Hypothesis: With technical optimizations, non-contrast DWI can detect clinically and mammographically occult breast cancer in women with dense breasts with high sensitivity and low false positive rate.
Aim 1: Improve the breast DWI technique to maximize spatial resolution, reduce distortion, and increase lesion contrast.
* Develop novel DWI acquisition to increase spatial resolution and reduce distortion (using reduced field of-view and/or multishot echo planar imaging techniques)
* Identify optimal diffusion sensitization (b-value) to maximize conspicuity of cancers in women with dense breasts
Aim 2: Develop interpretation tools to optimize diagnostic performance for detecting cancer on DWI.
* Determine quantitative DWI thresholds (contrast-to-noise ratio, apparent diffusion coefficient \[ADC\]) that best differentiate benign and malignant lesions (i.e. maximize sensitivity and specificity)
* Develop computer aided assessment tools to facilitate clinical implementation and optimize reader accuracy
Aim 3: Test the performance of the optimized DWI approach for detecting clinically and mammographically-occult cancer in women with dense breasts.
* Conduct a controlled reader study of non-contrast DWI alone for breast cancer detection
* Perform receiver operating characteristic (ROC) analysis and determine the sensitivity and specificity for detection of mammographically occult cancer
