Deep-learning synthesized pseudo-CT for MR high-resolution pediatric cranial bone imaging (MR-HiPCB)

被引:9
|
作者
Boroojeni, Parna Eshraghi [1 ]
Chen, Yasheng [2 ]
Commean, Paul K. [3 ]
Eldeniz, Cihat [3 ]
Skolnick, Gary B. [4 ]
Merrill, Corinne [4 ]
Patel, Kamlesh B. [4 ]
An, Hongyu [1 ,2 ,3 ]
机构
[1] Washington Univ, Dept Biomed Engn, St Louis, MO 63110 USA
[2] Washington Univ, Dept Neurol, St Louis, MO 63110 USA
[3] Washington Univ, Sch Med, Mallinckrodt Inst Radiol, St Louis, MO 63110 USA
[4] Washington Univ, Div Plast & Reconstruct Surg, St Louis, MO 63110 USA
基金
美国国家卫生研究院;
关键词
cranial bone imaging; craniosynostosis; deep learning; head trauma; MRI; pseudo-CT; TRAUMATIC BRAIN-INJURY; ATTENUATION CORRECTION; COMPUTED-TOMOGRAPHY; BLACK BONE; RADIATION-EXPOSURE; 3-DIMENSIONAL RECONSTRUCTION; CRANIOFACIAL SKELETON; CHILDREN; SEGMENTATION; PET/MRI;
D O I
10.1002/mrm.29356
中图分类号
R8 [特种医学]; R445 [影像诊断学];
学科分类号
1002 ; 100207 ; 1009 ;
摘要
Purpose CT is routinely used to detect cranial abnormalities in pediatric patients with head trauma or craniosynostosis. This study aimed to develop a deep learning method to synthesize pseudo-CT (pCT) images for MR high-resolution pediatric cranial bone imaging to eliminating ionizing radiation from CT. Methods 3D golden-angle stack-of-stars MRI were obtained from 44 pediatric participants. Two patch-based residual UNets were trained using paired MR and CT patches randomly selected from the whole head (NetWH) or in the vicinity of bone, fractures/sutures, or air (NetBA) to synthesize pCT. A third residual UNet was trained to generate a binary brain mask using only MRI. The pCT images from NetWH (pCT(NetWH)) in the brain area and NetBA (pCT(NetBA)) in the nonbrain area were combined to generate pCT(Com). A manual processing method using inverted MR images was also employed for comparison. Results pCT(Com) (68.01 +/- 14.83 HU) had significantly smaller mean absolute errors (MAEs) than pCT(NetWH) (82.58 +/- 16.98 HU, P < 0.0001) and pCT(NetBA) (91.32 +/- 17.2 HU, P < 0.0001) in the whole head. Within cranial bone, the MAE of pCT(Com) (227.92 +/- 46.88 HU) was significantly lower than pCT(NetWH) (287.85 +/- 59.46 HU, P < 0.0001) but similar to pCT(NetBA) (230.20 +/- 46.17 HU). Dice similarity coefficient of the segmented bone was significantly higher in pCT(Com) (0.90 +/- 0.02) than in pCT(NetWH) (0.86 +/- 0.04, P < 0.0001), pCT(NetBA) (0.88 +/- 0.03, P < 0.0001), and inverted MR (0.71 +/- 0.09, P < 0.0001). Dice similarity coefficient from pCT(Com) demonstrated significantly reduced age dependence than inverted MRI. Furthermore, pCT(Com) provided excellent suture and fracture visibility comparable to CT. Conclusion MR high-resolution pediatric cranial bone imaging may facilitate the clinical translation of a radiation-free MR cranial bone imaging method for pediatric patients.
引用
收藏
页码:2285 / 2297
页数:13
相关论文
共 50 条
  • [1] MR Cranial Bone Imaging: Evaluation of Both Motion-Corrected and Automated Deep Learning Pseudo-CT Estimated MR Images
    Linkugel, Andrew D.
