Abstract
Background:
Elderly women with clinically node-positive (cN+) breast cancer (BC) often have comorbidities that limit life expectancy and complicate treatment. We sought to determine whether the number of lymph nodes (LNs) retrieved among older women with node-positive breast cancer was associated with overall survival (OS).
Methods:
Using the National Cancer Database (2010–2015), women 70–90y with cN+ BC and ≥1 LN removed were categorized by treatment sequence: upfront surgery or neoadjuvant chemotherapy (NAC). Multivariable Cox proportional hazards models with restricted cubic splines characterized the functional association of LN retrieval with OS; threshold values of LN retrieval were estimated. Cox proportional hazards models were used to estimate the association of LN retrieval groups with OS.
Results:
In the upfront surgery cohort, a non-linear association was identified between LNs retrieved and OS. In the NAC cohort, no association was identified. For the upfront surgery cohort, the optimal threshold value of LN retrieval was 21 LN (90% CI 18–23). Based on this estimate, LN retrieval groups were created: <6, 6–11, 12–17, 18–23, and >23 LNs. After adjustment, retrieval of <12 LNs in the upfront surgery group was associated with a worse OS. No differences were observed in the NAC group.
Conclusions:
For elderly women receiving upfront surgery, there is no survival benefit to removing more than 12 LN, and for those receiving NAC, there is no association between number of LNs removed and survival. In older women who present with cN+ BC, aggressive surgery to remove more than 12 LNs may not be necessary.
Keywords: elderly, breast cancer, node-positive, axillary lymph node dissection
INTRODUCTION
In 2017, women ≥70y made up a third of all new invasive breast cancer cases1, and women ≥70y have a 6.7% probability of developing invasive breast cancer throughout the remainder of their lifetime2. Despite the prevalence of breast cancer in this age group, and 11% of the population in the United States being ≥70y, elderly women with breast cancer are often under-represented or excluded from clinical trials3,4. Furthermore, factors such as life expectancy, competing comorbidities, quality of life, and patient preference are often more relevant considerations when making treatment decisions in this elderly population5. Current guidelines are often the result of a multidisciplinary taskforce between providers in geriatrics and oncology5, and most are created by extrapolating research conducted on either the population as a whole or on younger cohorts6. Without well-studied standardized recommendations, the elderly population is less likely to receive guideline-concordant care and is at risk for being both undertreated and overtreated6–8.
In general, breast cancer treatments are becoming more personalized than ever, and surgical management of the axilla continues to evolve as we strive to do what is necessary but nothing more. However, most research has focused on women with clinically negative nodes or limited axillary nodal involvement9–12. Currently, the standard of care for women, including those ≥70y, with clinically node-positive (cN+) breast cancer is an axillary lymph node dissection (ALND)5. ALND is important for axillary staging and may offer some local-regional control. However, it is also associated with a variety of potential complications, including lymphedema, numbness, pain, reduced upper extremity range of motion, and reduced strength13–15. To date, no studies have clearly elucidated whether less aggressive axillary management may be adequate in this elderly population with unique considerations. Therefore, we sought to determine if there exists a threshold for lymph node (LN) removal that optimizes treatment planning and is associated with improved survival in an elderly population with cN+ breast cancer.
MATERIALS AND METHODS
Using the 2015 National Cancer Data Base (NCDB) Participant User File (PUF), we identified women (70–90y) diagnosed with cN+ invasive breast cancer from 2010–2015. Patients with metastatic disease, treated with neoadjuvant radiation, and/or with missing data regarding type of staging, surgery, treatment sequence, or survival were excluded. Given the potential selection bias in who may be offered neoadjuvant endocrine therapy and the variation in treatment response (compared to those receiving neoadjuvant chemotherapy [NAC]), patients receiving neoadjuvant endocrine therapy were excluded. Our study population was then divided into two cohorts based on treatment sequence: upfront surgery or NAC.
Patient and tumor characteristics were extracted from the NCDB, including the Charlson/Deyo Comorbordity Score, which is a numerical score reflecting the presence of comorbid conditions (e.g., diabetes, myocardial infarction, dementia, etc.)16. Patient characteristics were summarized with N (%) for categorical variables and median (interquartile range, IQR) for continuous variables for the entire study population and by treatment sequence. The two cohorts were compared using Chi-square or Fisher’s exact tests for categorical variables, and t-tests or Wilcoxon Rank Sum tests for continuous variables. After stratifying each cohort by pN stage, subgroup analyses summarizing the number of LNs removed at the time of surgery, the number of pathologically positive nodes, and select treatment variables were conducted.
For each cohort (upfront surgery and NAC), separate multivariable Cox proportional hazards model with restricted cubic splines (RCS) were created to characterize the functional association of the number of LNs removed with OS, after adjustment for known covariates17,18. The RCS method allows for a nonlinear relationship between the continuous dependent variable, LN retrieval, and OS, while restricting the two tails of the function to be linear to prevent instability17,18. The number of knots used to fit the functional association with RCS was selected based on Akaike Information Criteria19 with 3-, 4-, and 5-knot models as candidate models. The 5-knot model (knots at 5th, 28th, 50th, 73rd, and 95th percentiles) and 3-knot model (knots at 10th, 50th, and 90th percentiles) were selected for the upfront surgery and NAC cohorts, respectively. If the RCS methodology revealed a non-linear association, a bootstrap simulation of that particular cohort with a Monte Carlo Markov Chain procedure was used to estimate the critical point that marked the directional change of the log-hazard ratio function across 1,000 iterations. The critical point, or the threshold value, and the 90% confidence intervals (CIs) were estimated from the 1,000 iterations as mean, 5th percentile, and 95th percentile. All RCS models were adjusted for known covariates, including pT, pN, age, race/ethnicity, grade, ER (estrogen receptor) status, PR (progesterone receptor) status, HER2 (human epidermal growth factor receptor 2) status, facility type, facility location, insurance, surgery type, chemotherapy, radiation, and endocrine therapy. Based on the thresholds and CIs identified, the study population was separated into LN retrieval groups defined by the number of LNs removed.
