Baseline staging imaging for distant metastasis in women with stages I, II, and III breast cancer

Background In Ontario, there is no clearly defined standard of care for staging for distant metastasis in women with newly diagnosed and biopsy-confirmed breast cancer whose clinical presentation is suggestive of early-stage disease. This guideline addresses baseline imaging investigations for women with newly diagnosed primary breast cancer who are otherwise asymptomatic for distant metastasis. Methods The medline and embase databases were systematically searched for evidence from January 2000 to April 2019, and the best available evidence was used to draft recommendations relevant to the use of baseline imaging investigation in women with newly diagnosed primary breast cancer who are otherwise asymptomatic. Final approval of this practice guideline was obtained from both the Staging in Early Stage Breast Cancer Advisory Committee and the Report Approval Panel of the Program in Evidence-Based Care. Recommendations These recommendations apply to all women with newly diagnosed primary breast cancer (originating in the breast) who have no symptoms of distant metastasis Staging tests using conventional anatomic imaging [chest radiography, liver ultrasonography, chest–abdomen– pelvis computed tomography (ct)] or metabolic imaging modalities [integrated positron-emission tomography (pet)/ct, integrated pet/magnetic resonance imaging (mri), bone scintigraphy] should not be routinely ordered for women newly diagnosed with clinical stage i or stage ii breast cancer who have no symptoms of distant metastasis, regardless of biomarker status. In women newly diagnosed with stage iii breast cancer, baseline staging tests using either anatomic imaging (chest radiography, liver ultrasonography, chest–abdomen–pelvis ct) or metabolic imaging modalities (pet/ct, pet/ mri, bone scintigraphy) should be considered regardless of whether the patient is symptomatic for distant metastasis and regardless of biomarker profile.


INTRODUCTION
More than 7000 women will develop breast cancer (bca) each year in the province of Ontario 1 . The incidence of distant metastatic disease in even the most common metastatic sites-such as lung, liver, and bone-is exceedingly rare (<1% in all patients with early-stage bca), questioning the need for universal intensive staging at baseline [2][3][4][5] .
A recent population-based study of patients with early-stage bca in Ontario demonstrated significant overuse of diagnostic imaging tests for the purposes of staging, with approximately 80% of patients receiving such tests 6 . Additional imaging tests expose patients to potentially harmful radiation, psychological distress, heightened anxiety, and possibly, delays to treatment.
Health care policy initiatives such as the Choosing Wisely Campaign and the increasing focus on value-based care through programs such as Quality-Based Procedures at Ontario Health (Cancer Care Ontario) [oh(cco)] 7 to enhance awareness among physicians and patients and to use knowledge translation to increase adherence to recommendations.
To standardize clinical practice across the province of Ontario and to expedite, in cancer centres, the assessment and treatment of patients with biopsy-confirmed early-stage primary bca, the Staging in Early Stage Breast Cancer Working Group developed the present guideline, which addresses the research question "Should women with newly diagnosed primary bca receive imaging staging tests to rule out distant metastases? If so, when should those tests be performed? And what are the optimal imaging modalities for staging?"

