The evaluation of breast health concerns and the diagnosis of breast cancer have changed significantly over the past 20 years due to advances in imaging technology and pathology techniques as well as increased interdisciplinary management of breast problems.

The referral pathways for diagnostic breast imaging address both the evaluation of a presenting breast complaint in the symptomatic patient and the evaluation of an abnormality identified on screening mammography in an asymptomatic patient. The diagnostic imaging workup for each requires careful risk assessment, use of the Breast Imaging Reporting and Data System (BI-RADS) lexicon, and concordance assessment of the physical examination, imaging, and pathology results. If these components of the patient’s workup are not concordant, further investigation is necessary to rule out malignancy.

History and physical examination
When patients present with breast concerns, a detailed clinical history should be obtained and a physical examination performed to clarify the symptoms and physical findings and determine the need for further investigation with diagnostic breast imaging. Palpable lumps or areas of thickening, nipple discharge, and breast pain are among the most common presenting complaints.

A lump or area of thickening should be assessed for any features of concern for breast cancer, such as hard or irregular texture and apparent fixation to the chest wall, as well as any associated nipple discharge, overlying skin or nipple changes, and the presence of palpable axillary or supraclavicular lymphadenopathy. 
Physiologic nipple discharge should be differentiated from pathologic nipple discharge. Physiologic discharge is bilateral, involves multiple ducts, is milky/yellow/green in color, and is nonspontaneous. This type of discharge does not require diagnostic medical imaging. Pathologic discharge is unilateral, involves a single duct, is clear/bloody/serosanguinous in color, and is spontaneous.

A physical examination is the first step in the evaluation of breast pain. The location and characteristics of the pain help to differentiate breast pain from chest wall pain. Chest wall pain is commonly located at the costochondral junctions and over the lateral chest wall rather than in the breast itself. Physiologic breast pain is characterized as cyclical, nonfocal, and bilateral, and is typically due to hormonal or age-related changes. Physiologic breast and chest wall pain do not require diagnostic breast imaging as pain of this nature has not been shown to correlate with the presence of breast cancer.[1] Breast pain that is persistent, noncyclical, unilateral, and, most importantly, focal, may indicate underlying breast pathology.

Diagnostic breast imaging is indicated when a palpable breast lump or thickening, suspicious skin changes, pathologic nipple discharge, or focal breast pain are confirmed on physical examination. In the absence of suspicious physical examination findings, screening mammography is recommended.

If an abnormality is identified on screening mammography, a physical examination of the breast should be performed to determine if the finding is image-detected only or associated with findings on examination. Ideally, breast examination should occur prior to a core needle biopsy (CNB) to allow for accurate clinical assessment of the abnormality in question.

Diagnostic breast imaging 
There are several reasons a patient may be referred for diagnostic breast imaging. This specialized imaging is used to evaluate a physical examination finding or to further characterize an abnormality identified on a screening mammogram in an asymptomatic patient. In addition, diagnostic breast imaging should be scheduled on a routine surveillance basis for patients with breast prostheses and previously treated breast cancers or high-risk lesions.

Bilateral diagnostic mammography and targeted ultrasound focused on the area of clinical concern are used to evaluate the majority of breast health concerns. Digital breast tomosynthesis, a form of 3-D or volumetric mammography, is becoming increasingly available for use in diagnostic breast imaging and permits better differentiation between normal and abnormal tissue.[2] Magnetic resonance imaging (MRI) may be used in limited circumstances, such as for screening patients with a known BRCA1 or BRCA2 mutation or a history of mantle radiation between age 10 and 30, for assessment of breast prostheses that may have ruptured, for assessment of indeterminate lesions, for assessment of pathologic nipple discharge, and for staging of breast cancers, particularly if there is multifocality or lobular histology.[3]

Categories of risk
The Breast Imaging Reporting and Data System,[4] published by the American College of Radiology (ACR) and summarized in (Table 1 [19]), provides standardized descriptions of findings that communicate the likelihood of breast cancer based on the imaging workup. The BI-RADS assessment of risk ultimately determines the need for an image-guided needle biopsy.

Patients with imaging studies categorized as BI-RADS 1 (negative) or BI-RADS 2 (conclusively benign) do not require further imaging workup. Lesions assessed as BI-RADS 3 have an estimated malignancy risk of less than 2% and follow-up imaging is recommended at 6, 12, and 24 months to ensure stability. Lesions assessed as BI-RADS 4 have an estimated malignancy risk of between 2% and 95% and biopsy is recommended. With this large range of risk, many radiologists subdivide BI-RADS 4 into three categories based on positive predictive values: 4A (less than 2% to less than or equal to 10% risk), 4B (greater than 10% to less than or equal to 50% risk), and 4C (greater than 50% to less than or equal to 95% risk). Lesions assessed as BI-RADS 5 have an estimated risk of malignancy greater than 95%, require core needle biopsy, and should have surgical excision even if the pathology from the biopsy does not confirm malignancy.