The potential association of the LN retrieval groups and OS were explored using Cox proportional hazards regression analyses, which were adjusted for covariates including patient age, pT stage, pN stage, tumor grade, ER status, PR status, HER2 status, facility type, facility location, insurance, receipt of surgery, chemotherapy, endocrine therapy, and/or radiation. To account for the correlation of patients treated at the same hospital, a robust sandwich covariance estimator was included in these models. A p-value <0.05 was considered statistically significant, and no adjustments were made for multiple comparisons. Statistical analyses were conducted using SAS, version 9.4 (SAS Institute, Cary, NC). Due to use of de-identified data, our institutional review board granted the study exempt status.
RESULTS
Study Population and Tumor Characteristics
9,026 patients were identified: 941 (10.4%) received NAC, and 8,085 (89.6%) underwent upfront surgery (Figure 1). The median age of the entire study population was 76y (IQR 72–81). The median follow-up was 44.5 months (95% CI: 43.8–45.5), and the median survival was 83.8 months (95% CI 79.1 – non-estimable due to low number of events; Table 1). Compared to those who received NAC, women who underwent upfront surgery were significantly older (77y vs 73y, p<0.001) and more likely to have comorbidity scores > 0 (29% vs 21.5%, p<0.001). The NAC cohort had higher cT and cN stage, higher grade tumors, higher pT stage, lower pN stage, and more triple-negative tumors (Table 2, all p<0.001). Women in the NAC cohort were less likely to receive a mastectomy without radiation than the upfront surgery group (20% vs 35.8%, p<0.001), but were more likely to receive a lumpectomy plus radiation or mastectomy plus radiation (34.3% vs 31.3%, 39.9% vs 24.4%, both p<0.001; Table 2). At the time of axillary surgery, a median of 11–12 LNs were removed, and the median number of positive LNs was 2 for both cohorts (Table 2).
Figure 1. Patient selection diagram based on data from the National Cancer Database (NCDB) from 2010–2015, applying the defined inclusion and exclusion criteria.
PUF: participant user file. M: metastasis status. cT: clinical tumor size stage. cN: clinical nodal stage. WHO: World Health Organization. ER: estrogen receptor. PR: progesterone receptor. HER2: human epidermal growth factor receptor 2.
Table 1. Summary of select patient demographics and facility characteristics.
Analysis based on a cohort of elderly women aged 70–90y with cT1–2 and cN1–3, and at least 1 LN removed at time of surgery; National Cancer Database, 2010–2015. Data presented as N (%) or Median (IQR) unless otherwise specified. IQR: interquartile range. LN: lymph node. NE: non-estimable due to low number of events.
| All Patients N=9,026 (100%) | Neoadjuvant Chemotherapy N=941 (10.4%) | Upfront Surgery N=8,085 (89.6%) | P-Value | |
|---|---|---|---|---|
| Age | ||||
| Median (IQR) | 76 (72 – 81) | 73 (71 – 77) | 77 (73 – 82) | <0.001 |
| Median Follow-up Time | ||||
| Months (95% CI) | 44.5 (43.8–45.5) | 38.4 (36.7–41.3) | 45.3 (44.3–46.2) | |
| Median Survival | ||||
| Months (95% CI) | 83.8 (79.1 - NE) | NE | 83.8 (79.1 - NE) | |
| Charlson/Deyo Comorbidity Score | ||||
| 0 | 6,480 (71.8%) | 739 (78.5%) | 5,741 (71%) | <0.001 |
| 1 | 1,965 (21.8%) | 162 (17.2%) | 1,803 (22.3%) | |
| ≥ 2 | 448 (5%) | 34 (3.6%) | 414 (5.1%) | |
| Race/Ethnicity | ||||
| Hispanic | 371 (4.1%) | 52 (5.5%) | 319 (3.9%) | <0.001 |
| Non-Hispanic Black | 981 (10.9%) | 146 (15.5%) | 835 (10.3%) | |
| Non-Hispanic Other | 255 (2.8%) | 33 (3.5%) | 222 (2.7%) | |
| Non-Hispanic White | 7,007 (77.6%) | 686 (72.9%) | 6,321 (78.2%) | |
| Insurance Status | ||||
| Government | 7,903 (87.6%) | 810 (86.1%) | 7,093 (87.7%) | 0.26 |
| Not Insured | 49 (0.5%) | 6 (0.6%) | 43 (0.5%) | |
| Private | 984 (10.9%) | 117 (12.4%) | 867 (10.7%) | |
| Facility Type | ||||
| Academic | 2,230 (24.7%) | 283 (30.1%) | 1,947 (24.1%) | <0.001 |
| Community | 1,130 (12.5%) | 90 (9.6%) | 1,040 (12.9%) | |
| Comprehensive | 4,664 (51.7%) | 439 (46.7%) | 4,225 (52.3%) | |
| Integrated Network | 1,002 (11.1%) | 129 (13.7%) | 873 (10.8%) | |
Table 2. Summary of select disease characteristics and treatments.