METHODS
The present guidance document, produced by the Program in Evidence-Based Care (pebc) and approved by oh(cco)'s Staging in Early Stage Breast Cancer Advisory Committee, was developed through a systematic review of the available evidence using the methods of the practice guidelines development cycle 9,10 . The pebc is editorially independent of the Ontario Ministry of Health.
The guidance document was prepared in 3 planned stages, including a search for existing guidelines, followed by a search for systematic reviews and primary literature.
The electronic search for existing guidelines focused on baseline imaging investigations for distant metastases in the electronic databases medline (Ovid) and embase (Ovid) and in the Standards and Guidelines Evidence Directory of Cancer Guidelines. That search was undertaken before any search for systematic reviews or primary literature. The goal was to identify existing guidelines for adaptation or endorsement so as to avoid duplication of guideline development efforts across jurisdictions.
Subsequently, the Cochrane Database of Systematic Reviews and medline (Ovid) and embase (Ovid) were searched from January 2000 to May 2017 for systematic reviews. Any systematic reviews identified were assessed for quality using amstar 11 , and the results of the amstar assessment were used to determine whether the existing systematic review should be included as part of the evidence base.
Assuming that no existing guidelines or systematic reviews were identified, a systematic review of the primary literature was also planned. If a suitable guideline or systematic review were to have been found, a systematic review of the primary literature would be conducted from the date of the previously reported search, only to update the evidence that informed the existing guideline or that appeared in any identified systematic reviews. The search strategy included a logical combination of terms for the condition (breast tumour, metastasis), the intervention (imaging modalities), and studies of interest [systematic reviews, clinical trials, and nonrandomized prospective (30 participants minimum) or retrospective (50 participants minimum) studies]. Relevant articles were assessed by 3 reviewers (NPV, AE, AA), and the reference lists from those sources were searched for additional trials. A data audit procedure was conducted by 2 independent individuals (Ananya Nair, Megan Smyth) to verify the accuracy of the information obtained from the studies included in the guideline.
The identified systematic review was published in 2012 by the Screening and Diagnostic Test Evaluation Program (step) established within the Sydney School of Public Health and funded by the National Health and Medical Research Council in Australia. That review not only significantly overlapped in scope with the objectives of the present work, but also provided a comprehensive summary, to June 2011, of the best available evidence concerning imaging used for staging investigations to detect asymptomatic distant metastases in women with newly diagnosed bca 12 . It was assumed by the members of the Working Group that any relevant document published entirely within the review's search dates (1995 to June 2011) would have been identified. Therefore, the step systematic review was determined to be the main evidence source for the accompanying guideline, to be supplemented with additional data from relevant studies identified in the primary literature search. Only primary literature published from June 2011 onward (corresponding to the end date of the search in the 2012 step systematic review) was considered.
The step systematic review is summarized in the next subsection, and Table i depicts the characteristics of the newly identified observational studies.

STEP Systematic Review
The 2012 step systematic review included twenty-two studies: nine reporting on conventional imaging only (one prospective and eight retrospective studies); eight reporting on fluorodeoxyglucose (fdg)-pet or fdg-pet/ct, or both (five prospective, two retrospective, and one with an unreported study design); and five reporting on both conventional imaging and fdg-pet or fdg-pet/ct.
The study population included women with all presentations of bca: locally advanced (three studies), inflammatory (two studies), and large-tumour bca (>30 mm in diameter, one study), and a mixed population of stages and presentations (eighteen studies). Characteristics of BASELINE STAGING IMAGING FOR DISTANT METASTASIS, Arnaout et al.             the studies such as author, publication year, timeframe, study design, mean or median age, stage distribution, and the outcomes of interest were summarized and presented in evidence tables. All studies reported data about the overall prevalence of asymptomatic distant metastasis and the prevalence of metastasis by site and stage of disease.
Based on the amstar criteria, the methodologic quality of the systematic review was considered to be good.

Primary Literature
The quality of the primary literature was assessed using the quadas-2 tool. For all studies, concerns about applicability were judged to be low. For domains relating to bias, three studies were unclear about whether they avoided the inclusion of patients with symptoms of distant metastasis, and there is therefore an unknown risk of bias for patient selection 37,42,43 .
The reference standard was considered on the basis of clinical and short-term follow-up imaging of metastatic lesions, or on the judgment of a multidisciplinary tumour board when biopsy or histopathology was not feasible, or both, because no "gold standard" for the detection of real metastases has been established. Fifteen studies did not provide enough information to determine whether the results of the reference standard test were blinded to the results of the index test 15,17,19,20,23,26,[30][31][32]34,[38][39][40][41]43 , and the risk for bias is related to the potential influence of previous knowledge on test interpretation 44 .
Seven studies were identified to have concerns about flow and timing. One study was judged to have high risk of bias because only a proportion of suspicious findings received confirmation of the diagnosis by the test used as the reference standard, which might lead to verification bias 32 . The other six studies did not provide sufficient information to determine whether suspicious findings were confirmed by the reference standard or whether all patients with suspected metastasis received the same reference standard 21,25,30,31,40,42 .
Overall, the evidence quality was considered to range from low to moderate because it was derived mainly from retrospective studies with bias concerns.

