10.1245/s10434-006-9092-y
Annals of Surgical Oncology 14:652-659 (2007)
© 2007 Society of Surgical Oncology
Selective Application of Axillary Node Dissection in Elderly Women with Early Breast Cancer
Dalal Aziz, MD1,
Sandra Gardner, MMath2,
Kathleen Pritchard, MD2,3,4,
Lawrence Paszat, MD, MPH4,5,6 and
Claire M. B. Holloway, MD, PhD1,7
1 Department of Surgery, University of Toronto, 100 College Street, Toronto, Canada M5G 1L5
2 Department of Clinical Trials and Epidemiology, Sunnybrook Health Sciences Centre, 2075 Bayview Avenue, Toronto Canada M4N 3M5
3 Department of Medical Oncology, Sunnybrook Health Sciences Centre, 2075 Bayview Avenue, Toronto, Canada M4N 3M5
4 Department of Medicine, University of Toronto, 190 Elizabeth Street, Toronto, Canada M5G 2C4
5 Department of Radiation Oncology, Sunnybrook Health Sciences Centre, 2075 Bayview Avenue, Toronto, Canada M4N 3M5
6 Institute for Clinical Evaluative Sciences, University of Toronto, 2075 Bayview Avenue, Toronto, Canada M4N 3M5
7 Department of Surgery, Sunnybrook Health Sciences Centre, 2075 Bayview Avenue, Toronto, Canada M4N 3M5
Correspondence: Address correspondence and reprint requests to: Claire M. B. Holloway, MD, PhD; E-mail: claire.holloway{at}sunnybrook.ca
 |
ABSTRACT
|
|---|
Background: Routine use of axillary lymph node dissection (ALND) has been questioned in elderly women. This study examines whether selective application of ALND in early stage breast cancer affects breast cancer-related survival.
Methods: From the Surveillance, Epidemiology, and End Results (SEER) database, records of women 
70 years of age with stage I or II breast cancer diagnosed between 1990 and 1995 were reviewed. Hazard ratios (HR) of cause-specific survival (CSS) between women receiving ALND and those who did not were compared. To minimize the potential for bias in the selection of women to receive ALND, HR of CSS was compared within propensity analysis deciles.
Results: 20,151 women entered the analysis. Median follow up was 6 years (interquartile range 4.33–7.67 years). Seventy-five percent underwent ALND. Women with higher risk disease and younger age were more likely to undergo ALND. Five year unadjusted CSS in women who did and did not receive ALND was 92.1% and 90.6%, respectively, with a HR of 0.85, P = 0.002. Using the propensity analysis method, the adjusted HR for CSS associated with ALND was 0.89, P = 0.066.
Discussion: After adjusting for differences in the probability of receiving ALND, no clinically or statistically significant difference in survival was observed among women who received ALND when compared with those who did not, although we could not account for differences in co-morbidity or use of systemic therapy between groups.
Conclusion: Surgeons select elderly women with early stage breast cancer for ALND with a negligible impact on CSS.
Key Words: Axillary lymph node dissection • SEER database • Breast cancer • Age • Survival
|
INTRODUCTION
|
|---|
Axillary lymph node dissection (ALND) has been the standard method of staging the axilla in women with early breast cancer for the last half century. Whether ALND confers a survival advantage is controversial,1,2 but it has been routinely employed to determine prognosis and guide decisions about adjuvant therapy. ALND provides accurate staging and local disease control,3 but it is commonly associated with arm edema, paresthesia of the medial arm and axilla, and decreased range of motion.4 The more widespread use of systemic therapy in node-negative women and the advent of sentinel lymph node biopsy (SLNB) has brought the routine use of ALND for staging into question.5,6 Axillary radiation has been proposed as an alternative to ALND for axillary therapy after staging with sentinel node biopsy.7–10
The importance of ALND in the management of breast cancer is particularly relevant in women older than 70 years of age. Adjuvant chemoendocrine therapy in this age group has not been shown to provide more quality-adjusted survival time than endocrine treatment alone,11 reducing the importance of axillary staging in the selection of adjuvant therapy.
