Abstract
Purpose
The prognostic relevance of the detection mode for regional lymph node recurrence (LNR) remains unclear. We aimed to identify the risk factors associated with isolated regional LNR and to evaluate the impact of the mode of detection on distant metastasis-free survival (DMFS) and overall survival (OS).
Methods
This retrospective cohort study included women who underwent surgery for stage I–III invasive breast cancer between January 2011 and December 2019, with follow-up until December 2024. Isolated regional LNR was categorized according to the detection method (imaging vs. symptoms). Risk factors were identified using the Fine–Gray sub-distribution hazard model, and DMFS and OS were compared using Kaplan–Meier analysis and log-rank tests.
Results
Among 13,406 women (median age, 49 years; range, 20–87 years), 126 (0.9%) developed isolated regional LNR during a median follow-up of 6.3 years. Of these, 118 (93.7%) were detected through imaging and 8 (6.3%) through symptoms. Independent risk factors for isolated regional LNR included higher T stages (T2: hazard ratio [HR], 1.71; 95% confidence interval [CI], 1.11–2.64; T3: HR, 3.08; 95% CI, 1.68–5.65), higher N stages (N2: HR, 2.97; 95% CI, 1.54–5.75; N3: HR, 3.12; 95% CI, 1.36–7.13), triple-negative subtype (HR, 1.71; 95% CI, 1.09–2.67), and omission of adjuvant radiotherapy (HR, 4.35; 95% CI, 2.86–6.67). Radiological patterns varied by detection, with ultrasound usually detecting level I nodes (51.6%) and computed tomography detecting internal mammary nodes (58.1%). Symptom-detected recurrence frequently occurred in the supraclavicular region (50.0%). Symptom-detected LNRs were associated with significantly worse DMFS (p < 0.05) than imaging-detected LNRs; however, OS did not differ between the groups.
Conclusion
The detection mode was associated with DMFS, but not OS, suggesting that long-term outcomes are driven more by disease biology than by how recurrence is detected.
Keywords: Breast Neoplasms, Lymph Nodes, Lymphatic Metastasis, Prognosis, Ultrasonography
INTRODUCTION
Locoregional recurrence occurs in approximately 10%–30% of patients with breast cancer following initial treatment and includes local therapy, systemic adjuvant therapies, and/or radiotherapy [1,2,3]. It is strongly associated with an increased risk of concurrent or subsequent distant metastases and a poor overall prognosis. Isolated regional lymph node recurrence (LNR), defined as recurrence limited to regional lymphatics without simultaneous local or distant metastases, is a rare event with an incidence of 0.5%–4.6% [4].
For surveillance after initial treatment, the American Society of Clinical Oncology and the National Comprehensive Cancer Network recommend annual mammography for patients with a personal history of breast cancer (PHBC) [5,6]. In asymptomatic patients without specific clinical findings, additional imaging such as ultrasound (US) or computed tomography (CT) is not recommended for screening for recurrence or metastases, and there are no standardized surveillance guidelines for routine axillary ultrasound (AUS) screening. In some studies on breast cancer screening in women without a PHBC, such as the ACRIN 6666 protocol, AUS was considered optional [7]. However, the utility of AUS surveillance for detecting axillary recurrence in women with a PHBC remains uncertain. Although some studies have indicated that ipsilateral LNR predicts distant metastases and that lymph node evaluation during breast US is beneficial for early LNR detection [8], another study has found that AUS is inefficient for diagnosing axillary recurrence, detecting only 0.2 cancers per 1,000 examinations [9]. Despite conflicting reports on the diagnostic performance and clinical value of AUS, its impact on prognosis remains unknown.
The mode of detection is a well-recognized prognostic factor for breast cancer [10,11,12]. For example, interval breast cancers, detected between scheduled mammograms because of palpability or other clinical symptoms, exhibit more aggressive features and worse survival than screening-detected cancers [13]. Whether a similar effect applies to isolated regional LNR remains unclear. In South Korea, intensive imaging surveillance, typically semiannual breast imaging with AUS and periodic CT, is widely implemented for breast cancer survivors, even in the absence of symptoms. Although this practice differs from the major guidelines that discourage routine imaging in asymptomatic patients [5,6], it raises an important clinical question: Is imaging-based detection of isolated regional LNR associated with improved outcomes compared with symptom-based detection? Given the rarity of isolated regional LNR and the limited evidence on this issue, we aimed to identify the risk factors for isolated regional LNR and evaluate whether survival outcomes differ according to the detection method (imaging vs. symptoms) in patients with a PHBC.
METHODS
Study design and population
This study was approved by the Institutional Review Board (IRB) of Samsung Medical Center (IRB 2024-05-007). The requirement for informed consent was waived.
We retrospectively reviewed our institutional surgical database and identified 15,895 consecutive women who underwent surgery for stage I–III invasive breast cancer between January 2011 and December 2019. Staging was performed in accordance with the 7th edition of the AJCC Cancer Staging Manual [14]. Patients with distant metastases at primary breast cancer (PBC) diagnosis (n = 55), those who were lost to follow-up for > 1 year (n = 1,152), those with preexisting malignancies in other organs (n = 121), those with synchronous bilateral breast cancer at diagnosis (n = 378), and those who underwent surgery for recurrent breast cancer (n = 783) were excluded. Finally, we included 13,406 patients (median age, 49 years; range, 20–87 years) (Figure 1) in the study.
Figure 1. Flowchart of patient selection.
Imaging surveillance
For patients with a PHBC, semiannual imaging surveillance is partially covered by the Korean National Health Insurance Service for up to 5 years after cancer surgery. In accordance with this national guideline, our tertiary referral hospital offers annual two-dimensional mammography (Lorad Selenia; Hologic, Bedford, USA, or Senographe 2000D; GE Healthcare, Milwaukee, USA) and whole-breast US (Aixplorer; SuperSonic Imagine, Aix-en-Provence, France, or iU22; Philips Medical Systems, Bothell, USA) supplemented with surveillance using various imaging methods at 6-month intervals for the first 5 years and annually thereafter.
