Patient characteristics
A total of 44755 elderly patients were included in this study (Fig. 1), among which 30375 (67.9%) patients underwent breast conserving surgery (BCS) and 14380 (32.1%) patients underwent mastectomy. It was observed that elderly breast cancer patients had larger proportion of early disease (smaller tumor and less positive lymph nodes) and luminal A subtype (72.2%) thus without receiving adjuvant radiotherapy after BCS (35.7%) or receiving adjuvant chemotherapy (55.8%).
There was significant distribution in age, race, TNM stage, subtype, axillary dissection, radiotherapy and chemotherapy among BCS and mastectomy group (all p < 0.001). After propensity score matching, 7222 patients undergoing mastectomy were matched with 7222 patients undergoing BCS. And balance in above characteristics was achieved in matching cohort (all p > 0.05) in two surgical groups. Details of related variables were shown in Table 1.
Table 1
Clinicopathological characteristics of study cohort
| Before PSM | | After PSM |
| BCS (n = 30375) | Mastectomy (n = 14380) | p value | | BCS (n = 7222) | Mastectomy (n = 7222) | p value |
Age (years) 70–74 75–79 80–84 85+ | 12631 (41.6) 8603 (28.3) 5398 (17.8) 3743 (12.3) | 5405 (37.6) 4028 (28.0) 2838 (19.7) 2109 (14.7) | < 0.001 | | 2803 (38.8) 1983 (27.5) 1350 (18.7) 1086 (15.0) | 2810 (38.9) 2008 (27.8) 1342 (18.6) 1062 (14.7) | 0.928 |
Race Unknown American Indian Asian Black White | 134 (0.4) 121 (0.4) 1710 (5.6) 2364 (7.8) 26046 (85.7) | 42 (0.3) 77 (0.5) 1205 (8.4) 1344 (9.3) 11712 (81.4) | < 0.001 | | 34 (0.5) 30 (0.4) 522 (7.2) 651 (9.0) 5985 (82.9) | 17 (0.2) 26 (0.4) 545 (7.5) 588 (8.1) 6046 (83.7) | 0.205 |
TNM stage 0 Ⅰ Ⅱ Ⅲ | 13 (0.0) 21137 (69.6) 8525 (28.1) 700 (2.3) | 11 (0.1) 5385 (37.4) 6840 (47.6) 2144 (14.9) | < 0.001 | | 1 (0.0) 3958 (54.8) 2778 (38.5) 485 (6.7) | 7 (0.1) 3950 (54.7) 2778 (38.5) 487 (6.7) | 0.211 |
Subtype HR+/HER2- HR+/HER2+ HR-/HER2+ HR-/HER2- | 25529 (84.0) 1898 (6.2) 622 (2.0) 2326 (7.7) | 10734 (74.6) 1350 (9.4) 627 (4.4) 1669 (11.6) | < 0.001 | | 5689 (78.8) 603 (8.3) 211 (2.9) 719 (10.0) | 5631 (78.0) 638 (8.8) 251 (3.5) 702 (9.7) | 0.175 |
Axillary dissection No Yes | 3932 (12.9) 26443 (87.1) | 637 (4.4) 13743 (95.6) | < 0.001 | | 583 (8.1) 6639 (91.9) | 581 (8.0) 6641 (92.0) | 0.951 |
Radiotherapy No Yes | 10845 (35.7) 19530 (64.3) | 12176 (84.7) 2204 (15.3) | < 0.001 | | 6315 (87.4) 907 (12.6) | 6316 (87.5) 906 (12.5) | 0.980 |
Chemotherapy No Yes | 19651 (64.7) 10724 (35.3) | 8707 (60.5) 5673 (39.5) | < 0.001 | | 4019 (55.6) 3203 (44.4) | 4006 (55.5) 3216 (44.5) | 0.828 |
Surgical option on survival
Median follow-up time was 67.0 months in all patients and 68.0 months in PSM cohort. There were altogether 8288 (21.1%) death occurred in the unmatched cohort, of which 5875 of patients had BCS and 2413 had mastectomy with significant difference on overall survival (OS) (p < 0.001, Fig. 2A and Table 2). And 3806 death (26.4%) in the matched cohort, of which 1843 had BCS and 1963 had mastectomy with little difference on OS (p = 0.135, Fig. 2B and Table 3). According to Kaplan-Meier survival curve, patients received BCS had a significantly superior survival than those receiving mastectomy, while this advantage disappeared when PSM was applied (Fig. 2, Tables 2 and 3).
