Patterns of Popular Artifacts in QSM and χ-separation (chi-separation)

2292 

Abstract Submission 

Hayeon Lee¹, Kyeongseon Min¹, Sooyeon Ji¹, Jonghyo Youn¹, Taechang Kim¹, Beomseok Sohn², Woo Jung Kim^3,4, Chae Jung Park⁵, Soohwa Song⁶, Dong Hoon Shin⁶, Kyung Won Chang⁷, Na-Young Shin⁸, Phil Hyu Lee⁹, Yangsean Choi¹⁰, Yoonho Nam¹¹, Koung Mi Kang¹², Agnieszka Burzynska^13,14, Catherine Lebel¹⁵, Jongho Lee¹

¹Department of Electrical and Computer Engineering, Seoul National University, Seoul, Korea, Republic of, ²Department of Radiology, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Korea, Republic of, ³Institute of Behavioral Sciences in Medicine, Yonsei University College of Medicine, Seoul, Korea, Republic of, ⁴Department of Psychiatry, Yongin Severance Hospital, Yonsei University College of Medicine, Yongin, Korea, Republic of, ⁵Department of Radiology, Yongin Severance Hospital, Yonsei University College of Medicine, Yongin, Korea, Republic of, ⁶Heuron Co., Ltd, Seoul, Korea, Republic of, ⁷Department of Neurosurgery, Severance Hospital, Yonsei University College of Medicine, Seoul, Korea, Republic of, ⁸Department of Radiology, Severance Hospital, Yonsei University College of Medicine, Seoul, Korea, Republic of, ⁹Department of Neurology, Severance Hospital, Yonsei University College of Medicine, Seoul, Korea, Republic of, ¹⁰Department of Radiology and Research Institute of Radiology, University of Ulsan College of Medicine, Seoul, Korea, Republic of, ¹¹Division of Biomedical Engineering, Hankuk University of Foreign Studies, Yongin, Korea, Republic of, ¹²Department of Radiology, Seoul National University Hospital, Seoul, Korea, Republic of, ¹³Department of Human Development and Family Studies, Colorado State University, Fort Collins, CO, United States, ¹⁴Department of Molecular, Cellular and Integrative Neurosciences, Colorado State University, Fort Collins, CO, United States, ¹⁵Department of Pediatrics, University of Calgary, Calgary, Alberta, Canada

Hayeon Lee  
Department of Electrical and Computer Engineering, Seoul National University
Seoul, Korea, Republic of

Kyeongseon Min  
Department of Electrical and Computer Engineering, Seoul National University
Seoul, Korea, Republic of

Sooyeon Ji  
Department of Electrical and Computer Engineering, Seoul National University
Seoul, Korea, Republic of

Jonghyo Youn  
Department of Electrical and Computer Engineering, Seoul National University
Seoul, Korea, Republic of

Taechang Kim  
Department of Electrical and Computer Engineering, Seoul National University
Seoul, Korea, Republic of

Beomseok Sohn  
Department of Radiology, Samsung Medical Center, Sungkyunkwan University School of Medicine
Seoul, Korea, Republic of

Woo Jung Kim  
Institute of Behavioral Sciences in Medicine, Yonsei University College of Medicine|Department of Psychiatry, Yongin Severance Hospital, Yonsei University College of Medicine
Seoul, Korea, Republic of|Yongin, Korea, Republic of

Chae Jung Park  
Department of Radiology, Yongin Severance Hospital, Yonsei University College of Medicine
Yongin, Korea, Republic of

Soohwa Song  
Heuron Co., Ltd
Seoul, Korea, Republic of

Dong Hoon Shin  
Heuron Co., Ltd
Seoul, Korea, Republic of

Kyung Won Chang  
Department of Neurosurgery, Severance Hospital, Yonsei University College of Medicine
Seoul, Korea, Republic of

Na-Young Shin  
Department of Radiology, Severance Hospital, Yonsei University College of Medicine
Seoul, Korea, Republic of

Phil Hyu Lee  
Department of Neurology, Severance Hospital, Yonsei University College of Medicine
Seoul, Korea, Republic of

Yangsean Choi  
Department of Radiology and Research Institute of Radiology, University of Ulsan College of Medicine
Seoul, Korea, Republic of

Yoonho Nam  
Division of Biomedical Engineering, Hankuk University of Foreign Studies
Yongin, Korea, Republic of