Whole-breast US routinely includes nodal screening of the axillary, internal mammary, and supraclavicular fossae. The lymph nodes were considered suspicious based on the radiologist’s overall assessment, considering features such as new or enlarging nodes, cortical thickening, hilar effacement, round morphology, indistinct margins, or abnormal non-hilar/peripheral vascularity [15]. US-guided fine-needle aspiration (FNA) was performed when lymphadenopathy was suspected.
In addition to annual mammography, supplemental breast/AUS and cross-sectional imaging are frequently used in clinical practice. In particular, chest and abdominal CT scans are commonly obtained at approximately 1–2 years to evaluate for regional or distant metastases. However, the exact interval and selection of additional imaging modalities were not strictly standardized and could vary depending on clinician judgment and individual patient circumstances. All CT scans were interpreted by board-certified radiologists with subspecialty training in thoracic and abdominal imaging. Lymph nodes detected on CT or 18F-FDG positron emission tomography (PET)/CT were reevaluated using targeted US, and FNA was performed when indicated.
When regional LNR was diagnosed on imaging or physical examination, systemic staging, including whole-body bone scan, chest CT, abdominal CT, or 18F-FDG PET/CT, was performed as clinically indicated to assess for synchronous distant metastasis.
Regional LNR: definition and data collection
Radiologic reports and medical records were retrospectively reviewed by two breast radiologists (KES and KMK) to identify factors associated with isolated regional LNR. The collected data included demographics (age, menopausal status, and body mass index), family history, BRCA status, surgical method, PBC pathology (histological type, grade, hormone receptor status, human epidermal growth factor receptor 2 [HER2] status, Ki-67, and stage), treatment details, key dates (surgery, recurrence, last follow-up, or death), and US and CT imaging characteristics of the regional LNR.
Hormone receptor positivity was defined as ≥ 1% of tumor cells staining positive for estrogen or progesterone receptors in immunohistochemical assays. HER2 status was evaluated according to the guidelines of the American Society of Clinical Oncology and the College of American Pathologists [16]. The tumor subtypes were categorized as hormone receptor-positive/HER2-negative, hormone receptor-positive/HER2-positive, hormone receptor-negative/HER2-positive, or triple-negative.
Regional LNR was defined as recurrence involving the ipsilateral axillary, supraclavicular, infraclavicular, or internal mammary lymph nodes in the absence of ipsilateral breast parenchymal recurrence. Other types of recurrence, including ipsilateral or contralateral breast recurrence and distant metastases, were considered as previously described [17]. For the detection mode, LNRs were categorized into imaging- and symptom-detected. Imaging-detected LNR was defined as recurrence identified on surveillance imaging, such as US or CT. Symptom-detected LNR was defined as recurrence identified through patient-reported symptoms or physical examination by a clinician. To evaluate the effect of isolated regional LNR on prognoses, cases of regional LNR with synchronous distant metastases were classified as distant metastases. The treatment for an isolated regional LNR was determined by a multidisciplinary tumor board comprising breast surgeons, medical oncologists, radiation oncologists, pathologists, plastic surgeons, and radiologists. Treatment details were collected from the medical records.
Statistical analyses
Continuous variables are reported as means ± standard deviations or medians (interquartile ranges [IQRs]), as appropriate, and categorical variables are presented as counts and percentages. For group comparisons according to the mode of detection, analysis of variance or the Kruskal–Wallis test was applied for continuous variables, as appropriate, based on distributional assumptions. For categorical variables, Pearson’s chi-squared (χ2) test was used.
To evaluate the risk factors for isolated regional LNR, we applied the Fine–Gray sub-distribution hazard model, treating ipsilateral or contralateral breast recurrence and distant metastases as competing events. Univariable analyses were first conducted, and variables with a two-sided p-value < 0.05 were entered into the multivariable Fine–Gray model. Variables that violated the proportional hazards assumption or exhibited multicollinearity were excluded from the analysis. Sub-distribution hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated. In addition, we reviewed the imaging characteristics of lymph nodes according to the mode of detection.
Survival analyses were performed to assess distant metastasis-free survival (DMFS) and overall survival (OS). To address potential lead-time bias, both outcomes were evaluated from two distinct time origins: 1) date of PBC surgery and 2) date of isolated regional LNR diagnosis. DMFS was defined as the interval until the first occurrence of distant metastases, and OS was defined as the interval until death from any cause. Survival distributions were estimated using the Kaplan–Meier method, and differences between groups were assessed using the log-rank test. Additionally, among the patients who developed isolated regional LNR, the time from PBC surgery to isolated regional LNR diagnosis was compared between the detection groups using a log-rank test.
All statistical analyses were performed using R version 3.6.3 (R Foundation for Statistical Computing, Vienna, Austria). Statistical significance was defined as a two-sided p-value < 0.05.
RESULTS
Isolated regional LNR incidence and characteristics
In a cohort of 13,406 women with a PHBC (mean age, 49.8 ± 10.0 years), during a median follow-up of 6.3 years (IQR, 4.9–8.6 years), 173 (1.3%) patients (mean age, 46.9 ± 11.6 years) developed regional LNR. The clinicopathological characteristics of patients with regional LNR, other recurrences, and the overall cohort are summarized in Supplementary Table 1. Of these, 47 patients in whom distant metastases were detected during subsequent whole-body evaluation were excluded. Finally, during a median follow-up of 3.1 years (IQR, 1.6–4.6 years), 126 (0.9%) isolated regional LNRs were detected. Among these, 118 and 8 were identified based on imaging and clinical symptoms, respectively. Among the imaging-detected LNRs, 89 and 29 were identified on surveillance breast US and CT, respectively. All symptom-detected LNRs presented as palpable masses, except in one patient who presented with chest discomfort. The anatomical distribution and treatment of isolated regional LNRs are described in detail in Supplementary Data 1. Table 1 presents the clinicopathological characteristics of patients with isolated regional LNRs according to the mode of detection. Isolated regional LNR was significantly associated with high T stages (p = 0.025) and neoadjuvant chemotherapy (p = 0.044).
