Accurate, Rapid and Inexpensive MRI Protocol for Breast Cancer Screening

Overview

The purpose of this study is to test an innovative MRI breast cancer screening method in women with mammographically dense breasts as well as other women with moderately increased cancer risk. MRI, combined with other methods of risk assessment, has potential to significantly improve sensitivity to cancer in dense breasts and detect cancer in all cases at a much earlier stage, with far fewer interval cancers than mammography. Previous tests of MRI sensitivity show that this screening could significantly increase the likelihood of detecting invasive cancers resulting in decreased mortality from breast cancer. Suspicious lesions will be defined by the clinical interpretation of the breast MRI images performed by the attending breast radiologists. Based on the radiologist determination that the MRI findings are suspicious (these findings include masses, non-mass enhancement and foci), suspicious lesions will be assigned a Bi-Rads code specifying whether additional work up or biopsy is necessary. These are Bi-Rads codes 0, 4 and 5. False positive diagnosis should be minimized as all attending physicians reading breast MRI at this institution are fellowship trained in breast imaging.

1. BACKGROUND X-ray mammography has significantly reduced breast cancer mortality, yet many cancers are still detected at advanced stages. Early detection remains the best approach for improving outcomes. MRI offers high sensitivity and can detect cancers-especially aggressive subtypes and those in dense breasts-years earlier than mammography. In areas like south Chicago, with disproportionately high rates of aggressive breast cancer, MRI screening could dramatically reduce morbidity and mortality if implemented effectively. Mammography is less effective for women with dense breasts and does not always detect cancer until tumors are relatively large. Women with dense breasts or those at increased risk for aggressive cancers (e.g., triple-negative) often lack adequate screening options. MRI has consistently shown superior sensitivity, and its performance is not diminished by breast density. Concerns remain, however, about MRI's specificity, potential false positives, and perceived cost for general screening. Nevertheless, newer techniques and protocols may mitigate these limitations. Routine mammography improves survival by detecting cancers early, but sensitivity can be reduced by 30-50% in dense breasts. Since these women are also at higher risk for aggressive cancers, better screening tools are needed. Earlier detection with MRI could significantly decrease both morbidity and mortality. \_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_ 2. PURPOSE / HYPOTHESIS We propose a novel MRI-based breast cancer screening protocol for women with dense breasts and/or moderately increased risk. Our goal is to establish a fast, quantitative, and cost-effective MRI exam that enhances sensitivity while maintaining acceptable specificity and feasibility for clinical use. Prior data suggest MRI screening may improve early detection of invasive cancers and reduce interval cancers \[1,8,9\]. The University of Chicago is uniquely positioned to lead this effort, particularly in underserved areas with high breast cancer burden. Specific Aims: 1. Develop an abbreviated (\<15-minute) quantitative MRI screening protocol and evaluate reproducibility in 10 volunteers with dense breasts. 2. Scan \~50 women with mammographic or sonographic findings requiring biopsy using the short MRI protocol to define thresholds that differentiate benign from malignant lesions. 3. Recruit 150 women with dense breasts and/or intermediate breast cancer risk for short MRI screening. 4. Conduct a reader study using data from women with biopsy-confirmed benign/malignant lesions to evaluate false positive rates. 5. Analyze quantitative metrics from both standard and ultrafast DCE-MRI, including: * Ktrans (contrast uptake rate) * Initial enhancement time in parenchyma and lesions * Novel approaches to Ktrans and arterial input function * Vessel count/size and enhancement rate near suspicious lesions These markers, measurable only via ultrafast sequences, will be assessed for false positive rates using thresholds established in Aim 2. 6. Lesions deemed suspicious (BI-RADS 0, 4, 5) by attending breast radiologists will be documented. All readers are fellowship-trained to minimize false positives. \_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_ 3\. ELIGIBILITY Inclusion Criteria: * Women with mammographic/sonographic findings requiring image-guided biopsy * OR * Women aged 30-70 with \>10% lifetime risk or dense breasts without current cancer diagnosis The two groups (screening vs. diagnostic) will be matched for parity, menopausal status, and age. Exclusion Criteria: * History of contrast media reaction * Pregnant women * Those at risk for nephrogenic systemic fibrosis (NSF) \_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_ 4. STUDY DESIGN Enrollment: * Recruit \~50 women with suspicious findings on mammography (expected: 30 malignant, 20 benign based on biopsy). * Continue enrollment until 30 cancers are confirmed. * Additionally, recruit 150 women with dense breasts or intermediate risk. MRI Protocol: • \<15-minute scan including calibration, bilateral T2, and DCE-MRI sequences. Ground Truth: * For diagnostic group: Biopsy results (surgical pathology) will serve as gold standard. * For screening group: 1. No suspicious MRI findings = no cancer 2. Suspicious lesions biopsied = pathology determines status 3. Suspicious but unbiopsied lesions = considered benign unless progression occurs at 6-month follow-up * READER STUDY Radiologists will independently review T2 and post-contrast T1 images, as well as ultrafast sequences, assigning malignancy probability (scale 1-10). ROC curves will assess inter-reader variability and compare reader performance to quantitative metrics. This exploratory aim is not powered for definitive conclusions but will guide future studies. \_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_\_ * STATISTICAL POWER With 30 confirmed cancers and 170 controls (including benign biopsy cases), we will evaluate diagnostic accuracy of several parameters (e.g., Ktrans, initial enhancement time, vascular metrics). ROC analysis will determine optimal thresholds-defined by maximum sensitivity + specificity with specificity \>80%. We aim to detect an AUC ≥74% for Ktrans and initial enhancement time at 5% significance. These thresholds will inform a larger R01-funded validation study, which will include covariate-adjusted ROC analysis across risk strata (e.g., Gail model, family history). Preliminary ROC curves from the reader study will assess whether visual interpretation of abbreviated MRI aligns with quantitative results, although this aim remains exploratory and not statistically powered.