Table 2
Univariate and multivariable analysis of potential factors associated with overall survival
| Univariate analysis OS (%) (95% CI) | p value | Multivariable analysis HR* (95% CI) | p value |
Surgical option BCS Mastectomy | 80.7 (80.2–81.1) 72.9 (71.9–73.8) | < 0.001 | 1 (reference) 1.074 (0.992–1.105) | 0.098 | |
Age (years) 70–74 75–79 80–84 85+ | 89.6 (89.1–90.1) 82.8 (82.1–83.5) 69.9 (68.8–71.0) 48.7 (47.3–50.1) | < 0.001 | 1 (reference) 1.597 (1.495–1.776) 2.603 (2.438–2.778) 4.173 (3.898–4.467) | < 0.001 < 0.001 < 0.001 | |
Race Unknown American Indian Asian Black White | 95.7 (92.5–98.8) 78.3 (71.8–84.7) 85.9 (84.6–87.3) 76.3 (74.8–77.8) 78.5 (78.1–78.9) | < 0.001 | 0.216 (0.103–0.454) 1.355 (0.972–1.890) 0.659 (0.592–0.733) 1.136 (1.052–1.226) 1 (reference) | < 0.001 0.073 < 0.001 0.001 | |
TNM stage 0 Ⅰ Ⅱ Ⅲ | 76.2 (56.3–96.1) 83.7 (83.2–84.2) 71.3 (70.5–72.1) 59.1 (56.4–61.8) | < 0.001 | 1 (reference) 1.246 (0.517–3.005) 2.307 (0.957–5.563) 4.460 (1.843–10.979) | 0.624 0.063 0.001 | |
Subtype HR+/HER2- HR+/HER2+ HR-/HER2+ HR-/HER2- | 80.3 (79.9–80.7) 77.4 (75.8–79.0) 73.4 (70.6–76.1) 67.7 (66.1–69.3) | < 0.001 | 1 (reference) 1.170 (1.076–1.272) 1.270 (1.121–1.439) 1.586 (1.484–1.695) | < 0.001 < 0.001 < 0.001 | |
Axillary dissection No Yes | 56.1 (54.7–57.6) 81.9 (81.5–82.3) | < 0.001 | 1.754 (1.655–1.858) 1 (reference) | < 0.001 | |
Radiotherapy No Yes | 71.9 (71.3–72.5) 85.4 (84.9–85.9) | < 0.001 | 1.488 (1.412–1.568) 1 (reference) | < 0.001 | |
Chemotherapy No Yes | 83.7 (83.2–84.2) 70.6 (69.9–71.3) | < 0.001 | 1 (reference) 1.314 (1.254–1.377) | < 0.001 | |
*Hazard ratio (HR) > 1 represents higher odds of death |
Table 3
Univariate and multivariable analysis of potential factors associated with overall survival after PSM
| Univariate analysis OS (%) (95% CI) | p value | Multivariable analysis HR* (95% CI) | p value |
Surgical option BCS Mastectomy | 74.5 (73.5–75.5) 72.8 (71.8–73.8) | 0.135 | 1 (reference) 1.062 (0.997–1.132) | 0.063 | |
Age (years) 70–74 75–79 80–84 85+ | 86.4 (85.5–87.3) 78.1 (76.8–79.4) 65.0 (63.2–66.8) 42.8 (40.7–44.9) | < 0.001 | 1 (reference) 1.605 (1.456–1.769) 2.523 (2.289–2.780) 4.055 (3.677–4.472) | < 0.001 < 0.001 < 0.001 | |
Race Unknown American Indian Asian Black White | 96.1 (90.6-101.6) 71.4 (59.2–83.6) 82.1 (79.8–84.4) 72.0 (69.5–74.5) 73.0 (72.2–73.8) | < 0.001 | 0.173 (0.043–0.693) 1.773 (1.084-2.900) 0.672 (0.580–0.777) 1.126 (1.007–1.259) 1 (reference) | 0.013 0.023 < 0.001 0.037 | |
TNM stage 0 Ⅰ Ⅱ Ⅲ | 62.5 (19.2-105.8) 81.2 (80.3–82.1) 65.2 (64.0-66.5) 60.7 (57.6–63.8) | < 0.001 | 1 (reference) 0.722 (0.230–2.262) 1.331 (0.425–4.170) 2.337 (0.743–7.352) | 0.576 0.623 0.147 | |
Subtype HR+/HER2- HR+/HER2+ HR-/HER2+ HR-/HER2- | 75.4 (74.6–76.2) 74.0 (71.5–76.4) 70.3 (66.2–74.5) 60.5 (58.0-63.1) | < 0.001 | 1 (reference) 1.133 (1.009–1.271) 1.239 (1.042–1.473) 1.589 (1.448–1.742) | 0.034 0.015 < 0.001 | |
Axillary dissection No Yes | 42.9 (40.0-45.7) 76.3 (75.6–77.0) | < 0.001 | 1.853 (1.691–2.029) 1 (reference) | < 0.001 | |
