 had a significantly lower 5-year cumulative IBTR rate (1.3%) than did patients treated before 1994 (5.7%) (P = 0.001). The 5-year IBTR-free survival rate also improved from 94% before 1994 to 98% from 1994 to 1996 (P = 0.004).
|Total % (n = 1355)||1970-1993 % (n = 974)||1994-1996 % (n = 381)||P value|
|Median yrs (range, min-max)||7.4 (0.05–28.4)||8.8 (0.4–28.4)||6.0 (0.05–9.3)||< 0.001|
|5-year IBTR-free survival rate||95||94||98||0.004|
|5-year cumulative||4.5 (61 of 1355)||5.7 (56 of 974)||1.3 (5 of 381)||< 0.001|
|Patients age ≤ 50||7.4 (45 of 610)||9.1 (43 of 470)||1.4 (2 of 140)||0.001|
|Age < 35||11.2 (13 of 116)||11.3 (11 of 97)||10.5 (2 of 19)||0.999|
|Age 35–50||6.5 (32 of 494)||8.6 (32 of 373)||0 (0 of 121)||0.0001|
|Patients age > 50||2.1 (16 of 745)||2.6 (13 of 504)||1.2 (3 of 241)||0.291|
|Age 50–65||2.6 (13 of 495)||3.5 (12 of 345)||0.7 (1 of 150)||0.122|
|Age > 65||1.2 (3 of 250)||0.6 (1 of 159)||2.1 (2 of 91)||0.300|
|5-year DSS rate||95||93||98||0.007|
In our previous work,9 we found that age ≤ 50 years was an independent predictor of locoregional recurrence. Therefore, we reanalyzed the IBTR rates after stratifying the patients by age (≤ 50 yrs vs. > 50 yrs) to ensure that the improvement in IBTR rate during the later period was not solely attributable to the increase in the proportion of patients older than 50 years during that period. Interestingly, the reduction in IBTR rate was attributable to a decline in IBTR rates among patients ≤ 50 years old (1.4% for those treated 1994–1996 vs. 9.1% for those treated 1970–1993, P = 0.001). Among patients older than 50 years, the 5-year cumulative IBTR rate also proved to be lower during the latter period, but this apparent difference was not statistically significant (1.2% for those treated 1994–1996 vs. 2.6% for those treated 1970–1993, P = 0.291).
Factors Affecting IBTR-Free Survival
Consistent with our previous findings,9 in this study the 5-year IBTR-free survival rate for patients aged 50 years or younger (92%) was significantly lower than that for patients older than 50 years (98%) (P < 0.001) (Table 4). On multivariate analysis, age > 50 (hazard rate [HR] [vs. age ≤ 50] = 0.401; 95% confidence interval [CI], 0.248–0.649; P = 0.0001), presence of negative surgical margins (HR [vs. positive or unknown margins] = 0.574; 95% CI, 0.364–0.906; P = 0.017) and use of adjuvant hormonal therapy (HR [vs. no hormonal therapy] = 0.402; 95% CI, 0.159–1.019; P = 0.05) were independent predictors of improved IBTR-free survival. Notably, among the patients treated from1970 to 1993, patients older than 50 years had a significantly lower 5-year cumulative IBTR rate than did patients ≤ 50 years (2.6% vs. 9.1%, P < 0.0001). However, because of the drop in IBTR rates among younger patients during the later period, no significant difference was found in IBTR rates for those treated during 1994–1996 according to age > 50 years (1.2%) vs. ≤ 50 years (1.4%) (P = 0.999).
|Characteristic||Total 5-Year IBTR rate, %||P value||Age ≤ 50||Age > 50|
|5-Year IBTR rate, %||P Value||5-Year IBTR rate, %||P value|
|Age (yrs)||< 0.001||—||—|
|≤ 2 cm||96||92||99|
|> 2 cm||93||90||95|
|Invasive lobular ± other||95||92||98|
|Lymph node involvement||0.203||0.114||0.602|
|Positive or unknown||92||88||95|
|Radiation therapy boost||0.018||0.074||0.34|
As we had shown a significant drop in IBTR rates among younger patients, we then evaluated the determinants of 5-year-IBTR free survival separately for patients ≤ 50 years versus those > 50 years (Table 4). Among patients older than 50 years, being white, having a tumor ≤ 2 cm, having negative surgical margins, and receiving adjuvant hormonal therapy were determinants of improved 5-year IBTR-free survival (Table 4). On multivariate analysis, white race (HR [vs. other] = 0.285; 95% CI, 0.125–0.649; P = 0.003), tumor size ≤ 2 cm (HR [vs. > 2 cm] = 0.434; 95% CI, 0.210–0.899; P = 0.025), and presence of negative surgical margins (HR [vs. positive or unknown margins] = 0.433; 95% CI, 0.205–0.915; P = 0.028) were independent predictors of improved 5-year IBTR-free survival.