Systematic Review (All Stages):
The systematic review by Brennan and Houssami 12 reported a low median prevalence of distant metastasis in women initially diagnosed with stages i and ii bca, with a much higher prevalence in those initially diagnosed with stage iii disease. For stage i, the median prevalence from seven studies, all based on conventional imaging alone, was 0.2% (range: 0%-5%). For stage ii bca, the overall median prevalence was reported to be 1.2% (range: 0%-34%): 1.1% with conventional imaging alone (seven studies), 3.3% with pet/ct (one study), and 34.3% with both (one study). Patients who were clinically node-negative, with stage T1 tumours, were included postoperatively if, after sentinel lymph node biopsy, they were positive for malignant cells.         For women initially diagnosed with stage iii bca, the median prevalence was reported to be 8% (six studies) with conventional imaging, 26% (four studies) with pet or pet/ct, and 34% (one study) with both. Two studies that included only cases of inflammatory bca reported a prevalence of 30.5% (conventional imaging) and 48.8% (pet or pet/ct).
For conventional imaging, the median from two studies that reported detection of metastasis by site was 2.5% for bone, 1.0% for liver, and 0.5% for lung 38,40 . Only one study of pet/ct in 19 women initially diagnosed with stage i bca reported distant metastasis detection rates of 5% for bone, 0% for liver, and 0% for lung 41 . Figure 1 depicts the detection of distant metastasis by imaging modality in women initially diagnosed with stage i bca, including studies from the systematic review by Brennan and Houssami 12 . In two studies of pet/ct reporting by biomarker status in primary bca, unsuspected distant metastasis was detected in 7%, 0%, and 0% of patients with er+, her2-; her2+; and triple-negative bca respectively 27,29 .
As expected, survival or disease-free survival, or both, were reported to be significantly shorter for patients with distant metastasis than for those without distant metastasis 21,22,29 .
The median prevalence of metastasis from three studies of pet/ct was 1.0% in bone (range: 0%-21%), 1.0% in liver (range: 0%-4.0%), and 0% in lung (range: 0%-2%) 15,20,41 . In two studies of conventional imaging, the median prevalence was 1.4% for bone, 0.4% for liver, and 0.5% for lung 38,40 . Figure 2 depicts the detection of distant metastasis by imaging modality in women initially diagnosed with stage ii bca, including studies from the systematic review Prevalence of distant metastasis detected by conventional imaging and by integrated positron-emission tomography/computed tomography (PET/CT) ranged from 0% to 5% and from 0% to 8.8% respectively. Conventional imaging and PET/CT combined (one study) detected a prevalence of 0.3% (95% CI: 0.1% to 1.8%). The overall prevalence of distant metastasis ranged from 0% to 8.8%. Moderate to high levels of heterogeneity were observed between the studies (I 2 : 52% for PET/CT and >75% for conventional imaging).

FIGURE 2
Plot of individual studies and pooled prevalence of distant metastasis, by imaging modality, in women initially diagnosed with stage II breast cancer, including 95% confidence intervals (95% CIs). * From the systematic review by Brennan and Houssami, 2012 12 . Prevalence of distant metastasis detected by conventional imaging and by integrated positron-emission tomography/computed tomography (PET/CT) ranged from 0% to 5.4% and from 0% to 33% respectively. Conventional imaging and PET/CT combined (one study) detected a prevalence of 1.8% (95% CI: 1% to 3.2%). The overall prevalence of distant metastasis ranged from 0% to 33%. The included studies were statistically heterogeneous (I 2 : 67% for PET/CT and >75% for conventional imaging). by Brennan and Houssami 12 . In two studies of pet/ct reported by biomarker, unsuspected distant metastasis was detected in 10% of each of these groups: er+, her2-; her2+; and triple-negative bca 27,29 .
The median prevalence of distant metastases detected by studies of pet/ct was 26% (range: 13%-64%). Studies reporting on the detection of distant metastasis reported median detection rates of 21% for conventional imaging (range: 3%-31%) and 16% when both conventional imaging and pet/ct were used ( Table ii).
The median prevalence of metastasis from three studies of pet/ct was 11% in bone (range: 7.5%-43%), 5% in liver (range: 1.9%-14%), and 10% in lung (range: 3.8%-14%) 5,20,41 . The median prevalences of metastasis from two studies of conventional imaging was 7.6% in bone, 7.7% in liver, and 12.1% in lung 38,40 . Figure 3 depicts the detection of distant metastasis by imaging modality in women initially diagnosed with stage iii bca, including studies from the systematic review by Brennan and Houssami 12 .

