Axillary lymph node failure in patients treated with accelerated partial breast irradiation
Data on the risk of axillary failure (AF) after accelerated partial breast irradiation (APBI) are limited. In this study, the authors determined the rate of AF and regional lymph node failure (RNF) in patients who received various forms of APBI and identified factors that were associated with its occurrence.
In total, 534 patients with early stage breast cancer were treated at William Beaumont Hospital with APBI, including 466 patients (87%) with invasive breast cancer and 68 patients (13%) with ductal carcinoma in situ. Clinical variables (patient age, tumor location), pathologic variables (tumor size, grade, estrogen receptor status, margin status, lymph node status), and treatment-related variables (receipt of hormone and systemic chemotherapy) were analyzed to determine which factors were associated with AF and RNF. The median length of follow-up was 63 months (range, 1-201 months).
The 5-year actuarial AF rate was 0.19%. Three patients (0.56%) developed RNF (all patients initially had invasive breast cancer) with a 5-year actuarial rate of 0.37%. Two of the regional recurrences were in the supraclavicular fossa, and 1 was in the axilla. No variables were associated with AF. However, patient numbers were very small. The median survival after RNF was 0.8 years (range, 0.3-1.7 years), and 2 of the 3 patients died of disease.
The rate of AF and RNF after APBI was low and appeared to be similar to the rate observed with whole-breast irradiation. No variables were associated with a higher rate of AF after APBI. Cancer 2012;. © 2011 American Cancer Society.
The role of radiation therapy after breast-conserving surgery has been well established in several large, prospective randomized trials with extensive follow-up.1, 2 Radiation therapy (RT) after lumpectomy reduces the risk of local recurrence in the ipsilateral breast and has been associated with a small but significant overall survival benefit.1, 3, 4 Standard whole-breast irradiation (WBI) involves daily treatments for approximately 5 weeks followed, in most patients, by a supplemental boost of RT to the tumor bed. Despite the undisputed efficacy of this treatment approach, efforts have been focused on further improving the quality of life in these patients by reducing overall treatment times and, in so doing, potentially increasing the feasibility of breast-conserving therapy for more patients.
Accelerated partial breast irradiation (APBI) is 1 proposed treatment approach that not only reduces the overall treatment time to less than 1 week but also limits the RT to the region of the tumor bed. Preliminary data suggest that this approach is efficacious in highly selected, low-risk patients.5, 6 Despite the apparent advantages of APBI, some concerns regarding its application continue to persist. In contrast to WBI, APBI does not treat the lower axilla to any significant degree. Whether or not this results in a higher rate of axillary failure (AF) than that observed after WBI remains inconclusive. Typical WBI employs tangential fields that variably cover the lower levels of the axilla, potentially controlling micrometastatic disease in lymph nodes within the field. In some patients, it is believed that this provides additional local control benefits over axillary surgery alone. To date, few trials have analyzed the rates of axillary or regional lymph node failures after APBI with long-term follow-up. The current study was conducted to evaluate the rates of AF and RNF in patients who received APBI at a single institution who had long-term follow-up available and to determine any factors that may be associated with the occurrence of AF or RNF.
MATERIALS AND METHODS
In total, 534 patients with early stage breast cancer were treated at William Beaumont Hospital with APBI. The APBI techniques included interstitial brachytherapy (n = 221), balloon-based brachytherapy (n = 207), and 3-dimensional conformal RT (3-D CRT) (n = 106). Patients were treated from April 1993 to January 2011. Patients were entered prospectively into an institutional review board (IRB)-approved database with pertinent patient, pathologic, treatment, and outcome data, including follow-up data through January 2011. The interstitial population consisted of 221 patients, including 158 women who were enrolled prospectively on 1 of 3 IRB-approved APBI protocols. For the first 2 years after treatment, patients are followed in our department every 6 to 9 months, depending on the follow-up schedules with their medical oncologist and breast surgeon. After 2 years, alternating visits are spaced out every 6 months; after 5 years, patients are followed yearly. Patients are followed in our department for ≥10 years, depending on their comfort level. Then, they are discharged, but they may choose to continue their follow-up with a physician who works primarily with breast cancer survivors. If patients are discharged or stop follow-up in our department, then follow-up is maintained through chart review on our health system-based electronic medical records system. Mammograms are obtained 6 months after treatment and yearly thereafter.
