The total number of patients meeting all inclusion and no exclusion criteria was 360: 221 (61.4%) in the RFID group and 139 (38.6%) in the wire group. Procedures were performed by nine different surgeons. The first RFID localization occurred in January 2019, however, the majority (n = 214, 97%) took place beginning in October 2019. The last wire localization was done in August 2020, with the majority (n = 123, 88%) occurring prior to October 2019.
The two localization groups were found to be well matched in terms of patient demographics and tumor characteristics (Table 1). The RFID group had a higher proportion of white patients than the wire group (87.7% vs. 78.3%, p = 0.0048). Individual comorbidities were compared between the RFID and wire localization patients but were found not to differ between the groups. The most common comorbidities included hypertension (RFID: 46.6%, wire: 51.8%, p = 0.3864), hyperlipidemia (RFID: 35.3%, wire: 43.2%, p = 0.1484), and arthritis (RFID: 15.4%, wire: 15.1%, p = 1.000).
Table 1
Characteristics of Patients by RFID Tag vs. Wire Localization
Characteristic | RFID n = 221 | Wire n = 139 | p-value |
Age | 65.1 (11.3) | 65.0 (10.8) | 0.8487 |
White Black Asian Hispanic Other | 193 (87.7%) 24 (10.9%) 2 (0.9%) 0 1 (0.5%) | 108 (78.3%) 19 (13.8%) 8 (5.8%) 2 (1.5%) 1 (0.7%) | 0.0048 |
Never Smoker Former Smoker Current Smoker Missing (n = 89) | 109 (65.7%) 36 (21.7%) 21 (12.7%) | 75 (71.4%) 18 (17.1%) 12 (11.4%) | 0.5980 |
Postmenopausal Menopausal Perimenopausal Premenopausal Missing (n = 211) | 94 (92.2%) 1 (1.0%) 2 (2.0%) 5 (4.9%) | 44 (93.6%) 0 2 (4.3%) 1 (2.1%) | 0.8023 |
Co-morbidities Atrial fibrillation / flutter Arthritis Asthma GERD OSA DM HLD HTN CAD COPD Hypothyroidism Anxiety Depression | 11 (5.0%) 34 (15.4%) 13 (5.9%) 34 (15.4%) 12 (5.4%) 28 (12.7%) 78 (35.3%) 103 (46.6%) 17 (7.7%) 15 (6.8%) 29 (13.1%) 15 (6.8%) 15 (6.8%) | 7 (5.0%) 21 (15.1%) 6 (4.3%) 13 (9.4%) 8 (5.8%) 27 (19.4%) 60 (43.2%) 72 (51.8%) 16 (11.5%) 11 (7.9%) 11 (7.9%) 6 (4.3%) 8 (5.8%) | 1.000 1.000 0.6318 0.1097 1.000 0.0980 0.1484 0.3864 0.2611 0.6816 0.1678 0.3665 0.8261 |
Diagnostic imaging Ultrasound Mammography MRI | 199 (90.1%) 197 (89.1%) 58 (26.2%) | 121 (87.1%) 121 (87.1%) 30 (21.6%) | 0.3932 0.6137 0.3781 |
Days from Imaging to Surgery Median [25%ile – 75%ile] Mammogram to surgery1 US to surgery2 MRI to surgery3 | 43 [28–63] 39 [25–61] 28.5 [15.5–54] | 34.5 [27–55] 32 [24–50] 18[14–32] | 0.0166 0.0250 0.0823 |
Days from Biopsy to Surgery4 Median [25%ile – 75%ile] | 29 [18–50] | 26 [17–41] | 0.0903 |
ER positive ER negative | 193 (87.3%) 28 (12.7%) | 123 (89.8%) 14 (10.2%) | 0.6126 |
PR positive PR negative Missing (n = 4) | 174 (79.1%) 46 (20.9%) | 101 (74.3%) 35 (25.7%) | 0.3006 |
HER2 positive HER2 equivocal HER2 negative | 29 (15.1%) 4 (2.1%) 159 (83.2%) | 6 (4.9%) 3 (2.4%) 114 (92.7%) | 0.0165 |
Estrogen and progesterone receptor status was positive in the majority of patients and did not differ between localization groups (Table 1). The RFID group had a somewhat higher proportion of HER2 positive tumors (15.1% vs. 4.9%, p = 0.0165) and tumors with Ki-67 > 20% (31.7% vs. 21.0%, p = 0.0649) (Table 1). The COVID-19 pandemic may account for the slightly higher average number of days from diagnostic imaging to surgery in the RFID group (Table 1). Days from biopsy to surgery were not different between localization groups (Table 1, p = 0.0903).
The median (25%ile – 75%ile) time in days from RFID placement to breast conserving surgery was 1 (1–3). This is significantly different from a time of zero days for the wire group (p < 0.0001 by the Wilcoxon signed rank test). The range for RFID placement prior to surgery was zero to 26 days.
Table 2 provides the features and outcomes of breast conserving surgery. RFID and wire groups did not differ by individual surgeon (p = 0.7841). Procedures may have included global, selective, or no cavity shave margins. Approximately half of procedures involved global shave margins and did not differ by localization group (p = 0.8782, Table 2). Tumor size, cancer histology, pathological stage and nodal status did not differ by group (Table 2).
