Validation of the PSIR sequence for the determination of arrhythmogenic channels in ventricular ischemia

Abstract

In patients with ventricular tachycardia of ischemic origin, arrhythmogenic channels are the pathway of abnormal tissue activation and their determination in the substrate is a essential factor when treating these cases by radiofrequency ablation. Extracting this information from images obtained by magnetic resonance imaging has great advantages over other more invasive imaging techniques. The most commonly used reconstruction technique in MRI-2D is the Magnitude sequence. Recently, another sequence called Phase Sensitive Inversion Recovery (PSIR) is beginning to be established, which takes into account the polarity of the protons, apart from their magnitude, when generating the image. In this project the level of validity of the PSIR reconstruction sequence to determine the arrhythmogenic channels has been demonstrated, comparing data obtained using this sequence with data obtained using the strongly validated and referenced Magnitude sequence. Data from 21 patients with specific conditions have been used for this study. The images have been segmented and processed in order to extract the parameters that have allowed to solve the question raised by means of a statistical analysis of the information obtained. We have worked with ADAS-3D rendering software to study the cases and have determined the configuration that allows the highest quality in the visualization of PSIR images, specifically the setting of contrast thresholds. From the information provided by the ADAS-3D we selected the information considered relevant for the statistical analysis, descriptive information about the channels and characteristics of the tissue. These data, together with the contrast thresholds set in the study, have been statistically analysed with the RStudio programme. Valuable information has been obtained from the results. The ideal thresholds for studying PSIR images have been found and it has been concluded that there is a considerable similarity between both sequences when interpreting MRI images clinically, although not enough to validate it completely. Regarding the characterisation of channels, a high accuracy in the calculation of their mass has been determined, but a great inaccuracy in their counting. In terms of quantification, identification and classification of ventricular tissue, considerable correlation and acceptable measurement accuracy have been demonstrated.