Inclusion and exclusion criteria have been published previously.6 Treatment techniques included a low-dose-rate implant, which delivered 50 gray (Gy) over 96 hours at 0.52 Gy per hour, and a high-dose-rate implant, which delivered either 32 Gy in 8 fractions twice daily or 34 Gy in 10 fractions twice daily. Patients who received balloon-based brachytherapy were treated either with the MammoSite radiation therapy system (Hologic Inc., Bedford, Mass) or with the Contura Multi-Lumen Balloon catheter (SenoRx Inc., Aliso Viejo, Calif) using a dose fractionation scheme of 34 Gy in 10 fractions twice daily.7 Patients who received 3-D CRT were treated with 38.5 Gy in 10 fractions, as described in a previous publication.8 With regard to systemic therapy, all patients consulted a medical oncologist who determined whether the patient would require systemic or hormone therapy.
Patient characteristics were analyzed, including age, tumor location, tumor size, tumor grade, receptor status, margin status, lymph node status, adjuvant hormone therapy, adjuvant chemotherapy, and follow-up. Clinical outcomes analyzed included AF, RNF, and ipsilateral breast tumor recurrence (IBTR). RNF was defined as failure within the regional lymphatics (axillary, supraclavicular, or internal mammary lymph nodes). On univariate analysis, patient age, tumor size, receptor status, histology, grade, margin status, lymph node status, adjuvant hormone therapy, adjuvant chemotherapy, and APBI modality were analyzed for their associations with AF or RNF. Survival after RNF also was analyzed.
Statistical analyses were performed using the SAS statistical software package (SAS Institute, Cary, NC). A P value of ≤ .05 was considered significant. Univariate associations between clinical, pathologic, and treatment-related variables and recurrence event rates were analyzed by fitting parametric models to the recurrence time data and examining the significance of the parameter estimates.
In total, 534 patients underwent APBI, including 466 patients (87%) who had invasive histology and 68 patients (13%) who had ductal carcinoma in situ. Characteristics for the cohort are displayed in Table 1. The median follow-up was 5.3 years (range, 0.02-16.9 years). The mean age of our cohort was 65.2 years (range, 40-93 years); 180 patients (34%) were aged <60 years, including 116 patients (64%) in the cautionary group and 64 patients (36%) the unsuitable group according to American Society for Therapeutic Radiology and Oncology (ASTRO) Consensus groupings. Tumors were located primarily in the upper outer quadrant (56 %), and the next most common quadrant was the upper inner quadrant (16%). The mean tumor size was 10.7 mm (range, 0-35 mm) and 40 patients were lymph node-positive (8%). Among the lymph node-positive patients, 11 had a lymph node status of N1mic, and 29 had a status of N1; in these groups, 82% (n = 9) of patients with N1mic status and 100% of patients with N1 status underwent completion axillary lymph node dissection (ALND). Fifty-three percent of patients (n = 257) received adjuvant hormone therapy, and 21% (n = 61) received adjuvant chemotherapy. The lower percentage of patients receiving hormone therapy despite positive estrogen receptor status in 88% was secondary to the large number of patients who received interstitial treatment in the early 1990s, before the widespread use of adjuvant hormone therapy. With regard to surgical margin status, 78 patients (15%) had close margins, and 26 patients (5%) had positive margins. Using the ASTRO Consensus guidelines, 39% of patients (n = 206) were categorized as suitable, 43% (n = 230) were categorized as cautionary, and 18% (n = 98) were categorized as unsuitable.