Table 1. Clinicopathological characteristics of patients with isolated regional lymph node recurrence by mode of detection.
| Variables | Regional LNR (n = 126) | Imaging-detected LNR (n = 118) | Symptom-detected LNR (n = 8) | p-value | |
|---|---|---|---|---|---|
| Age (yr) | 47.35 ± 10.93 | 46.5 ± 11.09 | 48.88 ± 8.54 | 0.685 | |
| BMI (kg/m2) | 23.14 ± 3.28 | 23.21 ± 3.31 | 22.19 ± 3.83 | 0.397 | |
| Family history of breast cancer | > 0.999 | ||||
| No | 108 (85.7) | 101 (85.6) | 7 (87.5) | ||
| Yes | 18 (14.3) | 17 (14.4) | 1 (12.5) | ||
| BRCA mutation | 0.074 | ||||
| No | 27 (21.4) | 26 (22.0) | 1 (12.5) | ||
| Yes | 5 (4.0) | 4 (3.3) | 1 (12.5) | ||
| Not done | 94 (74.6) | 88 (74.6) | 6 (75.0) | ||
| Menopausal status | 0.938 | ||||
| Premenopausal | 73 (57.9) | 68 (57.6) | 5 (62.5) | ||
| Postmenopausal | 53 (42.1) | 50 (42.3) | 3 (37.5) | ||
| Breast surgery | 0.845 | ||||
| Breast conserving | 51 (40.5) | 47 (39.8) | 4 (50.0) | ||
| Mastectomy | 75 (59.5) | 71 (60.2) | 4 (50.0) | ||
| Axillary surgery | 0.411 | ||||
| SLNB | 57 (45.2) | 55 (46.6) | 2 (25.0) | ||
| ALND | 69 (54.8) | 63 (53.4) | 6 (75.0) | ||
| T stage | 0.025 | ||||
| T1 | 39 (31.0) | 38 (32.2) | 1 (12.5) | ||
| T2 | 62 (49.2) | 59 (50.0) | 3 (37.5) | ||
| T3 | 23 (18.3) | 20 (16.9) | 3 (37.5) | ||
| T4 | 2 (1.6) | 1 (0.8) | 1 (12.5) | ||
| N stage | 0.409 | ||||
| N0 | 48 (38.1) | 47 (39.8) | 1 (12.5) | ||
| N1 | 30 (23.8) | 28 (23.7) | 2 (25.0) | ||
| N2 | 31 (24.6) | 28 (23.7) | 3 (37.5) | ||
| N3 | 17 (13.5) | 15 (12.7) | 2 (25.0) | ||
| Stage | 0.030 | ||||
| Invasive stage I | 26 (20.6) | 25 (21.2) | 1 (12.5) | ||
| Invasive stage II | 47 (37.3) | 47 (39.9) | 0 (0) | ||
| Invasive stage III | 53 (42.1) | 46 (38.9) | 7 (87.5) | ||
| Histological type | 0.418 | ||||
| Invasive ductal carcinoma | 119 (94.4) | 112 (94.9) | 7 (87.5) | ||
| Invasive lobular carcinoma | 2 (1.6) | 2 (1.7) | 0 (0) | ||
| Other | 5 (4.0) | 4 (3.4) | 1 (12.5) | ||
| Nuclear grade* | 0.756 | ||||
| Low | 5 (4.0) | 5 (4.3) | 0 (0) | ||
| Intermediate | 61 (49.2) | 58 (50.0) | 3 (37.5) | ||
| High | 58 (46.8) | 53 (45.7) | 5 (62.5) | ||
| Histological grade† | 0.178 | ||||
| 1 | 5 (4.1) | 5 (4.3) | 1 (2.7) | ||
| 2 | 71 (57.7) | 68 (58.6) | 17 (45.9) | ||
| 3 | 47 (38.2) | 43 (37.1) | 19 (51.4) | ||
| Lymphovascular invasion | >0.999 | ||||
| Absent | 57 (45.2) | 53 (44.9) | 4 (50.0) | ||
| Present | 69 (54.8) | 65 (55.1) | 4 (50.0) | ||
| Extensive intraductal component | 0.301 | ||||
| Absent | 81 (64.3) | 74 (62.7) | 7 (87.5) | ||
| Present | 45 (35.7) | 44 (37.3) | 1 (12.5) | ||
| Resection margin‡ | 0.735 | ||||
| Negative | 62 (48.1) | 59 (50.9) | 3 (37.5) | ||
| Close (< 1 mm) | 57 (44.2) | 52 (44.8) | 5 (62.5) | ||
| Positive | 5 (3.9) | 5 (4.3) | 0 (0) | ||
| Ki-67 | 0.785 | ||||
| Low (< 20%) | 65 (51.6) | 60 (50.8) | 5 (62.5) | ||
| High (≥ 20%) | 61 (48.4) | 58 (49.2) | 3 (37.5) | ||
| Molecular subtype | 0.600 | ||||
| Hormone receptor+/HER2− | 79 (62.7) | 75 (63.6) | 4 (50.0) | ||
| Hormone receptor+/HER2+ | 12 (9.5) | 11 (9.3) | 1 (12.5) | ||
| Hormone receptor−/HER2+ | 8 (6.3) | 8 (6.8) | 0 (0) | ||
| TNBC | 27 (21.4) | 24 (20.3) | 3 (37.5) | ||
| Neoadjuvant chemotherapy | 0.044 | ||||
| No | 81 (64.3) | 79 (66.9) | 2 (25.0) | ||
| Yes | 45 (35.7) | 39 (33.1) | 6 (75.0) | ||
| Adjuvant chemotherapy | 0.075 | ||||
| No | 64 (50.8) | 57 (48.3) | 7 (87.5) | ||
| Yes | 62 (49.2) | 61 (51.7) | 1 (12.5) | ||
| Adjuvant radiotherapy | 0.055 | ||||
| No | 48 (38.1) | 48 (40.7) | 0 (0) | ||
| Yes | 78 (61.9) | 70 (59.3) | 8 (100.0) | ||
| Adjuvant endocrine therapy | > 0.999 | ||||
| No | 35 (27.8) | 33 (28.0) | 2 (25.0) | ||
| Yes | 91 (72.2) | 85 (72.0) | 6 (75.0) | ||
| HER2-targeted therapy | 0.489 | ||||
| No | 112 (88.9) | 105 (89.0) | 7 (87.5) | ||
| Yes | 14 (11.1) | 13 (11.0) | 1 (12.5) | ||
Data are presented as means ± standard deviations or number of patients with percentages in parentheses.
