HKHahnsung Kim

Hahnsung Kim

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Hahnsung Kim has a total of 24 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Quasi‐Steady‐State Equilibrium CEST Reconstruction Improves T 1 ‐Normalized Apparent Exchange‐Dependent Relaxation Analysis in Brain Tumor Patients at 3 Tesla

    by Hahnsung Kim, Julius Juhyun Chung, Jinsuh Kim, Joshua David Brown, Vahid Khalilzad Sharghi, Ranliang Hu, Phillip Zhe Sun

    This study evaluates quasi‐steady‐state (QUASS) equilibrium CEST reconstruction for improving T 1 ‐normalized apparent exchange‐dependent relaxation (AREX) analysis in quantitative CEST MRI, aiming to reduce confounding T 1 contributions in brain tumor imaging at a clinical 3T scanner. Two‐ and five‐pool CEST effects were simulated using Bloch–McConnell modeling in MATLAB. Phantom and in vivo imaging were performed on a 3T MRI system. Phantom experiments used MnCl 2 doping to vary T 1 relaxation times, and in vivo data were acquired from three healthy volunteers and one brain tumor patient. CESTR and AREX metrics, with and without QUASS correction, were evaluated using linear regression models to assess T 1 dependence. In phantoms, conventional AREX showed a significant T 1 dependence (slope = −0.75, p = 0.014), whereas QUASS‐corrected AREX showed no significant dependence on T 1 (slope = 0.18, p = 0.16). Analysis of covariance confirmed that T 1 dependence differed between the two approaches ( p = 0.0018). In vivo, QUASS correction reduced acquisition‐dependent variability and stabilized AREX contrast across regions with substantially different T 1 values. QUASS reconstruction effectively mitigates confounding T 1 effects in CEST measurements, improving the robustness of AREX‐based quantitative CEST analysis and supporting its use in clinical CEST MRI applications.

  • Cubic Millimeter High Resolution 3D Inner‐Volume GRASE ( IV ‐ GRASE ) CEST MRI Using T 1 ‐Integrated Variable Density CAIPI Sampling With Temporal Random Walk: A Feasibility Study

    by Hahnsung Kim, Suhyung Park, Julius Juhyun Chung, Jinsuh Kim, Kimberly B. Hoang, Ranliang Hu, Jeffrey Olson, David A. Feinberg, Phillip Zhe Sun

    Purpose: To develop a cubic‐millimeter, high‐resolution CEST MRI sequence with enhanced efficiency and quantitative stability. Methods: The sequence combined an inversion‐recovery T 1 module and a multi‐offset CEST module, both acquired with inner‐volume GRASE (IV‐GRASE) readout. Acquisition efficiency was improved using variable‐density random‐walk sampling, and reconstruction employed a self‐calibrated joint framework with sparsity and low‐rank constraints to leverage redundancy across T 1 and CEST datasets. Feasibility was evaluated in phantoms and preliminary patient studies. Results: The framework enabled simultaneous T 1 and CEST imaging with reduced artifacts, improved quantitative stability, and efficient high‐resolution reconstruction. Phantom experiments demonstrated accurate T 1 and Z‐spectral measurements with inherently co‐registered datasets. Preliminary patient imaging revealed heterogeneity within the diseased region and exchange‐dependent relaxation contrast, clearly differentiating it from contralateral normal tissue. Conclusion: The proposed hybrid T 1 –CEST MRI sequence provides efficient, artifact‐reduced, and quantitatively reliable high‐resolution imaging in a single acquisition, advancing the feasibility of quantitative CEST analysis.

  • Detection of Altered Muscle Glycogen‐ and NOE ‐Weighted CEST MRI Signals in an Acid Alpha‐Glucosidase‐Deficient Mouse Model of Pompe Disease Using QUASS ‐Enhanced Multi‐Pool Quantitative CEST ( QCEST ) Imaging

