Non-rigid motion compensation in MR prostate perfusion imaging

Introduction: Prostate cancer (PCa) is currently the most frequently diagnosed non-cutaneous cancer and the second most important reason for cancer mortality in men. Early diagnosis is imperative as cure can often be established. Dynamic Contrast enhance MRI (DCE-MRI) has been established as accurate method in detection and localization of PCa. DCE-MRI is performed by obtaining repeated, fast T1-weighted images before and up to a few minutes after intravenous injection of a gadolinium containing contrast agent. Three-dimensional datasets of the prostate are obtained every few seconds. This allows obtaining a per-voxel signal-intensity vs. time curve. From this, different pharmacokinetic parameters are modeled and used to differentiate cancerous from non-cancerous tissue. Drawbacks are that rectal peristalsis and patient movement often result in spatial-mismatching of serial 3D datasets and therefore incorrect enhancement curves and pharmacokinetic parameters. Hence, a definite need exists to correct these movement related inaccuracies. Therefore, in this work we present a new algorithm that exploits temporal movement patterns to compensate for this movement and enable a more reliable and automatic analysis of the pharmacokinetic data.