Koung Mi Kang  
Department of Radiology, Seoul National University Hospital
Seoul, Korea, Republic of

Agnieszka Burzynska  
Department of Human Development and Family Studies, Colorado State University|Department of Molecular, Cellular and Integrative Neurosciences, Colorado State University
Fort Collins, CO, United States|Fort Collins, CO, United States

Catherine Lebel, Ph.D.  
Department of Pediatrics, University of Calgary
Calgary, Alberta, Canada

Jongho Lee  
Department of Electrical and Computer Engineering, Seoul National University
Seoul, Korea, Republic of

QSM and χ-separation⁸ quantify brain tissue susceptibility for studying neurodegenerative diseases. However, artifacts in data acquisition and processing can compromise accuracy, impacting reliability. It is important to recognize these artifacts and implement appropriate correction methods. The study analysed data from healthy subjects and patients in various groups and from different vendors and explored commonly encountered artifacts. The effects, origins, and potential solutions for these artifacts were investigated to improve the application of QSM and χ-separation.

364 subjects (52.24 ± 25.78 years; 163 males and 201 females) from Parkinson's disease, Alzheimer's disease, hypertension, and alcohol-exposed adolescents development studies were scanned at six different 3T MRI scanners (Siemens Trio, Siemens Vida, Siemens Skyra, Philips Ingenia CX, Philips Ingenia Elition X, and GE Discovery 750w).
All data processing for QSM and χ-separation was performed using the χ-separation toolbox (https://github.com/SNU-LIST/chi-separation). Phases were unwrapped using a Laplacian-based algorithm⁷. V-SHARP⁹ was applied to remove background fields. QSM was calculated using QSMnet¹⁰. χ-sepnet-R2*⁶ and χ-sepnet-R2'⁶ were utilized calculating χ_para and χ_dia maps.

Motion resulted in ghosting and blurring in QSM and χ-separation results (Fig. 1a). Respiration-induced B0 fluctuations hinder QSM and χ-separation (Fig. 1b), affecting reproducibility and accuracy of QSM and χ-separation results¹. Post-acquisition correction for these artifacts is challenging without extra information such as B0-navigation. Incorrect coil combination introduced "phase singularities", and a localized high intensity region may appear (Fig. 1c), resolved by proper coil combination. GRAPPA algorithm reconstruction errors yield aliasing artifacts (Fig.1d), corrected by reprocessing raw data. Large slice thickness leads to QSM underestimation⁴, R2* overestimation³, and ultimately overestimation of both χ_para and χ_dia (Fig. 1e), mitigated by using isotropic voxels of 1 mm or less⁴. Thin slabs cause significant QSM underestimation^2,4 due to truncated non-local dipole, leading to underestimation of both dominant source in χ_para and χ_dia maps (Fig. 2a,b). Data from P*** vendor sometimes contained linear field bias, causing residual phase wraps in QSM and χ-separation results (Fig. 1g) as the nonlinear complex data fitting approach for echo combination and phase unwrapping were utilized. Unwrapping phase at each TE and combining multi-echo images can be the solution. QSMnet may mistakenly use a local field map in radians instead of Hz, resulting in intensity range errors (Fig. 1h). Failure to correct B0 orientation during processing causes misalignment of the magnetic dipole kernel with B0 direction and susceptibility errors⁵ (Fig. 2c). Unintentionally flipped images impact alignment between T2-weighted and GRE magnitude images, leading to incorrect R2' maps. This can manifest as a prominent bright pattern in the cortical region in χ_para and χ_dia (Fig. 2d). Vessel flow artifacts can impact adjacent regions, leading to erroneously high or low values in QSM and χ-separation results (Fig. 1k). To ensure accurate ROI analysis, it is advisable to exclude these affected regions.

We reported various artifacts in QSM and χ-separation from 364 subjects. Their origins and proposed practical solutions to mitigate their effects were investigated, thereby improving the precision and reliability of quantitative analysis in MRI studies. The inclusion of visual images as aids for artifact recognition benefits both researchers and practitioners, significantly enhancing data quality and interpretation across clinical and research studies.

Methods Development

Motion Correction and Preprocessing ²

Anatomical MRI ¹

Imaging Methods Other

Acquisition

Data analysis

MRI

Myelin

STRUCTURAL MRI

White Matter

Other - QSM