Authors: Kadhim Taqi (1Division of Surgical Oncology, Department of Surgery, University of Calgary, Calgary, AB T2N 1N4, Canada; 2Division of Surgical Oncology, Department of Surgery, Sultan Qaboos Comprehensive Cancer Care and Research Center, University Medical City, Muscat 123, Oman), Camelia Ursu (3Cumming School of Medicine, University of Calgary, Calgary, AB T2N 1N4, Canada), Susan Isherwood (4Department of Surgery, University of Calgary, Calgary, AB T2N 1N4, Canada), Rabia Raheel (4Department of Surgery, University of Calgary, Calgary, AB T2N 1N4, Canada), Julia Downey (5Department of Community Health Sciences, University of Calgary, Calgary, AB T2N 1N4, Canada), Jeffrey Q. Cao (6Division of Radiation Oncology, Arthur J.E. Child Comprehensive Cancer Centre, Calgary, AB T2N 1N4, Canada; 7Department of Oncology, University of Calgary, Calgary, AB T2N 1N4, Canada), Sasha Lupichuk (8Division of Medical Oncology, Department of Oncology, University of Calgary, Calgary, AB T2N 1N4, Canada), Omar Khan (8Division of Medical Oncology, Department of Oncology, University of Calgary, Calgary, AB T2N 1N4, Canada), Nancy Nixon (8Division of Medical Oncology, Department of Oncology, University of Calgary, Calgary, AB T2N 1N4, Canada), May Lynn Quan (1Division of Surgical Oncology, Department of Surgery, University of Calgary, Calgary, AB T2N 1N4, Canada; 5Department of Community Health Sciences, University of Calgary, Calgary, AB T2N 1N4, Canada; 7Department of Oncology, University of Calgary, Calgary, AB T2N 1N4, Canada), Alison Laws (1Division of Surgical Oncology, Department of Surgery, University of Calgary, Calgary, AB T2N 1N4, Canada; 7Department of Oncology, University of Calgary, Calgary, AB T2N 1N4, Canada)
Categories: Article, staging investigations, breast cancer, neoadjuvant chemotherapy, locally advanced, systemic staging
Source: Current Oncology
Authors: Kadhim Taqi, Camelia Ursu, Susan Isherwood, Rabia Raheel, Julia Downey, Jeffrey Q. Cao, Sasha Lupichuk, Omar Khan, Nancy Nixon, May Lynn Quan, Alison Laws
Radiological imaging is commonly used in patients with breast cancer to assess disease stage before starting chemotherapy, but its benefit varies by stage. Using population-wide data, we studied how often these tests detected cancer that had already spread in patients planned for chemotherapy before surgery. We found that cancer spread was very uncommon in patients with small tumors and no lymph node involvement, even when chemotherapy was planned, suggesting that routine staging scans can be safely avoided in this group. In contrast, patients with larger tumors or lymph node involvement had a higher likelihood of cancer spread. These results support more selective use of staging tests, which could reduce unnecessary scans, costs, patient anxiety, and treatment delays, while helping refine future guidelines and policy decisions.
Staging investigations alter the management of patients newly diagnosed with breast cancer when distant metastatic disease is identified. In patients with locoregionally advanced breast cancer (LABC), systemic staging is recommended due to the relatively high likelihood of detecting distant metastases [1,2,3]. Conversely, the diagnostic yield of staging is much lower for early breast cancer (EBC), and thus its use is more debated, with current guidelines varying in their recommendations. The National Comprehensive Cancer Network (NCCN), American Society of Clinical Oncology (ASCO) and Cancer Care Ontario discourage routine staging for asymptomatic patients with stage I or II disease [1,4,5]. In contrast, the European Society for Medical Oncology (ESMO) supports the use of staging for patients with node-positive disease, tumors larger than 5 cm (T3), or aggressive tumor biology (3), while the American College of Radiology (ACR) recommends staging for patients with stage IIB or higher disease [6].