    by Limin Wu, Hahnsung Kim, Phillip Zhe Sun

    Purpose: Chemical exchange saturation transfer (CEST) MRI is a non‐invasive imaging technique that detects dilute labile protons with enhanced sensitivity, aiding in the assessment of tissue properties. This study aimed to evaluate the diagnostic usefulness of CEST MRI in acid alpha‐glucosidase knockout (GAAKO) mice, a model of Pompe disease characterized by lysosomal glycogen accumulation in muscles. Methods: We conducted multiparametric MRI at a 7 T Bruker MRI scanner, including T1 and T2 relaxation mapping and densely sampled CEST Z‐spectral acquisition, on adult GAAKO and wild‐type control mice. Multi‐pool CEST signals were obtained from spinlock model‐based fitting of the quasi‐steady‐state reconstructed Z‐spectrum. Using the gold‐standard muscle glycogen assay as a biomarker of disease burden, we analyzed the correlation between relaxation and CEST signals and tissue glycogen levels. Results: Biochemical glycogen assays of excised muscle samples confirmed significantly elevated glycogen levels in GAAKO mice (556 ± 125 mg/kg) compared to wild‐type controls (60 ± 22 mg/kg). Among MRI‐derived parameters, glycogen and near‐resonance nuclear Overhauser enhancement (NOE) signals showed significant differences between control and GAAKO mice. Importantly, glycogen‐weighted CEST (R 2 = 0.45, p < 0.0001) and NOE signal at −1.2 ppm ( R 2 = 0.57, p < 0.0001) demonstrated significant correlations with glycogen assay measurement. Conclusion: Multi‐pool CEST MRI metrics show potential for characterizing muscle pathology in GAAKO mice. These findings support using CEST MRI to monitor tissue glycogen levels non‐invasively.

  • Demonstration of CEST MRI in an Animal Model of Acute Ischemic Stroke at 15. 2 T With Simple Linearization for B 1 Inhomogeneity Correction

    by Julius Juhyun Chung, Geun‐Ho Im, Hahnsung Kim, Seong‐Gi Kim, Phillip Zhe Sun

    Purpose: CEST at ultra‐high fields such as 15.2 T benefits significantly from increased signal and increased spectral resolution. However, inhomogeneous fields become a confounding issue for B 0 and B 1 . While B 0 inhomogeneity correction is widely studied and implemented, methods for B 1 correction have been limited to interpolation, which requires close to ground truth sampling across B 1 or sophisticated methods that increase data acquisition requirements. This work demonstrates a simple and easy‐to‐employ model‐based linearization that allows for reasonable B 1 correction with as low as two B 1 samples for implementation. Methods: Simulation and in vivo rodent brain CEST data at 15.2 T were tested for reasonable B 1 correction using a surface coil to demonstrate an exaggerated situation of B 1 inhomogeneity. Individual exchange fitting was performed, and quantified maps were examined for apparent residual B 1 dependencies, and quantitative values were examined for consistency across animals. Results: Simulation showed linearization significantly improved linear dependency over simple linear interpolation, even when multiple pools were involved. B 1 corrected Z‐spectra demonstrated reasonable consistency across animals, and parameter maps of simple quantifying CEST parameters demonstrated little B 1 dependency. Fitting CEST curves demonstrates similar consistency and independence from B 1 . Finally, comparison of parameters across animals demonstrates reasonable consistency after B 1 correction. Conclusion: The simple linearization described in this paper offers a facile way to rapidly correct for B 1 inhomogeneity with minimally required additional spectral sampling. This method can have significant advantages, particularly in scan‐time restricted situations where multiple Z‐spectra at more than two power levels are not feasible.


  • Higher spatial resolution and sensitivity in whole brain functional MRI at 7T using 3D EPI accelerated with variable density CAIPI sampling and temporal random walk

    by Suhyung Park, Alexander Beckett, Suvi Häkkinen, Erica Walker, Samantha J. Ma, Sugil Kim, Hahnsung Kim, David A. Feinberg

    Purpose: To develop an efficient 3D EPI encoding technique for high spatiotemporal resolution functional MRI. Methods: To exploit spatiotemporal fMRI data structure, we introduce a variable density 2D CAIPI sampling in the spatial domain combined with time‐wise extra random encoding in the time domain, thus achieving pseudo‐regular sampling with a regular blip while allowing incoherent sampling in a complementary manner across time. This enabled temporally regularized reconstruction of highly accelerated functional data acquisition. The encoding scheme was then validated against temporally invariant CAIPI encoding by applying to locally confined and whole‐brain around the primary visual cortex, respectively, with increasing the spatial resolutions. Results: For partial brain imaging, our proposed method achieved higher reconstruction accuracy, resulting in a substantial increase of SSIM compared to an alternative method for 0.64 mm‐isotropic resolution. When used for whole brain imaging at 0.56 mm‐isotropic resolution, our method showed a decreased spatial extent of activation and produced high‐quality images for a clear distinction between activated and non‐activated regions around calcarine fissure with high spatial specificity. Conclusion: The proposed 3D EPI encoding scheme, which exploits coherent and incoherent sampling properties, can significantly improve the image quality while providing a good balance between sensitivity and specificity in the activated regions.


