Performance Measures of Magnetic Resonance Imaging Plus Mammography in the High Risk Ontario Breast Screening Program

Anna M. Chiarelli; Kristina M. Blackmore; Derek Muradali; Susan J. Done; Vicky Majpruz; Ashini Weerasinghe; Lucia Mirea; Andrea Eisen; Linda Rabeneck; Ellen Warner


J Natl Cancer Inst. 2020;112(2):136-144. 

In This Article


Study Cohort

The cohort was identified from 29 602 asymptomatic women age 30–69 years referred by their physician to the High Risk OBSP between July 1, 2011, and June 30, 2015, with a final result to December 31, 2016 (Figure 1). Of these women, 2432 (8.2%) were previously known to meet one of the eligibility criteria and 24 811 (83.8%) completed genetic assessment. Of the 10 459 women eligible for screening, 1156 either declined (n = 403), deferred (n = 282), died (n = 3), had bilateral mastectomy (n = 146), or were not enrolled for reasons unknown (n = 322). Among the 9303 women who underwent screening, screening episodes that included only a mammogram (n = 577) or MRI (or ultrasound) (n = 689) were excluded because the other screening test was unavailable. Screening episodes that included an ultrasound and mammogram (n = 649) or breast cancers found incidentally on prophylactic mastectomy (n = 15) and those without a final result (n = 69) were also excluded. The study was approved by the University of Toronto Research Ethics Board, and informed consent was not required.

Figure 1.

Women age 30–69 years referred to the High Risk Ontario Breast Screening Program between July 1, 2011, and June 30, 2015, with final result to December 31, 2016. *Excludes 1156 women who declined (n = 403), deferred (n = 282), died (n = 3), had bilateral mastectomy (n = 146), or reasons unknown (n = 322). Excludes 521 women with mammogram-only screens (n = 577), MRI (or ultrasound)-only screens (n = 689), mammogram and ultrasound screens (n = 649), mammogram and MRI screens for breast cancers found incidentally on prophylactic mastectomy (n = 15), or screens without a final result (n = 69). MRI = magnetic resonance imaging.

Clinical Pathway

A clinical pathway was developed for the identification, referral, and genetic assessment of asymptomatic women at potential high risk.[37] High-risk criteria included: women with a deleterious mutation in BRCA1/2 or other gene(s) predisposing to a markedly elevated breast cancer risk; untested first-degree relatives of a gene mutation carrier; women with a family history and an estimated personal lifetime breast cancer risk of 25% or higher; and women who had radiation therapy to the chest (before age 30 years and at least 8 years previously). Women who met at least one of the high-risk criteria were eligible even if they had a prior history of breast and/or other cancers or breast implants or had a unilateral mastectomy.

Women were assessed for risk of breast cancer and referred to the program by their physician. If there was prior knowledge that the woman met at least one of the high risk criteria, she was enrolled and eligible for screening. If the woman was a first-degree relative of a mutation carrier and had not previously had genetic assessment or had a personal or family history of breast or ovarian cancer suggestive of a hereditary breast cancer syndrome, she was referred to a genetics clinic for further risk assessment by a genetic counsellor. Women found to be mutation carriers and/or assessed as having a 25% or greater personal lifetime risk of breast cancer based on the International Breast Cancer Intervention Study model[38] or the Breast and Ovarian Analysis of Disease Incidence of Carrier Estimation Algorithm[39] were enrolled. Women with contraindications to MRI were scheduled for breast ultrasound.[12,13,19]

At High Risk OBSP centers, quality assurance on the equipment met that specified by the Canadian Association of Radiologists' Mammography Accreditation Program, and radiologists and technologists were accredited under the Canadian Association of Radiologists' Mammography Accreditation Program. Centers conducted MRI-guided biopsy on site or had a partnership with a facility that did and were affiliated with a Breast Assessment Site. The minimum MRI standards included a 1.5 Tesla magnet, gadolinium enhancement (0.1–0.2 mmol/kg), and a dedicated breast coil with bilateral axial or sagittal acquisition.

Diagnostic interpretation of lesions detected on mammograms or MRIs was based on assessment categories of the American College of Radiology Breast Imaging Reporting and Data System 5th edition.[40] Screening or assessment results reported as "normal/benign" (categories 1 and 2) were considered negative results, whereas results reported as "probably benign" (category 3), "abnormal requiring further imaging" (category 0), or "suspicious for breast cancer" (category 4 or 5) were considered positive results.