Table 1. Clinical, Pathologic, and Treatment-Related Characteristics for Patients Treated With Accelerated Partial Breast Irradiation
|Age at diagnosis, y|| |
|Tumor size, mm|| |
|ER status|| |
| Positive||443 (88)|
| Negative||63 (12)|
|PR status|| |
| Positive||302 (70)|
| Negative||130 (30)|
| Negative||423 (80)|
| Positive||26 (5)|
| Close||78 (15)|
|Tumor classification|| |
| Tis||67 (12.5)|
| T1||419 (79)|
| T2||45 (8)|
| T2 >3.0 cm||3 (0.5)|
|Lymph node status|| |
| Negative||487 (92)|
| Positive||40 (8)|
|Adjuvant hormone therapy|| |
| Yes||257 (53)|
| No||226 (47)|
|Adjuvant chemotherapy|| |
| Yes||69 (21)|
| No||263 (79)|
|APBI type|| |
| Interstitial||221 (41)|
| Balloon-based||207 (39)|
| 3-D CRT||106 (20)|
|Tumor location|| |
| UOQ||299 (56)|
| UIQ||85 (16)|
| LOQ||51 (10)|
| LIQ||45 (8)|
|Follow-up, y|| |
|ASTRO Consensus category|| |
| Suitable||206 (39)|
| Cautionary||230 (43)|
| Unsuitable||98 (18)|
Failure Rates and Patterns
The 5-year AF rate was 0.19%, and 0.21% when noninvasive cancers were excluded. When we excluded patients who had <6 months of follow-up, the rates were 0.21% for all patients and 0.24% when noninvasive cancers were excluded. The sole AF occurred in a patient who developed an IBTR 15 years after treatment; she underwent mastectomy with ALND followed by chemotherapy and RT to the axilla. Currently, she is free of disease.
At 5 years, the actuarial RNF rate was 0.37%; in total, 3 regional failures developed at 1.6 years, 4.0 years, and 15.0 years. All regional failures occurred in patients who had invasive breast cancer, and the regional failure rate was 0.56% for patients with invasive disease. Of the 3 regional failures, 2 occurred in the supraclavicular region, and 1 occurred in the axillary region. One of the recurrences was isolated, 1 occurred concurrently with distant metastasis, and 1 occurred concurrently with a local recurrence. The median time to regional recurrence was 4.0 years (range, 1.6-15.0 years). Of the patients who developed RNF, 2 initially were lymph node-positive and underwent ALND, and 1 had a negative sentinel lymph node biopsy.
The treatment of RNF consisted of chemotherapy alone in the patient who presented with concurrent distant metastasis, mastectomy with ALND and regional radiation in the patient who presented with a concurrent local recurrence, and RT with hormone therapy for the patient who had with an isolated RNF. The median overall survival after RNF was 0.8 years (range, 0.3-1.7 years), and 2 patients died of disease, both of whom had supraclavicular recurrences. In total, 20 patients within the cohort developed an IBTR, for a 5-year actuarial rate of 1.9%.
Variables Associated With AF and RNF
Univariate analysis was performed to analyze the variables associated with AF or RNF. Factors that were included in the univariate analysis included APBI modality, receptor status, age, tumor size, tumor location, histology, grade, margin, lymph node status, adjuvant hormone therapy, and adjuvant chemotherapy. No significant factor was associated with AF on univariate analysis, probably because there was only 1 AF in the cohort. With regard to RNF, no association was observed between regional recurrence and treatment modality (P = .97), negative progesterone receptor status (P = .21), age (P = .99), tumor size (P = .12), histology (P = .98), grade (P = .24), margins status (P = .98), adjuvant hormone therapy (P = .58), or adjuvant chemotherapy (P = .98). Negative estrogen receptor status (P = .02) and positive lymph node status (P = .02) were associated with RNF (Table 2). Two of the 3 patients with RNF were estrogen receptor-negative (both had supraclavicular failures). In addition, 2 of the 3 patients with RNF had positive lymph nodes (both had supraclavicular failures).