Obesity and regional nodal irradiation have been found to be the only factors correlated with increased risk of long-term morbidity after ALND.12–14 Although it has not been reported whether the elderly are at an increased risk of complications from ALND compared with younger women, the impact of ALND in an older population with other co-morbidities may be more likely to interfere not only with arm function but also with activities of daily living. ALND is therefore omitted more frequently with advancing age and more favorable tumor characteristics.15
Despite the fact that 50% of breast cancer occurs in women older than 65 years of age, there is little data on the management of breast cancer in elderly women. They are often excluded from clinical trials and their management is often extrapolated from trials on a much younger population.16 Studies of ALND in women greater than 70 years of age have examined the impact of axillary staging on adjuvant treatment protocols, but very few have evaluated the impact of ALND on recurrence rate and survival.15,17 Furthermore, since any survival advantage associated with ALND in the elderly is likely to be small, some studies may lack the statistical power to identify a benefit.17
In summary, although ALND offers good local control in women with involved axillary lymph nodes, and although it may alter treatment protocols in older women if chemotherapy is an option, it is not clear whether the application of ALND, and the therapeutic decisions which may be taken as a result, affects survival or recurrence rates in elderly women with clinically negative axillae. We therefore sought to examine whether the selective use of ALND in elderly women with stage I and II breast cancer compromises survival.
|
MATERIALS AND METHODS
|
|---|
This study was approved by the Research Ethics Board of Sunnybrook Health Sciences Centre in keeping with Health Canada guidelines. Data were extracted from the Surveillance, Epidemiology, and End Results (SEER) database (SEER November 2002 Public-Use Database in ASCII Format 1973–2000), the program of the National Cancer Institute on cancer incidence and survival in the United States. The SEER program currently collects and publishes cancer incidence and survival data from 14 population-based cancer registries and three supplemental registries. It covers approximately 26% of the US population.
Data were collected on all women in the database greater than or equal to 70 years who were diagnosed with breast cancer between January 1, 1991 and December 31, 1995. We excluded patients with stage III and IV disease, patients with another cancer diagnosis within 60 months prior to the breast cancer diagnosis, and patients in registries that did not have data for each year of interest (three registries: Alaska, San Jose, and Los Angeles). We also excluded patients who had a concurrent diagnosis of another type of cancer or were missing data about survival or type of surgery. If a patient had a concurrent breast cancer diagnosis, we chose the record of the tumor that had the worse prognosis.
The raw SEER data text files were read in using SAS version 9.1.3. From 502,417 breast cancer records, 20,151 patients met the study criteria. A descriptive analysis of the extracted variables including age, year of diagnosis, race, marital status, registry, tumor size, tumor grade, ER status, PR status, stage, node positivity, surgical treatment, and receipt of radiotherapy was performed. SLNB was not in use during the time period in question; data on systemic therapy for this period is not present in the SEER database. Tumor stage was described according to the American Joint Commission on Cancer (AJCC) staging system third edition (1988), as modified by SEER.
We compared overall survival outcomes in those who underwent ALND with those who did not. In this population, most deaths were due to causes unrelated to breast cancer. Cause-specific survival (CSS) was therefore selected as the study outcome measure. Many important covariates that are related to survival such as age, size of tumor, stage, receipt of adjuvant systemic therapy, and co-morbidities are also related to whether a woman undergoes selective ALND. Recognizing the potential for bias in the survival data in this observational database, we performed a propensity analysis, using the method of Rubin and Rosenbaum.18,19 The data set was stratified based on the probability of a woman receiving ALND. Logistic regression was used to identify factors associated with receiving ALND. Hazard ratios (HR) for CSS using Cox Proportional Hazards method were compared within each stratum and pooled across strata.20 A multivariate Cox regression model was not done since this approach is limited by the lack of data regarding relevant covariates such as co-morbidity and adjuvant systemic therapy.