Among patients aged 50 years or younger, the presence of negative margins and use of chemotherapy (Fig. 1) were found to be associated with increased 5-year IBTR-free survival rates. However, on multivariate Cox regression analysis, use of chemotherapy was the only significant independent factor for improved IBTR-free survival (HR [vs. no chemotherapy] = 0.383; 95% CI, 0.221–0.663; P = 0.001).
IBTR in Patients Receiving Preoperative Chemotherapy
A larger proportion (13.4%) of patients treated in 1994–1996 underwent BCT after preoperative chemotherapy than the proportion (7.2%) of patients treated before 1994 (P = 0.001). Having excluded these patients from the study would have resulted in a significant increase in the proportion of patients with smaller breast carcinomas in the later period. Therefore, to ensure that the improvement in IBTR rate during the later period was not biased by the increase in the proportion of patients with smaller breast tumors, patients receiving preoperative chemotherapy who became pathologic Stage I and II before breast-conserving surgery were included in the study. Initial clinical stage distribution among patients receiving preoperative chemotherapy was as follows: Stage I, 2.5%; Stage II, 63.6%; Stage III, 31.4%. The cumulative 5-year IBTR rate was 4.1% (5 of 121). The cumulative 5-year IBTR rates declined from 5.7% (1988–1994) to 2.0% (1994–1996) in this group.
BCT is widely used as an alternative to mastectomy for patients with breast carcinoma and is the preferred treatment for patients with unifocal, early stage breast carcinoma. Our review of a 27-year institutional experience with BCT for women with pathologic Stages I and II invasive breast carcinoma showed that a multidisciplinary approach to BCT provided excellent ipsilateral breast tumor recurrence-free survival during the past decade. Our goal in this study was to identify factors associated with the improvement in IBTR rates in patients treated after 1994 by performing a descriptive, comparative analysis. Patients in the group treated from 1994 to 1996 were more likely to be older than 50 years, and age > 50 years was found in multivariate analyses to be associated with increased IBTR-free survival rates. To account for the possible confounding effects of this variable in the improvement in IBTR rates after 1994, we reanalyzed the data after stratifying the cases by age (≤ 50 yrs vs. > 50 yrs). We found that the improvement in local control after 1994 was attributable to a decrease in incidence of IBTR among patients ≤ 50 years old.
Young age has been found to be a risk factor for the development of local recurrence after BCT in some studies10, 11, 16–24 but not in others.25–27 A retrospective review of three randomized trials comparing BCT with mastectomy for predictors of local recurrence showed that young age was a factor.17–19 Leong et al.11 recently reported age < 35 years to be associated with a high risk of IBTR, and Stotter et al.28 and Mate et al.29 found that the risk of local recurrence was significantly increased among patients < 50 years of age compared with older patients. However, other studies have found no difference in local recurrence rates when patients younger than 50 years were compared with those older than 50 years.30 Our previous study9 and the study reported here both demonstrated that age > 50 years was independently associated with decreased risk of IBTR. However, we also demonstrate here that the IBTR rate among patients aged ≤ 50 years improved dramatically after 1994 and became similar to the IBTR rate among patients aged > 50 years. This finding could be a reflection of the increasing use of chemotherapy, as has been shown elsewhere.23, 31 Use of chemotherapy was found to be the only significant factor associated with increased IBTR-free survival rates in patients ≤ 50 years of age. These results indicate that age ≤ 50 years should not be considered a contraindication for BCT as long as appropriate local and systemic therapies are used. In our latter time period, we had too few patients (n = 19) to make a definite conclusion of impact of age < 35 years on IBTR rates. However, we did not see a difference in IBTR rates among women < 35 years of age treated between1994 and 1996 (10.5%) compared with women treated before 1994 (11.3%). Further study is needed to determine whether women < 35 years of age are indeed a high-risk group for IBTR despite contemporary multidisciplinary management.