Analysis based on a cohort of elderly women aged 70–90y with cT1–2 and cN1–3, and at least 1 LN removed at time of surgery; National Cancer Database, 2010–2015. Data presented as N (%) or Median (IQR). IQR: interquartile range. LN: lymph node. HR: hormone receptor status. HER2: human epidermal growth factor receptor 2. TNBC: triple-negative breast cancer.
| All Patients N=9,026 (100%) | Neoadjuvant Chemotherapy N=941 (10.4%) | Upfront Surgery N=8,085 (89.6%) | P-Value | |
|---|---|---|---|---|
| LNs Examined | ||||
| Median (IQR) | 11 (6 – 17) | 12 (6 – 16) | 11 (6 – 17) | 0.79 |
| Positive LNs | ||||
| Median (IQR) | 2 (1 – 5) | 2 (1 – 5) | 2 (1 – 5) | 0.04 |
| Histology | ||||
| Ductal | 7,987 (88.5%) | 854 (90.8%) | 7,133 (88.2%) | 0.002 |
| Lobular | 972 (10.8%) | 75 (8%) | 897 (11.1%) | |
| Other | 67 (0.7%) | 12 (1.3%) | 55 (0.7%) | |
| Grade | ||||
| 1 | 1,022 (11.3%) | 72 (7.7%) | 950 (11.8%) | <0.001 |
| 2 | 3,841 (42.6%) | 329 (35%) | 3,512 (43.4%) | |
| 3 | 4,163 (46.1%) | 540 (57.4%) | 3,623 (44.8%) | |
| Hormone Receptor Status | ||||
| HR+ | 7,046 (78.1%) | 575 (61.1%) | 6,471 (80%) | <0.001 |
| HR− | 1,980 (21.9%) | 366 (38.9%) | 1,614 (20%) | |
| Molecular Subtype | ||||
| HER2+ | 1,531 (17%) | 263 (27.9%) | 1,268 (15.7%) | <0.001 |
| HR+/HER2− | 6,066 (67.2%) | 418 (44.4%) | 5,648 (69.9%) | |
| TNBC | 1,429 (15.8%) | 260 (27.6%) | 1,169 (14.5%) | |
| Tumor Size (mm) | ||||
| Median (IQR) | 25 (17 – 35) | 28 (20 – 40) | 25 (17 – 35) | <0.001 |
| Clinical T Stage | ||||
| cT1 | 3,490 (38.7%) | 210 (22.3%) | 3,280 (40.6%) | <0.001 |
| cT2 | 5,536 (61.3%) | 731 (77.7%) | 4,805 (59.4%) | |
| Clinical N Stage | ||||
| cN1 | 7,552 (83.7%) | 732 (77.8%) | 6,820 (84.4%) | <0.001 |
| cN2 | 1,115 (12.4%) | 138 (14.7%) | 977 (12.1%) | |
| cN3 | 359 (4%) | 71 (7.5%) | 288 (3.6%) | |
| Pathological T Stage | ||||
| pT1 | 3,426 (38%) | 532 (56.5%) | 2,894 (35.8%) | <0.001 |
| pT2 | 5,083 (56.3%) | 343 (36.5%) | 4,740 (58.6%) | |
| pT3 | 401 (4.4%) | 51 (5.4%) | 350 (4.3%) | |
| pT4 | 116 (1.3%) | 15 (1.6%) | 101 (1.2%) | |
| Pathological N Stage | ||||
| pN1 | 5,195 (57.6%) | 451 (47.9%) | 4,744 (58.7%) | <0.001 |
| pN2 | 2,130 (23.6%) | 203 (21.6%) | 1,927 (23.8%) | |
| pN3 | 1,031 (11.4%) | 90 (9.6%) | 941 (11.6%) | |
| pN0 | 670 (7.4%) | 197 (20.9%) | 473 (5.9%) | |
| Local-regional Treatment | ||||
| Lumpectomy alone | 747 (8.3%) | 55 (5.8%) | 692 (8.6%) | <0.001 |
| Lumpectomy+radiation | 2,850 (31.6%) | 323 (34.3%) | 2,527 (31.3%) | |
| Mastectomy alone | 3,084 (34.2%) | 188 (20%) | 2,896 (35.8%) | |
| Mastectomy+radiation | 2,345 (26%) | 375 (39.9%) | 1,970 (24.4%) | |
| Endocrine Therapy | ||||
| Received, HR+ | 5,888 (65.2%) | 490 (52.1%) | 5,398 (66.8%) | <0.001 |
| Received, HR− | 47 (0.5%) | 5 (0.5%) | 42 (0.5%) | |
| Not received, HR+ | 1,158 (12.8%) | 85 (9%) | 1,073 (13.3%) | |
| Not received, HR− | 1,933 (21.4%) | 361 (38.4%) | 1,572 (19.4%) | |
Lymph Node Retrieval and Adjuvant Treatment
After stratification based on pN stage, the median number of LNs retrieved was 3–19 LNs for the surgery upfront cohort and 7–18 LNs for the NAC cohort. In order to capture 90% of women within a specific pN stage, the minimum number of nodes retrieved ranged from 14–30 LNs in the upfront surgery cohort and from 21–26 LNs in the NAC cohort (Table 3). The percentage of women who received adjuvant therapy differed based on pN stage, and the largest differences between pN stages were noted between stages pN1 and pN2. In the upfront surgery cohort, 35.4% of women with pN1 disease received adjuvant chemotherapy, compared to 50.4% of those with pN2 disease (absolute difference of 15%). In this same cohort, 26.1% of women with pN1 disease received post mastectomy radiotherapy (PMRT), and 59.8% of those with pN2 disease received PMRT (absolute difference of 33.7%). In the NAC cohort, similar differences were observed: 61.2% of women with pN1 disease received PMRT, and 81% of those with pN2 disease received PMRT (absolute difference of 19.8%) (Table 3).
