Patient characteristics
The characteristics of the 112 patients who were included are listed in Table 1. Their median age was 45 years (range, 26–72 years), with the numbers of patients who underwent PTR increasing over time (from 2000 to 2014). About half of the patients were clinical T2 stage (51.8%), and clinical N-stage was evenly distributed among N0 (27.7%), N1 (34.8%), and N3 (31.2%). The most common site of distant metastases was the bones (58.0%), followed by distant LNs (25.2%). About 40% of the patients had a single metastatic lesion, whereas 43.8% had oligometastatic disease. Besides, 59.8% of patients presented with luminal A (ER and/or PR positive, but HER2 negative), 12.5% with luminal B (ER and/or PR positive, and HER2 positive), and 14.3% with HER2-enriched (ER and PR negative, but HER2 positive) tumors. Furthermore, 24.1% of patients had a high (≥ 30%) Ki-67 labeling index (LI). Neoadjuvant chemotherapy was administered to 59.8% of the patients, resulting in overall down-staging of pathological T- and N-stages. Of the 67 patients who received neoadjuvant chemotherapy, 5 (7.5%) showed a complete response, 57 (85.1%) showed a partial response, and 5 (7.5%) showed stable disease. Mastectomy (67.0%) was the main type of PTR, with 76.8% of these patients also undergoing ALND. PORT was administered to 21.4% of the patients after BCO and to 33.9% after mastectomy. Hormonal therapy was administered to 73.2% of these patients, correlating with the percentages of patients diagnosed with luminal A or B type tumors (75.0%). Trastuzumab was utilized in 25.9% of the patients, which also correlated with the percentage of patients having HER2 positive tumors (luminal B and Her2-enriched types, 27.8%). Forty-five patients underwent surgery or radiotherapy of all metastatic lesions within 3 months after PTR or postoperative chemotherapy.
Table 1
Characteristics | Total (n = 112) n (%) | No PORT (n = 50) n (%) | PORT (n = 62) n (%) | p-value |
Median age (range), years <45 ≥45 | 45 (26–72) 54 (48.2) 58 (51.8) | 45 (26–72) 22 (44.0) 28 (56.0) | 46 (26–70) 32 (51.6) 30 (48.4) | 0.541 |
Period of primary tumor resection 2000–2005 2006–2010 2011–2014 | 20 (17.9) 28 (25.0) 64 (57.1) | 14 (28.0) 18 (36.0) 18 (36.0) | 6 (9.7) 10 (16.1) 46 (74.2) | < 0.001 |
Laterality Left Right | 62 (55.4) 50 (44.6) | 29 (58.0) 21 (42.0) | 33 (53.2) 29 (46.8) | 0.753 |
Histologic grade Low (1–2) High (3) Unknown | 63 (56.3) 47 (42.0) 2 (1.7) | 26 (52.0) 24 (48.0) 0 (0.0) | 37 (59.7) 23 (37.1) 2 (3.2) | 0.408 |
Clinical T-stage T1/2 T3/4 Tx | 77 (68.8) 33 (29.5) 2 (1.7) | 37 (74.0) 12 (24.0) 1 (2.0) | 40 (64.5) 21 (33.9) 1 (1.6) | 0.357 |
Clinical N-stage N0/1 N2/3 Nx | 70 (62.5) 41 (36.6) 1 (0.9) | 37 (74.0) 13 (26.0) 0 (0.0) | 33 (53.2) 28 (45.2) 1 (1.6) | 0.050 |
Metastatic site Bone only Bone and other Other (No bone) | 55 (49.1) 10 (8.9) 47 (42.0) | 24 (48.0) 7 (14.0) 19 (38.0) | 31 (50.0) 3 (4.8) 28 (45.2) | 0.227 |
Metastatic burden* Single Oligo Disseminated | 45 (40.2) 49 (43.8) 18 (16.1) | 13 (26.0) 20 (40.0) 17 (34.0) | 32 (51.6) 29 (46.8) 1 (1.6) | < 0.001 |