Radiotherapy No Yes | 73.4 (72.6–74.2) 75.2 (73.2–77.2) | 0.002 | 1.396 (1.251–1.558) 1 (reference) | < 0.001 | |
Chemotherapy No Yes | 79.5 (78.6–80.4) 66.3 (65.1–67.5) | < 0.001 | 1 (reference) 1.283 (1.197–1.375) | < 0.001 | |
*Hazard ratio (HR) > 1 represents higher odds of death |
Survival analysis on unmatched and matched cohort
Overall, the univariate analysis showed that surgical option, age, race, TNM stage, subtype, axillary dissection, radiotherapy and chemotherapy were factors that all affected OS (p < 0.001, Table 2). After adjusting for these factors in the Cox proportional risk regression multivariable analysis, older age (75–79 years old, HR = 1.597, 95% CI 1.495–1.776, p < 0.001; 80–84 year old, HR = 2.603, 95% CI 2.438–2.778, p < 0.001; HR = 4.173, 95% CI 3.898–4.467, p < 0.001), Black (HR = 1.136, 95% CI 1.052–1.226, p < 0.001), stage Ⅲ (HR = 4.460, 95% CI 1.843–10.979, p = 0.001), HER2-positive (HR+/HER2+, HR = 1.170, 95% CI 1.076–1.272, p < 0.001; HR-/HER2+, HR = 1.270, 95% CI 1.121–1.439, p < 0.001) and triple-negative subtype (HR = 1.586, 95% CI 1.484–1.695, p < 0.001), no axillary dissection (HR = 1.754, 95% CI 1.655–1.858, p < 0.001), no adjuvant radiotherapy (HR = 1.488, 95% CI 1.412–1.568, p < 0.001) as well as adjuvant chemotherapy (HR = 1.314, 95% CI 1.254–1.377, p < 0.001) were still associated with poor prognosis (Table 2). However, OS did not differ between BCS and mastectomy in the multivariable analysis (HR = 1.074, 95% CI 0.992–1.105, p = 0.098).
After PSM, all potential factors including age, race, TNM stage, subtype, axillary dissection, radiotherapy and chemotherapy (p < 0.001, Table 3) except surgical option were still associated with OS in the univariate analysis. And multivariable analysis had similar result apart from surgical option (HR = 1.062, 95% CI 0.997–1.132, p = 0.063) and TNM stage (all p > 0.05, Table 3). These results demonstrated that OS did not significantly differ between patients who underwent BCS and mastectomy, no matter in whole elderly population nor in PSM cohort.
Subgroup analysis
According to above results, subgroup analysis was performed to explore the impact of different surgical choices on survival among patients with distinct characteristics or treatments. Before PSM, the benefit of BCS with regard to OS was still significant categorized by age, race, subtype, axillary dissection, radiotherapy and chemotherapy when compared to mastectomy (p < 0.05, Table 4, Figure S1). However, a different pattern showed in the matched cohort. There was no significant improvement of OS when comparing BCS and mastectomy in age subgroups (except 80–84 years old), race, all histological subtypes and whether receiving axillary dissection, no adjuvant radiotherapy and no chemotherapy (all p > 0.05). Only patients who aged 80–84 years old, receiving radiotherapy and receiving chemotherapy had improved survival results in the BCS group (p < 0.05, Table 4). Figure S2 shows survival plots for different subgroups in PSM cohort. In addition, there was no characteristic of better OS concerning mastectomy in either unmatched cohort or matched cohort.