Our findings here reflect four known major changes in the management strategy for breast conservation between 1994 and 1996 at our institution. First, the rate of positive or unknown margins has decreased significantly because of the routine use of detailed intraoperative and postoperative margin analysis. Second, the use of chemotherapy has increased both in adjuvant and preoperative settings; e.g., more patients with large operable or locally advanced breast carcinoma received preoperative chemotherapy and were subsequently treated with BCT. Third, the use of tamoxifen by women with hormone-receptor–positive tumors has increased in both premenopausal and postmenopausal patients. Finally, the delivery of a radiation boost to the tumor bed after whole-breast radiation declined during the study period. Our results mostly parallel those recently reported by Pass et al.32 and suggest that changes in the management of breast carcinoma may lead to further declines in IBTR rates after BCT.
Several studies have reported a reduction in local recurrence rates for patients undergoing BCT followed by adjuvant chemotherapy.33–35 The National Surgical Adjuvant Breast and Bowel Project (NSABP) B-06 trial compared mastectomy, BCS, and BCS with radiotherapy for Stage I and II breast carcinoma and included adjuvant chemotherapy for node-positive patients. Use of chemotherapy led to a marked reduction in local recurrence rates.33 Other studies also demonstrated that adjuvant chemotherapy significantly improved long-term local control in patients with node-negative disease who underwent BCT.34, 35 Freedman et al.36 further showed that for patients who underwent BCT with close or positive surgical margins, IBTR could be delayed by adjuvant systemic therapy. We demonstrated here that use of chemotherapy was significantly associated with improved 5-year IBTR-free survival for patients aged ≤ 50 years. Perez et al.10 recently reported similar findings: Among patients with T1 tumors, the IBTR rate was 3.7% in women aged ≤ 40 years who received chemotherapy and 15.6% for those who did not receive chemotherapy. All of these results suggest that receiving the most effective systemic therapy can reduce the risk of IBTR in patients who undergo BCT.
The risk-reducing effect of tamoxifen on IBTR of invasive breast carcinoma has been well established in large randomized trials, including the NSABP B-14 and NSABP B-21 trials.37, 38 Similarly, patients in our study who received adjuvant hormonal therapy were 60% less likely to develop IBTR than those who were not given hormonal therapy, a finding concordant with previous reports.9, 10 After 1994, the use of adjuvant tamoxifen, for premenopausal as well as postmenopausal women, has significantly increased at our institution; that increase could have contributed to the decline in IBTR rates reported in the current study. However, we were unable to show a significant improvement in IBTR rates among patients older than 50 years. At M. D. Anderson, IBTR rates in this population have remained < 5% from 1970 to 1996, and those rates have not been significantly affected by the institution-wide changes in the management of breast carcinoma. However, recent results from large randomized trials suggest that aromatase inhibitors, used alone or after tamoxifen, may be more effective than tamoxifen alone in improving disease-free survival for postmenopausal women with hormone-receptor–positive tumors.39–41 It will be interesting to see whether the increasing use of aromatase inhibitors as adjuvant hormonal therapy will lead to further declines in IBTR rates among postmenopausal women with invasive breast carcinoma undergoing BCT. This hypothesis is already being tested prospectively in the NSABP B-35 trial which includes patients with DCIS who are undergoing BCT.
Positive histologic margins have been shown to increase the risk of local recurrence and decrease disease-specific survival after BCT.11, 42, 43, 44 Consistent with these reports, in the study reported here, the presence of negative margins was associated with improved IBTR-free survival in both univariate and multivariate analyses. Therefore, our multidisciplinary emphasis on improving margin status may have contributed to the declining IBTR rates in our study. We showed that the rates of unknown or positive margins declined from approximately from 24% in our earlier time period, to less than 3% after 1994, which likely reflects the establishment of contemporary intraoperative margin analysis in routine use at our institution. We expect that our current clinical practice of reexcising not only positive margins but also unknown and close (< 2 mm) margins whenever feasible not only decreased the positive margin rate but also led us to achieve more widely negative margins in the later period of this study. However, because the width of the margin has not been routinely assessed in the past, the influence of wider margins on IBTR rate could not be specifically studied.
