We compared the oncologic outcomes of breast-conserving surgery plus radiation therapy (BCS+RT) and modified radical mastectomy (MRM) under anthracycline plus taxane-based (AT) regimens and investigated the role of adjuvant radiation therapy (RT) in patients with pathologic N1 (pN1) breast cancer treated by mastectomy.
We retrospectively reviewed the medical records of 2,011 patients with pN1 breast cancer who underwent BCS+RT or MRM alone at 12 institutions between January 2006 and December 2010. Two-to-one propensity score matching was performed for balances in variables between the groups.
The median follow-up duration for the total cohort was 69 months (range, 1 to 114 months). After propensity score matching, 1,074 patients (676 in the BCS+RT group and 398 in the MRM-alone group) were analyzed finally. The overall survival, disease-free survival, locoregional failure-free survival, and regional failure-free survival (RFFS) curves of the BCS+RT group vs. MRM-alone group were not significantly different. The subgroup analysis revealed that in the group with both lymphovascular invasion (LVI) and histologic grade (HG) III, the BCS+RT showed significantly superior RFFS (p=0.008). Lymphedema (p=0.007) and radiation pneumonitis (p=0.031) occurred more frequently in the BCS+RT group than in the MRM-alone group, significantly.
There are no differences in oncologic outcomes between BCS+RT and MRM-alone groups under the AT chemotherapy regimens for pN1 breast cancer. However, BCS+RT group showed superior RFFS to MRM-alone group in the patients with LVI and HG III. Adjuvant RT might be considerable for pN1 breast cancer patients with LVI and HG III.
Randomized trials comparing breast-conservation surgery (BCS) plus adjuvant radiation therapy (RT) with mastectomy alone in early-stage breast cancer have shown no differences in oncologic outcomes [
Anthracycline plus taxane-based (AT) chemotherapy is the currently recommended adjuvant treatment in the breast cancer, especially with nodal metastasis [
We retrospectively reviewed the medical records of 2,011 patients with pathologic N1 breast cancer who underwent BCS+RT or MRM alone at 12 institutions in South Korea between Jan 2006 and Dec 2010.
Information was obtained from medical records regarding pathologic tumor features, including molecular subtype, tumor size, resection margin, lymphovascular invasion (LVI), nuclear grade (NG), histologic grade (HG), number of lymph nodes (LN) with metastasis, and extracapsular extension (ECE). Patients were excluded because of non-AT chemotherapy (n=47), BCS without adjuvant RT (n=37) and PMRT (n=157), and insufficient medical records (n=296). Finally, 1,474 patients were included in the analysis.
BCS+RT was performed for 1,047 patients (71.0%). The median dose of adjuvant RT on the whole breast was 50 Gy (range, 45 to 50.4 Gy). Tumor bed boost with median dose of 10 Gy (range, 5 to 16 Gy) was applied to 1,026 patients. RT to the supraclavicular fossa with a median dose of 50 Gy (range, 45 to 50.4 Gy) was performed in 320 patients (30.6%), and 37 (3.5%) of them received RT to the internal mammary area with a median dose of 50.4 Gy (range, 45 to 50.4 Gy). RT plans followed the general principles of RT to the whole breast. Median number of axillary LN dissections performed was 16 (range, 1 to 47). Most patients (99.8%) were treated with an AC (adriamycin and cyclophosphamide) plus T (taxane) regimen, while others were administrated the FAC (fluorouracil, adriamycin, and cytoxan) plus T regimen (0.1%) or EC (epirubicin and cyclophosphamide) plus T regimen (0.1%).
MRM alone was performed for 427 patients (29.0%). All patients in the MRM-alone group received a median of 19 axillary LN dissections (range, 2 to 43). AC plus T was administered to most patients (99.8%) in the MRM group. Other patients received the FEC (fluorouracil, epirubicin, and cytoxan) plus T regimen (0.2%). None of this group underwent adjuvant RT.
