Marcus Carlsson

Background: Valvular diseases can be both a cause and complication of heart failure. It is important to understand how valve leakage and stenosis affect heart function in order to choose right treatment. Valvular diseases are usually examined with ultrasound at rest, but how the heart is affected by increased blood flow and during work is not known. Magnetic resonance imaging (MRI) with computer modeling has the potential to provide better information but is rarely used in healthcare today but could guide the selection of better treatment. Purpose and aims: To determine how outcome after intervention/surgery of mitral and aortic valve diseases can be predicted with MRI and modeling and how the heart is affected by increased blood flow and at work. The purpose is to guide appropriate therapy and thus increase the well-being of patients with heart failure caused by valvular disease. In addition, to demonstrate whether the cause of chest pain in valvular disease is due to hemodynamic influence or small vessel disease. We hypothesize that computer modeling based on patient specific MRI input can predict ventricular reverse remodeling after mitral valvular surgery, 2) regression of left ventricular hypertrophy after aortic stenosis surgery/transcatheter intervention. We also hypothesize that the dynamics of aortic insufficiency is load dependent and affected by peripheraral vascular resistacne and that this can be predicted by computer modeling. Finally we hypothesize that chest pain in aortic stenosis is of hemodynamic etiology. Work plan: Patients with severe aortic stenosis are examined before catheter intervention (n=40) or surgery (n=40) and followed up early (1 month) and late (1 year) with MRI at rest and hyperemia. Patients with significant aortic insufficiency (n=40) are examined with MRI during exercise and significant mitral insufficiency (n=90) with hyperemia before and after surgery. Computer modeling is used to predict the surgical outcomes and is compared to the actual outcome. Patient outcome in this study is compared with a larger population from national registries. Significance: This project opens up new opportunities to predict the outcome of surgery and catheter treatment in valvular diseases and can be used to better guide the right patients with valve leakage to be operated on at the right time. In addition, patients with aortic stenosis and chest pain will be able to receive the correct diagnosis and treatment.

Impaired cardiac function often results in heart failure which represents a major burden of morbidity and mortality. Cardiac magnetic resonance imaging (MRI) is considered to be the reference method for assessment of cardiac function. However, the relatively long scan times makes MRI sensitive to motion. Most current cardiac MRI techniques require multiple breath holds, which many patients with heart failure are unable to perform., resulting in non-diagnostic image quality. The aim of this project is to develop a novel framework for adaptive image acquisition, where the pulse sequence is controlled by online motion estimators. This will enable acquisition that is robust to respiratory motion and arrhythmia, as well as fetal motion during examination of fetal hearts. The methods will be evaluated both in healthy volunteers, heart failure patients, and in pregnant women. The pulse sequence and corresponding image reconstruction will be implemented directly on the clinical scanner platform, making it immediately available for clinical use. The research will take place immersed in a clinical environment at the Karolinska University Hospital, with full access to dedicated cardiac MRI scanners and relevant patient populations. The proposed approach is highly original, fills an important unmet clinical need in medicine, and involves novel approaches that push the technical limits of image acquisition and reconstruction beyond state-of-the-art.