Performance of a deep learning‐based CT image denoising method: Generalizability over dose, reconstruction kernel, and slice thickness
dc.contributor.author | Zeng, Rongping | |
dc.contributor.author | Lin, Claire Yilin | |
dc.contributor.author | Li, Qin | |
dc.contributor.author | Jiang, Lu | |
dc.contributor.author | Skopec, Marlene | |
dc.contributor.author | Fessler, Jeffrey A. | |
dc.contributor.author | Myers, Kyle J. | |
dc.date.accessioned | 2022-03-07T03:14:05Z | |
dc.date.available | 2023-03-06 22:14:03 | en |
dc.date.available | 2022-03-07T03:14:05Z | |
dc.date.issued | 2022-02 | |
dc.identifier.citation | Zeng, Rongping; Lin, Claire Yilin; Li, Qin; Jiang, Lu; Skopec, Marlene; Fessler, Jeffrey A.; Myers, Kyle J. (2022). "Performance of a deep learning‐based CT image denoising method: Generalizability over dose, reconstruction kernel, and slice thickness." Medical Physics 49(2): 836-853. | |
dc.identifier.issn | 0094-2405 | |
dc.identifier.issn | 2473-4209 | |
dc.identifier.uri | https://hdl.handle.net/2027.42/171892 | |
dc.description.abstract | PurposeDeep learning (DL) is rapidly finding applications in low‐dose CT image denoising. While having the potential to improve the image quality (IQ) over the filtered back projection method (FBP) and produce images quickly, performance generalizability of the data‐driven DL methods is not fully understood yet. The main purpose of this work is to investigate the performance generalizability of a low‐dose CT image denoising neural network in data acquired under different scan conditions, particularly relating to these three parameters: reconstruction kernel, slice thickness, and dose (noise) level. A secondary goal is to identify any underlying data property associated with the CT scan settings that might help predict the generalizability of the denoising network.MethodsWe select the residual encoder–decoder convolutional neural network (REDCNN) as an example of a low‐dose CT image denoising technique in this work. To study how the network generalizes on the three imaging parameters, we grouped the CT volumes in the Low‐Dose Grand Challenge (LDGC) data into three pairs of training datasets according to their imaging parameters, changing only one parameter in each pair. We trained REDCNN with them to obtain six denoising models. We test each denoising model on datasets of matching and mismatching parameters with respect to its training sets regarding dose, reconstruction kernel, and slice thickness, respectively, to evaluate the denoising performance changes. Denoising performances are evaluated on patient scans, simulated phantom scans, and physical phantom scans using IQ metrics including mean‐squared error (MSE), contrast‐dependent modulation transfer function (MTF), pixel‐level noise power spectrum (pNPS), and low‐contrast lesion detectability (LCD).ResultsREDCNN had larger MSE when the testing data were different from the training data in reconstruction kernel, but no significant MSE difference when varying slice thickness in the testing data. REDCNN trained with quarter‐dose data had slightly worse MSE in denoising higher‐dose images than that trained with mixed‐dose data (17%–80%). The MTF tests showed that REDCNN trained with the two reconstruction kernels and slice thicknesses yielded images of similar image resolution. However, REDCNN trained with mixed‐dose data preserved the low‐contrast resolution better compared to REDCNN trained with quarter‐dose data. In the pNPS test, it was found that REDCNN trained with smooth‐kernel data could not remove high‐frequency noise in the test data of sharp kernel, possibly because the lack of high‐frequency noise in the smooth‐kernel data limited the ability of the trained model in removing high‐frequency noise. Finally, in the LCD test, REDCNN improved the lesion detectability over the original FBP images regardless of whether the training and testing data had matching reconstruction kernels.ConclusionsREDCNN is observed to be poorly generalizable between reconstruction kernels, more robust in denoising data of arbitrary dose levels when trained with mixed‐dose data, and not highly sensitive to slice thickness. It is known that reconstruction kernel affects the in‐plane pNPS shape of a CT image, whereas slice thickness and dose level do not, so it is possible that the generalizability performance of this CT image denoising network highly correlates to the pNPS similarity between the testing and training data. | |