Table 2
Surgery Characteristics by RFID Tag vs. Wire Localization
Characteristic | RFID n = 221 | Wire n = 139 | p-value |
OR Time (min)a Mean (SD) Missing | 84.1 (43.7) 142 | 76.3 (29.4) 75 | 0.4089 |
Tumor size (mm) (greatest dimension) Missing | 12.4 (8.6) 32 | 13.2 (10.2) 33 | 0.6623 |
Cavity Shave margins None Selective Global | 31 (14.1%) 84 (38.2%) 105 (47.7%) | 17 (12.3%) 52 (37.7%) 69 (50.0%) | 0.8782 |
Intraoperative Rad Tx | 17 (8.1%) | 7 (5.3%) | 0.3898 |
# Nodes Taken Mean (SD) 0 1 2 3 ≥4 | 1.7 (1.8) 60 (27.3%) 57 (25.9%) 47 (21.4%) 30 (13.6%) 26 (11.8%) | 1.7 (1.8) 31 (22.3%) 33 (23.7%) 49 (35.3%) 17 (12.2%) 9 (6.5%) | 0.6983 |
# Nodes Positive Mean (SD) 0 1 2 3 4 | 0.2 (0.5) 125 (77.6%) 34 (21.1%) 2 (1.2%) 0 0 | 0.3 (0.6) 89 (79.5%) 20 (17.9%) 2 (1.8%) 0 1 (0.9%) | 0.7736 |
Breast Cancer Typeb DCIS DCIS only Invasive ductal carcinoma Invasive lobular carcinoma Invasive papillary carcinoma Invasive mucinous carcinoma Invasive metaplastic Invasive tubular carcinoma Invasive mammary carcinoma | 83 (37.6%) 29 (13.1%) 163 (73.8%) 25 (11.3%) 3 (1.4%) 3 (1.4%) 1 (0.5%) 2 (0.9%) 0 | 60 (43.2%) 28 (20.1%) 90 (64.8%) 15 (10.8%) 1 (0.7%) 3 (2.2%) 0 1 (0.7%) 1 (0.7%) | 0.3199 0.0762 1.000 1.000 0.6801 1.000 1.000 0.3861 |
Pathologic Stage T pTis pT0 ypT1 pT1 pT1a pT1b pT1c pT2 pT3 | 26 (11.8%) 3 (1.4%) 5 (2.3%) 4 (1.8%) 22 (10.0%) 50 (22.6%) 80 (36.2%) 31 (14.0%) 0 | 26 (18.8%) 2 (1.5%) 1 (0.7%) 4 (2.9%) 10 (7.3%) 25 (18.1%) 46 (33.3%) 23 (16.7%) 1 (0.7%) | 0.2881 |
Pathologic Stage N pNx pN0 ypN0 pN1 pN1a pN2a | 58 (26.2%) 125 (56.6%) 6 (2.7%) 14 (6.3%) 18 (18.1%) 0 | 28 (20.4%) 90 (65.7%) 1 (0.7%) 9 (6.7%) 8 (5.8%) 1 (0.7%) | 0.5460 |
Ki-67 Mean (SD) | 0.21 (0.20) N = 183 | 0.18 (0.19) N = 100 | 0.0412 |
aFrom first incision to close. |
bPatients could have more than one type |
The rate of obtaining positive margins was non-inferior for patients in the RFID group compared to the wire group (4.5% vs. 2.2%, p = 0.1392). Of the 13 patients with a positive margin, nine (69%) had DCIS with no invasive cancer. The difference in the positive margin rate between localization groups was 2.36%, with a one-sided 95% confidence limit for this difference of 5.44%, which was lower than the pre-established limit of 10%. Re-excision rates were also equivalent for patients in the RFID and wire groups (0.5% vs. 0.7%, p = 1.00). The margin status was negative after both re-excisions.
A subgroup analysis was done to examine if performing global, selective, or no cavity shave margins affected the positive margin rates in the RFID and wire groups (Table 3). None, selective and global groups were similar with respect to localization method, cancer histology, estrogen and progesterone receptor status, HER2 positivity, tumor size, and number of positive lymph nodes (Table 3). In addition, the rate of positive margins did not differ among the cavity shave groups (p = 0.2302, Table 3). Both re-excisions occurred in the selective group.
Table 3
Selected Variables by Cavity Shave
Variable | None n = 48 | Selective n = 136 | Global n = 174 | p-value |
RFID Wire | 31 (64.6%) 17 (35.4%) | 84 (61.8%) 52 (38.2%) | 105 (60.3%) 69 (39.7%) | 0.8782 |
DCIS only Intraductal carcinoma only DCIS & IDC Other pathologya | 5 (10.4%) 20 (41.7%) 14 (29.2%) 9 (18.8%) | 22 (16.3%) 58 (43.0%) 35 (25.9%) 20 (14.8%) | 29 (16.9%) 82 (47.7%) 36 (20.9%) 25 (14.5%) | 0.7543 |
ER positive | 42 (89.4%) | 120 (88.2%) | 152 (87.9%) | 0.9747 |
PR positive | 35 (74.5%) | 112 (83.0%) | 126 (73.3%) | 0.0993 |
Ki-67 Mean (SD) | 0.18 (0.17) N = 41 | 0.16 (0.15) N = 109 | 0.24 (0.23) N = 131 | 0.0191 |
Tumor size (mm) | 11.9 (8.9) | 12.8 (8.1) | 12.6 (9.9) | 0.6953 |
# LN removed | 1.8 (1.4) | 1.9 (1.8) | 1.6 (1.9) | 0.0212 |
# LN positive | 0.3 (0.5) | 0.2 (0.4) | 0.3 (0.6) | 0.6774 |
Positive margin | 1 (2.1%) | 8 (5.9%) | 4 (2.3%) | 0.2302 |
Re-excision | 0 | 2 (1.5%) | 0 | 0.3942 |
aLobular, papillary, metaplastic, mucinous, mammary, or tubular, sometimes in combination with DCIS or IDC |