OS, disease-free survival (DFS), locoregional failure-free survival (LRFFS), and regional failure-free survival (RFFS) were defined as the interval from surgery to death, cancer recurrence, locoregional recurrence, and regio-nal recurrence, respectively. The chi-square test or Fisher exact test was used to compare patient characteristics and patterns of failure between the BCS+RT and MRM-alone group. The Kaplan-Meier method was used to estimate survival curves. Log-rank tests were performed to compare survival between groups for various variables. Cox regression analysis was chosen for multivariate analysis to determine the independent prognostic factors for outcomes. A two-sided p-value of < 0.05 was considered statistically significant. Statistical analyses were performed with SPSS ver. 22.0 (IBM Corp., Armonk, NY). Analyses were also performed for subgroups defined according to the number of risk factors identified as significant in multivariate analyses for DFS.
Two-to-one propensity score matching was performed to eliminate imbalances in variables between the two treatment groups. Matching variables were age, menopause status, site, pathology, pathologic T staging, number of LNs with metastasis, LVI, NG, HG, molecular subtype, and ECE. R Statistical Software ver. 3.2.3 (The R foundation for Statistical Analyses, Vienna, Austria) was utilized in propensity score matching.
This study was approved by the Institutional Review Board (IRB) of each hospital and performed in accordance with the principles of the Declaration of Helsinki. Each IRB approved a waiver of informed consent.
The median follow-up duration for the total cohort was 69 months (range, 1 to 114 months). The 5-year rates of OS, DFS, LRFFS, and RFFS of the total cohort were 98.0%, 92.4%, 97.0%, and 97.6%, respectively. Among the total cohort, there were significant differences in age, menopausal status, pathology, pathologic T category, LVI, NG, HG, and ECE between the BCS+RT and MRM-alone groups. After propensity score matching, a total of 1,074 patients (676 in the BCS+RT group and 398 in the MRM-alone group) were included for analysis. The patient characteristics are summarized in
Among 1,074 patients, 92 (8.6%) experienced disease recurrence. The patterns of first failure were not significantly different between the groups (
Three subgroups were determined according to the number of risk factors (positive LVI and HG of III) identified in multivariate analysis for DFS and RFFS: the low risk group (n=346) had no risk factors, the intermediate risk group (n=512) had one risk factor, and the high risk group (n=216) had two risk factors. For the high risk group, BCS+RT showed significantly superior RFFS (99.5% vs. 96.8%, p=0.008) (
Lymphedema and radiation pneumonitis occurred more frequently in the BCS+RT group than in the MRM-alone group (
Randomized trials comparing BCS+RT with mastectomy in early-stage breast cancer have shown comparable oncologic outcomes by long-term follow-up data [
The oncologic outcomes were not significantly different between treatment modalities. Among population-based studies, two reported the outcome of the pN1 subgroup. Chen et al. [
The role of PMRT is also a controversial issue for pN1 non-metastatic breast cancer. The DBCG 82 B&C trial and EBCTCG revealed that PMRT reduced mortality rates regardless of the number of LN metastases [
There are several limitations in the current study. First, there is inevitable selection bias due to its retrospective nature. Although the propensity score matching method was utilized to balance treatment groups, some variables that were excluded from score matching might have been unevenly distributed. The pathologic data were collected from several different institutions, so bias regarding pathologic information also could not be excluded. Finally, the relatively short follow-up duration of the current study is also a limitation. The median follow-up duration of the total cohort was 69 months, which is not enough to detect late recurrence and treatment-related sequelae [
In conclusion, there are no differences in OS, DFS, LRFFS, and RFFS between BCS+RT and MRM alone under the AT chemotherapy regimen for patients with pN1 non-breast cancer. However, BCS+RT showed superior RFFS in patients with LVI and HG III, implying that adjuvant RT potentially has a role in reducing regional recurrence. Therefore, PMRT might be considerable for patients with LVI and HG 3.