  • Improving standardization and accuracy of in vivo omega plot exchange parameter determination using rotating‐frame model‐based fitting of quasi‐steady‐state Z‐spectra

    by Julius Juhyun Chung, Hahnsung Kim, Yang Ji, Dongshuang Lu, Iris Y. Zhou, Phillip Zhe Sun

    Purpose: Although Ω‐plot‐driven quantification of in vivo amide exchange properties has been demonstrated, differences in scan parameters may complicate the fidelity of determination. This work systematically evaluated the use of quasi‐steady‐state (QUASS) Z‐spectra reconstruction to standardize in vivo amide exchange quantification across acquisition conditions and further determined it in vivo.Methods: Simulation and in vivo rodent brain chemical exchange saturation transfer (CEST) data at 4.7 T were fit with and without QUASS reconstruction using both multi‐Lorentzian and model‐based fitting approaches. pH modulation was accomplished both in simulation and in vivo by inducing global ischemia via cardiac arrest. Amide parameters were determined via Ω‐plots and compared across methods.Results: Simulation showed that Ω‐plots using multi‐Lorentzian fitting could underestimate the exchange rate, with error increasing as conditions diverged from the steady state. In comparison, model‐based fitting using QUASS estimated the same exchange rate within 2%. These results aligned with in vivo findings where multi‐Lorentzian fitting of native Z‐spectra resulted in an exchange rate of 64 ± 13 s−1 (38 ± 16 s−1 after cardiac arrest), whereas model‐based fitting of QUASS Z‐spectra yielded an exchange rate of 126 ± 25 s−1 (49 ± 13 s−1).Conclusion: The model‐based fitting of QUASS CEST Z‐spectra enables consistent and accurate quantification of exchange parameters through Ω‐plot construction by reducing error due to signal overlap and nonequilibrium CEST effects.


  • 3D CEST MRI with an unevenly segmented RF irradiation scheme: A feasibility study in brain tumor imaging

    by Hahnsung Kim, Suhyung Park, Ranliang Hu, Kimberly B. Hoang, Phillip Zhe Sun

    Purpose: To integrate 3D CEST EPI with an unevenly segmented RF irradiation module and preliminarily demonstrate it in the clinical setting.Methods: A CEST MRI with unevenly segmented RF saturation was implemented, including a long primary RF saturation to induce the steady‐state CEST effect, maintained with repetitive short secondary RF irradiation between readouts. This configuration reduces relaxation‐induced blur artifacts during acquisition, allowing fast 3D spatial coverage. Numerical simulations were performed to select parameters such as flip angle (FA), short RF saturation duration (Ts2), and the number of readout segments. The sequence was validated experimentally with data from a phantom, healthy volunteers, and a brain tumor patient.Results: Based on the numerical simulation and l‐carnosine gel phantom experiment, FA, Ts2, and the number of segments were set to 20°, 0.3 s, and the range from 4 to 8, respectively. The proposed method minimized signal modulation in the human brain images in the kz direction during the acquisition and provided the blur artifacts‐free CEST contrast over the whole volume. Additionally, the CEST contrast in the tumor tissue region is higher than in the contralateral normal tissue region.Conclusions: It is feasible to implement a highly accelerated 3D EPI CEST imaging with unevenly segmented RF irradiation.


  • Demonstration of fast multi‐slice quasi‐steady‐state chemical exchange saturation transfer (QUASS CEST) human brain imaging at 3T

    by Hahnsung Kim, Lisa C. Krishnamurthy, Phillip Zhe Sun

    Purpose: To combine multi‐slice chemical exchange saturation transfer (CEST) imaging with quasi‐steady‐state (QUASS) processing and demonstrate the feasibility of fast QUASS CEST MRI at 3T.Methods: Fast multi‐slice echo planar imaging (EPI) CEST imaging was developed with concatenated slice acquisition after single radiofrequency irradiation. The multi‐slice CEST signal evolution was described by the spin‐lock relaxation during saturation duration (Ts) and longitudinal relaxation during the relaxation delay time (Td) and post‐label delay (PLD), from which the QUASS CEST was generalized to fast multi‐slice acquisition. In addition, numerical simulations, phantom, and normal human subjects scans were performed to compare the conventional apparent and QUASS CEST measurements with different Ts, Td, and PLD.Results: The numerical simulation showed that the apparent CEST effect strongly depends on Ts, Td, and PLD, while the QUASS CEST algorithm minimizes such dependences. In the L‐carnosine gel phantom, the proposed QUASS CEST effects (2.68 ± 0.12% [mean ± SD]) were higher than the apparent CEST effects (1.85 ± 0.26%, p < 5e‐4). In the human brain imaging, Bland‐Altman analysis bias of the proposed QUASS CEST effects was much smaller than the PLD‐corrected apparent CEST effects (0.03% vs. −0.54%), indicating the proposed fast multi‐slice CEST imaging is robust and accurate.Conclusions: The QUASS processing enables fast multi‐slice CEST imaging with minimal loss in the measurement of the CEST effect.