|
RESULTS
|
|---|
There were 20,151 women in this cohort. Median follow up was 6 years (interquartile range 4.33–7.67 years). Seventy-five percent of these women underwent ALND. The demographics and tumor characteristics of the two groups are presented in Table 1
. The nine SEER registries varied in size and geographical location. The five diagnosis years (1991–1995) were almost equally represented. Eighty-five percent of the women in the age range of 70–74 years underwent ALND, whereas only 45% of the women 85 years underwent ALND showing a trend toward omitting ALND with advancing age (P < 0.0001, Cochran--Armitage trend test). Most of the women were white (92%); however, race did not influence whether a women received ALND. Eighty-one percent of the married women underwent ALND compared to 71% of the single or widowed women. Women with stage II disease were more likely to be offered ALND (82%) than women with stage I breast cancer (70%), P < 0.0001. Also, women with larger tumors (2 cm) were more likely to be offered ALND (77%) compared to women who had tumors smaller than 2 cm (73%), P < 0.0001. High rates of missing values were reported for grade, ER, and PR. Grade was missing in 30% of the records; ER and PR status were missing in 20 and 22%, respectively. Women with more aggressive tumors were significantly more likely to undergo ALND compared to women with less aggressive tumors: ALND was performed in 83% of women with grade IV tumors compared with 70% with grade I tumors, in 80% with ER negative tumors compared with 76% with ER positive tumors, and in 78% of women with PR negative tumors compared 76% with PR positive tumors.
The characteristics of the cohort by nodal status are presented in Table 2. Of those who underwent ALND, 73% had negative nodes and 25% had positive nodes; nodal status was unknown in 2%. Nine percent of the patients who did not have ALND had their nodal status reported. Seven percent had negative nodes while 2% had positive nodes. Twenty percent of patients who did not undergo ALND received mastectomy.
Locoregional treatment is summarized in Fig. 1. As expected, radiation was more common with partial mastectomies but it was given less frequently for older age groups, even as partial mastectomy was increasingly applied.
View larger version (14K):
[in this window]
[in a new window]
|
FIG. 1. Distribution of local treatment modalities by age group. The percentage distribution of different local treatment modalities across four age categories, 70–74, 75–79, 80–84, and >85 years, is shown. Radiotherapy was used predominantly after partial mastectomy. Axillary lymph node dissection (ALND) and radiotherapy were used less frequently with advancing age.
|
|
The main outcome of interest was survival. Overall survival was dominated by non-breast cancer related deaths; therefore CSS was felt to be a more appropriate measure of the therapeutic utility of ALND in this cohort. Most of the women (58%) were alive by the end of the follow up period. Twenty-three percent of the deaths were listed with a breast cancer cause of death. The median follow up time was approximately 3 years for women who died from breast cancer and 4 years for those who died from other causes. Five year actuarial unadjusted CSS was 92.1% for those receiving ALND and 90.6% for those who did not (Fig. 2). Because of concern about selection bias for the use of ALND in this cohort, the data were stratified based on the probability of a woman receiving ALND using the propensity analysis method of Rubin and Rosenbaum.18,19 The covariates used for this model were age, size, registry, year of diagnosis, marital status, and race (Table 3). Stage was excluded because coded values for this covariate could change for those receiving ALND. The propensity analysis does not adjust for other important, but unavailable, covariates such as co-morbidities. Women were separated into 10 groups based on their propensity to undergo ALND. Table 4 summarizes the basic characteristics of the propensity groups. The first group had the lowest propensity for ALND. The groups were ordered by increasing probability of receiving ALND. Year of diagnosis, marital status, and race were balanced across the groups. Patient age, tumor size, and registry were reasonably, but not perfectly, balanced across strata. Note that where age did not balance, the mean difference in age within the strata was much less than the mean difference across strata. It is apparent that the primary determinant of the probability of receiving ALND was younger age. With the exception of the fifth and eighth strata, there was no significant survival benefit associated with ALND across strata (Table 4). The unadjusted HR for CSS between those who did and did not receive ALND was 0.85, P = 0.002; however, after adjusting for propensity groups, the HR was 0.89, P = 0.066 (Tables 5 and 6).