LNR = lymph node recurrence; BMI = body mass index; SLNB = sentinel lymph node biopsy; ALND = axillary lymph node dissection; HER2 = human epidermal growth factor receptor 2; TNBC = triple-negative breast cancer.
*Data are missing for two cancers. †Data are missing for three cancers. ‡Data are missing for two cancers.
Among 126 patients with isolated regional LNRs, 36 (28.6%) developed distant metastases, and 22 (17.5%) died. Of the eight patients with symptom-detected regional LNRs, five (62.5%) subsequently developed distant metastases, and two (25.0%) died. Additionally, in the overall cohort of 13,406 patients, 1,008 (7.5%) experienced other types of recurrences, including 189 ipsilateral breast recurrences, 256 contralateral breast recurrences, and 563 distant metastases.
The imaging characteristics of the regional LNRs based on the detection mode are presented in Table 2. The median long-axis diameter of lymph nodes was larger in the CT-detected group (13 mm; IQR, 11.75–18 mm) than in the US-detected group (10 mm; IQR, 8–14 mm). The long-to-short axis ratio was lower in the US-detected group (1.72; IQR, 1.44–2.22) than in the other detection modes. The distribution of the nodal regions differed between the groups. Level I nodes were most frequently detected using US (51.6%), whereas internal mammary nodes were more often detected using CT (58.1%) (Figure 2, Supplementary Figure 1). Symptom-detected recurrences were most frequently located in the supraclavicular region (50.0%) (Supplementary Figure 2). Non-circumscribed margins and hypervascularity were more frequently noted in symptomatic nodes on US (n = 125). In CT (n = 84), necrosis was more common in symptom-detected nodes (100%) than in CT- (42.1%) or US-detected nodes (30.0%).
Table 2. Radiologic characteristics of regional lymph node recurrence according to detection mode.
| Variables | Mode of detection | ||||
|---|---|---|---|---|---|
| Imaging-detected | Symptom-detected (n = 8) | ||||
| US-detected (n = 89) | CT-detected (n = 29)* | ||||
| Nodal region† | |||||
| Level I | 47 (51.6) | 5 (16.1) | 3 (37.5) | ||
| Levels II–III | 9 (9.9) | 2 (6.5) | 0 (0) | ||
| IMN | 21 (23.1) | 18 (58.1) | 1 (12.5) | ||
| SCN | 14 (15.4) | 6 (19.4) | 4 (50.0) | ||
| US finding (n = 125)* | |||||
| Lymph node size (long axis, mm) | 10 (8–14) | 13 (11.75–18) | 12.5 (11.75–18.75) | ||
| Lymph node size (short axis, mm) | 6 (4–8) | 5 (5–8) | 6.5 (4.13–12.25) | ||
| L/S ratio | 1.72 (1.44–2.22) | 2.67 (1.85–2.62) | 2.39 (1.98–2.63) | ||
| Cortical thickness > 3 mm | |||||
| Yes | 82 (92.1) | 25 (89.3) | 8 (100.0) | ||
| No | 7 (7.9) | 3 (10.7) | 0 (0) | ||
| Cortical thickening | |||||
| Diffuse | 69 (77.5) | 18 (64.3) | 4 (50.0) | ||
| Focal or eccentric | 20 (22.5) | 10 (35.7) | 4 (50.0) | ||
| Fatty hilum | |||||
| Preserved | 3 (3.4) | 0 (0) | 0 (0) | ||
| Partial loss | 38 (42.7) | 14 (50.0) | 5 (62.5) | ||
| Complete loss | 48 (53.9) | 14 (50.0) | 3 (37.5) | ||
| Margin | |||||
| Circumscribed | 66 (74.2) | 20 (71.4) | 3 (37.5) | ||
| Non-circumscribed | 23 (25.8) | 8 (28.6) | 5 (62.5) | ||
| Vascularity | |||||
| Yes | 48 (53.9) | 18 (64.3) | 7 (87.5) | ||
| No | 23 (25.9) | 8 (28.6) | 1 (12.5) | ||
| Not performed‡ | 18 (20.2) | 2 (7.1) | 0 (0) | ||
| CT finding (n = 84)§ | 60 | 20 | 4 | ||
| Visibility | |||||
| Yes | 50 (83.3) | 19 (95.0) | 3 (75.0) | ||
| No | 10 (16.7) | 1 (5.0) | 1 (25.0) | ||
| Necrosis | |||||
| Yes | 15 (30.0) | 8 (42.1) | 3 (100.0) | ||
| No | 35 (70.0) | 11 (57.9) | 0 (0) | ||
| Extranodal extension | |||||
| Yes | 4 (8.0) | 2 (10.5) | 1 (33.3) | ||
| No | 46 (92.0) | 17 (89.5) | 2 (66.7) | ||
Data are presented as number of patients with percentages in parentheses or median (interquartile ranges).
US = ultrasound; CT = computed tomography; IMN = internal mammary node; SCN = supraclavicular node; L/S ratio = long-axis to short-axis ratio.
*One lesion that was undetectable on US because of its deep location was excluded and later diagnosed by excisional biopsy.
†Cases with simultaneous involvement of multiple nodal regions (e.g., one axillary level I and IMN detected on US or three axillary levels I and SCN, with two detected on US and one on CT) were counted separately according to each involved location.
‡Assessment of vascularity was limited to lymph nodes located in the deep portions.
§CT findings are described when CT scans are available.
Figure 2. Ultrasound-detected axillary lymph node recurrence. Regional lymph node recurrence detected by ultrasound in a 57-year-old woman, 18 months after neoadjuvant chemotherapy, total mastectomy, and sentinel lymph node biopsy of the left breast, followed by radiotherapy and chemotherapy for invasive ductal carcinoma (ER−/PR−/HER2−, pT3N0). Transverse ultrasound imaging revealed round, hypoechoic lymph nodes measuring 0.4 cm or less (solid arrows) with irregular margins in the right level I, which were confirmed as metastatic by fine-needle aspiration. Despite palliative chemotherapy and radiotherapy, lung and mediastinum lymph node recurrence was diagnosed 12 months later, but the patient remained alive at follow-up.