During the diagnostic breast imaging appointment, the radiologist will integrate elements from this detailed assessment by correlating the findings on the different imaging modalities. The radiologist will then prepare a report providing a final BI-RADS assessment score and guidance. Typically, the report will make one of five recommendations:

1.    The imaging is incomplete and has not as yet permitted thorough assessment of the presenting clinical concern (BI-RADS 0). Additional imaging is required with final management options to be detailed in a subsequent report.
2.    The findings are negative (BI-RADS 1) or benign (BI-RADS 2) and the patient can be discharged from undergoing further diagnostic breast imaging. If the patient is older than 40, routine screening mammography will be recommended. 
3.    The findings are estimated to have a greater than 98% likelihood of being benign (BI-RADS 3) and follow-up imaging in 6, 12, and 24 months is required to demonstrate stability, at which point the findings will be considered definitively benign.
4.    An abnormality identified is estimated to have a greater than 2% likelihood of being malignant (BI-RADS 4 or BI-RADS 5), and an image-guided biopsy is required for definitive pathologic diagnosis. 
5.    Depending on the presenting compliant, referral to a breast specialist or surgeon may be recommended.

A very small subset of breast cancers will not be detected on a mammogram or breast ultrasound, and negative findings on imaging should not be cause for a delay in diagnosis of breast cancer. Any suspicious findings from a physical examination require further imaging and biopsy. In this situation, referral to a surgeon is advised.

Image-detected breast abnormalities
Nonpalpable breast lesions (masses, calcifications, asymmetries, and architectural distortion) can be identified on screening mammograms as well as incidentally on diagnostic breast imaging done for other indications. The evaluation of such abnormalities focuses on assessing the risk for malignancy. Patients should have an examination to determine whether there is an associated physical finding; however, in cases where there are no associated physical findings to guide the treating clinician, the BI-RADS reporting system and radiology-pathology concordance statements are particularly useful in determining the need for further management and follow-up.

Core needle biopsy
Core needle biopsy (CNB) of breast lesions is the preferred initial invasive diagnostic procedure,[5-9] and image guidance by ultrasound, mammography, or MRI is recommended to increase the accuracy of targeting.[8,10] CNB is able to accurately diagnose most breast lesions and results in an improved cosmetic outcome for the patient and cost savings to the health care system. As well, CNB can decrease the number of operations that a patient requires.[7,8] CNB is safe and does not cause dissemination of cancer.[7]

The guidance modality chosen depends on the conspicuity of the lesion. Most masses are seen clearly on ultrasound, and can therefore be biopsied with ultrasound guidance. For lesions seen on mammography, but not visible on ultrasound, biopsy with stereotactic or tomosynthesis guidance is used. Either way, the procedures are well tolerated under local anesthesia. CNB of breast lesions is now the standard of care and surgical biopsy should be performed only when core needle biopsy is not possible.[5-9,11] Obtaining preoperative core needle biopsy diagnoses of breast lesions in more than 90% to 95% of patients is a quality indicator in breast surgery internationally.[5,6,9] CNB pathology results are classified as benign, high-risk, or malignant.[8,12]

Concordance assessment 
Concordance assessment is an integral part of core needle biopsy. For an image-guided CNB, “concordance refers to the agreement of imaging and histopathological findings such that the histopathology satisfactorily explains the imaging findings.”[12,13] Discordance refers to occasions when a breast CNB demonstrates benign histology while the imaging findings indicate possible malignancy. Discordance indicates that further evaluation is required, with options including (1) repeating CNB, possibly with a larger gauge or vacuum-assisted device, (2) surgical excisional biopsy, or (3) clinical and imaging surveillance.[8,12]

Failure to resolve discordance between radiology and pathology results can lead to a delay in breast cancer diagnosis. The specifics of radiologic-pathologic correlation should be provided by the radiologist performing the biopsy,[12,13] and this is typically done by issuing a radiology-pathology addendum report that includes the histopathologic diagnosis and outlines whether further investigation, follow-up, surgical excision, or multidisciplinary review is recommended.

For symptomatic patients, there should be correlation with the physical examination findings, diagnostic breast imaging, and pathology[12] by the clinician who has ordered and reviewed the investigations. If the physical examination results remain suspicious for malignancy, but imaging and/or core needle biopsy results are benign, further investigations such as skin biopsy, surgical biopsy, repeat core needle biopsy, or further imaging are necessary to rule out malignancy.