View larger version (12K):
[in this window]
[in a new window]
|
FIG. 2. Unadjusted cause-specific survival (CSS) in elderly women by receipt of axillary lymph node dissection (ALND). Five year unadjusted CSS among elderly women who received ALND (92%) was marginally greater than am those who did not (91%) (P = 0.0021, log-rank test).
|
|
|
DISCUSSION
|
|---|
Axillary lymph node dissection, and more recently, sentinel lymph node biopsy, provide staging information in women with early breast cancer, and thereby influence adjuvant treatment decisions, particularly the use of chemotherapy. Data from older series, when the use of systemic therapy was less frequent, have also suggested a survival benefit for ALND.2 This study was performed to determine the impact of ALND, and any subsequent therapies provided as a consequence of ALND, on survival in elderly women, a population in which the use of chemotherapy is low.21
We report one of the largest studies examining the value of ALND in elderly women. Our data indicate a trend toward survival benefit in elderly women with stage I or II breast cancer who undergo ALND, but the trend is not statistically significant after adjusting for differences in the probability of receiving ALND. Other authors have not identified any survival advantage associated with ALND in older women.15,17 Feigelson et al.17 found no axillary recurrences and no difference in survival between those receiving or not receiving ALND in a small study of ALND in women 70 years and older with T1 tumors. A large retrospective study from British Columbia similarly found no significant difference in breast cancer-specific survival between women over 75 who did and did not receive AND.15 Any survival benefit that may be evident in our study is of negligible clinical significance. Most deaths among women in our cohort were unrelated to breast cancer, and the absolute improvement in unadjusted cause specific survival was only 1.5%.
The trend toward a survival advantage for women undergoing ALND observed in our study may be due to our inability to separate immeasurable and unknown prognostic factors from the receipt of ALND itself. We could not validate the cause of death, and the SEER database was missing data on potential covariates such as ER/PR status, tumor grade, lymph node status in the group not receiving ALND, co-morbidities, and use of chemotherapy and endocrine therapy. A randomized trial would be required to balance these known and unknown prognostic factors. To the extent that the propensity analysis attempts to balance such factors, it reduced the hazard ratio for CSS between the two groups, such that the adjusted CSS is not statistically significant. This analysis supports the conclusion that ALND alone is unlikely to confer an important survival advantage for women greater than 70 years, and any perceived survival benefit associated with ALND in the elderly is probably related to patient selection.
The apparent lack of benefit of ALND found in our study could be related to variations in the rates of prognostic or predictive factors, such as the presence of involved lymph nodes or receipt of chemotherapy or endocrine therapy among those women who did or did not receive ALND. These data are unavailable in the SEER database, but the use of chemotherapy in 1995 among women in the SEER registry who were 65 years or older and had stage I or II breast cancer has been reported at 3.4 or 18.7%, respectively.21 Furthermore, the use of chemotherapy in such women was less frequent with advancing age, such that women over 80 years were one-twelfth as likely to receive chemotherapy as women aged 65–69 years.21 In 1995, tamoxifen therapy was used by approximately half of the women in the SEER database 70 years and older with stage I or II node negative breast cancer, and just over 60% of such women with stage II node positive or stage III disease.22 That there is little difference in tamoxifen use across disease stages suggests that tamoxifen use is minimally informed by the results of ALND in this age group. Thus, although the use of systemic therapy is not known for individuals in our study cohort, it is unlikely that differences in use of systemic therapy among elderly women who did or did not receive ALND is masking a significant survival difference between them.
