Development of a Rapid and Non-Contrast Low-Field MRI-Based Assessment of Pulmonary Embolism - The goal of this project is to develop an efficient 3D, contrast-free, low-field Magnetic Resonance Imaging (MRI) pipeline to detect pulmonary embolism (PE), leveraging the unique physics of low-field MRI and denoising via deep learning. PE is a major cause of acute cardiovascular mortality worldwide, surpassed only in frequency by myocardial infarction and stroke. When patients present in the Emergency Department (ED) with PE symptoms, immediate diagnosis is critical to enable prompt and appropriate treatment. Computed Tomography Pulmonary Angiography (CTPA) is the current imaging gold standard for PE diagnosis, but its use is limited in settings where exposure to ionizing radiation and the associated cancer risk is severely undesirable (i.e. for children or pregnant patients). CTPA also requires the injection of iodinated contrast media which is contraindicated in a significant fraction of the population. Magnetic Resonance Imaging (MRI) is a radiation-free alternative for diagnostic imaging, although the lungs are traditionally not imaged using MRI due to poor image quality. However, MRI does not require ionizing radiation, and vascular evaluation often can be performed without the need for contrast agents. An efficient, simple, contrast-free, whole-lung MRI method for the assessment of the pulmonary vasculature would enable the detection of PEs without the disadvantages of CT. Recent advancements in MRI hardware, primarily the availability of FDA-cleared lower-field 0.55T MRI systems, have revitalized lung MRI as a field of study. The proposed project aims to develop a fast, contrast-free, 3D whole-lung low- field MRI protocol for PE detection that is both comfortable for the patient and simple to perform: Aim 1: Develop and optimize a data collection and image reconstruction pipeline for rapid, free-breathing, high- resolution, contrast-free, 3D whole-lung MRI at 0.55T with high vessel conspicuity for PE detection We will develop a 3D balanced steady-state free-precession MRI (bSSFP) acquisition which can be used during free- breathing for whole-lung imaging (covering a field-of-view of approximately 400x400x300mm) with a resolution of 1mm3 in an acquisition time of less than 10 minutes. Following optimization of the acquisition/reconstruction pipeline, images of the lung vasculature will be collected in 10 healthy subjects and assessed by three cardiothoracic radiologists for visibility of first and second-order pulmonary branches and the ability to track vessels and distinguish pulmonary arteries from veins. Aim 2: Compare the PE detectability of low-field MRI to that of CTPA in a cohort of ED patients We hypothesize that the low-field, contrast-free 3D whole-lung MRI imaging pipeline developed in Aim 1 will enable PE detection in the first and second-order pulmonary branches comparable to that of CTPA. A cohort of 40 patients presenting in the ED with suspicion of PE scheduled to undergo a CTPA scan will be recruited for an additional low-field MRI scan. Three radiologists will compare the CTPA and MRI images to assess the number of clots, and their location and size; they will also evaluate diagnostic confidence using a Likert scale. Outcome: When successful, this project will provide preliminary data showing the potential for rapid, contrast-free lung PE screening using a low-field MRI technique that is easy to perform in the ED.