Table 3. Summary of LNs retrieved and treatments received by pathological nodal (pN) stage.
Analysis based on a cohort of elderly women aged 70–90y with cT1–2 and cN1–3, and at least 1 LN removed at time of surgery; National Cancer Database, 2010–2015. Data presented as N (%) or Median (IQR). IQR: interquartile range. XRT: radiation therapy.
| Upfront Surgery | Neoadjuvant Chemotherapy | |||||||
|---|---|---|---|---|---|---|---|---|
| pN0 | pN1 | pN2 | pN3 | pN0 | pN1 | pN2 | pN3 | |
| Total N (%) | 459 (5.7%) | 4,745 (58.7%) | 1,942 (24.0%) | 939 (11.6%) | 135 (14.3%) | 488 (51.9%) | 216 (22.9%) | 102 (10.8%) |
| Median # nodes retrieved (IQR) | 3 (2–7) | 10 (4–15) | 13 (10–18) | 19 (15–24) | 7 (3–14) | 10 (5–15) | 13 (9.5–17) | 18 (14–22) |
| Minimum # nodes retrieved to capture 90% of pN stage | 14 | 20 | 22 | 30 | 21 | 20 | 22 | 26 |
| Adjuvant chemotherapy N (%) | 96 (20.9%) | 1,679 (35.4%) | 979 (50.4%) | 536 (57.1%) | - | - | - | - |
| Lumpectomy+adjuvant radiation N (%) | 173 (75.2%) | 1,762 (79.9%) | 425 (75.2%) | 167 (76.6%) | 48 (78.7%) | 189 (87.9%) | 62 (83.8%) | 24 (85.7%) |
| Mastectomy+adjuvant radiation N (%) | 15 (6.5%) | 664 (26.1%) | 823 (59.8%) | 468 (64.9%) | 33 (44.6%) | 167 (61.2%) | 115 (81.0%) | 60 (81.1%) |
Lymph Node Retrieval and Overall Survival
In the upfront surgery cohort, the functional association between LN retrieval and OS was significant and non-linear (non-linearity p<0.001; Figure 2). However, in the NAC cohort, there was no overall association between LN retrieval and OS (overall association p=0.80; Figure 3). The estimated threshold for LN retrieval associated with a marked change in OS was 21 LNs (90% CI 18–23). Based on this threshold and the associated 90% CI, 5 LN retrieval groups were created around this inflection point of 21 LNs: <6, 6–11, 12–17, 18–23, and >23 LNs. Notably, retrieval of <18 LNs was further divided into <6, 6–11, and 12–17 LNs to allow for more granularity in comparison to 18–23 LNs (the estimated 90% CI of the threshold).
Figure 2. Functional association of LN retrieval with OS based on multivariable Cox proportional hazards modeling with restricted cubic splines for the upfront surgery cohort (N=7995).
The optimal value of LN retrieval associated with a significant change in OS is indicated by the black arrow and vertical dotted line (21 LN, 90% CI 18–23). The median number of LNs retrieved (11) was used as the reference level for all HR estimates. Analysis based on a cohort of elderly women aged 70–90y with cT1–2 and cN1–3, and at least 1 LN removed at time of surgery; National Cancer Database, 2010–2015. Model adjusted for pT, pN, age, race/ethnicity, grade, ER (estrogen receptor) status, PR (progesterone receptor) status, HER2 (human epidermal growth factor receptor 2) status, facility type, facility location, insurance, surgery type, chemotherapy, radiation, endocrine. LN: lymph nodes. CI: confidence interval. HR: hazard ratio.
Figure 3. Functional association of LN retrieval with OS based on multivariable Cox proportional hazards modeling with restricted cubic splines for the neoadjuvant chemotherapy cohort (N=932).
No specific value of LN retrieval associated with a significant change in OS. The median number of LNs retrieved (12) was used as the reference level for all HR estimates. Analysis based on a cohort of elderly women aged 70–90y with cT1–2 and cN1–3, and at least 1 LN removed at time of surgery; National Cancer Database, 2010–2015. Model adjusted for pT, pN, age, race/ethnicity, grade, ER (estrogen receptor) status, PR (progesterone receptor) status, HER2 (human epidermal growth factor receptor 2) status, facility type, facility location, insurance, surgery type, chemotherapy, radiation, endocrine. LN: lymph nodes. HR: hazard ratio.
In the unadjusted analysis, a significant difference in OS was noted between LN retrieval groups in both the upfront surgery and the NAC cohorts (Supplemental Figure 1). However, in the upfront surgery cohort, the unadjusted OS differed between the LN retrieval groups by only 0–2% at 1 year and 0–4% at 5 years (Table 4). In the NAC cohort, this difference ranged from 0–6% and 1–17% at 1 and 5 years, respectively (Table 4). After adjustment, in the upfront surgery cohort, retrieval of <6 LNs (HR 1.44, 95% CI 1.21–1.70) or 6–11 LNs (HR 1.30, 95% CI 1.11–1.51) were both associated with a higher risk of death as compared to retrieval of 18–23 LNs (both p<0.001), while removing 12–17 or >23 LNs were not associated with a significantly different risk of death (both p>0.05; Table 5). In the NAC cohort, there was no significant association with OS for any of the LN retrieval groups after adjustment (all p>0.05, Table 5).
