Abstract

Abstract: There is growing public health concern over the benefits of X-ray mammography in breast cancer screening. Though it accomplishes early detection with high sensitivity, the consequent large number of false positive readings leads to a high fraction (typically, 70%) of biopsy procedures yielding benign pathology reports. These biopsies, unnecessary in this sense, have negative (morbid) consequences: pain, anxiety, possible harm, and expense (biopsy procedures cost typically ten times more than mammography or sonography). It has long been wished that MRI provide a more accurate diagnosis of breast disease. For 15 years a particular protocol, quantitative (Dynamic-Contrast-Enhanced) DCE -MRI, has held promise because it is a form of generalized “functional MRI ,” focusing on tumor vascular properties. However, DCE -MRI hasn’t proven significantly better than mammography and, since it is also more expensive, has not come into wide practice. The mathematical model generally used for quantitative DCE -MRI analysis was adopted from well-known tracer pharmacokinetic derivations. However, its unedited translation for DCE -MRI has the unreasonable consequence of treating all inter-compartmental equilibrium water exchange processes as effectively infinitely fast. This is because, though the DCE -MRI contrast agent (CA, a monomeric Gd(III) chelate) plays the tracer molecule role, the signal molecule is water, and these have different tissue compartmentalizations. We have modified the model to accommodate well-known water inter-compartmental transfer rate constants. We conduct an analysis designed to detect tumor foci exhibiting precisely water exchange kinetics effects (“shutter-speed” effects) on pharmacokinetic parameters, particularly the CA extravasation rate constant, Ktrans. In a population now reaching 137 positive-screening breast lesions, we have found that malignant tumor capillaries are sufficiently more CA-permeable to trigger focal Ktrans biomarker shutter-speed effects (ΔKtrans) that are essentially zero for benign lesions. The molecular basis for this is quite reasonable. Lesion region-of-interest and pixel-by-pixel ΔKtrans mapping allows complete specificity between biopsy/pathology-proven malignant (32) and benign (105) tumors. The biopsy procedures on the latter could have been avoided, and with cost savings since they are more expensive than MRI . This discrimination is independent of the screening modality, the MRI instrument vendor (platform/software), and the magnetic field strength. A significant sub-population of 95 lesions occurred in 93 high-risk women who were mammographically-negative but positively screened by clinical MRI approaches. Twenty malignant tumors in this population were missed by mammography. The MR shutter-speed is a fundamental concept that applies not only in quantitative MRI studies of many different cancers, but also of other pathologies and of physiological conditions.