Patients referred for neoadjuvant chemotherapy (NAC) represent a distinct subgroup encompassing both early and locally advanced disease. These patients often present with higher-risk features that render them eligible for NAC such as node positivity or more aggressive biologic subtypes (even when presenting with a lower anatomic clinical stage). The diagnostic yield and clinical utility of staging investigations in this population is not well-characterized. Multiple studies have shown high rates of staging in patients with EBC, especially before NAC initiation [7,8,9]. Overuse of staging investigations is costly [7,10] and may expose patients to harms such as unnecessary radiation, invasive follow-up procedures, and anxiety over false-positive findings [11]. Further, staging workups can delay the timely initiation of therapy, potentially impacting treatment outcomes [12]. Nevertheless, accurate assessment of disease extent remains crucial for prognostication and to guide treatment strategies, including avoidance of chemotherapy toxicity for some patients when metastatic disease is confirmed.
Across the six tertiary and regional cancer centers in the province of Alberta, Canada, distant staging investigations are performed routinely for patients with breast cancer planned for NAC. The objective of this study was to evaluate the rate of de novo metastatic (M1) disease detected in staging investigations among patients referred for NAC, stratified by EBC and LABC, and to identify factors associated with M1 disease.
Institutional review board approval was obtained from the Health Research Ethics Board of Alberta–Cancer Committee (HREBA-CC), and a waiver of consent was granted given the minimal-risk nature of the study. All patients with breast cancer referred to medical oncology for consideration of NAC between 2020 and 2022 were identified from the Alberta Cancer Registry, which captures all new cancers in the province of approximately five million individuals. Patients who did not undergo any distant staging investigations were excluded. Two cohorts were generated based on the clinical T/N stage (cT/cN) at presentation according to the American Joint Committee on Cancer (AJCC) 8th edition anatomic staging system. The EBC cohort was defined as stage I-II (cT1-2 N0-1 or cT3N0) and the LABC cohort as stage III (cT3N1, any cT4, and any cN2-3).
The primary outcome of interest was final metastatic status after staging (M0 vs. M1), obtained from the Alberta Cancer Registry and confirmed in the medical record. Details of the staging workup were obtained from medical chart review. Each distant staging investigation was documented, including computed tomography (CT) of the chest and abdomen with or without pelvis (CA ± P), bone scan, brain imaging (CT or Magnetic Resonant Imaging (MRI)) and positron emission tomography (PET). Staging results according to the reporting radiologist’s interpretation were categorized as negative (any findings considered likely benign), indeterminate (findings could not be confidently classified as benign or malignant) or suspicious (findings were concerning for malignancy and considered likely metastatic). Further workup was defined as additional imaging tests or biopsies prompted by findings on the baseline staging scans.
Remaining demographic, disease and treatment characteristics were obtained from medical chart review. Histological type, grade, estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor-2 (HER2) status were obtained from the pre-treatment core biopsy. Hormone receptor positivity (HR+) and HER2 positivity (HER2+) were defined according to the American Society of Clinical Oncology and the College of American Pathologists (ASCO/CAP) guidelines [13] and were used to construct biologic subtype categories of HR+/HER2−, HER2+ and triple-negative (TN).
The EBC and LABC cohorts were analyzed separately. Descriptive statistics were used to summarize demographic and disease characteristics, with continuous variables reported as medians with interquartile ranges (IQRs) and categorical variables as frequencies and proportions. Time from pathologic diagnosis (biopsy date) to NAC initiation was compared by staging result using Mann–Whitney U tests. Associations between clinical factors (age, cT and cN, histologic type, grade, and biologic subtype) and M1 status were evaluated using univariable logistic regression. Factors with statistically significant associations were included in multivariable logistic regression models. For the EBC cohort, the multivariable model included cT/N category, grade and biologic subtype (histologic type was excluded due to a small sample size of lobular cancers). For the LABC cohort, samples sizes were small for individual cT/N categories. As such, two models were generated, one including cT and cN as separate variables and one using AJCC subgroups of stage III (A vs. B vs. C), along with age and grade. For both cohorts, we reported absolute rates of M1 disease for pre-planned subgroups of cT/N category and biologic subtype, reflecting groups for which staging investigations have been recommended in some guidelines. All analyses were performed using StataIC version 16.1 (StataCorp LLC, College Station, TX, USA). A p-value of <0.05 was considered statistically significant.