Consistent with our results, the British Columbia study found that ALND was omitted more frequently with advancing age and favorable tumor characteristics.15 Although co-morbidity data was not available for the patients in our study, it is probable that ALND is also omitted more frequently in the presence of significant co-morbidity, as suggested by its reduced application in the oldest age group. The lack of a clinically or statistically significant survival advantage for ALND when applied selectively, as in our study, supports this practice.
Sentinel lymph node biopsy has become the method of choice for staging of the axilla in many institutions. As SLNB was not in general use during the period of our study, we could not evaluate the survival effect of SLNB. Our study shows no survival benefit for ALND in women over 70 years; therefore, it is unlikely that SLNB would offer a survival benefit in this population. It may, however, be an important tool for the management of individual patients over 70 years. Although results of our study indicate that axillary staging does not lead to improved survival in the majority of elderly women, it often provides important information for decision making around adjuvant therapy. Chua et al.23 concluded that ALND influenced choice of adjuvant therapy in elderly women with high risk disease while Di Fronzo et al.,24 in a small study of 73 elderly women, concluded that nodal status was the variable most significantly associated with changes in adjuvant radiotherapy and systemic therapy. As SLNB provides the same staging information as ALND with less morbidity, it would seem appropriate to offer it to women over 70 whose choice of adjuvant therapy would be influenced by the results, such as the fit woman with hormone-receptor negative disease who wishes to consider systemic therapy or the woman with a favorable tumor and multiple co-morbidities who would prefer to avoid all systemic therapy, including endocrine therapy.
Although our study showed no survival benefit for ALND in the entire cohort, it does not exclude the possibility that ALND may offer a survival benefit for those with node positive disease. This study could not determine the value of completion axillary dissection in elderly women with a positive sentinel node (SN); therefore, we cannot advocate a different approach to the positive SN than that which is recommended for younger women. As we were unable to separate the effects of ALND from the use of systemic therapies that may have been used subsequently, we cannot exclude a potential survival benefit associated with ALND alone. Furthermore, the effect of ALND on regional recurrence rates could not be determined, as SEER does not collect this data.
In this large study of ALND in elderly women, it would appear that ALND, and any therapies which are offered as a consequence of receiving ALND, confers no survival benefit over simple lumpectomy/mastectomy in women over 70 years with clinical stage I or II breast cancer. A selective approach to ALND in the elderly as practiced in the United States is not associated with significant adverse effects on breast cancer related survival.
Received for publication June 29, 2006.
Accepted for publication June 29, 2006.
|
REFERENCES
|
|---|
- Fisher B, Jeong JH, Anderson S, Bryant J, Fisher ER, Wolmark N. Twenty-five-year follow-up of a randomized trial comparing radical mastectomy, total mastectomy, and total mastectomy followed by irradiation. N Engl J Med 2002; 347:567–75.[Abstract/Free Full Text]
- Orr RK. The impact of prophylactic axillary node dissection on breast cancer survival—a Bayesian meta-analysis. Ann Surg Oncol 1999; 6:109–16.[Medline]
- Chua B, Ung O, Taylor R, Boyages J. Is there a role for axillary dissection for patients with operable breast cancer in this era of conservatism? ANZ J Surg 2002; 72:786–92.[CrossRef][Medline]
- Roses DF, Brooks AD, Harris MN, Shapiro RL, Mitnick J. Complications of level I and II axillary dissection in the treatment of carcinoma of the breast. Ann Surg 1999; 230:194–201.[CrossRef][Medline]
- Silverstein MJ, Gierson ED, Waisman JR, Senofsky GM, Colburn WJ, Gamagami P. Axillary lymph node dissection for T1a breast carcinoma. Is it indicated? Cancer 1994; 73:664–7.[CrossRef][Medline]
- Mincey BA, Bammer T, Atkinson EJ, Perez EA. Role of axillary node dissection in patients with T1a and T1b breast cancer: Mayo Clinic experience. Arch Surg 2001 (July); 136:779–82.[Abstract/Free Full Text]
- Bourez R, Rutgers E, Van De Velde C. Will we need lymph node dissection at all in the future? Clin Breast Cancer 2002; 3:315–22.[Medline]
- Bohler F, Eiter H, Rhomberg W. Is axillary dissection in clinically lymph node-negative breast carcinoma further indicated? Strahlenther Onkol 1998; 174:605–12.[Medline]
- Buchholz TA, Strom EA, McNeese MD, Hunt KK. Radiation therapy as an adjuvant treatment after sentinel lymph node surgery for breast cancer. Surg Clin N Am 2003; 83: 911–30.
