Automated Image Registration and Perfusion Sorting Algorithms for PREFUL MRI

Abstract

Respiratory diseases are the leading cause of death and disabilities worldwide. Current clinical approaches are based on computed tomography and positron emission tomography which use harmful ionizing radiation and cannot be used often. Because of that, proton MRI is a promising tool for functional lung assessment. PREFUL MRI method was shown to yield promising results for future clinical use, however, no standard imaging protocols or computer programs which do not require human supervision exist. Therefore, the purpose of this study was to make a step toward automating and improving robustness of the PREFUL method. Several algorithms were designed including phase sorting for respiratory and heart cycles, image registration and lung segmentation. Ten healthy volunteers underwent PREFUL MRI study and maps of fractional ventilation (FV) and perfusion (Q_quant) were calculated. The maps showed no sign of any pathology and among all healthy volunteers the mean values of FV and Q_quant were 0.21 ± 0.08 and 460 ± 140 ml/min/100 ml, respectively. The obtained results are well agreed with known research data. Thus, our designed automated algorithms for PREFUL MRI can be implemented for assessing ventilation and perfusion of the lung.