Among 529 eligible patients with EBC, 14 patients had no staging investigations and were excluded. The remaining 515 (97.4%) underwent CT CA +/− P and/or bone scan and were included for analysis. Most patients (n = 469, 91.1%) had guideline-concordant staging with both CT CA ± P and bone scan. Additional brain imaging was performed for 48 patients (9.3%) and an additional PET scan for 4 patients (0.8%). Detailed staging results are shown in Figure 1. Overall, M1 disease was identified in 28/515 patients (5.4%; 95% CI, 3.8–7.7%). Sites of distant metastases were as lung (n = 14, 50.0%), bone (n = 12, 42.9%), liver (n = 11, 39.3%), brain (n = 1, 3.6%), and other (n = 9, 32.2%). Among M0 patients, those who required additional workup (n = 114, 22.1%) after baseline staging scans experienced a median delay of five days to NAC initiation (43 vs. 38 days; p = 0.03).
Clinical characteristics stratified by M0 versus M1 disease are presented in Table 1. Clinical T/N stage, histology, grade and biologic subtype were all significantly associated with M1 status (all p < 0.05) in univariable analyses. In multivariable analysis, only the cT1N1 (OR 5.31; 95% CI 1.05–27.0; p = 0.044) and cT2N1 (OR 4.59; 95% CI 1.02–20.67; p = 0.047) stage remained significantly associated with higher odds of M1 disease (Table 2). While not statistically significant, TN or HER2+ subtypes and grade 3 tumors showed lower odds of M1 disease. Rates of M1 disease in pre-specified subgroups of cT/N category and biologic subtype are presented in Table 3. Most notably, the M1 rate was 1.1% (2/174) for cT1-2N0, 5.4% (2/37) for cT3N0 and 7.9% (24/302) for cT1-2N1.
Among 320 eligible patients with LABC, all (100%) underwent CT CA +/− P and/or bone scan and were included for analysis. Most patients (n = 293; 91.5%) received guideline-concordant staging with both CT CA ± P and a bone scan. Additional brain imaging was performed in 40 patients (12.5%) and an additional PET scan in 7 patients (2.2%). Detailed staging results are shown in Figure 1. Overall, M1 disease was identified in 73/320 patients (22.8%; 95% CI, 18.6–27.7%), a significantly higher rate than the EBC cohort (p < 0.001). Sites of distant metastases were as bone (n = 41, 56.2%), lung (n = 39, 53.4%), liver (n = 29, 39.7%), brain (n = 3, 4.1%) and other (n = 25, 34.2%). Among M0 patients, those who required additional workup after baseline staging scans did not experience any delay to NAC initiation (median 36 days for both, p = 0.54).
Clinical characteristics stratified by M0 versus M1 disease are shown in Table 4. Age, cT and cN categories as well as overall AJCC stage, and grade were significantly associated with M1 disease (all p < 0.05) in univariable analyses. In multivariable analysis, cT4 and cN3 disease remained significantly associated with higher odds of M1 disease, whereas stage IIIA disease (vs. IIIB or IIIC) and grade 3 were associated with significantly lower odds of M1 (all p < 0.05, Supplementary Tables S1 and S2). Notably, all cT, cN, and grade subgroups for the LABC cohort had absolute M1 rates ≥ 10%. The absolute rate of M1 disease for stage IIIA was 9.2% (12/130); rates for each T/N category are shown in Table 3.
