Projection Space Implementation of Deep Learning–Guided Low-Dose Brain PET Imaging Improves Performance over Implementation in Image Space

Our purpose was to assess the performance of full-dose (FD) PET image synthesis in both image and sinogram space from low-dose (LD) PET images and sinograms without sacrificing diagnostic quality using deep learning techniques. Methods: Clinical brain PET/CT studies of 140 patients were retrospectively used for LD-to-FD PET conversion. Five percent of the events were randomly selected from the FD list-mode PET data to simulate a realistic LD acquisition. A modified 3-dimensional U-Net model was implemented to predict FD sinograms in the projection space (PSS) and FD images in image space (PIS) from their corresponding LD sinograms and images, respectively. The quality of the predicted PET images was assessed by 2 nuclear medicine specialists using a 5-point grading scheme. Quantitative analysis using established metrics including the peak signal-to-noise ratio (PSNR), structural similarity index metric (SSIM), regionwise SUV bias, and first-, second- and high-order texture radiomic features in 83 brain regions for the test and evaluation datasets was also performed. Results: All PSS images were scored 4 or higher (good to excellent) by the nuclear medicine specialists. PSNR and SSIM values of 0.96 ± 0.03 and 0.97 ± 0.02, respectively, were obtained for PIS, and values of 31.70 ± 0.75 and 37.30 ± 0.71, respectively, were obtained for PSS. The average SUV bias calculated over all brain regions was 0.24% ± 0.96% and 1.05% ± 1.44% for PSS and PIS, respectively. The Bland–Altman plots reported the lowest SUV bias (0.02) and variance (95% confidence interval, −0.92 to +0.84) for PSS, compared with the reference FD images. The relative error of the homogeneity radiomic feature belonging to the gray-level cooccurrence matrix category was −1.07 ± 1.77 and 0.28 ± 1.4 for PIS and PSS, respectively. Conclusion: The qualitative assessment and quantitative analysis demonstrated that the FD PET PSS led to superior performance, resulting in higher image quality and lower SUV bias and variance than for FD PET PIS.

[1]  Yang-Ming Zhu,et al.  Full-Dose PET Image Estimation from Low-Dose PET Image Using Deep Learning: a Pilot Study , 2018, Journal of Digital Imaging.

[2]  Tianyu Ma,et al.  An investigation of quantitative accuracy for deep learning based denoising in oncological PET , 2019, Physics in medicine and biology.

[3]  Maurizio Conti,et al.  Quantitative Accuracy and Lesion Detectability of Low-Dose 18F-FDG PET for Lung Cancer Screening , 2017, The Journal of Nuclear Medicine.

[4]  Konstantin Nikolaou,et al.  Towards tracer dose reduction in PET studies: Simulation of dose reduction by retrospective randomized undersampling of list-mode data. , 2016, Hellenic journal of nuclear medicine.

[5]  H. Zaidi,et al.  Advances in PET Image Reconstruction. , 2007, PET clinics.

[6]  W. Wong,et al.  The calculation of posterior distributions by data augmentation , 1987 .

[7]  Xiao Han,et al.  MR‐based synthetic CT generation using a deep convolutional neural network method , 2017, Medical physics.

[8]  F Schoonjans,et al.  MedCalc: a new computer program for medical statistics. , 1995, Computer methods and programs in biomedicine.

[9]  Division on Earth Health Risks from Exposure to Low Levels of Ionizing Radiation: BEIR VII Phase 2 , 2006 .

[10]  Qiu Huang,et al.  Enhancing the Image Quality via Transferred Deep Residual Learning of Coarse PET Sinograms , 2018, IEEE Transactions on Medical Imaging.

[11]  Mark Sandler,et al.  MobileNetV2: Inverted Residuals and Linear Bottlenecks , 2018, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[12]  Jiahe Tian,et al.  PET image denoising using unsupervised deep learning , 2019, European Journal of Nuclear Medicine and Molecular Imaging.

[13]  Dinggang Shen,et al.  Deep auto-context convolutional neural networks for standard-dose PET image estimation from low-dose PET/MRI , 2017, Neurocomputing.

[14]  Guoyan Zheng,et al.  Novel adversarial semantic structure deep learning for MRI-guided attenuation correction in brain PET/MRI , 2019, European Journal of Nuclear Medicine and Molecular Imaging.

[15]  Huafeng Liu,et al.  Low dose PET reconstruction with total variation regularization , 2014, 2014 36th Annual International Conference of the IEEE Engineering in Medicine and Biology Society.

[16]  François Chollet,et al.  Keras: The Python Deep Learning library , 2018 .

[17]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[18]  D. Townsend,et al.  Low dose positron emission tomography emulation from decimated high statistics: A clinical validation study. , 2019, Medical physics.

[19]  Chih-Chieh Liu,et al.  Higher SNR PET image prediction using a deep learning model and MRI image , 2019, Physics in medicine and biology.

[20]  Z. Jane Wang,et al.  Joint correction of attenuation and scatter in image space using deep convolutional neural networks for dedicated brain 18F-FDG PET , 2018, Physics in medicine and biology.

[21]  Dinggang Shen,et al.  3D conditional generative adversarial networks for high-quality PET image estimation at low dose , 2018, NeuroImage.

[22]  Thomas Brox,et al.  U-Net: Convolutional Networks for Biomedical Image Segmentation , 2015, MICCAI.

[23]  Greg Zaharchuk,et al.  Ultra-low-dose PET reconstruction using generative adversarial network with feature matching and task-specific perceptual loss. , 2019, Medical physics.

[24]  Habib Zaidi,et al.  Improvement of image quality in PET using post-reconstruction hybrid spatial-frequency domain filtering , 2018, Physics in medicine and biology.

[25]  Junshen Xu,et al.  200x Low-dose PET Reconstruction using Deep Learning , 2017, ArXiv.

[26]  Thomas Brox,et al.  3D U-Net: Learning Dense Volumetric Segmentation from Sparse Annotation , 2016, MICCAI.

[27]  Thomas J. Fuchs,et al.  DeepPET: A deep encoder–decoder network for directly solving the PET image reconstruction inverse problem , 2018, Medical Image Anal..

[28]  E. Topol,et al.  A comparison of deep learning performance against health-care professionals in detecting diseases from medical imaging: a systematic review and meta-analysis. , 2019, The Lancet. Digital health.

[29]  Hamed Hamid Muhammed,et al.  Comparison of pre- and post-reconstruction denoising approaches in positron emission tomography , 2016, 2016 1st International Conference on Biomedical Engineering (IBIOMED).

[30]  Yan Wang,et al.  Predicting standard-dose PET image from low-dose PET and multimodal MR images using mapping-based sparse representation , 2015, MLMI.

[31]  Bram van Ginneken,et al.  A survey on deep learning in medical image analysis , 2017, Medical Image Anal..

[32]  John M Pauly,et al.  Ultra-Low-Dose 18F-Florbetaben Amyloid PET Imaging Using Deep Learning with Multi-Contrast MRI Inputs. , 2019, Radiology.

[33]  Yaozong Gao,et al.  Prediction of standard-dose brain PET image by using MRI and low-dose brain [18F]FDG PET images. , 2015, Medical physics.

[34]  Irène Buvat,et al.  LIFEx: A Freeware for Radiomic Feature Calculation in Multimodality Imaging to Accelerate Advances in the Characterization of Tumor Heterogeneity. , 2018, Cancer research.