CMR-guidance of passively tracked endomyocardial biopsy in an in vivo porcine model.

Endomyocardial biopsy (EMB) is considered to be the diagnostic gold-standard in detection of myocardial-inflammation. EMB is usually conducted under fluoroscopy without any specific target information. Specific target-information provided by cardiovascular magnetic resonance (CMR) may improve specificity of EMB. The aim was to investigate feasibility and safety of CMR-guided and targeted EMB in a preclinical-model using passively-tracked devices. Procedures were performed on a MRI-System equipped with an Interventional Software-Platform for real-time imaging. Ex vivo experiments were conducted to optimize visibility of the guide-sheath. In vivo experiments were conducted in 2 pigs for technical feasibility assessment and in 4 pigs after acute myocardial infarction to test feasibility of guided and lesion targeted EMB. For anatomical real-time imaging a single-shot-balanced-SSFP-sequence was applied. Myocardial targets were identified under real-time imaging (single-shot-T2 (sshT2) and single-shot Late-Gadolinium-Enhancement (sshLGE) sequences). Ex vivo experiments demonstrated best visibility of continuously labelled guide-sheath. CMR-guided EMB was feasible in all cases without major complications. Likewise, lesion-targeting endomyocardial biopsy was feasible in two cases. Biopsies exhibited appropriate sizes and qualities. Real-time lesion sequences revealed comparable CNR values to clinical-protocols. Real-time imaging of lesions showed following signal- and contrast-to-noise ratios (SNR/CNR): SNR of sshT2- and sshLGE was 124 ± 35 and 67 ± 51 respectively, whereas CNR was 81 ± 30 and 57 ± 44. This study demonstrates feasibility and safety of CMR-guided and basically targeted EMB with passively-tracked devices. Signal-to-noise ratios of real-time sequences is non-inferior to standard sequences for lesion detection. CMR-guidance may improve diagnostic accuracy of EMB since CMR can detect myocardial-targets under real-time-imaging.