Table 2. Univariate Analysis of Axillary Failure and Regional Lymph Node Failure
|Tumor size (continuous variable)||.98||.12|
|Histology (invasive vs DCIS)||.98||.98|
|Initial tumor grade||.98||.24|
|Margins (negative vs close/positive)||.98||.98|
|Lymph node status (positive vs negative)||.98||.02a|
|Tumor quadrant location||.98||.39|
|ASTRO Consensus category||.98||.24|
Over the past 5 to 10 years, APBI has been investigated as a potential option for the delivery of adjuvant RT in many patients with early stage breast cancer who receive breast-conserving therapy. In fact, in a recent of study on the use of brachytherapy in the United States, approximately 5% of patients in a nationwide database of 6882 Medicare beneficiaries who had private supplemental insurance received brachytherapy (APBI) as the sole RT modality after breast-conserving surgery. Moreover, this trend appears to be increasing.9 One of the potential concerns expressed with this increased use of APBI to treat women with early stage breast cancer is the possibility of higher rates of AF and/or RNF. This concern is based on the observation that the lower axilla generally is not included in the RT target volume with APBI, as it often can be with WBI.10 However, to our knowledge, no recent APBI study has yet to report significantly higher rates of AF or RNF. The low AF rate observed in the current trial (0.19% 5-year actuarial rate) does not suggest that this should become a significant problem with longer follow-up. Furthermore, the sole AF in our review developed in the setting of an IBTR 15 years after initial treatment, suggesting a potential second primary with regional lymph node involvement rather than a true AF. Of course, only additional data (and mature phase 3 trial data) will conclusively establish the impact of APBI or WBI on the RNF rate and specifically on AF.
The results presented here are concordant with the American Society of Breast Surgeons MammoSite Registry Trial which found a 3 year AF rate of 0.36% with no RNF data presented.11 Radiation Therapy Oncology Group trial 0319, which evaluated the efficacy of 3-D CRT as a method for APBI, reported a 4-year lymph node failure rate of 2%, which is somewhat higher than the rate in our series.12 However, a previous review of 199 patients who received interstitial APBI from our institution with 10-year follow-up produced RNF rates of 0%, 1.7%, and 5.9% in patients who were categorized as suitable, cautionary, and unsuitable, respectively, which was slightly higher but comparable to our current findings.13 Evaluating higher risk patients who received APBI has yielded similar findings. A review of 176 patients who were categorized as unsuitable (according to ASTRO Consensus guidelines) and received APBI on the American Society of Breast Surgeons trial indicated a 5-year RNF rate of 0.63%.14 Even when evaluating triple-negative patients who received APBI, a recent review from William Beaumont Hospital indicated a 0% RNF rate at 5 years.15
Although negative estrogen receptor status and positive axillary lymph nodes were associated with the development of RNF in the current analysis, the number of events was quite small. Further follow-up and additional events will be needed to determine whether these or any other clinical, pathologic, or treatment-related factors should be avoided in patients who are not considered optimal candidates for APBI. Nonetheless, these findings represent 1 of the largest data sets addressing this issue and should provide some assurance regarding the efficacy of APBI and its impact on RNF and AF.
AF and the Use of WBI and APBI
It has been demonstrated that standard WBI covers the lower levels of the axilla to various degrees, as discussed above. Studies that have specifically addressed the typical range of doses of radiation delivered to axillary levels I, II, and III using standard, tangential, WBI fields (not designed to encompass the axilla) suggest that 66%, 44%, and 31% of the prescribed dose reaches these levels, respectively.10 A modification of the typical whole-breast tangential fields (referred to as “high tangents”) can be used to increase these doses if desired. By using these “high tangent” fields, the percentages to the levels of the axilla increase to 86%, 71%, and 73% of the prescribed dose for levels I, II, and III, respectively.10 However, the reproducibility of these modified tangents in covering these lymph nodes is unknown. Consequently, the degree to which this axillary irradiation contributes to the control of regional disease is uncertain.