ER = estrogen receptor; PR = progesterone receptor; HER2 = human epidermal growth factor receptor 2.
Risk factors for isolated regional LNR
Factors associated with the risk of isolated regional LNR are shown in Table 3. In the multivariable analysis, higher T stages (T2: HR, 1.71; 95% CI, 1.11–2.64; p = 0.014; T3: HR, 3.08; 95% CI, 1.68–5.65; p < 0.001), higher N stages (N2: HR, 2.97; 95% CI, 1.54–5.75; p = 0.001; N3: HR, 3.12; 95% CI, 1.36–7.13; p = 0.007), triple-negative subtypes (HR, 1.71; 95% CI, 1.09–2.67; p = 0.019), and lack of adjuvant radiotherapy (HR, 4.35; 95% CI, 2.86–6.67; p < 0.001) were identified as significant predictors of increased risk of isolated regional LNR.
Table 3. Univariable and multivariable Fine–Gray competing risk models for isolated regional lymph node recurrence.
| Variables | Univariable | Multivariable | |||
|---|---|---|---|---|---|
| HR (95% CI) | p-value | HR (95% CI) | p-value | ||
| Age | 0.98 (0.96–1.00) | 0.026 | |||
| Family history of breast cancer | |||||
| No | |||||
| Yes | 0.91 (0.55–1.50) | 0.710 | |||
| BRCA mutation | |||||
| No | Ref. | ||||
| Yes | 0.83 (0.25–2.73) | 0.760 | |||
| Menopausal status | |||||
| Postmenopausal | Ref. | ||||
| Premenopausal | 1.04 (0.73–1.48) | 0.820 | |||
| Breast surgery | |||||
| Breast conserving | Ref. | ||||
| Mastectomy | 3.22 (2.26–4.60) | < 0.001 | |||
| Axillary surgery | |||||
| SLNB | Ref. | Ref. | |||
| ALND | 2.85 (2.01–4.05) | < 0.001 | 1.68 (1.00–2.83) | 0.051 | |
| Not done* | NA | NA | NA | NA | |
| T stage | < 0.001 | 0.003 | |||
| T1 | Ref. | Ref. | |||
| T2 | 2.45 (1.64–3.66) | < 0.001 | 1.71 (1.11–2.64) | 0.014 | |
| T3 | 5.37 (3.21–8.99) | < 0.001 | 3.08 (1.68–5.65) | < 0.001 | |
| T4 | 3.00 (0.72–12.47) | 0.130 | 1.28 (0.27–6.03) | 0.750 | |
| N stage | < 0.001 | < 0.001 | |||
| N0 | Ref. | Ref. | |||
| N1 | 1.52 (0.96–2.40) | 0.072 | 0.92 (0.49–1.73) | 0.790 | |
| N2 | 4.06 (2.58–6.37) | < 0.001 | 2.97 (1.54–5.75) | 0.001 | |
| N3 | 4.53 (2.60–7.90) | < 0.001 | 3.12 (1.36–7.13) | 0.007 | |
| Stage | < 0.001 | ||||
| Invasive stage I | Ref. | ||||
| Invasive stage II | 1.99 (1.23–3.22) | 0.005 | |||
| Invasive stage III | 5.42 (3.39–8.67) | < 0.001 | |||
| Histological type | 0.257 | ||||
| Invasive ductal carcinoma | Ref. | ||||
| Invasive lobular carcinoma | 0.34 (0.09–1.39) | 0.130 | |||
| Other | 0.73 (0.30–1.78) | 0.480 | |||
| Nuclear grade | < 0.001 | ||||
| Low | Ref. | ||||
| Intermediate | 2.82 (1.13–7.03) | 0.026 | |||
| High | 4.82 (1.94–11.98) | 0.001 | |||
| Histologic grade | < 0.001 | ||||
| 1 | Ref. | ||||
| 2 | 6.80 (2.75–16.83) | < 0.001 | |||
| 3 | 7.78 (3.11–19.48) | < 0.001 | |||
| Lymphovascular invasion | |||||
| Absent | Ref. | ||||
| Present | 3.23 (2.27–4.59) | < 0.001 | |||
| Extensive intraductal component | |||||
| Absent | Ref. | ||||
| Present | 1.41 (0.98–2.03) | 0.064 | |||
| Resection margin | |||||
| Negative | Ref. | ||||
| Positive or close | 1.31 (0.92–1.87) | 0.140 | |||
| Ki-67 | |||||
| Low (< 20%) | Ref. | ||||
| High (≥ 20%) | 1.43 (1.01–2.04) | 0.043 | |||
| Subtype | 0.017 | 0.004 | |||
| Hormone receptor+/HER2− | Ref. | Ref. | |||
| Hormone receptor+/HER2+ | 0.86 (0.47–1.57) | 0.610 | 0.65 (0.35–1.19) | 0.160 | |
| Hormone receptor−/HER2+ | 0.67 (0.32–1.39) | 0.280 | 0.53 (0.25–1.10) | 0.089 | |
| TNBC | 1.77 (1.15–2.75) | 0.010 | 1.71 (1.09–2.67) | 0.019 | |
| Neoadjuvant chemotherapy | |||||
| Yes | Ref. | ||||
| No | 0.29 (0.20–0.42) | < 0.001 | |||
| Adjuvant chemotherapy | |||||
| Yes | Ref. | ||||
| No | 0.93 (0.66–1.32) | 0.690 | |||
| Adjuvant radiotherapy | |||||
| Yes | Ref. | Ref. | |||
| No | 2.27 (1.59–3.23) | < 0.001 | 4.35 (2.86–6.67) | < 0.001 | |
| Adjuvant endocrine therapy | |||||
| Yes | Ref. | ||||
| No | 1.35 (0.92–2.00) | 0.120 | |||
| HER2-targeted therapy | |||||
| Yes | Ref. | ||||
| No | 1.45 (0.76–2.70) | 0.260 | |||
HRs were estimated from subdistribution hazard models by considering ipsilateral or contralateral breast tumor recurrence and distant metastasis as competing events.