Table 4. Unadjusted 1-year and 5-year overall survival rates stratified by LN retrieval groups.
Analysis based on a cohort of elderly women aged 70–90y with cT1–2 and cN1–3, and at least 1 LN removed at time of surgery; National Cancer Database, 2010–2015. LN: lymph node. CI: confidence interval.
| Upfront Surgery | Neoadjuvant Chemotherapy | |||
|---|---|---|---|---|
| LN Retrieval Groups | 1-year survival (95% CI) | 5-year survival (95% CI) | 1-year survival (95% CI) | 5-year survival (95% CI) |
| < 6 | 0.96 (0.95–0.97) | 0.68 (0.65–0.71) | 0.99 (0.98–1) | 0.72 (0.62–0.84) |
| 6–11 | 0.95 (0.94–0.96) | 0.64 (0.62–0.67) | 0.98 (0.97–1) | 0.71 (0.64–0.79) |
| 12–17 | 0.95 (0.94–0.96) | 0.64 (0.62–0.67) | 0.97 (0.95–0.99) | 0.58 (0.5–0.68) |
| 18–23 | 0.95 (0.94–0.96) | 0.66 (0.63–0.7) | 0.98 (0.95–1) | 0.75 (0.67–0.85) |
| > 23 | 0.94 (0.93–0.96) | 0.64 (0.6–0.69) | 0.93 (0.87–0.99) | 0.63 (0.51–0.78) |
| Total | 0.95 (0.95–0.95) | 0.65 (0.64–0.67) | 0.98 (0.97–0.99) | 0.68 (0.63–0.72) |
Table 5. Cox proportional hazard model of overall survival.
(upfront surgery: N=7,626, event=1,968; neoadjuvant chemotherapy: N=909, event=201). Model adjusted for age, pT, pN, grade, ER (estrogen receptor) status, PR (progesterone receptor) status, HER2 (human epidermal growth factor receptor 2) status, facility type, facility location, insurance, surgery receipt, chemotherapy receipt, endocrine therapy receipt, and radiation receipt. LN: lymph node. HR: hazard ratio. CI: confidence interval
| Upfront Surgery | Neoadjuvant Chemotherapy | |||||
|---|---|---|---|---|---|---|
| LN Retrieval Groups | HR (95% CI) | P-value | Overall p-value | HR (95% CI) | P-value | Overall p-value |
| 18–23 | -REF- | <0.001 | -REF- | 0.58 | ||
| <6 | 1.44 (1.21 – 1.70) | <0.001 | 0.91 (0.48 – 1.72) | 0.76 | ||
| 6–11 | 1.30 (1.11 – 1.51) | <0.001 | 1.12 (0.65 – 1.95) | 0.68 | ||
| 12–17 | 1.10 (0.95 – 1.27) | 0.19 | 1.25 (0.74 – 2.09) | 0.41 | ||
| > 23 | 1.08 (0.88 – 1.33) | 0.46 | 1.30 (0.69 – 2.45) | 0.42 | ||
DISCUSSION
For elderly women with cN+ breast cancer receiving upfront surgery in our study, removing more LNs resulted in identification of more positive LNs and thus higher pN stage. Furthermore, we demonstrated that receipt of adjuvant therapy was associated with higher nodal stage disease. The largest absolute differences in adjuvant therapy receipt were noted between stages pN1 and pN2, suggesting that establishing whether a patient is truly pN1 or pN2 may be the most relevant delineation in this elderly population. Our findings also suggest that many treatment decisions are likely influenced by the findings from the surgical evaluation of the axilla, and they are consistent with the National Comprehensive Cancer Network (NCCN) breast cancer guidelines, where pN stage is utilized for systemic adjuvant therapy decisions20. When extrapolating these guidelines to the elderly population in particular, age is not a contraindication for PMRT nor adjuvant chemotherapy21. However, there is currently insufficient data to make global recommendations on systemic adjuvant therapy in the elderly, and decisions surrounding therapy are usually assessed on an individual level7,21. Therefore, it is important to be mindful of the impact nodal information may have on adjuvant treatment decisions. If treatment decisions may be altered based on the identification of more positive nodes and the final pN stage, then performing a thorough ALND (clearly defining all true borders of the axilla during the dissection) may be an important consideration when determining the appropriate surgical and post-surgical management.
However, our data also suggest that regardless of the treatments received, removal of more LNs may not be associated with improved survival. This was most notably apparent for the NAC cohort, where no association was identified between number of LNs removed and survival. In the upfront surgery cohort, we demonstrated that removing <12 LNs was associated with a worse OS, although removing more than 12 LNs did not improve survival, suggesting that the minimum cutoff for determining adequacy of axillary surgery in this elderly population with cN+ breast cancer may be 12 LNs. Currently, the NCCN defines an adequate ALND as removal of ≥10 LNs20. However, this number is largely based on historic trials from the 1990s that sought to confirm that the axilla was truly node negative22,23. More recently, Rosenberger et al. evaluated 129,685 patients with node-positive breast cancer in the NCDB and demonstrated that removal of approximately 20 LNs may result in more accurate nodal staging and subsequently impact adjuvant therapy decisions, which may improve survival24. However, the median age of this population was 56y and excluded those >75y, which may account for some of the differences noted between our studies.
