Quantitative Tracer Based MRI Perfusion : Potentials and Limitations

• Main Author: Morell, Arvid
• Format: Doctoral Thesis
• Language: English
• Published: Uppsala universitet, Enheten för radiologi 2012
• Subjects: MRI; Perfusion; Contrast agent; Relaxivity
• Online Access: http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-171901; http://nbn-resolving.de/urn:isbn:978-91-554-8330-2

Tracer based MRI perfusion measurements is a clinically useful tool to assess regional distributions of tissue blood flow and volume. The method may be based on any of the three relaxation mechanisms T1, T1 and T2*, the latter denoted DSC-MRI being the most common. The primary aim of this work was to study the feasibility of obtaining quantitative estimates using these methods. 1) Feasibility of DSC-MRI for kidneys using an iron oxide based contrast agent and the influence of secondary relaxation effects on the results, part of a clinical phase II trial: The method proved feasible and the underestimation induced by secondary relaxation can be corrected for by using a double echo sequence. 2) Influence of blood flow rate on risk factors for developing cerebral ischemia during cardio pulmonary bypass, measurements in pig with gadolinium based DSC-MRI: The results indicated an ischemic threshold level at a blood flow rate of approximately 6 ml/kg/min. 3) The ability of gadolinium based DSC-MRI to detect changes in global blood flow, experimental measurements in pig and numerical simulations: The results support that DSC-MRI can discriminate between global flow levels in the same subject given that all other parameters are kept constant. The results also indicate that calculated perfusion values are highly sensitive to the arterial deconvolution procedure. 4) Influence of differences in blood/tissue relaxivity and secondary relaxation for a gadolinium based contrast agent, measurements in pig and numerical simulations: The blood/tissue relaxivity ratio is not unity and the situation is complicated by secondary relaxation effects. Deconvolution regularization appears to partly counteract the overestimation induced by difference in blood/tissue relaxivity for DSC-MRI. In summary, the fundamental assumption of equal blood and tissue relaxivity is experimentally shown to be invalid and the influence of this discrepancy is substantial. Several factors contribute to measurement errors, a combination of these factors can incidentally lead to additive errors or error cancellation based on a variety of experimental and analysis conditions. Given that the differences in blood/tissue relaxivity cannot readily be accounted for in a clinical setting, absolute perfusion quantification by tracer based MRI remains challenging if not impossible.