Table 4
Survival analysis of overall survival stratified by different factors
| Before PSM OS (%) (95% CI) | | After PSM OS (%) (95% CI) |
| BCS | Mastectomy | p value | | BCS | Mastectomy | p value |
Age (years) 70–74 75–79 80–84 85+ | 90.6 (90.1–91.2) 84.0 (83.2–84.8) 72.7 (71.6–73.9) 50.7 (49.1–52.3) | 85.9 (84.7–87.1) 78.9 (77.2–80.5) 60.8 (58.5–63.1) 42.8(40.1–45.5) | < 0.001 < 0.001 < 0.001 < 0.001 | | 87.2 (86.0-88.4) 76.9 (75.0-78.7) 69.1 (66.6–71.6) 44.0 (41.1–47.0) | 85.7 (84.4–87.0) 79.2 (77.5–81.0) 61.0 (58.3–63.6) 41.6 (38.7–44.6) | 0.276 0.081 < 0.001 0.519 |
Race Unknown American Indian Asian Black White | 96.3 (93.0-99.5) 78.5 (71.1–85.9) 87.5 (85.9–89.1) 77.6 (75.9–79.3) 80.4 (79.9–80.9) | 92.6 (82.0-103.2) 77.5 (64.0–91.0) 82.7 (80.2–85.3) 72.0 68.7–75.3() 71.7 (70.7–72.8) | 0.336 0.826 0.001 0.005 < 0.001 | | 97.1 (91.1–103.0) 70.0 (52.6–87.4) 82.6 (79.3–85.8) 70.8 (67.3–74.3) 74.1 (73.0-75.2) | 94.1 (81.6-106.6) 73.1 (54.8–91.3) 81.7 (78.4–84.9) 73.3 (69.7–76.9) 71.9 (70.8–73.0) | 0.538 0.927 0.641 0.086 0.050 |
Subtype HR+/HER2- HR+/HER2+ HR-/HER2+ HR-/HER2- | 81.8 (81.3–82.3) 79.0 (77.1–80.8) 76.5 (73.2–79.9) 70.8 (68.9–72.6) | 74.9 (73.8–75.9) 73.5 (70.4–76.6) 67.9 (63.0-72.7) 60.2 (57.2–63.3) | < 0.001 < 0.001 0.007 < 0.001 | | 76.4 (75.3–77.5) 76.1 (72.7–75.9) 71.1 (64.9–77.3) 58.8 (55.2–62.4) | 74.4 (73.2–75.5) 71.9 (68.4–75.4) 69.7 (64.0-75.4) 62.3 (58.7–65.8) | 0.084 0.085 0.946 0.160 |
Axillary dissection No Yes | 58.3 (56.7–59.8) 84.0 (83.5–84.4) | 41.6 (37.6–45.6) 75.1 (74.1–76.0) | < 0.001 < 0.001 | | 44.1 (40.0-48.1) 77.2 (76.1–78.2) | 41.7 (37.6–45.7) 75.5 (74.5–76.6) | 0.542 0.151 |
Radiotherapy No Yes | 71.1 (70.2–72.0) 86.0 (85.5–86.5) | 72.9 (72.0-73.9) 72.2 (69.3–75.2) | < 0.001 < 0.001 | | 73.9 (72.9–75.0) 78.3 (75.6–81.0) | 72.9 (71.8–74.0) 72.2 (69.3–75.3) | 0.826 < 0.001 |
Chemotherapy No Yes | 84.9 (84.4–85.4) 72.9 (72.1–73.8) | 79.2 (78.1–80.3) 63.5 (62.0-65.1) | < 0.001 < 0.001 | | 79.9 (78.4–80.9) 68.0 (66.4–69.6) | 79.3 (78.1–80.6) 64.7 (63.1–66.4) | 0.696 0.013 |
*Hazard ratio (HR) > 1 represents higher odds of death |