When considering the elderly population more specifically, Poodt et al.’s retrospective cohort study of women ≥75y with breast cancer demonstrated that the 10-year survival rate of women who received incomplete nodal staging did not statistically differ from women who received an ALND. Furthermore, only 14% of women who had incomplete axillary staging died of breast cancer or of complications from breast-cancer-related treatment. While most women in the incomplete nodal staging cohort had cN0 disease, a large proportion of the cohort had pN+ disease (44%), and yet, OS was not affected when compared to those receiving ALND25. While elderly women with cN+ disease are not well studied, NCCN guidelines do recognize the uniqueness of this patient population. If information from an ALND will not contribute to adjuvant treatment decisions, then performing less aggressive axillary surgery in this population may be a reasonable option20. Similar to the findings of Poodt et al., our results suggest that performing less aggressive axillary surgery in select circumstances may not significantly impact OS in this elderly population with node-positive breast cancer.
There has been some research examining sentinel lymph node biopsy (SLNB) versus ALND in the general population of women with node-positive disease. Bonneau et al. evaluated 9,521 patients in the SEER (Surveillance, Epidemiology, and End Results) database with T1–2 tumors and ≥ 3 positive LNs who underwent upfront surgery (median age 56y). They noted no statistical difference in OS between those receiving SLNB vs ALND26. On the other hand, Park et al. evaluated women with cT1–3, cN2–3 disease (median age 56y) and noted an association between ALND and improved OS for patients with cN2 and cN3 disease27. This difference in the association between ALND and OS may be attributed to the size of the tumors in each study population: Bonneau et al. examined patients with relatively small tumors (T1–2) compared to the study population of Park et al., which included T1–3 tumors. Our study population, like Bonnneau et al., consisted of women with relatively small tumors, cT1–2, which may explain the similarity to our findings.
Another important variable that has been shown to be associated with OS is patient age. Bonneau et al. and Park et al. both included women representative of the general population (median age 56y in both studies), while our study population focused on women ≥70y (median age 76y). Notably, the leading cause of death in people ≥65y is heart disease, suggesting that older women with a history of breast cancer may not actually die from breast cancer, but rather may be more likely to succumb to other conditions/diseases28. Competing comorbidities in the elderly make “relative breast cancer survival” an important concept to consider when selecting the appropriate breast cancer management5,29–31. While multiple studies have assessed SLNB vs ALND in the general population with conflicting results, limiting our study to older women with smaller tumors may at least partially explain why our findings did not demonstrate a survival benefit to removing >12 LNs. However, further investigation of the axillary management in elderly women with breast cancer and various tumor sizes may be warranted.
Interestingly, we found no association between the number of LNs removed and OS in elderly women who received NAC. While previous studies have explored omission of ALND in patients who receive NAC if the SLNB is negative32, critics emphasize the high false negative rate of SLNB after NAC (12.6% in women with cN1)33,34. In our NAC cohort, 85.7% had persistently positives nodes, and presumably many would have had a positive SLNB after NAC. Regardless, LN retrieval was not associated with a significant change in OS. One possible explanation is that adjuvant chemotherapy and radiation recommendations may be driven more by the presence of persistent nodal positivity after neoadjuvant chemotherapy (rather than an exact number of nodes involved after ALND)35–37. Thus, despite a high false negative rate and a positive axilla after NAC, women ≥70y with T1–2 disease may not benefit from extensive axillary surgery if detailed nodal information will not impact adjuvant treatment decisions. While no published prospective trials have analyzed omission of ALND in women who receive NAC, the Alliance A11202 trial (ClinicalTrials.gov identifier: NCT01901094) is an ongoing trial randomizing women with a positive SLNB after NAC to receive ALND versus axillary radiotherapy. The National Surgical Adjuvant Breast and Bowel Project (NSABP) B-51/Radiation Therapy Oncology Group (RTOG) 1304 trial (ClinicalTrials.gov identifier: NCT01872975) is an ongoing trial of women with clinically node-positive breast cancer who convert to node negative after NAC, who are randomized to receive or not receive regional nodal radiation38. The results of these studies will add further insight into how to best optimize axillary management in this elderly population with node-positive breast cancer who receive NAC.
When considering how our findings may impact patient care, it is important to note that a true ALND for breast cancer (the surgical procedure itself) is defined by the patient’s anatomy and not the number of nodes removed, as surgeons typically remove the level I and II axillary tissue en bloc and do not dissect out individual LNs while in the operating room. However, surgeons do have some influence on how aggressively they clear this tissue from the axilla and potentially influence the number of nodes removed, particularly in the upfront surgery population. Similar to how level III nodes are no longer included in an ALND for breast cancer surgery, the dissection could be further limited to only level I nodes, and the medial border of the dissection would then be the pectoralis minor muscle. For patients receiving NAC where the number of nodes removed may be even less critical, one could alternatively consider a dual-tracer SLNB or targeted axillary dissection (TAD, retrieving a previously biopsied and clipped node and performing a SLNB)34,39, which could still provide information for pathological nodal staging while minimizing the potential sequelae of axillary surgery. Furthermore, a combination of these strategies (level I dissection and tracer-based lymph node removal) could also be employed to help minimize the dissection.
There were a few limitations of our study, most of which are due to its retrospective design and the inherent limitations associated with the NCDB. However, utilizing a national database allows for larger sample sizes of elderly women with cN+ breast cancer and potentially increases diversity in multiple ways at the patient, facility, and disease-levels. One important limitation is the lack of data on the type of axillary surgery being performed (SLNB vs ALND). While the NCDB provides information on the number of nodes removed, the specific NCDB data set used for this study did not include information of the surgeon’s intent. In addition, the NCDB only provides information related to OS and not breast cancer specific survival. However, in this population of women with competing comorbidities, OS is likely an appropriate outcome measure. It is also important to note that the NCDB only includes data from Commission on Cancer-accredited facilities, which may not be fully representative of the population, although numerous institutions are included in the NCDB. As with most retrospective analyses, there was likely some degree of selection bias that we could not account for in our analysis. For example, women who were healthier and able to withstand treatments like chemotherapy, were probably more likely to receive NAC, as suggested by the lower comorbidity scores in that cohort. Lastly, it is also important to recognize that an ALND may result in significant complications, such as lymphedema, pain, and decreased range of motion, consideration of which should be balanced with the benefits of doing an ALND.