    Wang, Tongyao
    Boroojeni, Parna Eshraghi
    Eldeniz, Cihat
    Chen, Yasheng
    Skolnick, Gary B.
    Commean, Paul K.
    Merrill, Corinne M.
    Strahle, Jennifer M.
    Goyal, Manu S.
    An, Hongyu
    Patel, Kamlesh B.
    AMERICAN JOURNAL OF NEURORADIOLOGY, 2024, 45 (09) : 1284 - 1290
  • [2] Dixon-VIBE Deep Learning (DIVIDE) Pseudo-CT Synthesis for Pelvis PET/MR Attenuation Correction
    Torrado-Carvajal, Angel
    Vera-Olmos, Javier
    Izquierdo-Garcia, David
    Catalano, Onofrio A.
    Morales, Manuel A.
    Margolin, Justin
    Soricelli, Andrea
    Salvatore, Marco
    Malpica, Norberto
    Catana, Ciprian
    JOURNAL OF NUCLEAR MEDICINE, 2019, 60 (03) : 429 - 435
  • [3] Deep learning-based pseudo-CT synthesis from zero echo time MR sequences of the pelvis
    Getzmann, Jonas M.
    Deininger-Czermak, Eva
    Melissanidis, Savvas
    Ensle, Falko
    Kaushik, Sandeep S.
    Wiesinger, Florian
    Cozzini, Cristina
    Sconfienza, Luca M.
    Guggenberger, Roman
    INSIGHTS INTO IMAGING, 2024, 15 (01):
  • [4] HIGH-RESOLUTION MR IMAGING OF JUXTASELLAR MENINGIOMAS WITH CT AND ANGIOGRAPHIC CORRELATION
    YEAKLEY, JW
    KULKARNI, MV
    MCARDLE, CB
    HAAR, FL
    TANG, RA
    AMERICAN JOURNAL OF NEURORADIOLOGY, 1988, 9 (02) : 279 - 285
  • [5] High-resolution 3D MR Fingerprinting using parallel imaging and deep learning
    Chen, Yong
    Fang, Zhenghan
    Hung, Sheng-Che
    Chang, Wei-Tang
    Shen, Dinggang
    Lin, Weili
    NEUROIMAGE, 2020, 206
  • [6] Radiotherapy-Specific Imaging Prior Knowledge and Its Impact in Deep Learning Based Cranial Pseudo-CT Generation
    Law, M.
    Yuan, J.
    Lo, G.
    Wong, O.
    Zhou, Y.
    Yu, S.
    MEDICAL PHYSICS, 2019, 46 (06) : E510 - E510
  • [7] TEMPORAL BONE REGION - HIGH-RESOLUTION MR IMAGING USING SURFACE COILS
    KOENIG, H
    LENZ, M
    SAUTER, R
    RADIOLOGY, 1986, 159 (01) : 191 - 194
  • [8] A robust two-step deep learning method for pseudo-CT generation for PET-MR attenuation correction in brain
    Krokos, G.
    Mackewn, J.
    Yakub, S.
    Srinivasan, R.
    Jeljeli, S.
    Kinsella, S.
    Marsden, P. K.
    Hammers, A.
    EUROPEAN JOURNAL OF NUCLEAR MEDICINE AND MOLECULAR IMAGING, 2024, 51 : S131 - S131
  • [9] High-Resolution Isotropic Three-Dimensional MR Imaging of the Extraforaminal Segments of the Cranial Nerves
    Wen, Jessica
    Desai, Naman S.
    Jeffery, Dean
    Aygun, Nafi
    Blitz, Ari
    MAGNETIC RESONANCE IMAGING CLINICS OF NORTH AMERICA, 2018, 26 (01) : 101 - +
  • [10] Segmental analysis of cochlea on three-dimensional MR imaging and high-resolution CT
    Akiba, H
    Tamakawa, M
    Hyodoh, H
    Hyodoh, K
    Nakamura, N
    Yama, N
    RADIOLOGY, 2001, 221 : 135 - 135