Whether patients undergoing BCS should be given a boost radiation dose to the tumor bed in addition to radiation treatment of the entire breast is controversial. Some studies, including NSABP B-06, have reported excellent results without the boost for patients with negative margins after BCS.1, 45–47 However, a prospective randomized trial48 of patients with negative margins showed that the addition of a boost dose to whole-breast irradiation resulted in a 20% reduction in the local recurrence rate at 5 years (4.5% for the no-boost group vs 3.6% for the boost group, P = 0.044). In the study reported here, use of a radiation boost significantly declined for patients treated from 1994 to 1996. The most common indication for using a boost during that period was the presence of close, positive, or unknown margins, as has been reported by other studies.45, 46 The use of a boost had no effect on IBTR-free survival rates in multivariate analyses. However, our study is limited because of its retrospective nature and its small sample size compared with the sample size of a prospective trial reported by Romestaing et al.48 Because Romestaing's randomized trial showed a benefit in improving local control, our contemporary approach changed to use a boost routinely in every patient undergoing breast conservation after 1997.
The major limitation of our study is its short follow-up. As expected, the median follow-up in the patients treated before 1994 was longer than that in the “more contemporary” 1994–1996 cohort (8.8 yrs vs. 6.0 yrs). To minimize the bias introduced from this discrepancy in follow-up, we evaluated the short-term (5 yrs) IBTR rate in the two study populations. Of note, in the NSABP B-06 trial, the cumulative 20-year IBTR rate was 39.2% among women who underwent BCS without irradiation, and 73.2% of these events occurred within the first 5 years after surgery.1 However, in the same study, the 20-year IBTR rate among women treated with BCS followed by breast irradiation was 14.3%, but only 39.7% of recurrences were detected within the first 5 years. Although IBTRs that occur later are more likely to represent new ipsilateral primary tumors rather than true recurrences, we cannot exclude that we are underestimating the occurrence of true local recurrences because of our short follow-up. Further, it is possible that the increased use of systemic therapy in the later time period in our study simply led to a delay in the appearence of IBTRs, rather than an actual decrease in development of IBTRs. Longer follow-up is needed to exclude this possibility.
In summary, our study suggests that there is a substantial decline in the incidence of IBTR after BCT and that this is largely attributable to a decline in IBTR among patients aged < 50 years at diagnosis. The reasons for this finding are likely to be multifactorial, reflecting changes in diagnostic imaging, surgical techniques, pathologic evaluation, and, most importantly, the increased use of systemic therapy. Our findings have significant clinical implications because the often-quoted in-breast recurrence rates of 9–15% may no longer be applicable to patient counseling. With multidisciplinary management of breast carcinoma, the low current risk of IBTR is likely to result in more patients choosing BCT over mastectomy.
- 14Processing and evaluation of breast excision specimens. A clinically oriented approach. Am J Clin Pathol. 1992; 981: 25–137., .
- 15Surgical margin evaluation in patients treated with breast-conserving therapy. In: SingletarySE, RobbGL, HortobagyiGN, editors. Advanced therapy of breast disease. 2nd ed. Ontario: BC Decker Inc, 2004: 341–348..
- 23High local recurrence risk after breast conserving therapy in node negative premenopausal breast cancer patients is greatly reduced by one course of perioperative chemotherapy: a European Organization for Research and Treatment of Cancer Breast Cancer Cooperative Group Study. J Clin Oncol. 2000; 18: 1075–1083., , , et al.
- 24Factors associated with breast recurrence in young females treated with breast conservation therapy: how young is young. Int J Radiat Oncol Biol Phys. 2003; 57(2 Suppl ): S356–357., , , et al.
- 27Factors influencing local relapse and survival and results of salvage treatment after breast-conserving therapy in operable breast cancer: EORCT trial 10801, breast conversation compared with mastectomy in TNM stage I and II breast cancer. Eur J Cancer. 1992; 28A: 801–805., , , et al.
- 36Patients 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.
- 39The ATAC (Arimidex, Tamoxifen alone or in Combination) Trialists' Group. Anastrozole alone or in combination with tamoxifen versus tamoxifen versus tamoxifen alone for adjuvant treatment of postmenopausal women with early breast cancer: first results of the ATAC randomised trial. Lancet. 2002; 359: 2131–2139.