Evaluation of fully automated motion corrected first pass myocardial perfusion MRI with semi quantitative perfusion parameter maps in patients with ischemic heart disease

Abstract

Summary: The purpose of the study is to evaluate a fully automated motion corrected first pass myocardial perfusion (FPMP) MRI with semi quantitative perfusion parameter maps in patients with suspected ischemic heart disease. Background: Coronary heart disease is the leading cause of death and disability in the US. FPMP MRI is increasingly used to assess ischemic heart disease; however respiratory motion is one of the major problems for myocardial blood flow quantification. An algorithm for motion correction, surface coil correction, temporal denoising and robust pixelwise parameter map generation model was previously desribed [Xue H et al MICCAI 2009).This work evaluates automated workflow in a clinical setting to diagnose ischemic heart disease comparing free breathing (Figure presented) and motion correcteted images and corresponding pixel-wise parameter map. Methods: Stress and rest FPMP images were acquired using a 1.5T scanner (MAGNETOM Avanto, Siemens Healthcare) in 39 patients with suspected ischemic heart disease. Short axis slices were acquired during infusion of 0.075 mMol/kg of Gadolinium (Magnevist, Bayer HealthCare Pharmaceuticals, USA) and adenosine (Adenoscan, AstellasPharma, USA) infusion (0.14 mg/kg/min; duration: 4 min) was administrated to induce stress. Free breathing, motion-corrected images and corresponding perfusion maps were assessed by 2 radiologists independently using the AHA 16 model and evaluated using a four point Likert scale (poor to excellent) to evaluate image quality and confidence level in presence or absence of hypo-perfusion regions. Upslope index of both free breathing and motion corrected images during stress and rest were manually calculated in non-ischemic and ischemic areas and compared to the corresponding pixel-wise parameter map generated based on motion corrected images. FPMP MRI results were subsequently compared to coronary angiogram, stress echocardiography, or SPECT. Results: Perfusion defects were detected in 25 patients (representative patient shown in Fig. 1). Mean image quality score for motion corrected images (3.48 ± 0.50) and confidence level (3.31 ± 0.41) were significantly higher (p<0.001) than free breathing images (mean image score of 2.51 ± 0.63 and confidence level of 2.83 ± 0.58). Upslope index of non ischemic and ischemic areas and semi quantitative perfusion parameter maps values were comparable (P> 0.05). Sensitivity and specificity for each technique, as well as the inclusion of perfusion parameter maps, are shown in Table 1. Although inline processed results showed higher (Table presented) sensitivity and specifity than standard images, differences were not significant (p< 0.05). Conclusions: A fully automated motion corrected first pass myocardial perfusion (FPMP) MRI with semi quantitative perfusion parameter maps showed comparable accuracy for detection of significant coronary artery disease in patients with ischemic disease.