In summary, we demonstrated that in women ≥70y with cN+ breast cancer, pN stage likely impacts adjuvant treatment decisions. However, in women who underwent surgery first, our data suggest that removing >12 LNs may not improve survival, and a procedure that removes at least 12 LNs may allow for adequate axillary staging to best tailor adjuvant treatment decisions. For elderly women who received NAC, there was no association between LN retrieval and OS, perhaps due to the fact that adjuvant treatment decisions after NAC are driven more by knowledge of persistent nodal positivity (rather than an exact number). Therefore, discussions at the patient’s first clinic visit between the surgeon and the oncology team should include not only the surgical options, but also how the information from surgery may be used to determine subsequent therapies. Our work highlights the importance of multidisciplinary care and a more personalized approach to node-positive breast cancer, particularly in this elderly population.
Supplementary Material
DISCLOSURES
• Dr. O. Fayanju is supported by the National Institutes of Health (NIH) under Award Number 1K08CA241390 (PI: Fayanju).
• This work is also supported by the Duke Cancer Institute through NIH grant P30CA014236 (PI: Kastan)
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Presentation: Podium presentation at the Academic Surgical Congress in February 2020.
REFERENCES
- 1.Breast Cancer Facts & Figures 2017–2018. American Cancer Society;2017. [Google Scholar]
- 2.Cancer Facts & Figures 2019. Atlanta: American Cancer Society;2019. [Google Scholar]
- 3.American Community Survey 1-year estimates. http://censusreporter.org/profiles/01000us-united-states. Published 2018. Accessed.
- 4.Freedman RA, Foster JC, Seisler DK, et al. Accrual of Older Patients With Breast Cancer to Alliance Systemic Therapy Trials Over Time: Protocol A151527. J Clin Oncol. 2017;35(4):421–431. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Biganzoli L, Wildiers H, Oakman C, et al. Management of elderly patients with breast cancer: updated recommendations of the International Society of Geriatric Oncology (SIOG) and European Society of Breast Cancer Specialists (EUSOMA). Lancet Oncol. 2012;13(4):e148–160. [DOI] [PubMed] [Google Scholar]
- 6.Swaminathan D, Swaminathan V. Geriatric oncology: problems with under-treatment within this population. Cancer Biol Med. 2015;12(4):275–283. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Mislang AR, Cheung KL, Hamaker ME, et al. Controversial issues in the management of older adults with early breast cancer. J Geriatr Oncol. 2017;8(6):397–402. [DOI] [PubMed] [Google Scholar]
- 8.Fang P, He W, Gomez DR, et al. Influence of Age on Guideline-Concordant Cancer Care for Elderly Patients in the United States. Int J Radiat Oncol Biol Phys. 2017;98(4):748–757. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Liang S, Hallet J, Simpson JS, Tricco AC, Scheer AS. Omission of axillary staging in elderly patients with early stage breast cancer impacts regional control but not survival: A systematic review and meta-analysis. J Geriatr Oncol. 2017;8(2):140–147. [DOI] [PubMed] [Google Scholar]
- 10.Giuliano AE, Ballman KV, McCall L, et al. Effect of Axillary Dissection vs No Axillary Dissection on 10-Year Overall Survival Among Women With Invasive Breast Cancer and Sentinel Node Metastasis: The ACOSOG Z0011 (Alliance) Randomized Clinical Trial. Jama. 2017;318(10):918–926. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Galimberti V, Cole BF, Viale G, et al. Axillary dissection versus no axillary dissection in patients with breast cancer and sentinel-node micrometastases (IBCSG 23–01): 10-year follow-up of a randomised, controlled phase 3 trial. Lancet Oncol. 2018;19(10):1385–1393. [DOI] [PubMed] [Google Scholar]
- 12.Donker M, van Tienhoven G, Straver ME, et al. Radiotherapy or surgery of the axilla after a positive sentinel node in breast cancer (EORTC 10981–22023 AMAROS): a randomised, multicentre, open-label, phase 3 non-inferiority trial. Lancet Oncol. 2014;15(12):1303–1310. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Hidding JT, Beurskens CH, van der Wees PJ, van Laarhoven HW, Nijhuis-van der Sanden MW. Treatment related impairments in arm and shoulder in patients with breast cancer: a systematic review. PLoS One. 2014;9(5):e96748. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Crane-Okada R, Wascher RA, Elashoff D, Giuliano AE. Long-term morbidity of sentinel node biopsy versus complete axillary dissection for unilateral breast cancer. Ann Surg Oncol. 2008;15(7):1996–2005. [DOI] [PubMed] [Google Scholar]
- 15.Mandelblatt JS, Edge SB, Meropol NJ, et al. Sequelae of axillary lymph node dissection in older women with stage 1 and 2 breast carcinoma. Cancer. 2002;95(12):2445–2454. [DOI] [PubMed] [Google Scholar]
- 16.Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. Journal of clinical epidemiology. 1992;45(6):613–619. [DOI] [PubMed] [Google Scholar]
- 17.Desquilbet L, Mariotti F. Dose-response analyses using restricted cubic spline functions in public health research. Stat Med. 2010;29(9):1037–1057. [DOI] [PubMed] [Google Scholar]
- 18.Harrell FE. Regression Modeling Strategies: with Applications to Linear Models, Logistic Regression, and Survival Analysis. Switzerland: Springer International Publishing; 2015. [Google Scholar]
- 19.Akaike H Information Theory and an Extension of the Maximum Likelihood Principle: Selected Papers of Hirotogu Akaike. New York, NY: Springer; 1998. [Google Scholar]
- 20.Gradishar WJ, Anderson BO, Abraham J, et al. NCCN Clinical Practice Guidelines in Oncology: Breast Cancer. Online 3/14/2019 2019. [Google Scholar]
- 21.Network NCC. Older Adult Oncology (Version 1.2019). https://www.nccn.org/professionals/physician_gls/pdf/senior.pdf. Accessed October 27, 2019.