HR = hazard ratio; CI = confidence interval; SLNB = sentinel lymph node biopsy; ALND = axillary lymph node dissection; NA = not applicable; HER2 = human epidermal growth factor receptor 2; TNBC = triple-negative breast cancer.
*Excluded from the analysis because of the absence of recurrence.
Survival outcomes according to the mode of detection
The median time from PBC surgery to isolated regional LNR diagnosis was 1.77 years (95% CI, 1.63–NA) in the symptom-detected group and 3.10 years (95% CI, 2.68–3.55) in the imaging-detected group, but the difference was not statistically significant (p = 0.100) (Supplementary Figure 3). When survival outcomes were analyzed from the time of isolated regional LNR diagnosis, the 5-year DMFS rate was significantly lower in the symptom-detected than in the imaging-detected group (30.0%, 95% CI, 9.54%–94.3% vs. 66.5%, 95% CI, 57.1%–77.5%; p = 0.007), whereas the 5-year OS rate was similar in the two groups (85.7%, 95% CI, 63.3%–100.0% vs. 89.7%, 95% CI, 84.1–95.7%; p = 0.900) (Figure 3). Similar results were obtained when survival outcomes were analyzed at the time of PBC surgery; a significant between-group difference in DMFS persisted (p = 0.010), with no significant difference in OS (p = 0.600) (Supplementary Figure 4).
Figure 3. Survival outcomes after isolated regional lymph node recurrence diagnosis according to mode of detection. (A) Distant metastasis-free survival and (B) overall survival.
DISCUSSION
In this study of 13,406 patients with stage I–III invasive breast cancer, we identified 126 cases of isolated regional LNR over a median follow-up of 6.3 years. Isolated regional LNR was associated with a higher tumor stage, triple-negative subtype, and omission of adjuvant radiotherapy. Modality-specific detection patterns were evident: US most often identified level I axillary nodes and tended to detect smaller, rounder nodes; CT preferentially detected deep nodal regions, particularly internal mammary nodes, with larger long-axis dimensions; and symptom-detected events were more frequent in the supraclavicular fossa. The mode of detection of isolated regional LNR significantly affected DMFS but had no impact on OS.
Interval breast cancers are associated with adverse clinicopathological features, such as larger size, more nodal involvement, higher grade, and more advanced stage, and they more often show triple-negative or HER2-positive status; accordingly, survival is worse than that for screening-detected cancers [13]. In our setting of intensive imaging surveillance, symptom-detected isolated regional LNR can be viewed as an “interval” event. Two caveats limit this interpretation. First, we evaluated nodal recurrence rather than new primary tumors. Second, both detection groups (imaging- and symptom-detected LNR) represent patients with regional recurrence, which is associated with an increased risk of systemic disease. Moreover, although the interval from PBC surgery to LNR diagnosis did not differ significantly between groups, symptom-detected cases tended to recur earlier, which is consistent with greater biological aggressiveness.
In analyses stratified by detection mode, the anatomical distribution of nodal recurrences differed. US most often identified level I axillary nodes that are readily accessible to US, whereas CT more commonly detected internal mammary disease, which is frequently outside of or obscured on routine breast US. Symptom-detected events more often involved the supraclavicular fossa, which is consistent with a palpable superficial presentation. US tended to detect smaller and rounder nodes (median long axis of 10 mm and lower long-to-short axis ratio), whereas CT- and symptom-detected recurrences generally had larger long-axis dimensions; importantly, the superficial location of supraclavicular nodes means they may be detected clinically even when they are relatively small. On US, non-circumscribed margins were observed more often in symptom-detected nodes, whereas on CT, extranodal extension and necrosis were more frequent in the symptom-detected group, a pattern compatible with a more aggressive morphology. Taken together, these observations suggest that the modality-specific visibility and anatomical accessibility of specific nodal regions determine where and how nodal recurrences are detected, whereas the higher rates of necrosis and extranodal extension in symptom-detected cases may indicate a more aggressive disease biology in that subset.
AUS performed during routine breast US is technically simple, takes only 1–2 minutes, and can detect asymptomatic regional recurrences [8]. Therefore, it is commonly incorporated into routine surveillance in Korea to facilitate timely diagnosis and improve patient outcomes. However, the reported diagnostic performance is heterogeneous, with sensitivities ranging from 24% to 94% in a meta-analysis [18]. A recent study examining AUS performance in patients with a PHBC reported an axillary recurrence rate of 0.2 per 1,000 AUS examinations (1 in 4,430), with an interval axillary recurrence rate of 0.9 per 1,000 (4 in 4,402) and a sensitivity of 20.0% [9]. However, in our cohort, approximately two-thirds of the isolated regional LNRs were identified on surveillance US, suggesting a more substantial role for imaging in detecting clinically occult nodal recurrences within an intensively surveyed population.
In this context, although symptom-detected LNRs were associated with worse DMFS, imaging-based detection was not associated with improved OS during follow-up. This discrepancy may be explained by the small number of symptom-detected cases and the potential impact of effective salvage and systemic therapies, which may mitigate survival differences despite the more aggressive disease biology. Notably, symptom-detected cases were more frequently associated with higher T stages at the initial diagnosis, which is a well-established adverse prognostic factor. Thus, the worse DMFS observed in the symptom-detected group may, at least in part, reflect differences in the baseline tumor burden and disease biology rather than the mode of detection itself. Accordingly, our findings indicate an association between the detection mode and recurrence behavior within an intensively surveyed cohort, rather than a causal effect on survival outcomes. Because our study lacked a non-surveillance comparator, we could not determine whether routine imaging improved OS in asymptomatic survivors. Nevertheless, within an intensively surveyed program, imaging identified the most isolated regional nodal recurrences and influenced diagnostic and planning pathways. These observations support risk-adapted rather than uniform surveillance and motivate program-level endpoints (e.g., interval events, nodal burden, time to treatment, and cost-effectiveness) for future studies. Prospective multicenter studies, ideally incorporating risk-adapted surveillance strategies and cost-effectiveness analyses, are needed to clarify the role of AUS in healthcare settings.