Conflict of interest relevant to this article was not reported.
This study was supported by a grant from the National R&D Program for Cancer Control, Ministry of Health & Welfare, Republic of Korea (1720170).
Survival curves according to treatment group. Overall survival (OS) (A), disease-free survival (DFS) (B), locoregional failure-free survival (LFFS) (C), and regional failure-free survival (RFFS) (D). BCS+RT, breast conserving surgery plus radiation therapy; MRM, modified radical mastectomy.
Subgroup analyses according to risk group for overall survival (OS) (A), disease-free survival (DFS) (B), and regional failure-free survival (RFFS) (C). BCS+RT, breast conserving surgery plus radiation therapy; MRM, modified radical mastectomy; NA, not applicable, a)Hazard ratio was not calculated for RFFS in the low risk group because there was no regional failure in that subgroup.
Baseline characteristics of study patients
Characteristic | Before matching |
After matching |
||||
---|---|---|---|---|---|---|
BCS+RT (n=1,047) | MRM alone (n=427) | p-value | BCS+RT (n=676) | MRM alone (n=398) | p-value | |
47.4±8.6 | 48.9±9.4 | 0.004 | 48.1±8.9 | 48.7±9.3 | 0.319 | |
Premenopause | 676 (64.6) | 251 (58.8) | 0.043 | 415 (61.4) | 238 (59.8) | 0.604 |
Postmenopause | 371 (35.4) | 176 (41.2) | 261 (38.6) | 160 (40.2) | ||
Left | 504 (48.1) | 217 (50.8) | 0.380 | 330 (48.8) | 199 (50.0) | 0.732 |
Right | 543 (51.9) | 210 (49.2) | 346 (51.2) | 199 (50.0) | ||
IDC | 984 (94.0) | 413 (96.7) | 0.044 | 648 (95.9) | 384 (96.5) | 0.601 |
Non-IDC | 63 (6.0) | 14 (3.3) | 28 (4.1) | 14 (3.5) | ||
1 | 507 (48.4) | 160 (37.5) | 0.000 | 280 (41.4) | 160 (40.2) | 0.897 |
2 | 532 (50.8) | 265 (62.1) | 393 (58.1) | 236 (59.3) | ||
3 | 8 (0.8) | 2 (0.5) | 3 (0.4) | 2 (0.5) | ||
1 | 603 (57.6) | 231 (54.1) | 0.070 | 381 (56.4) | 218 (54.8) | 0.722 |
2-3 | 444 (42.4) | 196 (45.9) | 295 (43.6) | 180 (45.2) | ||
SLNB only | 78 (7.4) | 24 (5.6) | 0.258 | 46 (6.8) | 22 (5.5) | 0.438 |
ALND | 969 (92.6) | 403 (94.4) | 630 (93.2) | 376 (94.5) | ||
≤ 0.1 | 506 (48.3) | 233 (54.6) | 0.034 | 338 (50.0) | 219 (55.0) | 0.101 |
> 0.1 | 541 (51.7) | 194 (45.4) | 338 (50.0) | 179 (45.0) | ||
No | 370 (35.3) | 273 (63.9) | 0.000 | 397 (58.7) | 244 (61.3) | 0.068 |
Yes | 677 (64.7) | 154 (36.1) | 279 (41.3) | 154 (38.7) | ||
1-2 | 599 (57.2) | 200 (46.8) | 0.000 | 324 (47.9) | 200 (50.3) | 0.412 |
3 | 448 (42.8) | 227 (53.2) | 352 (52.1) | 198 (49.7) | ||
I-II | 642 (61.3) | 216 (50.6) | 0.000 | 301 (44.5) | 185 (46.5) | 0.482 |
III | 405 (38.7) | 211 (49.4) | 375 (55.5) | 213 (53.5) | ||
Luminal A | 459 (43.8) | 192 (45.0) | 0.736 | 375 (55.4) | 216 (54.3) | 0.736 |
Non-luminal A | 588 (56.2) | 235 (55.0) | 301 (44.6) | 182 (45.7) | ||
No | 578 (55.2) | 267 (62.5) | 0.012 | 382 (56.5) | 239 (60.1) | 0.247 |
Yes | 469 (44.8) | 160 (37.5) | 294 (43.5) | 159 (39.9) |
Values are presented as mean±standard deviation or number (%). BCS, breast conserving surgery; RT, radiation therapy; MRM, modified radical mastectomy; IDC, invasive ductal carcinoma; LN, lymph node; SLNB, sentinel lymph node biopsy; ALND, axillary lymph node dissection; LVI, lymphovascular invasion; ECE, extracapsular extension.