- 22.Kiricuta CI, Tausch J. A mathematical model of axillary lymph node involvement based on 1446 complete axillary dissections in patients with breast carcinoma. Cancer. 1992;69(10):2496–2501. [DOI] [PubMed] [Google Scholar]
- 23.Veronesi U, Luini A, Galimberti V, Marchini S, Sacchini V, Rilke F. Extent of metastatic axillary involvement in 1446 cases of breast cancer. Eur J Surg Oncol. 1990;16(2):127–133. [PubMed] [Google Scholar]
- 24.Rosenberger LH, Ren Y, Thomas SM, et al. Axillary Lymph Node Dissection in Node-Positive Breast Cancer: Are 10 Nodes Adequate and When is Enough, Enough? BCRT. 2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Poodt IGM, Schipper RJ, Vugts G, et al. The rationale for and long-term outcome of incomplete axillary staging in elderly women with primary breast cancer. Eur J Surg Oncol. 2018;44(11):1714–1719. [DOI] [PubMed] [Google Scholar]
- 26.Bonneau C, Hequet D, Estevez JP, Pouget N, Rouzier R. Impact of axillary dissection in women with invasive breast cancer who do not fit the Z0011 ACOSOG trial because of three or more metastatic sentinel lymph nodes. Eur J Surg Oncol. 2015;41(8):998–1004. [DOI] [PubMed] [Google Scholar]
- 27.Park TS, Thomas SM, Rosenberger LH, et al. The Association of Extent of Axillary Surgery and Survival in Women with N2–3 Invasive Breast Cancer. Ann Surg Oncol. 2018;25(10):3019–3029. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Heron M Deaths: Leading Causes for 2017. Hyattsville, MD: National Center for Health Statistics;2019. [PubMed] [Google Scholar]
- 29.Patnaik JL, Byers T, Diguiseppi C, Denberg TD, Dabelea D. The influence of comorbidities on overall survival among older women diagnosed with breast cancer. J Natl Cancer Inst. 2011;103(14):1101–1111. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Vaeth PA, Satariano WA, Ragland DR. Limiting comorbid conditions and breast cancer stage at diagnosis. J Gerontol A Biol Sci Med Sci. 2000;55(10):M593–600. [DOI] [PubMed] [Google Scholar]
- 31.Schonberg MA, Marcantonio ER, Ngo L, Li D, Silliman RA, McCarthy EP. Causes of death and relative survival of older women after a breast cancer diagnosis. J Clin Oncol. 2011;29(12):1570–1577. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Kuerer HM, Sahin AA, Hunt KK, et al. Incidence and impact of documented eradication of breast cancer axillary lymph node metastases before surgery in patients treated with neoadjuvant chemotherapy. Ann Surg. 1999;230(1):72–78. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Ozcan LC, Giuliano AE. Is Axillary Lymph Node Dissection Necessary After a Positive Sentinel Lymph Node Biopsy? Adv Surg. 2017;51(1):165–178. [DOI] [PubMed] [Google Scholar]
- 34.Boughey JC, Suman VJ, Mittendorf EA, et al. Sentinel lymph node surgery after neoadjuvant chemotherapy in patients with node-positive breast cancer: the ACOSOG Z1071 (Alliance) clinical trial. Jama. 2013;310(14):1455–1461. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Masuda N, Lee SJ, Ohtani S, et al. Adjuvant Capecitabine for Breast Cancer after Preoperative Chemotherapy. N Engl J Med. 2017;376(22):2147–2159. [DOI] [PubMed] [Google Scholar]
- 36.von Minckwitz G, Huang CS, Mano MS, et al. Trastuzumab Emtansine for Residual Invasive HER2-Positive Breast Cancer. N Engl J Med. 2019;380(7):617–628. [DOI] [PubMed] [Google Scholar]
- 37.Recht A, Comen EA, Fine RE, et al. Postmastectomy Radiotherapy: An American Society of Clinical Oncology, American Society for Radiation Oncology, and Society of Surgical Oncology Focused Guideline Update. Ann Surg Oncol. 2017;24(1):38–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.ClinicalTrials.gov. National Library of Medicine (NLM) at the National Institutes of Health (NIH). https://clinicaltrials.gov/. Published 2019. Accessed 1/15/2019, 2019.
- 39.Caudle AS, Yang WT, Krishnamurthy S, et al. Improved Axillary Evaluation Following Neoadjuvant Therapy for Patients With Node-Positive Breast Cancer Using Selective Evaluation of Clipped Nodes: Implementation of Targeted Axillary Dissection. J Clin Oncol. 2016;34(10):1072–1078. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.