This study had some limitations. First, the single-center retrospective design may have introduced a selection bias. Second, US examinations were performed by multiple radiologists with varying experience and operator dependence, and differences in the sonographic accessibility of specific nodal regions (e.g., the internal mammary chain vs. axillary level I) may have affected the results. Third, the nodal burden at recurrence could not be uniformly quantified, and surrogate imaging measures (size, long-to-short axis ratio, and necrosis) were used, leaving potential residual confounding factors. Fourth, the isolated regional LNR and symptom-detected subgroups were small, limiting statistical power. Fifth, CT surveillance was not standardized, which may have introduced selection bias and limited generalizability. Sixth, trastuzumab administration was influenced by the treatment era and clinical factors; therefore, confounding factors, such as indication, era, and selection bias, cannot be excluded. Seventh, detailed data on the organ-specific distribution and treatment of distant metastases are not uniformly available. In addition, because most recurrent lesions were nodal recurrences confirmed by cytology or limited biopsy, detailed pathological and biomarker features were frequently unavailable. Therefore, differences in metastatic burden, treatment, and tumor biology may have confounded the association between detection mode and survival outcomes. Finally, because this study compared outcomes by detection mode within an intensive surveillance setting, we could not estimate the absolute survival benefit of surveillance compared to no surveillance. Rather, it characterizes associations within a surveyed cohort. In this analysis, OS did not differ between the detection modes.
In conclusion, within an intensive surveillance setting, most isolated regional LNRs were identified during routine imaging; however, OS did not differ according to the detection mode. Taken together, these findings suggest that long-term outcomes are driven more by disease biology than by how recurrence is detected. The incremental survival benefit of routine imaging-based surveillance (e.g., AUS) in asymptomatic survivors remains uncertain. Prospective multicenter evaluations of risk-adapted surveillance strategies are warranted.
Footnotes
Conflict of Interest: The authors declare that they have no competing interests.
Data Availability: Requests for data can be made by contacting the corresponding author.
- Conceptualization: Kim MK, Kim SW, Choi JS, Ko EY, Han BK, Yu J, Ko ES.
- Data curation: Kim MK, Kim SW, Ahn JH, Kim H, Choi JS, Ko EY, Han BK, Yu J, Ko ES.
- Formal analysis: Ahn JH.
- Investigation: Kim MK, Kim H, Yu J, Ko ES.
- Methodology: Kim MK, Yu J, Ko ES.
- Project administration: Kim MK, Yu J, Ko ES.
- Resources: Yu J, Ko ES.
- Supervision: Yu J, Ko ES.
- Validation: Kim MK, Kim SW.
- Visualization: Kim MK, Yu J, Ko ES.
- Writing - original draft: Kim MK, Yu J, Ko ES.
- Writing - review & editing: Kim MK, Kim SW, Kim H, Choi JS, Ko EY, Han BK, Yu J, Ko ES.
SUPPLEMENTARY MATERIALS
Anatomic distribution and treatments of isolated regional lymph node recurrences
Clinicopathological characteristics of patients
Chest computed tomography-detected internal mammary lymph node recurrence with initially negative ultrasound findings. Regional lymph node recurrence detected by chest CT in a 47-year-old woman, 37 months after total mastectomy of the left breast followed by radiotherapy, chemotherapy, and antihormonal therapy for invasive ductal carcinoma (ER+/PR−/HER2−, pT2N0). (A) Despite a negative surveillance ultrasound, contrast-enhanced chest CT obtained 3 days later revealed a 1.2-cm enhancing lymph node (dashed arrow) in the left second intercostal area. (B, C) Targeted transverse and longitudinal ultrasound imaging revealed a hypoechoic lymph node (solid arrows) in the same location that was confirmed to be metastatic by fine-needle aspiration. The patient subsequently underwent lymph node excision and chemotherapy, remained alive at follow-up, and experienced contralateral breast cancer recurrence 7 years later.
Symptom-detected supraclavicular lymph node recurrence. Regional lymph node recurrence detected by palpation in a 54-year-old woman, 20 months after total mastectomy of the left breast following neoadjuvant chemotherapy, radiotherapy, and antihormonal therapy for invasive ductal carcinoma (ER+/PR+/HER2−, pT2N2). Targeted ultrasound (US) imaging of the palpable site revealed an enlarged lymph node in the left supraclavicular area, with a round morphology (solid arrow). US-guided core-needle biopsy confirmed metastasis. The patient was diagnosed with bone metastasis 1 year after recurrence and died 4 years later.
Time to isolated regional lymph node recurrence after primary breast cancer surgery according to the mode of detection.
Survival outcomes after primary breast cancer surgery according to the mode of detection. (A) Distant metastasis-free survival and (B) overall survival.