Patterns of the first failure according to field of radiotherapy
Sites of the first failure | BCS+RT (n=676) | MRM alone (n=398) | p-value |
---|---|---|---|
10 (1.5) | 6 (1.5) | 0.971 | |
Local only | 5 (0.7) | 2 (0.5) | 1.000 |
Regional only | 5 (0.7) | 4 (1.0) | 0.733 |
28 (4.1) | 23 (5.8) | 0.223 | |
12 (1.8) | 13 (3.3) | 0.118 | |
50 (7.4) | 42 (10.6) | 0.074 |
Values are presented as number (%). BCS, breast conserving surgery; RT, radiation therapy; MRM, modified radical mastectomy.
Univariate and multivariate analysis (OS)
Characteristic | Univariate analysis |
Multivariate analysis |
||
---|---|---|---|---|
5-Year OS rate (%) | p-value | HR (95% CI) | p-value | |
≤ 55 | 97.7 | 0.828 | - | - |
> 55 | 97.5 | - | ||
Premenopause | 98.6 | 0.072 | 0.592 (0.295-1.190) | 0.141 |
Postmenopause | 96.0 | - | ||
Left | 97.3 | 0.737 | - | - |
Right | 97.9 | - | ||
IDC | 97.6 | 0.725 | - | - |
Non-IDC | 97.7 | - | ||
1 | 97.9 | 0.731 | - | - |
2-3 | 97.4 | - | ||
1 | 97.8 | 0.443 | - | - |
2 | 97.4 | - | ||
No | 98.0 | 0.342 | - | - |
Yes | 97.2 | - | ||
1-2 | 98.9 | 0.002 | 0.544 (0.174-1.704) | 0.296 |
3 | 96.2 | - | ||
I-II | 99.0 | 0.004 | 0.773 (0.269-2.220) | 0.632 |
III | 95.9 | - | ||
Luminal A | 99.1 | 0.001 | 0.327 (0.129-0.825) | 0.018 |
Non-luminal A | 96.3 | - | ||
No | 97.6 | 0.771 | - | - |
Yes | 97.7 | - | ||
BCS+RT | 98.6 | 0.088 | 0.597 (0.297-1.200) | 0.147 |
MRM alone | 96.1 | - |
OS, overall survival; HR, hazard ratio; CI, confidence interval; IDC, invasive ductal carcinoma; LN, lymph node; LVI, lymphovascular invasion; ECE, extracapsular extension; BCS, breast conserving surgery; RT, radiation therapy; MRM; modified radical mastectomy.