Molecular subtypes Luminal A Luminal B HER2-enriched Triple negative Unknown (Luminal A or B) | 67 (59.8) 14 (12.5) 16 (14.3) 12 (10.7) 3 (2.7) | 26 (52.0) 8 (16.0) 8 (16.0) 7 (14.0) 1 (2.0) | 41 (66.1) 6 (9.7) 8 (12.9) 5 (8.1) 2 (3.2) | 0.548 |
Ki-67 LI status Low (< 30%) High (≥ 30%) Unknown | 46 (41.1) 27 (24.1) 39 (34.8) | 17 (41.7) 11 (24.3) 22 (34.0) | 29 (46.8) 16 (25.8) 17 (27.4) | 0.943 |
Pathological T-stage T0/is/1/2 T3/4 | 87 (77.7) 25 (22.3) | 40 (80.0) 10 (20.0) | 47 (75.8) 15 (24.2) | 0.763 |
Pathological N-stage N0/1 N2/3 Nx | 61 (54.5) 45 (40.2) 6 (5.3) | 28 (56.0) 17 (34.0) 5 (10.0) | 33 (53.2) 28 (45.2) 1 (1.6) | 0.524 |
Chemotherapy Neoadjuvant Postoperative Both None | 53 (47.3) 34 (30.4) 14 (12.5) 11 (9.8) | 13 (26.0) 20 (40.0) 7 (14.0) 10 (20.0) | 40 (64.5) 14 (22.6) 7 (11.3) 1 (1.6) | < 0.001 |
Type of primary tumor resection Breast conserving Mastectomy | 37 (33.0) 75 (67.0) | 13 (26.0) 37 (74.0) | 24 (38.7) 38 (61.3) | 0.223 |
Method of axillary node evaluation ALND SLNB None | 86 (76.8) 20 (17.9) 6 (5.4) | 36 (72.0) 5 (10.0) 9 (18.0) | 50 (80.6) 1 (1.6) 11 (17.7) | 0.142 |
Postoperative radiotherapy Post-breast conserving Post-mastectomy None | 24 (21.4) 38 (33.9) 50 (44.7) | 0 (0.0) 0 (0.0) 50 (100.0) | 24 (38.7) 38 (61.3) 0 (0.0) | NA |
Hormone therapy Postoperative Neoadjuvant/postoperative None | 75 (67.0) 7 (6.2) 30 (26.8) | 29 (58.0) 4 (8.0) 17 (34.0) | 46 (74.2) 3 (4.8) 13 (21.0) | 0.194 |
Trastuzumab Neoadjuvant Postoperative Both None | 1 (0.9) 12 (10.7) 16 (14.3) 83 (74.1) | 0 (0.0) 6 (12.0) 7 (14.0) 37 (74.0) | 1 (1.6) 6 (9.7) 9 (14.5) 46 (74.2) | 0.813 |
Intervention to all metastatic lesions† Yes No | 45 (40.2) 67 (59.8) | 7 (14.0) 43 (86.0) | 38 (61.3) 24 (38.7) | < 0.001 |
Abbreviations: ALND = axillary lymph node dissection; HER2 = human epidermal growth factor 2; LI = labeling index; NA = not applicable; PORT = postoperative radiotherapty; SLNB = sentinel lymph node biopsy. |
*Oligometastases defined according to Kobayashi et al.(15) |
†All interventions were performed within 3 months after surgery or postoperative chemotherapy. |
Specific information on chemotherapy and radiotherapy is provided in Supplementary Table 1. Patients received various chemotherapy regimens in both neoadjuvant and postoperative settings. Commonly used agents included doxorubicin, cyclophosphamide, docetaxel, and paclitaxel. Similar to PTR, the number of patients who received PORT increased over time (from 2000 to 2015). The supraclavicular LN (SCLN) area was irradiated in 87.1% of patients, with 27.4% also undergoing irradiation of the internal mammary LN (IMLN). Median dose irradiated to breast/chest wall and SCLN was 50 Gy. After breast/chest wall irradiation, a 10-Gy boost was applied to the tumor bed and/or scar in 31 patients, with 24 previously undergoing BCO. The median intervals between surgery and PORT were 1 month (range, 1–2 months) in patients who did not receive postoperative chemotherapy, and 7 months (range, 3–9 months) in patients who did receive postoperative chemotherapy.
