Project Summary Motivation: In the United States, approximately 1 in 300 non-Hispanic Caucasians have hereditary hemochromatosis with varying rates in other ethnicities and races. Transfusion hemosiderosis often develops in cancer patients who have received repeated blood transfusions for cancer-related anemia treatment. Iron overload, if left untreated, can cause fatal organ damage. Treatment of iron overload aims at reducing body iron stores with phlebotomy or chelation therapy to maintain sufficiently low body iron levels while minimizing adverse effects. Liver iron concentration is widely accepted for the best indicator of total body iron stores; therefore, accurate liver iron quantification improves clinical management of iron overload and minimizes side effects of chelator administration. MRI-based R2* relaxometry (R2*-MRI) is a noninvasive clinical standard for liver iron quantification with no ionizing radiation due to its sensitivity to the presence of tissue iron. R2*-MRI acquires images at multiple time points to estimate R2* (= 1/T2*) relaxation rates at each voxel. However, clinically available R2*-MRI is prone to patient movement such as respiratory motion that breath-holding is commonly practiced during scan. Breath-holding is challenging for children and some adults, and unsatisfactory breath- holding leads to poor R2* values. The other challenge with the current R2*-MRI is caused by signal loss due to rapid T2* decay in the presence of a high concentration of iron. Severe iron loading is commonly seen in patients with transfusion-induced iron overload, and the current R2*-MRI does not reliably capture such a rapid signal decay, resulting in inaccuracy in R2* measurements. This project addresses these challenges to provide novel R2*-MRI methods for accurate and consistent liver iron quantification. Approach: The project has two development aims that are validated on clinical studies. Aim 1 will enable rapid free-breathing ultrashort TE MRI, which acquires images at multiple time points for R2* measurements. Such acquisition is achieved by incorporating a 3D center-out k-space trajectory with multiple data readouts with self- navigator without scan time overhead, facilitating parallel imaging compressed sensing. Aim 2 will enable retrospectively motion-corrected image reconstruction that makes use of a low-rank tensor structure of the acquired 3D spatiotemporal volumetric data. We will develop strategies for data parallelism and distributed computing for computationally demanding tensor-based multidimensional reconstruction. Aim 3 will determine the performance of the innovations in a clinical setting. Significance: This work will lead to rapid, robust, free-breathing abdominal MRI for more accurate assessment of liver iron overload in children and adults. The techniques will facilitate widespread application in quantitative body imaging.