In a recently published review on the rates of AF in patients who received standard WBI after a negative sentinel lymph node dissection, 0.4% of patients developed an axillary recurrence at a median follow-up of 22 months. It is noteworthy that the review identified an association between treatment with WBI and a reduced AF rate (despite small patient numbers).16 A separate study of locoregional recurrences in patients who had lymph node-negative breast cancer at diagnosis in 5 National Surgical Adjuvant Breast and Bowel Project (NSABP) trials reported 5-year RNF rates or ipsilateral chest wall recurrence rates that ranged between 0.8% and 2.1%.17 Finally, recently published data from the American College of Surgeons Oncology Group Z011 randomized trial suggest a possible, significant impact of WBI on axillary disease control. In that trial, the 5-year AF rate was <1% in patients who had sentinel lymph node metastases irrespective of whether they underwent sentinel lymph node dissection alone or ALND.18 Concerns about that study relate to its small sample size and few events. Nonetheless, the results are provocative. In the context of all of these differing experiences, the AF rate achieved with the various forms of APBI in the current study argues for the safety of this approach. The 5-year AF rate of 0.19% with APBI appears to be 1 of the lowest numbers reported even after standard WBI.
Prognosis After AF
An additional concern related to the concept of AF and the role of RT and surgery in preventing this event is the impact of axillary recurrence on survival. Not surprisingly, AF and RNF frequently are signals of aggressive disease and are associated with very poor outcomes.17, 19-21 In patients enrolled in 5 NSABP trials (B-13, B-14, B-19, B-20, and B-23), the 5-year overall survival rate was 34.9% after RNF or ipsilateral chest wall recurrence.17 A retrospective analysis of patients in the British Columbia Cancer Agency Breast Cancer Outcomes Unit database reported a 5-year survival rate after AF of only 49.3%.19 Our data are consistent with the poor survival after lymph node failure with a median survival of 0.8 years in patients with RNF after APBI. Clearly, techniques must be optimized (whether surgical or radiotherapeutic) to limit RNF in conjunction with careful patient selection, regardless of which local therapy approach is used.
Limitations are present in our study. The patients who received APBI represented a relatively low-risk population of patients with early stage breast cancer; therefore, the RNF and AF rates would have been greater if a higher risk population of early stage breast cancer patients had been included. However, our study did include 18% unsuitable patients and 43% cautionary patients (according to ASTRO Consensus guidelines). Another limitation is that, because of the small number of events, the univariate analysis and analysis of median survival after RNF were limited. Nonetheless, the primary purpose of this study was to highlight the extremely low incidence of RNF and AF (<1%) among hundreds of patients.
APBI was designed to be used in patients who underwent surgical staging of their axilla (negative sentinel lymph node dissection or ALND with positive lymph nodes), as observed in our series, in which 100% of patients with N1 disease and 82% of patients with N1mic disease underwent completion ALND. Therefore, our results cannot be extrapolated to patients who have 1 to 3 positive lymph nodes and do not undergo completion ALND. It is noteworthy that results from the American College of Surgeons Oncology Group Z0011 trial recently were published and indicated no difference in regional control with or without completion ALND in patients who had 1 to 3 positive sentinel lymph nodes who received WBI without regional field treatment.22 Currently, there are limited data to support the use of APBI in lymph node-positive patients who do not undergo completion ALND; however, increasing numbers of patients with N1mic and N0i-positive disease are not undergoing completion ALND. In patient subsets that are eligible for APBI, outcomes with APBI and no completion ALND should be evaluated.
In the current analysis, the rates of AF and RNF after APBI were low and appeared to be similar to the rate observed with WBI. Variables potentially associated with a higher rate of RNF after APBI include negative estrogen receptor status and positive axillary lymph nodes. No variables were associated with AF, however. It is anticipated that the NSABP B-39/Radiation Therapy Oncology Group 0413 trial, in which patients with stages 0 through II breast cancer are assigned randomly to partial mastectomy with WBI or APBI, will identify the factors associated with AF and RNF.
No specific funding was disclosed.
CONFLICT OF INTEREST DISCLOSURES
The authors made no disclosures.