Survival analyses of all patients
At a median follow-up time of 48.9 months (range, 3.5–183.4 months), the median duration of OS was 54.9 months (range, 5.3–185.9 months), and the 5-year OS rate was 59.6%. (Fig. 1A) Univariate and multivariate analyses were performed to identify factors predictive of OS (Table 2). Because of the strong multi-collinearities between molecular subtypes and hormonal and trastuzumab therapy, hormonal therapy and trastuzumab were not utilized in these analyses. Univariate analysis showed that age, the period of operation, clinical T-stage, metastatic burden, molecular subtypes, pathologic T-stage, postoperative chemotherapy, PORT, and intervention at all metastatic lesions within 3 months after surgery or postoperative chemotherapy were significant predictors of OS. On multivariate analysis, HER2-enriched, triple negative molecular subtypes, and omission of PORT were significant predictors of poorer OS. Median survival was 61.3 months (range, 5.9–185.9 months) in patients with luminal A or B type tumors, compared with 27.6 months (range, 5.3–103.1 months) in patients with HER2-enriched or triple negative molecular type tumors. Median survival in patients who did and did not receive PORT was 56.3 months (range, 5.9–179.4 months) and 46.6 months (range, 5.3–185.9 months), respectively.
Table 2
Univariate and multivariate analyses of factors predictive of overall survival in the entire patient population
Characteristics | Univariate analysis | Multivariate analysis |
HR (95% CI) | p-value | HR (95% CI) | p-value |
Age | < 45 years (vs. ≥45 years) | 0.57 (0.32–0.99) | 0.048 | | |
Period of primary tumor resection | 2006–2010 (vs. 2000–2005) 2011–2014 (vs. 2000–2005) | 0.57 (0.28–1.13) 0.38 (0.20–0.74) | 0.106 0.004 | | |
Histologic grade | Grade 1/2 (vs. Grade 3) | 0.58 (0.34–1.00) | 0.050 | | |
Clinical T-stage | T1/2 (vs. T3/4) | 0.56 (0.32–0.98) | 0.041 | | |
Clinical N-stage | N0/1 (vs. N2/3) | 0.97 (0.55–1.72) | 0.919 | | |
Metastatic site | Bone and others (vs. Bone only) Others (vs. Bone only) | 1.79 (0.72–4.43) 1.44 (0.81–2.55) | 0.207 0.209 | | |
Metastatic burden | Single (vs. Disseminated) Oligo (vs. Disseminated) | 0.40 (0.19–0.82) 0.53 (0.27–1.03) | 0.012 0.062 | | |
Molecular subtypes | Luminal B (vs. Luminal A) HER2-enriched (vs. Luminal A) Triple negative (vs. Luminal A) Luminal A or B (vs. Luminal A) | 1.44 (0.62–3.35) 3.95 (1.93–8.09) 3.62 (1.60–8.16) 0.59 (0.08–4.65) | 0.392 < 0.001 0.002 0.618 | 7.96 (2.25–28.22) 6.19 (2.04–18.81) | 0.001 0.001 |
Ki-67 LI status | < 30% (vs. ≥30%) | 0.50 (0.22–1.09) | 0.082 | | |
Pathological T-stage | T0/is/1/2 (vs. T3/4) | 0.49 (0.28–0.87) | 0.014 | | |
Pathological N-stage | N0/1 (vs. N2/3) | 0.62 (0.35–1.09) | 0.096 | | |
Neoadjuvant chemotherapy | Yes (vs. No) | 0.62 (0.36–1.07) | 0.086 | | |
Type of primary tumor resection | Mastectomy (vs. Breast conserving) | 1.79 (0.96–3.36) | 0.068 | | |
Method of axillary node evaluation | None (vs. ALND) SLNB (vs. ALND) | 0.77 (0.26–2.31) 0.80 (0.36–1.79) | 0.644 0.595 | | |
Postoperative chemotherapy | Yes (vs. No) | 2.03 (1.17–3.51) | 0.011 | | |
Postoperative radiotherapy | Yes (vs. No) | 0.43 (0.25–0.74) | 0.003 | 0.41 (0.17–0.99) | 0.048 |
Interventions to all metastatic lesions* | Yes (vs. No) | 0.53 (0.29–0.97) | 0.038 | | |
Abbreviations: ALND = axillary lymph node dissection; CI = confidence interval; HER2 = human epidermal growth factor 2; LI = labeling index; SLNB = sentinel lymph node biopsy. |
*All interventions were performed within 3 months after surgery or postoperative chemotherapy. |
The 5 year LRRFS rate was 79.0% (Fig. 1B), and the median LRRFS was 45.1 months (range, 3.5–178.2 months). Six patients experienced breast/chest wall recurrence alone, 14 experienced regional nodal failure only, and three experienced both. Of 23 patients with LRR, 14 (60.9%) were simultaneously diagnosed with distant progression, which preceded regional failure in three patients. Administration of PORT was the sole factor predicting good prognosis on univariate analysis (HR, 0.32; 95% CI, 0.14–0.76; p = 0.010; Fig. 2A) and multivariate analysis (HR, 0.36; 95% CI, 0.15–0.86; p = 0.021). The 5 year LRRFS rates were 85.8% and 74.9% in the group that did and did not receive PORT, respectively. In addition, seven patients (11.3%) who received PORT experienced in-field recurrence.