Timing of Baseline Staging: Pre-Compared with Post-Treatment:
Two studies addressed the issue of the timing of staging investigations in the evaluation of patients newly diagnosed with bca 13,41 .
In the nonrandomized study by Evangelista et al. 13 , 275 patients with stages i-iii triple-negative or her2+ bca were staged either before neoadjuvant systemic therapy and surgery (54%), or after surgery (45%). Almost one quarter of the patients with stage iii disease receiving pretreatment staging were upstaged to stage iv. Outcomes were worse in all patients who were upstaged before treatment compared with those who were not upstaged. Change in treatment was reported in 15 patients: 1 patient received a more aggressive surgical approach, 12 patients received systemic treatment only, and 2 patients received a combination of systemic and local treatment. For patients who underwent staging imaging after surgery, the upstaging rate was lower (10%), and no difference in prognosis was observed in those who were upstaged compared with those who were not.
The retrospective study by Gunalp et al. 41 retrospectively examined 341 patients who were referred for pet/ct staging after a diagnosis of bca. Patients had clinical stages i-iv bca and underwent pet/ct pre-or postoperatively. The paper did not indicate whether any of the patients received neoadjuvant systemic therapy, and the specific distribution of clinical stages in the pre-and postoperative groups was not reported.
Given the design limitations of the two studies, no conclusions can be drawn about the value of pre-compared with posttreatment staging. Because many patients in Ontario with clinical stage iii disease will receive neoadjuvant systemic therapy with curative intent 45 , it makes sense to perform staging investigations for that group before treatment initiation.

PET/CT Considerations in Stage III Disease:
As identified in the present review, the prevalence of distant metastases in patients with clinical stage iii bca who undergo pet/ct is high and greater than that seen with conventional imaging. Because upstaging patients to stage iv would likely alter treatment intent in their cases, it is important to accurately identify the presence of distant metastases. In Ontario, pet/ ct is not currently funded for the staging of patients with bca on the basis that the existing evidence consists largely of observational, retrospective, single-institution studies. To generate better-quality evidence, the Ontario Clinical Oncology Group initiated a randomized trial of pet/ct compared with conventional imaging in patients who present with clinical stage iii invasive ductal cancer. In the same study, a cohort of patients with similarly staged invasive lobular cancer will be staged using both modalities. The primary outcome of the study will be the proportion of patients who are upstaged to stage iv disease. Secondary Plot of individual studies and pooled prevalence of distant metastasis, by imaging modality, in women initially diagnosed with stage III breast cancer, including 95% confidence intervals (95% CIs). * From the systematic review by Brennan and Houssami, 2012 12 . Prevalence of distant metastasis detected by conventional imaging and by integrated positron-emission tomography/computed tomography (PET/ CT) ranged from 2.5% to 31.3% and from 8.3% to 64% respectively. Conventional imaging and integrated PET/CT combined (one study) detected a prevalence of 15.6% (95% CI: 11.7% to 20.6%). The overall prevalence of distant metastasis ranged from 2.5% to 64.3%. The included studies were statistically heterogeneous (I 2 : 74.4% for PET/ CT and >75% for conventional imaging). endpoints include final treatment intent, rates of additional testing generated by the staging tests, survival, prediction of response to treatment, and economic analysis.
The guideline Working Group members believe that, although the existing data are suggestive for a benefit of staging with pet/ct in clinical stage iii disease, high-quality evidence related to pet/ct will be generated by the randomized trial, and it would be prudent to wait for the results before adopting pet/ct as the standard of practice.

