Abstract
PURPOSE: To introduce a method for measuring an arterial input function (AIF) from breast ultrafast dynamic contrast enhanced (DCE) MRI with a very low dose (15% of standard dose) of contrast media, and fit it with a new empirical mathematical model (EMM) to produce a population AIF. METHODS: Total 18 patients were enrolled in this study. Axial DCE-MRI data were acquired on a Philips 3-Tesla scanner using a standard 16-element breast coil with a temporal resolution of 1.6 s/image. The AIF was extracted from the descending aorta in four steps by eliminating pixels that did not satisfy specific criteria. The final AIF was averaged from five connected slices and fitted with a new eight-parameter EMM-AIF. Additionally, the EMM-AIF was used to fit Parker's population AIF calculated by using the average parameter values. RESULTS: The AIF can be accurately extracted from breast DCE-MRI using our method. The EMM-AIF accurately fits the extracted AIF for both the 1st and 2nd passes of the contrast agent bolus. The EMM-AIF accurately fits Parker's population AIF. Our low dose AIF has significantly (p < 0.01) higher uptake slope, narrower full width at half maximum of the 1st pass, and lower ratio between the peak heights of the 1st and 2nd pass than Parker's AIF. CONCLUSION: Our method to extract AIF can be easily implemented for breast DCE-MRI data. The eight-parameter EMM-AIF introduced here can be used to fit the low dose AIF. Our population AIF can be easily scaled and applied to general pharmacokinetic analysis.