Survival outcomes were analyzed in subgroups of patients who underwent mastectomy and BCO. Of the 75 patients who underwent a mastectomy, 38 (50.7%) received PMRT. PMRT, however, was not significantly predictive of LRRFS on univariate analysis (HR, 0.52; 95% CI, 0.19–1.42; p = 0.203; Fig. 2B) and multivariate analysis. Of the 37 patients who underwent BCO, 24 (64.9%) received PORT. PORT was significantly predictive of LRRFS on univariate analysis (HR, 0.10; 95% CI, 0.01–0.87; p = 0.037; Fig. 2C) but not on multivariate analysis.
The 5 year DPFS rate was 34.3% (Fig. 1C), and the DPFS median was 30.7 months (range, 3.5–183.4 months). Distant progression was the major failure pattern observed in 74 patients, with 29 (39.2%) receiving palliative chemotherapy. Multivariate analysis showed that the latest period of PTR, single/oligo metastatic burden, lower pathologic T-stage, and low histologic grade were significantly predictive of improved DPFS, whereas PORT was not.
Effects of PORT in patients without disseminated metastases
In general, the percentage of patients with a favorable prognosis was higher among those who did than those who did not receive PORT (Table 1). Although the percentage of patients with clinical N2/3 tended to be higher in the PORT group, the percentages of patients with disseminated metastases (34.0% vs. 1.6%, p < 0.001) and who did not receive chemotherapy (20.0% vs. 1.6%, p < 0.001) were significantly higher in patients who did not receive PORT. By contrast, interventions to treat distant metastatic lesions were significantly more frequent in the PORT group (14.0 vs. 61.3%, p < 0.001).
To compensate for the differences between patients who did and did not receive PORT, patients with disseminated metastases were excluded, and the effects of PORT were analyzed in the 94 patients without disseminated metastases (Table 3). Median follow-up time was 51.1 months (range, 3.5–183.4), and median OS was 54.9 months (range, 5.3–185.9 months). The 5 year OS, LRRFS, and DPFS rates were 65.6%, 83.8%, and 38.3%, respectively. Period of operation, receipt of chemotherapy, and treatment of all metastatic lesions within 3 months after surgery or postoperative chemotherapy differed significantly between patients who did and did not receive PORT. The between-group difference in chemotherapy rates was mainly due to the increased use of neoadjuvant chemotherapy in the PORT group.