DISCUSSION
Although appropriate staging investigations in patients with newly diagnosed bca can aid in expediting appropriate care, overuse can lead to unnecessary invasive biopsies, unnecessary exposure to potentially harmful radiation from the imaging, psychological distress, heightened anxiety, and possible delays to treatment 46,47 . We sought to answer the question of which groups of patients diagnosed with asymptomatic primary bca should routinely undergo staging investigations, and what the optimal imaging modalities are.
Our systematic review of more than 5600 articles resulted in 32 studies for analysis. All analyzed studies reported an overall prevalence of asymptomatic distant metastases. The median prevalence was 14%, with the most common sites of distant metastasis being bone, lung, and liver, in that order. Excluding pet/ct, the detection of distant metastasis with anatomic imaging for staging in patients with stages i and ii bca was 1.0% (range: 0%-1.9%) and 1.9% (range: 1.9% -2.1%) respectively. Those exceedingly low rates of distant metastasis in stages i and ii disease do not warrant routine use of staging imaging.
Results were more significant for asymptomatic patients with stage iii disease, with a median prevalence of distant metastases reported by conventional imaging of 21% (range: 3%-31%)-which is why routine systemic imaging is recommended. Overall, our recommendations agree with those published by the U.S. National Comprehensive Cancer Network (updated in 2018) 48 , the European Society for Medical Oncology (2015) 49 , and the American Society of Clinical Oncology's Choosing Wisely guidelines 50 , in that routine systemic imaging in asymptomatic patients should be considered only in patients who present with locally advanced (stage iii, T3N1-3) disease.
Our current recommendations differ from the earlier oh(cco) guideline published in 2011 in that "we no longer recommend routine bone scan for stage ii patients, even if they have node positive disease." As more prospective studies became available, the low incidence of bone, lung, and liver metastasis was confirmed such that we no longer felt the need for routine body imaging in the initial evaluation of women with stage ii bca who show no symptoms of distant metastasis. Our current guidelines also differ from the latest Alberta Health Services (2012) and Eastern Health (2011) staging guidelines, both of which recommend that a routine baseline bone scan and ct of chest and abdomen should be performed in all patients with node-positive disease.
With respect to pet/ct imaging, the data overall did show additional detection rates at all stages. However, for asymptomatic patients with stage i or ii disease, the added prevalence of metastatic disease detection was highly variable (ranging from an additional 1% to 10%), and no study was a randomized controlled trial. For asymptomatic patients with stage iii disease, the average prevalence of distant metastases in studies of pet/ct was more significant at 26% (range: 13%-64%). Given that finding, we felt that pet/ct could be considered as a method of staging for distant metastasis in patients stage iii disease. The results of the Ontario pet-abc study, a randomized controlled trial that investigated the routine use of pet/ct compared with conventional imaging in asymptomatic patients with stage iii disease supplements that recommendation.
On the other hand, for patients with stage ii disease, we struggled with whether to recommend routine use of pet/ ct, because some of us felt that a 10% prevalence of distant metastasis was not to be ignored. We therefore looked to the literature for guidance on the issue. Interestingly, although the American Society of Clinical Oncology considers pet/ct to be a credible imaging modality for patients with stage iii disease, it recommends against its use in asymptomatic patients with stage i or ii disease. The U.S. National Comprehensive Cancer Network panel recommended against its use in stages i-iii disease, citing the high false-negative rate for lesions that are small or low-grade (or both), the low probability of those patients having detectable metastatic disease, and the high rate of false positives. In contrast, they recommend the use of pet/ct only as an adjunct to conventional imaging modalities when findings are suspicious or equivocal, especially in the setting of locally advanced or metastatic disease. Furthermore, results from a prospective multicentre diagnostic accuracy study reported that pet is not sufficiently specific to accurately identify distant metastasis in asymptomatic patients with primary bca (stages i and ii) 51 .
Apart from staging investigations in patients with newly diagnosed bca, the diagnostic value of pet in detecting distant metastasis in the initial staging of bca was determined to be beyond the scope of the present guideline.
Interpretation of the data from the analyzed studies has associated limitations based on substantial heterogeneity in design and quality. In general, the evidence is sparse and drawn mainly from single-institution retrospective and prospective studies, reflecting the need for a prospective randomized controlled trial. Substantial variability was observed in the quality of the referencestandard test used to confirm suspected metastasis, because not all patients received histopathologic confirmation, and no form of reference-standard test was used to confirm negative results (misclassification bias). For many of the studies, it was unknown whether the clinicians interpreting the results of the reference test had been blinded to the results of the index test. Furthermore, when comparing imaging modalities, of the eight studies that examined the use of conventional imaging as staging tests, five used chest radiography or ultrasonography, two used ct, and one used either ultrasonography or ct. No study compared the outcomes of ct, ultrasonography, and chest radiography, and therefore, based on the evidence review, no explicit recommendation can be made about which modality to use.
We also focused on the imaging detection of systemic disease without the ability to determine to any meaningful degree whether the detection of metastasis affected outcome or treatment decisions, because information on treatment and survival by initial stage was not integrated into the analyzed imaging studies.
Finally, it should be noted that, for the purposes of our proposed imaging recommendations, staging can be based on clinical (in the patient undergoing neoadjuvant therapy) or pathologic or anatomic stage assessment (in the postoperative patient). The new 8th revision to the American Joint Committee on Cancer (ajcc) bca staging system has combined tumour biology (grade, er status, progesterone receptor status, and her2 status) with TNM categories into prognostic stage groups. Although this new prognostic staging system is supposed to be a better representation of prognosis and outcome, we have not incorporated it into our guidelines, simply because of a lack of available studies using the resulting classification. It should be noted that up to 30%-40% of patients can be reassigned to a different prognostic stage group than the one assigned on the basis of anatomic staging. We acknowledge that, for staging, the studies included in this review used the ajcc 7th edition, which was based solely on anatomic stage. We reviewed the ajcc 8th edition to determine whether new clinical and pathologic prognostic stage groupings would affect our recommendations. In the new staging system, some patients at anatomic stage ii would be reclassified to stage iii (for example, high-grade triple-negative disease). Additionally, some patients at anatomic stage iii (for example, low-grade er+ disease) would be downstaged to stage ii in the new classification. Thus, there is some risk that our recommendations for patients with stage ii disease would result in misunderstandings when using the new clinical and pathologic prognostic stage groupings. On the other hand, the evidence review of specific studies that considered biomarker profile in the selection of patients for distant metastasis staging did not, compared with anatomic staging alone, show a greater prevalence of metastasis. Until further studies delineating the evidence for staging under this new classification system are performed, differences between the ajcc 7th and 8th classification systems in clinical and pathologic staging should be taken into consideration when interpreting the guideline. We are aware that additional preoperative imaging that might not be routine (mri), if applied, would also have the potential to upstage patients. We look forward to adjusting our systemic imaging recommendations in the future as evidence emerges about the prevalence of distant metastasis with the new (8th edition) ajcc classification system and about additional preoperative imaging modalities.

