The patho-physiology of heart failure (HF) is fundamentally determined by the failure of the circulatory system to deliver oxygen sufficient for metabolic needs, and is best explained by a complex interplay between intrinsic abnormalities of ventricular pump function and extra-cardiac factors that reduce oxygen utilization in metabolically active tissues (1-4). In HF-patients, the abnormalities in skeletal muscle mitochondrial oxidative function are important determinants of exercise intolerance and may offer novel targets for therapy (5-6). We evaluated the plantar flexion post-exercise PCr resynthesis rate in lower leg muscle with a novel spectrally selective three-dimensional turbo spin echo (3D-TSE) ³¹P-MRI and conventional ³¹P-MRS technique (as a validated measure of skeletal muscle mitochondrial oxidative capacity).All the ³¹P-MRS and MRI experiments were performed on a 3T Siemens scanner using a dual-tuned ³¹P/¹H quadrature volume coils (Rapid MRI, Ohio). Six non-anemic NYHA Class II-III HF-patients (4 male, 2 female, mean age=56±7.3years) and five normal volunteers (all male, mean age=35±6.8years) were recruited. Plantar flexion exercises were performed using a custom built MR compatible ergometer. An exercise protocol was designed for the purpose of ensure enough PCr depletion while keep the pH change minimal, which includes 1-2 mins before the exercise (baseline), 1 min plantar flexions with interval of 1.5s, and recovery phase after the exercise till 10mins. ³¹P-MR data were collected during two separate full exercise protocols for MRS and MRI respectively. Parameters for unlocalized free induction decay (FID) were TR=6s and 2048 points with a spectral width of 3kHz, 100 measurements in 10mins. Spectrally selective PCr Imaging experiments were performed using a modified version of centric 3D-TSE sequence (7), which allows to acquire 2 partitions per TR. Parameters were as following: ETL=24; TE and echo-spacing=26ms; acquisition bandwidth=1.6kHz; matrix size=24x24x4; FOV=220x220x200mm (voxel size=4.2mL); TR=6s, resulting in 12s temporal resolution and 50 measurements in 10 mins. A 16ms Gaussian pulse was used for spectral selective excitation of PCr.All MR spectra were post-processed via zero-filling to 8192 points; zero and first order phase correction and baseline correction were done. After normalizing the acquired data, both MRS and MRI (mean signal intensity of all voxels within muscle), to the pre-exercise value for each subject, we then fitted them to a mono exponential recovery function according to the following equation, using a least squares minimization algorithm.$${PCr(t)=PCr_{0}+C\times(1-e^{-t/k})}[1]$$In Equation[1], PCr₀ is the PCr level at the end of exercise, C is the difference between the steady-state level and the PCr after recovery, and k is the time constant (1/rate constant) of PCr resynthesis in seconds. In addition, we manually segmented the different muscles (GM, GL and SOL) and fitted Equation[1] in each volume of interest in order to measure the regional time constant of PCr resynthesis.According to clinical report, muscle mass of these HF-patients are within normal range. However, the HF-patients demonstrated an increased rate of PCr depletion during exercise and a decreased rate of PCr resynthesis post-exercise compared to normal volunteers; and moreover, the skeletal muscle pH from spectra (distance between Pi and PCr peaks) of both groups showed not much change during the exercise protocol (Fig.1 and Fig.2). The spectrally selective 3D-PCr imaging (Fig.3) also demonstrated that the post-exercise rate of PCr resynthesis was decreased in HF-patients in both predominantly fast twitch (type II) gastrocnemius and predominantly slow twitch (type I) soleus muscle, as well as a summed average of the entire calf muscle volume (Fig.4). In normal volunteers, the Global PCr resynthesis time constant is 24.4±5.5s for MRS and 25.9±5.0s for MRI. Respectively in HF-patients, the Global PCr resynthesis time constant is 48.6±9.2s for MRS and 51.3±7.6s for MRI. While the two methods agreed with each other very well (no statistically significant differences), all the analysis shows significant differences between normal volunteers and HF-patients, using unpaired t-test with significance level of 0.05 (Fig.5). The Regional PCr resynthesis time constants ranged from 25.8±4.0s to 29.8±4.s for volunteers, and from 56.1±9.0s to 63.5±10.4s for HF-patients.The preliminary results suggest that the skeletal muscle mitochondrial dysfunction may contribute to exercise intolerance in HF-patients, and for the first time show that these abnormalities in heart failure are present in both fast twitch (type II) gastrocnemius and slow twitch (type I) soleus muscles. 3D-PCr imaging allows simultaneous measurement of PCr resynthesis rates in several distinct muscles during the post-exercise period to determine the spatial heterogeneities. ³¹P-MRS and MRI of lower leg muscle oxidative metabolism could be a potential useful imaging biomarker for staging and monitoring therapies in patients with heart failure.