Magnetic susceptibility is an intrinsic physical property of tissue. In the pulmonary tissue, strong-susceptibility-induced field perturbation is mostly a source of artifact resulting in severe distortion. Breathing artifacts pose further challenges, especially in pediatric populations. These artifacts coupled with the intrinsic low proton SNR make it extremely difficult to perform quantitative susceptibility mapping in the lung. One technique, called streaking artifact reduced quantitative susceptibility mapping (STAR-QSM), is a promising tool that can characterize local tissue susceptibility and can reduce the strong susceptibility effects that are found in the lungs. STAR-QSM (1) combined with the recently developed locally low rank parallel imaging method (2) could further reduce motion artifacts and allow free-breathing high resolution QSM of the lung. To our knowledge, this is the first study to examine susceptibility properties in human lungs in vivo.The datasets were acquired on a 3T GE MR750 scanner (GE Healthcare, Waukesha, WI) using a 32 channel cardiac coil using a 3D modified SPGR sequence with motion navigation (Butterfly), intermitted spectrally selective fat inversion pulses, and VDRad sapling patterns were used during contrast injection. The datasets were accelerated in both phase encoding directions with an overall reduction factor R 7.8~8 (2). The imaging parameters were: FOV = 280 x 224 x 192 mm³, reconstructed matrix = 320 x 256 x 192, spatial resolution = 0.875 x 0.875 x 1 mm³, coronal 192 slices. TR = 4.26 ms, TE = 0.98 ms, bandwidth = 62.5 kHz. Raw k-space data was reconstructed from under-sampled data using coil compression technique (3) and locally low rank ESPIRiT for dynamic contrast enhanced (DCE) images (2) integrated in the BART toolbox (https://mikgroup.github.io/bart/). Then 3D complex data was separated into magnitude and phase. The resulting magnitude images were used to obtain the mask of the pulmonary regions. The phase was unwrapped using Laplacian-based unwrapping and background phase was removed using V-SHARP (4). The QSM images were reconstructed using a two-level streaking artifact reduction for QSM algorithm (1).The representative susceptibility maps between thresholded k-space Division (TKD) (5), iLSQR (6) and STAR-QSM are shown in Fig. 1. For the TKD and iLSQR methods, the streaking artifacts, especially form the lung/liver interface, strongly contaminate the lung tissues. These artifacts were larger eliminated by STAR-QSM (indicated by black arrows) and more vessels were revealed by STAR-QSM as indicated by red arrows. With the improved image quality from STAR-QSM, coronal slice susceptibility maps were compared with corresponding magnitude images as illustrated in Fig. 2. The anatomical structure revealed by QSM, e.g., major blood vessels, matches well with the magnitude image. Some of the smaller blood vessels show paramagnetic susceptibility. These blood vessels are not visible in the magnitude (indicated by yellow arrow in Fig. 2). Moreover, pulmonary tissues parenchyma is more visible in the QSM image than magnitude (red arrows in Fig. 2). The major pulmonary vessels show a paramagnetic susceptibility of 2.8 ± 0.2 ppm while the pulmonary tissues have a diamagnetic susceptibility of -0.8 ± 0.46 ppm. Volume rendering by magnitude and magnitude overlaid with QSM are shown in Fig. 3. The blood vessel network is revealed by magnitude. Moreover, details of the pulmonary tissue can be seen when QSM is overlaid on the magnitude image. For example, tissues with a more diamagnetic susceptibility are visible near the lung boundaries (red arrow in Fig. 3).One of the main challenges with imaging pulmonary tissues with MRI is the low proton density and the fast signal decay (short T2*) due to tissue-air interfaces (7). QSM takes advantage of the field perturbation of tissue susceptibility in the presence of an externally applied magnetic field. QSM provides a novel contrast that can be achieved at short TE given the strong susceptibility variations in the lung, in which the voxel intensity is linearly proportional to the underlying tissue magnetic susceptibility. QSM combined with low-rank reconstruction allows highly undersampled free-breathing lung susceptibility mapping, which is useful for quantifying endogenous tissue property and external contrast agents.In this study, we have demonstrated the feasibility of obtaining QSM maps of pulmonary vessels and tissue. Different susceptibility values were observed for blood vessels and pulmonary tissues. QSM offers a new method for in vivo quantification of gadolinium or SPIO concentrations. e.g., DCE, pulmonary embolism or edema.