dc.publisher | Wiley Periodicals, Inc. | |
dc.publisher | US Food and Drug Administration | |
dc.subject.other | CT image denoising | |
dc.subject.other | deep learning | |
dc.subject.other | generalizability performance | |
dc.subject.other | image quality assessment | |
dc.title | Performance of a deep learning‐based CT image denoising method: Generalizability over dose, reconstruction kernel, and slice thickness | |
dc.type | Article | |
dc.rights.robots | IndexNoFollow | |
dc.subject.hlbsecondlevel | Medicine (General) | |
dc.subject.hlbtoplevel | Health Sciences | |
dc.description.peerreviewed | Peer Reviewed | |
dc.description.bitstreamurl | http://deepblue.lib.umich.edu/bitstream/2027.42/171892/1/mp15430_am.pdf | |
dc.description.bitstreamurl | http://deepblue.lib.umich.edu/bitstream/2027.42/171892/2/mp15430.pdf | |
dc.identifier.doi | 10.1002/mp.15430 | |
dc.identifier.source | Medical Physics | |
dc.identifier.citedreference | Zhang K, Zuo W, Chen Y, Meng D, Zhang L. Beyond a Gaussian denoiser: residual learning of deep CNN for image denoising. IEEE Trans Image Process. 2017; 26 ( 7 ): 3142 ‐ 3155. | |
dc.identifier.citedreference | Qi LP, Li Y, Tang L, et al. Evaluation of dose reduction and image quality in chest CT using adaptive statistical iterative reconstruction with the same group of patients. Br J Radiol. 2012; 85 ( 1018 ): e906 ‐ e911. | |
dc.identifier.citedreference | Chen H, Zhang Y, Kalra MK, et al. Low‐dose CT with a residual encoder‐decoder convolutional neural network. IEEE Trans Med Imaging. 2017; 36 ( 12 ): 2524 ‐ 2535. | |
dc.identifier.citedreference | Chen H, Zhang Y, Zhang W, et al. Low‐dose CT via convolutional neural network. Biomed Opt Express. 2017; 8 ( 2 ): 679 ‐ 694. | |
dc.identifier.citedreference | Yang Q, Yan P, Zhang Y, et al. Low‐dose CT image denoising using a generative adversarial network with Wasserstein distance and perceptual loss. IEEE Trans Med Imaging. 2018; 37 ( 6 ): 1348 ‐ 1357. | |
dc.identifier.citedreference | Kim B, Han M, Shim H, Baek J. A performance comparison of convolutional neural network‐based image denoising methods: the effect of loss functions on low‐dose CT images. Med Phys. 2019; 46 ( 9 ): 3906 ‐ 3923. | |
dc.identifier.citedreference | MacDougall RD, Zhang Y, Callahan MJ, et al. Improving low‐dose pediatric abdominal CT by using convolutional neural networks. Radiology. 2019; 1 ( 6 ): e180087. | |
dc.identifier.citedreference | Yang W, Zhang H, Yang J, et al. Improving low‐dose CT image using residual convolutional network. IEEE Access. 2017; 5: 24698 ‐ 24705. | |
dc.identifier.citedreference | Solomon J, Lyu P, Marin D, Samei E. Noise and spatial resolution properties of a commercially available deep learning‐based CT reconstruction algorithm. Med Phys. 2020; 47 ( 9 ): 3961 – 3971. | |
dc.identifier.citedreference | Lenfant M, Chevallier O, Comby P‐O, et al. Deep learning versus iterative reconstruction for CT pulmonary angiography in the emergency setting: improved image quality and reduced radiation dose. Diagnostics (Basel, Switzerland). 2020; 10 ( 8 ): 558. | |
dc.identifier.citedreference | Kawashima H, Ichikawa K, Takata T, et al. Performance of clinically available deep learning image reconstruction in computed tomography: a phantom study. J Med Imaging (Bellingham). 2020; 7 ( 6 ): 063503. | |
dc.identifier.citedreference | USFDA. Executive Summary for the Patient Engagement Advisory Committee Meeting: Artificial Intelligence (AI) and Machine Learning (ML) in Medical Devices. US Food and Drug Administration. Accessed October 22, 2020. https://www.fda.gov/media/142998/download. | |