Borderline or high-risk lesions 
Concordance assessment also involves assessing the risk of undersampling, which refers to the possibility of missing a significant finding in the adjacent tissue that was not removed by the CNB. When there is a significant probability of undersampling, the lesion is regarded as a borderline or high-risk lesion. Surgical excision of the area is typically recommended for high-risk lesions found on CNB to rule out malignancy. The risk of undersampling varies by diagnosis and is affected by factors such as the number and size of CNB specimens taken, the size of the lesion, and the institution.[8,12] The high-risk lesions for which surgical excision is usually recommended are shown in (Table 2 [20]). Institutions that have documented a low upgrade-to-breast-cancer rate following surgical resection for specific diagnoses may offer their patients short-term imaging follow-up rather than surgical excision.[8,12,14]

After surgical excision
Concordance assessment is recommended after surgical excision of breast lesions to ensure that the lesion of interest has been accurately targeted and adequately sampled. For image-detected abnormalities, specimen imaging of fine-wire-guided excisions is recommended and all breast surgical specimens should be oriented for the pathologist. In rare circumstances, the lesion localized for surgery with palpation or wire placement is different from the lesion biopsied under imaging guidance, and this is suggested by discordance between preoperative CNB pathology and surgical pathology. The presence of the CNB scar in the tissue excised suggests accurate targeting, but scarring can also be present in adjacent tissue. Discordance between the surgical pathology and the preoperative diagnosis indicates further investigations are required, usually further imaging, with possible repeat image-guided localization and surgical re-excision. Multidisciplinary review is recommended to guide management.

Navigating the diagnostic system 
With the development of more options for diagnostic breast imaging and the introduction of image-guided breast biopsy, patients may need to attend several appointments to complete their imaging workup. It is recommended that patients have their mammograms, breast ultrasound, and image-guided core needle biopsies performed by a single radiology group to facilitate comparison, limit duplication of examinations, and decrease wait times. When more specialized biopsy or imaging techniques are required, such as stereotactic core biopsy or MRI-guided biopsy, the patient may be referred to a different facility as access to these services is not universally available. All relevant imaging and reports, as well as any pathology reports, should be forwarded when a patient is referred to a new facility.

The potential for fragmented care and extended wait times has been recognized.[15,16] The 2012 Provincial Breast Health Strategy Summary Report[16] recommended a target wait time of 21 days to diagnosis without biopsy and a wait time of 28 days to diagnosis when biopsy is required. The report also recommended facilitating necessary investigations and increasing navigation resources for patients to achieve the goal of a prompt diagnosis. The Screening Mammography Program of BC introduced fast track booking for evaluation of screen-detected lesions in 2010,[17] and many breast diagnostic teams in the province have introduced care pathways and dedicated breast health clinics in an attempt to reduce wait times.[18,19]

Currently, each community has a slightly different system for arranging breast imaging and biopsies. Additional investigations are facilitated either by radiology or by the ordering physician, depending on the centre. With so many different approaches and pathways, problems can arise, particularly if patients have mammograms outside of their usual community or health authority. Referring physicians should be aware of these differences and the role they are expected to play in arranging investigations and follow-up for their patients.

Summary 
Methodical evaluation of breast health concerns is essential to ensure completion of a thorough imaging and pathology assessment in an appropriate time frame. This allows for malignancy to be excluded and the patient returned to the care of the referring physician, or for malignancy to be confirmed and definitive treatment arranged. The BI-RADS lexicon is needed to communicate the likelihood of malignancy in a standardized way, while concordance assessments between the clinical presentation, imaging results, and final pathology are crucial and should be employed throughout the evaluation. Finally, because of slight variations in local practice, referring physicians should be aware of the role they play in helping patients navigate the diagnostic system.

Acknowledgments
The authors would like to acknowledge the BC Cancer Surgical Oncology Network for facilitating a series of meetings for surgeons from all health authorities. We would also like to thank Dr Stacey Piche, a radiologist in Penticton, Dr Karen Seland, a radiologist in Prince George, and Dr Elaine Wai, a radiation oncologist in Victoria, for reviewing the article and participating in our discussions of current diagnostic evaluation of breast problems. 

Competing interests
None declared.

This article has been peer reviewed.

Drs Padilla-Thornton, Farrell, and Gordon are radiologists in Vancouver. Drs Padilla-Thornton and Farrell are also clinical instructors in the Department of Radiology at UBC, and Dr Gordon is a clinical professor. Dr Wijayanayagam is a surgeon in BC and Dr Roberts is a surgeon in Ottawa, and both are associated with the BC Cancer Surgical Oncology Network. Drs Warburton, Chiu, Cader, Sutter, Baliski, Cheifetz, and McKevitt are surgeons in BC associated with the BC Cancer Surgical Oncology Network. As well, Drs Chiu and Cader are clinical instructors in the Department of Surgery at UBC; Drs Baliski, Warburton, and Sutter are clinical assistant professors; Dr Cheifetz is an associate professor; and Dr McKevitt is a clinical associate professor.