Table 3
Characteristics of the patients without disseminated metastasis who did and did not receive postoperative radiotherapy
| Without disseminated metastasis (n = 94) | Luminal A or B type tumors without disseminated metastasis (n = 70) |
Characteristics | No PORT (n = 33) n (%) | PORT (n = 61) n (%) | p-value | No PORT (n = 22) n (%) | PORT (n = 48) n (%) | p-value |
Age, years <45 ≥45 | 16 (48.5) 17 (51.5) | 31 (50.8) 30 (48.4) | 1.000 | 14 (63.6) 8 (36.4) | 25 (52.1) 23 (47.9) | 0.519 |
Period of primary tumor resection 2000–2005 2006–2010 2011–2014 | 12 (36.4) 7 (21.2) 14 (42.4) | 6 (9.8) 10 (16.4) 45 (73.8) | < 0.003 | 8 (36.4) 4 (18.2) 10 (45.5) | 4 (8.3) 7 (14.6) 37 (77.1) | 0.010 |
Laterality Left Right | 20 (60.6) 13 (39.4) | 33 (54.1) 28 (45.9) | 0.697 | 17 (77.3) 5 (22.7) | 26 (54.2) 22 (45.8) | 0.478 |
Histologic grade Low (1–2) High (3) Unknown | 14 (42.4) 19 (57.6) 0 (0.0) | 36 (59.0) 23 (37.7) 2 (3.3) | 0.134 | 10 (45.5) 12 (54.5) 0 (0.0) | 17 (35.4) 30 (62.5) 1 (2.1) | 0.637 |
Clinical T-stage T1/2 T3/4 Tx | 25 (75.8) 7 (21.2) 1 (3.0) | 39 (63.9) 21 (34.4) 1 (1.6) | 0.287 | 18 (81.8) 4 (18.2) 0 (0.0) | 31 (66.0) 16 (34.0) 0 (0.0) | 0.285 |
Clinical N-stage N0/1 N2/3 Nx | 25 (75.8) 8 (24.2) 0 (0.0) | 32 (52.5) 28 (45.9) 1 (1.6) | 0.057 | 17 (77.3) 5 (22.7) 0 (0.0) | 28 (59.6) 19 (40.4) 0 (0.0) | 0.243 |
Metastatic site Bone only Bone and other Others (no bone) | 20 (60.6) 1 (3.0) 12 (36.4) | 30 (49.2) 3 (4.9) 28 (45.9) | 0.559 | 16 (72.7) 0 (0.0) 6 (27.3) | 25 (52.1) 3 (6.2) 20 (41.7) | 0.191 |
Metastatic burden* Single Oligo Disseminated | 13 (39.4) 20 (60.6) 0 (0.0) | 32 (52.5.) 29 (47.5) 0 (0.0) | 0.320 | 10 (45.5) 12 (54.5) 0 (0.0) | 26 (54.2) 22 (45.8) 0 (0.0) | 0.675 |
Molecular subtypes Luminal A Luminal B HER2-enriched Triple negative Unknown (Luminal A or B) | 16 (48.5) 5 (15.2) 6 (18.2) 5 (15.2) 1 (3.0) | 40 (65.6) 6 (9.8) 8 (13.1) 5 (8.2) 2 (3.3) | 0.572 | 16 (72.7) 5 (22.7) 0 (0.0) 0 (0.0) 1 (4.5) | 40 (83.3) 6 (12.5) 0 (0.0) 0 (0.0) 2 (4.2) | 0.543 |
Ki-67 LI status Low (< 30%) High (≥ 30%) Unknown | 13 (39.4) 7 (21.2) 13 (39.4) | 28 (45.9) 16 (26.2) 17 (27.9) | 1.000 | 11 (50.0) 3 (13.6) 8 (36.4) | 25 (52.1) 11 (22.9) 12 (25.0) | 0.768 |
Pathological T-stage T0/is/1/2 T3/4 | 28 (84.8) 5 (15.2) | 46 (75.4) 15 (24.6) | 0.422 | 20 (90.9) 2 (9.1) | 36 (75.0) 12 (25.0) | 0.221 |
Pathological N-stage N0/1 N2/3 Nx | 21 (67.7) 8 (24.2) 4 (12.1) | 32 (52.5) 28 (45.9) 1 (1.6) | 0.137 | 15 (68.2) 5 (22.7) 2 (9.1) | 26 (54.2) 21 (43.7) 1 (2.1) | 0.215 |
Chemotherapy Neoadjuvant Postoperative Both None | 11 (33.3) 11 (33.3) 6 (18.2) 5 (15.2) | 39 (63.9) 14 (23.0) 7 (11.5) 1 (1.6) | 0.009 | 10 (45.4) 6 (27.3) 2 (9.1) 4 (18.2) | 33 (68.7) 11 (22.9) 3 (6.3) 1 (2.1) | 0.069 |
Type of PTR Breast conserving Mastectomy | 8 (24.2) 25 (75.8) | 23 (37.7) 38 (62.3) | 0.273 | 6 (27.3) 16 (72.7) | 19 (39.6) 29 (60.4) | 0.466 |
Method of axillary node evaluation ALND SLNB None | 22 (66.7) 4 (12.1) 7 (21.2) | 49 (80.3) 1 (1.6) 11 (18.0) | 0.079 | 17 (77.3) 2 (9.1) 3 (13.6) | 39 (81.2) 1 (2.1) 8 (16.7) | 0.397 |
Hormone therapy Postoperative Neoadjuvant/postoperative No | 18 (54.5) 3 (9.1) 12 (36.4) | 45 (73.8) 3 (4.9) 13 (21.3) | 0.166 | 18 (81.8) 2 (9.1) 2 (9.1) | 44 (91.7) 3 (6.2) 1 (2.1) | 0.356 |