Mammograms were usually performed before MRI studies, within 1 month of each other. Although different radiologists may interpret mammogram and MRI studies, they were aware of the mammogram results before interpreting MRI studies as well as all previous imaging and clinical history. If either or both modalities were positive based on referral criteria, the radiologists indicated their recommendations for assessment. Masses were biopsied if either MRI and/or mammography features were suspicious for malignancy and were pathologically confirmed to be benign, ductal carcinoma in situ (DCIS), or invasive cancer.

Data Collection

Data were obtained from routine information collected for all women screened within the High Risk OBSP from Cancer Care Ontario's Integrated Client Management System database. The requisition for high-risk screening form included data on physician visit date, referral method, high-risk criteria, and medical history. For women referred for genetic assessment, the breast cancer genetic assessment results form collected data on genetic counseling and/or testing dates and results, and screening eligibility. Information on screening dates, type (initial, rescreen), and results were recorded on the screening report, and details about diagnostic imaging or biopsy dates and outcomes on the assessment results form. Women with interval cancers were identified from record linkage using AutoMatch[41] with the Ontario Cancer Registry estimated to be 98% complete for breast cancer.[42]

Tumor characteristics were collected from surgical and pathological reports obtained from the Ontario Cancer Registry for any primary DCIS or invasive breast cancer of any histologic type diagnosed after enrollment. Reports were reviewed by a trained abstractor and overseen by a breast pathologist (SJD). Tumor size was defined as the greatest diameter of the largest invasive carcinoma. Lymph node status was defined as positive if at least one sentinel or other axillary lymph node contained at least one cluster of malignant cells larger than 0.2 mm. For breast cancers treated with neoadjuvant therapy, tumor size and nodal status were defined at the time of diagnosis, before treatment initiation. Tumor size for neoadjuvant cases was estimated to be larger than 2 cm unless otherwise specified in the pathology and any referenced imaging report.

Tumor morphology was coded using the International Classification of Diseases for Oncology.[43] Stage was coded using the TNM classification scheme.[44] Tumors were graded using the Nottingham combined histologic grading system.[45] Estrogen, progesterone, and HER2 protein receptors were coded as positive or negative.[46–48]

Performance Measures

The performance measure definitions used for this study were primarily those adopted by the Canadian Partnership Against Cancer[49] and other international screening programs.[50,51] Abnormal recall rate was defined as the percentage of screens referred for further testing because of an abnormal screening result. Biopsy rate was defined as the percentage of screens with a core or open surgical biopsy. Cancer detection rate was defined as the number of screen-detected invasive or DCIS breast cancers per 1000 screening examinations. For screening episodes with a final result, screen-detected cancers (true positive) were defined as breast cancers diagnosed within 12 months of a positive screening MRI and/or mammogram. Interval cancers (false negative) were defined as cancers diagnosed within 12 months of a negative screening MRI and mammogram. Sensitivity was defined as the percentage of all cancers (true positive + false negative) that were screen-detected (true positive). Specificity was defined as the percentage of women without breast cancer (true negative + false positive) who had a true negative screening MRI and mammogram. To allow for a follow-up period of 12 months from the start of screening, sensitivity and specificity analysis included screening episodes up to December 31, 2015.

Statistical Analysis

The unit for analyses was the screening episode including MRI and mammography with a final result after recommended work-up and/or biopsy; women may have had more than one screening round during the study. Performance measures were examined by screening result based on the following definitions: "mammography" refers to an abnormal mammography result, irrespective of MRI result; "MRI" refers to an abnormal MRI result irrespective of mammography result; "MRI plus mammography" refers to at least one modality having an abnormal result. Cancer detection rate, sensitivity, and specificity were further stratified by risk criteria and age group at index screen. A generalized estimating equation model using a binomial distribution with logit link function and independent working correlation matrix was used to compare performance measures by screening result overall and in stratified analyses with MRI plus mammography as the reference. Multivariable models adjusted for factors related to women and screen type.[52] The inversed logits were used to estimate performance measures and approximate 95% confidence intervals (CIs) by screening result based on the least squares means of fixed effects.

Prognostic features of screen-detected primary breast cancers were examined in three mutually exclusive groups. "Mammography only" refers to an abnormal mammography result and a normal MRI result. "MRI only" refers to an abnormal MRI result and a normal mammography result. "MRI and mammography" refers to an abnormal MRI and mammogram result. Fisher exact test was performed to compare prognostic characteristics of screen-detected cancers by screening result.

Analyses were conducted using SAS version 9.4.[53] A two-tailed 5% statistical significance level was used for statistical tests.