Univariate and multivariate analysis (DFS)
Characteristic | Univariate analysis |
Multivariate analysis |
||
---|---|---|---|---|
5-Year DFS rate (%) | p-value | HR (95% CI) | p-value | |
≤ 55 | 91.9 | 0.987 | - | - |
> 55 | 92.3 | - | ||
Premenopause | 92.5 | 0.248 | - | - |
Postmenopause | 91.0 | - | ||
Left | 90.7 | 0.171 | 1.239 (0.818-1.875) | 0.311 |
Right | 93.2 | - | ||
IDC | 91.9 | 0.598 | - | - |
Non-IDC | 93.2 | - | ||
1 | 95.0 | 0.001 | 0.665 (0.415-1.064) | 0.089 |
2-3 | 89.6 | - | ||
1 | 93.2 | 0.126 | 0.823 (0.545-1.244) | 0.356 |
2-3 | 90.4 | - | ||
No | 94.6 | < 0.001 | 0.472 (0.304-0.731) | 0.001 |
Yes | 88.7 | - | ||
1-2 | 95.4 | < 0.001 | 0.715 (0.389-1.314) | 0.280 |
3 | 88.0 | - | ||
I-II | 96.4 | < 0.001 | 0.254 (0.134-0.481) | 0.001 |
III | 86.1 | - | ||
Luminal A | 94.4 | 0.024 | 0.849 (0.544-1.326) | 0.472 |
Non-luminal A | 90.0 | - | ||
No | 91.8 | 0.771 | - | - |
Yes | 92.0 | - | ||
BCS+RT | 93.3 | 0.107 | 0.702 (0.463-1.065) | 0.096 |
MRM alone | 89.7 | - |
DFS, disease-free survival; HR, hazard ratio; CI, confidence interval; IDC, invasive ductal carcinoma; LN, lymph node; LVI, lymphovascular invasion; ECE, extracapsular extension; BCS, breast conserving surgery; RT, radiation therapy; MRM; modified radical mastectomy.
Univariate and multivariate analysis (RFFS)
Characteristic | Univariate analysis |
Multivariate analysis |
||
---|---|---|---|---|
5-Year RFFS rate (%) | p-value | HR (95% CI) | p-value | |
≤ 55 | 97.2 | 0.915 | - | - |
> 55 | 98.0 | - | ||
Premenopause | 97.3 | 0.571 | - | - |
Postmenopause | 97.4 | - | ||
Left | 96.9 | 0.571 | - | - |
Right | 97.8 | - | ||
IDC | 97.2 | 0.218 | - | - |
Non-IDC | 100 | - | ||
1 | 98.2 | 0.120 | 0.890 (0.417-1.900) | 0.764 |
2-3 | 96.7 | - | ||
1 | 97.1 | 0.638 | - | - |
2-3 | 97.7 | - | ||
No | 98.6 | 0.004 | 0.340 (0.159-0.726) | 0.005 |
Yes | 95.9 | - | ||
1-2 | 98.8 | 0.021 | 0.527 (0.205-1.356) | 0.184 |
3 | 95.7 | - | ||
I-II | 99.4 | < 0.001 | 0.138 (0.047-0.408) | < 0.001 |
III | 94.6 | - | ||
Luminal A | 98.5 | 0.193 | 0.902 (0.431-1.887) | 0.785 |
Non-luminal A | 96.4 | - | ||
No | 97.3 | 0.503 | - | - |
Yes | 97.3 | - | ||
BCS+RT | 98.3 | 0.102 | 0.523 (0.262-1.044) | 0.066 |
MRM | 95.7 | - |
RFFS, regional failure-free survival; HR, hazard ratio; CI, confidence interval; IDC, invasive ductal carcinoma; LN, lymph node; LVI, lymphovascular invasion; ECE, extracapsular extension; BCS, breast conserving; RT, radiation therapy; MRM; modified radical mastectomy.
Treatment-related toxicities
Morbidity | BCS+RT (n=676) |
MRM alone (n=398) |
p-value | ||||
---|---|---|---|---|---|---|---|
Grade 1 | Grade ≥ 2 | Total | Grade 1 | Grade ≥ 2 | Total | ||
Lymphedema | 69 (10.2) | 22 (3.3) | 91 (13.5) | 26 (6.5) | 6 (1.5) | 32 (8.0) | 0.007 |
Pneumonitis | 8 (1.2) | 1 (0.1) | 9 (1.3) | 0 | 0 | 0 | 0.031 |
Values are presented as number (%). BCS, breast conserving surgery; RT, radiation therapy; MRM, modified radical mastectomy.