CONCLUSIONS
This guideline is intended to provide recommendations for the use of imaging to detect distant metastases in women with newly diagnosed bca who are otherwise asymptomatic. Unless a patient has clinical or pathologic stage iii bca, this evidence-based guideline recommends against the routine use of imaging for staging investigations, regardless of biomarker profile.

REVIEW PROCESS
The health research methodologist (NPV), in collaboration with the lead author (AA), wrote the initial recommendations and qualifying statements pertaining to the use of imaging tests to detect distant metastases in women newly diagnosed with bca. The guideline was circulated to the members of the Staging in Early Stage Breast Cancer Working Group and discussed during a teleconference, after which the draft recommendations were generated. The ensuing guideline was reviewed by the pebc's Report Approval Panel (scientific director, the pebc assistant director, and two health research methodologists) to ensure that the guideline development was methodologically rigorous and that the evidence-based recommendations are indeed supported by the evidence in a transparent way. The refined guideline was then presented to the Staging in Early Stage Breast Cancer Advisory Committee to ensure the clinical relevance and utility of the recommendations, and to obtain a final approval.
After internal review, feedback on the approved draft guideline was obtained from content experts and the target users through two processes. In the targeted peer review, two individuals with content expertise were asked to review and provide feedback on the guideline document. In the professional consultation, 26 relevant care providers and other potential users of the guideline provided feedback on the guideline recommendations through a brief online survey. The latter consultation was intended to facilitate the dissemination of the final guideline to Ontario practitioners.
Practice guidelines and recommendation reports developed by the pebc are reviewed and updated as needed. Please visit the oh(cco) Web site (https://www.cancercare ontario.ca/) for the full guideline and subsequent updates.