References
- 1.Fisher B, Anderson S, Bryant J, Margolese RG, Deutsch M, Fisher ER, et al. Twenty-year follow-up of a randomized trial comparing total mastectomy, lumpectomy, and lumpectomy plus irradiation for the treatment of invasive breast cancer. N Engl J Med. 2002;347:1233–1241. doi: 10.1056/NEJMoa022152. [DOI] [PubMed] [Google Scholar]
- 2.Veronesi U, Cascinelli N, Mariani L, Greco M, Saccozzi R, Luini A, et al. Twenty-year follow-up of a randomized study comparing breast-conserving surgery with radical mastectomy for early breast cancer. N Engl J Med. 2002;347:1227–1232. doi: 10.1056/NEJMoa020989. [DOI] [PubMed] [Google Scholar]
- 3.Blichert-Toft M, Nielsen M, Düring M, Møller S, Rank F, Overgaard M, et al. Long-term results of breast conserving surgery vs. mastectomy for early stage invasive breast cancer: 20-year follow-up of the Danish randomized DBCG-82TM protocol. Acta Oncol. 2008;47:672–681. doi: 10.1080/02841860801971439. [DOI] [PubMed] [Google Scholar]
- 4.Baek JY, Choi DH, Park W, Kim H, Cho WK, Yoo GS. Survival and prognostic factors for breast cancer patients with regional oligo-recurrence. J Breast Cancer. 2020;23:622–634. doi: 10.4048/jbc.2020.23.e66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Runowicz CD, Leach CR, Henry NL, Henry KS, Mackey HT, Cowens-Alvarado RL, et al. American Cancer Society/American Society of Clinical Oncology Breast Cancer Survivorship Care Guideline. J Clin Oncol. 2016;34:611–635. doi: 10.1200/JCO.2015.64.3809. [DOI] [PubMed] [Google Scholar]
- 6.National Comprehensive Cancer Network. NCCN Clinical Practice Guidelines in Oncology: Breast Cancer. Version 2. Plymouth Meeting: National Comprehensive Cancer Network; 2025. [Google Scholar]
- 7.Berg WA, Blume JD, Cormack JB, Mendelson EB, Lehrer D, Böhm-Vélez M, et al. Combined screening with ultrasound and mammography vs mammography alone in women at elevated risk of breast cancer. JAMA. 2008;299:2151–2163. doi: 10.1001/jama.299.18.2151. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Moon HJ, Kim MJ, Kim EK, Park BW, Youk JH, Kwak JY, et al. US surveillance of regional lymph node recurrence after breast cancer surgery. Radiology. 2009;252:673–681. doi: 10.1148/radiol.2523081977. [DOI] [PubMed] [Google Scholar]
- 9.Shin SU, Chang JM, Park J, Lee HB, Han W, Moon WK. The usefulness of ultrasound surveillance for axillary recurrence in women with personal history of breast cancer. J Breast Cancer. 2022;25:25–36. doi: 10.4048/jbc.2022.25.e3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Porter PL, El-Bastawissi AY, Mandelson MT, Lin MG, Khalid N, Watney EA, et al. Breast tumor characteristics as predictors of mammographic detection: comparison of interval- and screen-detected cancers. J Natl Cancer Inst. 1999;91:2020–2028. doi: 10.1093/jnci/91.23.2020. [DOI] [PubMed] [Google Scholar]
- 11.Joensuu H, Lehtimäki T, Holli K, Elomaa L, Turpeenniemi-Hujanen T, Kataja V, et al. Risk for distant recurrence of breast cancer detected by mammography screening or other methods. JAMA. 2004;292:1064–1073. doi: 10.1001/jama.292.9.1064. [DOI] [PubMed] [Google Scholar]
- 12.Domingo L, Salas D, Zubizarreta R, Baré M, Sarriugarte G, Barata T, et al. Tumor phenotype and breast density in distinct categories of interval cancer: results of population-based mammography screening in Spain. Breast Cancer Res. 2014;16:R3. doi: 10.1186/bcr3595. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Houssami N, Hunter K. The epidemiology, radiology and biological characteristics of interval breast cancers in population mammography screening. NPJ Breast Cancer. 2017;3:12. doi: 10.1038/s41523-017-0014-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Edge SB, Byrd DR, Compton CC, Fritz AG, Greene FL, Trotti A. AJCC Cancer Staging Manual. 7th ed. New York: Springer; 2010. [Google Scholar]
- 15.Chung HL, Le-Petross HT, Leung JWT. Imaging updates to breast cancer lymph node management. Radiographics. 2021;41:1283–1299. doi: 10.1148/rg.2021210053. [DOI] [PubMed] [Google Scholar]
- 16.Wolff AC, Hammond MEH, Allison KH, Harvey BE, Mangu PB, Bartlett JMS, et al. Human epidermal growth factor receptor 2 testing in breast cancer: American Society of Clinical Oncology/College of American Pathologists Clinical Practice Guideline Focused Update. J Clin Oncol. 2018;36:2105–2122. doi: 10.1200/JCO.2018.77.8738. [DOI] [PubMed] [Google Scholar]
- 17.Wapnir IL, Anderson SJ, Mamounas EP, Geyer CE, Jr, Jeong JH, Tan-Chiu E, et al. Prognosis after ipsilateral breast tumor recurrence and locoregional recurrences in five National Surgical Adjuvant Breast and Bowel Project node-positive adjuvant breast cancer trials. J Clin Oncol. 2006;24:2028–2037. doi: 10.1200/JCO.2005.04.3273. [DOI] [PubMed] [Google Scholar]
- 18.Le Boulc’h M, Gilhodes J, Steinmeyer Z, Molière S, Mathelin C. Pretherapeutic imaging for axillary staging in breast cancer: a systematic review and meta-analysis of ultrasound, MRI and FDG PET. J Clin Med. 2021;10:1543. doi: 10.3390/jcm10071543. [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.
Supplementary Materials
Anatomic distribution and treatments of isolated regional lymph node recurrences
Clinicopathological characteristics of patients
Chest computed tomography-detected internal mammary lymph node recurrence with initially negative ultrasound findings. Regional lymph node recurrence detected by chest CT in a 47-year-old woman, 37 months after total mastectomy of the left breast followed by radiotherapy, chemotherapy, and antihormonal therapy for invasive ductal carcinoma (ER+/PR−/HER2−, pT2N0). (A) Despite a negative surveillance ultrasound, contrast-enhanced chest CT obtained 3 days later revealed a 1.2-cm enhancing lymph node (dashed arrow) in the left second intercostal area. (B, C) Targeted transverse and longitudinal ultrasound imaging revealed a hypoechoic lymph node (solid arrows) in the same location that was confirmed to be metastatic by fine-needle aspiration. The patient subsequently underwent lymph node excision and chemotherapy, remained alive at follow-up, and experienced contralateral breast cancer recurrence 7 years later.
Symptom-detected supraclavicular lymph node recurrence. Regional lymph node recurrence detected by palpation in a 54-year-old woman, 20 months after total mastectomy of the left breast following neoadjuvant chemotherapy, radiotherapy, and antihormonal therapy for invasive ductal carcinoma (ER+/PR+/HER2−, pT2N2). Targeted ultrasound (US) imaging of the palpable site revealed an enlarged lymph node in the left supraclavicular area, with a round morphology (solid arrow). US-guided core-needle biopsy confirmed metastasis. The patient was diagnosed with bone metastasis 1 year after recurrence and died 4 years later.
Time to isolated regional lymph node recurrence after primary breast cancer surgery according to the mode of detection.
Survival outcomes after primary breast cancer surgery according to the mode of detection. (A) Distant metastasis-free survival and (B) overall survival.