Trastuzumab Neoadjuvant Postoperative Both None | 0 (0.0) 5 (15.2) 5 (15.2) 23 (69.7) | 1 (1.6) 6 (9.8) 9 (14.8) 45 (73.8) | 0.775 | 0 (0.0) 2 (9.1) 4 (18.2) 16 (72.7) | 0 (0.0) 4 (8.3) 4 (8.3) 40 (83.3) | 0.472 |
Intervention to all metastatic lesionsa Yes No | 7 (21.2) 26 (78.8) | 38 (62.3) 23 (37.7) | < 0.001 | 5 (22.7) 17 (77.3) | 32 (66.7) 16 (33.3) | 0.002 |
Abbreviations: ALND = axillary lymph node dissection; HER2 = human epidermal growth factor 2; LI = labeling index; PORT = postoperative radiotherapy; PTR = primary tumor resection; SLNB = sentinel lymph node biopsy. |
*All interventions were performed within 3 months after surgery or postoperative chemotherapy. |
Univariate analysis showed that the period of operation, molecular subtypes, and PORT were significant predictors of OS in patients without disseminated metastases (Table 4; Fig. 3A). By contrast, multivariate analysis showed that higher clinical T-stage and Ki-67 index were independent predictors of poor prognosis. Multivariate analysis showed that PORT was significantly predictive of longer LRRFS (HR, 0.31; 95% CI, 0.11–0.91; p = 0.033; Fig. 3B), whereas higher pathological N-stage was significantly predictive of poorer LRRFS (HR, 3.31; 95% CI, 1.12–9.79; p = 0.031). Higher pathological T-stage was the sole predictor of DPFS on multivariate analysis, whereas PORT did not affect DPFS (Fig. 3C).
Table 4
Univariate and multivariate analyses of factors predictive of overall survival in patients without disseminated metastasis
Characteristics | Without disseminated metastasis | Luminal A or B type tumors without disseminated metastasis |
Univariate analysis | Multivariate analysis | Univariate analysis | Multivariate analysis |
HR (95% CI) | p-value | HR (95% CI) | p-value | HR (95% CI) | p-value | HR (95% CI) | p-value |
Age <45 years (vs. ≥45 years) | 0.69 (0.36–1.32) | 0.257 | | | 1.00 (0.42–2.38) | 0.996 | | |
Period of primary tumor resection 2006–2010 (vs. 2000–2005) 2011–2014 (vs. 2000–2005) | 0.36 (0.14–0.88) 0.34 (0.16–0.70) | 0.026 0.003 | | | 0.17 (0.04–0.78) 0.62 (0.24–1.60) | 0.023 0.325 | 0.10 (0.02–0.48) | 0.004 |
Histologic grade Grade 1/2 (vs. Grade 3) | 0.60 (0.32–1.13) | 0.114 | | | 0.76 (0.34–1.70) | 0.501 | | |
Clinical T-stage T1/2 (vs. T3/4) | 0.60 (0.31–1.14) | 0.113 | 0.35 (0.13–0.93) | 0.036 | 0.56 (0.25–1.23) | 0.150 | | |
Clinical N-stage N0/1 (vs. N2/3) | 0.95 (0.49–1.85) | 0.890 | | | 0.63 (0.27–1.47) | 0.289 | | |
Metastatic site Bone and others (vs. Bone only) Others (vs. Bone only) | 0.70 (0.09–5.25) 1.53 (0.81–2.88) | 0.728 0.190 | | | Not available due to infinite value | |
Metastatic burden Single (vs. Oligo) | 0.56 (0.71–2.56) | 0.353 | | | 0.56 (0.25–1.28) | 0.170 | | |
Molecular subtypes Luminal B (vs. Luminal A) HER2-enriched (vs. Luminal A) Triple negative (vs. Luminal A) Luminal A or B (vs. Luminal A) | 1.24 (0.42–3.68) 4.16 (1.86–9.30) 3.17 (1.24–8.10) 0.67 (0.08–5.33) | 0.700 0.001 0.016 0.701 | | | 1.31 (0.44–3.94) | 0.625 | | |
Ki-67 LI status <30% (vs. ≥30%) | 0.53 (0.21–1.35) | 0.182 | 0.12 (0.12–0.92) | 0.033 | 0.56 (0.17–1.89) | 0.349 | | |
Pathological T-stage T0/is/1/2 (vs. T3/4) | 0.56 (0.28–1.09) | 0.085 | | | 0.49 (0.19–1.27) | 0.141 | | |