- Reed DR, Lindsley SK, Mann GN, Austin-Seymour M, Korssjoen T, Anderson BO, Moe R. Axillary lymph node dose with tangential breast irradiation. Int J Radiat Oncol Biol Phys 2005; 61:358–64.[CrossRef][Medline]
- Gelber RD, Cole BF, Goldhirsch A, Rose C, Fisher B, Osborne CK, Boccardo F, et al. Adjuvant chemotherapy plus tamoxifen compared with tamoxifen alone for postmenopausal breast cancer: meta-analysis of quality-adjusted survival. Lancet 1996; 347:1066–71.[CrossRef][Medline]
- Flickinger D. Arm edema after lumpectomy and breast irradiation. Am J Clin Oncol 2003; 26:229–31.[CrossRef][Medline]
- Coen JJ, Taghian AG, Kachnic LA, Assaad SI, Powell SN. Risk of arm lymphedema after regional nodal irradiation with breast conservation therapy. Int J Radiat Oncol Biol Phys 2003; 55:1209–15.[CrossRef][Medline]
- Chua B, Ung O, Boyages J. Competing considerations in regional nodal treatment for early breast cancer. Breast J 2002; 8:15–22.[CrossRef][Medline]
- Truong PT, Bernstein V, Wai E, Chua B, Speers C, Olivotto IA. Age-related variations in the use of axillary dissection: a survival analysis of 8038 women with T1-ST2 breast cancer. Clin Invest 2002; 54:794–803.
- Hutchins LF, Unger JM, Crowley JJ, Coltman CA Jr, Albain KS. Underrepresentation of patients 65 years of age or older in cancer-treatment trials. N Engl J Med 1999; 341: 2061–7.[Abstract/Free Full Text]
- Feigelson BJ, Acosta JA, Feigelson HS, Findley A, Saunders EL. T1 breast carcinoma in women 70 years of age and older may not require axillary node dissection. Am J Surg 1996; 172:487–90.[CrossRef][Medline]
- Rosenbaum PR, Rubin DR. Reducing bias in observational studies using subclassification on the propensity score. J Am Stat Assoc 1984; 79:516–24.
- Breast Cancer Versus Mastectomy. US General Accounting Office-Program Evaluation and Methodology Division Report #GAO/PEMB-95-9.
- Snedecor GW, Cochran WG. Statistical Methods. Seventh Edition. The Iowa State University Press, Ames, Iowa 1980.
- Du X, Goodwin JS. Increase of chemotherapy use in older women with breast carcinoma from 1991 to 1996. Cancer 2001; 92:730–7.[CrossRef][Medline]
- Mariotto A, Feuer EJ, Harlan LC, Wun LM, Johnson KA, Abrams J. Trends in use of adjuvant multi-agent chemotherapy and tamoxifen for breast cancer in the United States: 1975–1999. J Natl Cancer Inst 2002; 94:1626–34.[Abstract/Free Full Text]
- Chua B, Ung O, Taylor R, Boyages J. Is information from axillary dissection relevant to patients with clinically node-negative breast cancer? Breast J 2003; 9:478–84.[CrossRef][Medline]
- Di Fronzo LA, Hansen NM, Stern SL, Brennan MB, Giuliano AE. Does sentinel lymphadenectomy improve staging and alter therapy in elderly women with breast cancer? Ann Surg Oncol 2000; 7:406–10.[Abstract]
TOP
ABSTRACT
INTRODUCTION