This study uniquely uses a large population-based cohort of breast cancer patients referred for NAC with unselected use of staging investigations, allowing for an accurate and generalizable assessment of the diagnostic yield of staging in this population. We found a low rate of M1 disease among patients with cT1–T2N0 disease (1% overall), supporting the omission of routine staging in this population even when NAC is planned, which is consistent with current guidelines for all-comers with early breast cancer [3,4,5]. In contrast, the incidence of M1 disease was higher for EBC patients with cT3N0 (5%) and N1 (8–9%) disease, suggesting that staging may be justified in these subgroups. Notably, a trend toward higher M1 rates was also observed among patients with ER+/HER2− subtype and grade 1–2 tumors. While these did not reach statistical significance in multivariable analysis, the low event rate may have limited power. Nonetheless, our findings do not clearly support preferential use of staging in aggressive subtypes (e.g., TN and HER2+) even at a lower clinical stage of disease, as some guidelines have suggested, though this observation should be interpreted with caution. In contrast, patients with LABC exhibited a significantly higher burden of metastatic disease, with even the lowest-risk subgroups demonstrating M1 rates of ≥9–10%, reinforcing guideline recommendations for routine staging in this population.
Most studies evaluating the diagnostic yield of staging investigations in breast cancer included all-comers and were stratified by stage. In such studies, the overall rate of M1 disease is low (0–6%) for EBC populations [11,13,14] but significant for LABC (18–36%) [8,10,15,16,17]. The likelihood of M1 disease has also been generally consistent across tumor subtypes [13,14]. Cohorts of high-risk subgroups such as patients referred for NAC are more limited. A retrospective review from Memorial Sloan Kettering Cancer Centre [7] included 303 NAC patients with stage I-II disease, of whom 85% had staging investigations, mostly with PET/CT. The M1 rate detected by staging was 4.9%, almost all (n = 14/15) in patients with cN1 disease. They concluded that pre-NAC staging is not necessary for patients with EBC. A study from the United Kingdom evaluated 163 patients referred for NAC where staging was routine. The overall rate of M1 disease was 9.3%, but in the cohort with T1-3N0, the rate was only 1.6% [9]. Our findings are aligned with these prior studies, but our larger sample size allowed for further stratification by stage and biologic subtype to develop recommendations more tailored to individual tumor characteristics.
In our cohort, staging investigations resulted in small delays to treatment (median 5 days) for patients who required additional workup after baseline scans. While time-to-treatment benchmarks are lacking for NAC patients specifically, delays of ≥4–8 weeks from breast cancer diagnosis to treatment are associated with worse outcomes in those undergoing surgery first [18,19,20]. As such, the differences in time to treatment observed in our study are likely clinically insignificant. Other studies have found delays of 6–14 days for patients who undergo staging [21], so this issue warrants consideration in the local context.
Several limitations of this study should be acknowledged. Despite the large overall sample of 835 patients, stratification by clinical T/N stage and biologic subtype resulted in small subgroup sizes, limiting the precision of some estimates. Similarly, multivariable analyses—particularly in the EBC cohort—may have been underpowered given the relatively low event rate. Given the retrospective nature of this study, there is a possibility of selection bias in staging use and potential for heterogeneity in the type of modality chosen, both of which could influence diagnostic yield. However, exclusions due to non-performance of staging were very minimal (n = 14 patients, 1.6%) and type of staging was quite homogenous, with 91% of patients in both EBC and LABC cohorts undergoing guideline-concordant CT CA +/− P and bone scan. Collectively, this supports the minimal risk of these issues. PET scan was infrequently used in this population, limiting the generalizability of our findings to centers where this modality is routinely incorporated. PET has been shown to increase the detection of M1 disease and is being increasingly incorporated into clinical practice guidelines, primarily for LABC [1,4,22]. In addition, clinical staging was used for group classification, which may be subject to misclassification, particularly with respect to nodal assessment. However, this reflects the reality of available information before NAC is initiated. Finally, with any retrospective medical chart review, there is a risk of variable misclassification, though variables included in this study were generally clearly documented.
Based on our findings, we advocate for the omission of routine staging investigations for asymptomatic patients with cT1-2N0 disease planned for NAC, in keeping with current guideline recommendations. EBC patients with cT3N0 or N1 disease may be considered for staging, and staging should be routine for those with LABC. Staging guidelines in our province are currently under review and will be informed by this study, and our results are widely generalizable to other jurisdictions.