Pathological N-stage N0/1 (vs. N2/3) | 0.57 (0.30–1.10) | 0.093 | | | 0.98 (0.38–5.50) | 0.966 | | |
Neoadjuvant chemotherapy Yes (vs. No) | 0.69 (0.37–1.31) | 0.262 | | | 0.79 (0.35–1.79) | 0.578 | | |
Type of primary tumor resection Breast conserving (vs. Mastectomy) | 0.60 (0.29–1.22) | 0.158 | | | 0.47 (0.19–1.18) | 0.115 | | |
Method of axillary node evaluation None (vs. ALND) SLNB (vs. ALND) | 0.74 (0.20–2.67) 0.74 (0.29–1.92) | 0.640 0.540 | | | 0.54 (0.11–2.62) 0.28 (0.04–2.10) | 0.441 0.216 | | |
Postoperative chemotherapy Yes (vs. No) | 2.18 (1.15–4.14) | 0.017 | | | 1.39 (0.62–3.12) | 0.419 | | |
Postoperative radiotherapy Yes (vs. No) | 0.53 (0.28–0.99) | 0.046 | | | 0.76 (0.34–1.70) | 0.501 | | |
Interventions to all metastatic lesions* Yes (vs. No) | 0.64 (0.33–1.23) | 0.178 | | | 0.62 (0.27–1.44) | 0.267 | | |
Abbreviations: ALND = axillary lymph node dissection; CI = confidence interval; HER2 = human epidermal growth factor 2; LI = labeling index; SLNB = sentinel lymph node biopsy. |
*All interventions were performed within 3 months after surgery or postoperative chemotherapy. |
The effect of PORT was also evaluated in patients subgrouped by the molecular characteristics of their tumors. Of the 94 patients without disseminated metastases, 70 had luminal A or B type tumors and 24 had HER2-enriched and triple negative molecular types. Of the 70 patients with luminal A or B type tumors, 48 received PORT, and 22 did not (Table 3). The Median follow-up time of these 70 patients was 54.9 months (range, 3.5–183.4 months), and median OS was 61.9 months (range, 5.9–185.9 months). The 5 year OS, LRRFS, and DPFS rates in this group were 75.5%, 87.3%, and 40.1%, respectively. Period of operation and treatment of all metastatic lesions within 3 months after surgery or postoperative chemotherapy differed significantly between patients who did and did not receive PORT. Period of operation (2006–2010) was the only significant predictor of OS on univariate and multivariate analyses. PORT was not a significant prognostic factor for OS (Fig. 4A). PORT was a significant predictor of LRRFS on univariate analysis (Fig. 4B) but not on multivariate analysis, with the latter showing that higher pathological N-stage was the sole independent predictor of LRRFS. Similar to OS, the period of operation (2006–2010) was the only significant predictor of DPFS, whereas PORT did not affect DPFS (Fig. 4C).
Of the 24 patients having HER2-enriched and triple negative molecular types without disseminated metastases, 13 received PORT and 11 did not, with no significant differences between these subgroups (Supplementary Table 2). Median follow-up time was 28.9 months (range, 5.3–98.4 months), and median OS was 29.9 months (range, 5.3–103.1 months). Five-year OS, LRRFS, and DPFS rates were 37.0%, 71.1%, and 34.1%, respectively. Period of operation and PORT were significant predictors of OS on univariate analysis (Supplementary Fig. 1A), but PORT was not on multivariate analysis (Supplementary Table 3). Univariate analysis showed no significant predictors of LRRFS, including PORT (HR, 0.69; 95% CI, 0.11–4.37; p = 0.689; Supplementary Fig. 1B). In addition, PORT was not a significant predictor of DPFS (Supplementary Fig. 1C).