PRACTICE GUIDELINE
Evidence from a systematic search of the primary literature, consensus of expert opinion, feedback obtained through a review process, and final approval given by the Staging in Early Stage Breast Cancer Advisory Committee and the pebc's Report Approval Panel collectively form the basis of this guideline, completed in October 2019.

Target Population
The target population for this guideline is women with newly diagnosed primary bca (originated in the breast) who have no symptoms of distant metastasis.

Recommendation 1
Staging tests using conventional anatomic (chest radiography, liver ultrasonography, chest-abdomen-pelvis ct) or metabolic imaging modalities (pet/ct, pet/mri, bone scintigraphy), or both, should not be ordered routinely for women newly diagnosed with clinical stage i or ii bca who have no symptoms of distant metastasis, regardless of biomarker status.

Qualifying Statements
Baseline conventional anatomic imaging modalities (chest radiography, liver ultrasonography, bone scan, BASELINE STAGING IMAGING FOR DISTANT METASTASIS, Arnaout et al. chest-abdomen-pelvis ct) should not be ordered routinely in women with newly diagnosed stage i or ii bca because this population shows an extremely low prevalence of asymptomatic distant metastasis.
Although pet/ct might improve the detection rate, the prevalence of distant metastasis in women with early stage i or ii bca is very low, and pet/ct might unnecessarily increase anxiety and resource use. The use of pet/ct as part of baseline staging in women clinically diagnosed with early-stage bca (i, ii) and with no symptoms of distant metastasis is therefore not recommended at this time.
Although women with triple-negative and her2+ bca have an increased risk of disease recurrence, the association of distant metastasis with biomarker profile in early-stage bca has not been adequately assessed in prospective studies of staging investigation. The benefits and risks of the routine use of biomarker profiles to assess for distant metastasis is still unclear, and thus its use to guide decisions about staging imaging for clinical early-stage bca is not recommended regardless of whether the patient will be receiving neoadjuvant therapy.

Recommendation 2
In women newly diagnosed with stage iii bca, baseline staging tests using either anatomic (chest radiography, liver ultrasonography, chest-abdomen-pelvis ct) or metabolic (pet/ct, pet/mri, bone scintigraphy) imaging modalities, or both, should be considered regardless of whether the patient is symptomatic for distant metastasis and regardless of biomarker profile.

Qualifying Statements
Staging tests should be considered at initial diagnosis so that appropriate treatment recommendations can be made.
A prospective randomized trial (see NCT02751710 at https://ClinicalTrials.gov/) of pet/ct compared with conventional anatomic imaging in patients with clinical stage iii disease who will receive neoadjuvant therapy is currently underway in Ontario. The goal of the trial is to determine the rate of upstaging to stage iv with each modality. Given that the existing evidence is based largely on retrospective, observational, and single-institution studies, members of the Working Group believe that it is prudent to wait for the results of the trial before making a recommendation on the choice between anatomic or functional imaging modalities as the standard of practice for staging in such patients.

ACKNOWLEDGMENTS
The members of the Working Group thank these individuals for their assistance in developing this guideline: Melissa Brouwers, Sheila McNair, and Jonathan Sussman for providing feedback on draft versions; the Staging in Early Stage Breast Cancer Advisory Committee for reviewing and formally approving the final version of this document; Ananya Nair and Megan Smyth for conducting a data audit; and Sara Miller for copyediting.
The pebc at oh(cco) is sponsored by the Ontario Ministry of Health. The full guideline is available on the oh(cco) Web site, in the Guidelines and Advice section at https://www.cancercareontario. ca/en/guidelines-advice/types-of-cancer/1096.