dc.identifier.citedreference | McCollough CH, Bartley AC, Carter RE, et al. Low‐dose CT for the detection and classification of metastatic liver lesions: results of the 2016 Low Dose CT Grand Challenge. Med Phys. 2017; 44 ( 10 ): e339 ‐ e352. | |
dc.identifier.citedreference | Zeng R, Lin CY, Li Q, Jiang L, Fessler JA, Myers KJ. Generalizability test of a deep learning‐based CT image denoising method. The 6th International Conference on Image Formation in X‐Ray Computed Tomography, 2020. | |
dc.identifier.citedreference | Huber NR, Missert AD, Yu L, Leng S, McCollough CH. Evaluating a convolutional neural network noise reduction method when applied to CT images reconstructed differently than training data. J Comput Assist Tomogr. 2021; 45 ( 4 ): 544 ‐ 551. | |
dc.identifier.citedreference | He K, Zhang X, Ren S, Sun J. Deep residual learning for image recognition. IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 2016; pp. 770 ‐ 778. | |
dc.identifier.citedreference | Ronneberger O, Fischer P, Brox T. U‐Net: convolutional Networks for Biomedical Image Segmentation. Medical Image Computing and Computer‐Assisted Intervention – MICCAI. 2015: 234 ‐ 241. | |
dc.identifier.citedreference | Goodfellow I, Pouget‐Abadie J, Mirza M, et al. Generative adversarial nets. Proceedings of the International Conference on Neural Information Processing Systems. 2014: 2672 – 2680. | |
dc.identifier.citedreference | Zeng R, Divel S, Li Q, Myers KJ. Performance evaluation of deep learning methods applied to CT image reconstruction. Med Phys. 2019;46: E162 ‐ E162. | |
dc.identifier.citedreference | Richard S, Husarik DB, Yadava G, Murphy SN, Samei E. Towards task‐based assessment of CT performance: system and object MTF across different reconstruction algorithms. Med Phys. 2012; 39 ( 7 ): 4115 ‐ 4122. | |
dc.identifier.citedreference | Vaishnav JY, Jung WC, Popescu LM, Zeng R, Myers KJ. Objective assessment of image quality and dose reduction in CT iterative reconstruction. Med Phys. 2014; 41 ( 7 ): 071904. | |
dc.identifier.citedreference | Goldman LW. Principles of CT: multislice CT. J Nucl Med Technol. 2008; 36 ( 2 ): 57 ‐ 68. | |
dc.identifier.citedreference | Goodfellow I, Bengio Y, Courville A. DeepLearning. The MIT Press; 2016. | |
dc.identifier.citedreference | Shorten C, Khoshgoftaar TM. A survey on image data augmentation for deep learning. J Big Data. 2019; 6 ( 1 ): 60. | |
dc.identifier.citedreference | Prabhat KcP, Zeng R, Farhangi MM, Myers KJ. Deep neural networks‐based denoising models for CT imaging and their efficacy. Proc SPIE 2021: 11595. | |
dc.identifier.citedreference | Zeng R, Petrick N, Gavrielides MA, Myers KJ. Approximations of noise covariance in multi‐slice helical CT scans: impact on lung nodule size estimation. Phys Med Biol. 2011; 56 ( 19 ): 6223 ‐ 6242. | |
dc.identifier.citedreference | Li K, Tang J, Chen G‐H. Statistical model based iterative reconstruction (MBIR) in clinical CT systems: experimental assessment of noise performance. Med Phys. 2014; 41 ( 4 ): 041906 ‐ 041906. | |
dc.identifier.citedreference | Söderberg M, Gunnarsson M. Automatic exposure control in computed tomography—an evaluation of systems from different manufacturers. Acta Radiol. 2010; 51 ( 6 ): 625 ‐ 634. | |
dc.identifier.citedreference | Huck SM, Fung GSK, Parodi K, Stierstorfer K. The z‐sbDBA, a new concept for a dynamic sheet‐based fluence field modulator in X‐ray CT. Med Phys. 2020; 47 ( 10 ): 4827 ‐ 4837. | |
dc.identifier.citedreference | Elbakri IA, Fessler JA. Statistical image reconstruction for polyenergetic X‐ray computed tomography. IEEE Trans Med Imaging. 2002; 21 ( 2 ): 89 ‐ 99. | |
dc.working.doi | NO | en |
dc.owningcollname | Interdisciplinary and Peer-Reviewed |
Files in this item
Remediation of Harmful Language
The University of Michigan Library aims to describe library materials in a way that respects the people and communities who create, use, and are represented in our collections. Report harmful or offensive language in catalog records, finding aids, or elsewhere in our collections anonymously through our metadata feedback form. More information at Remediation of Harmful Language.
Accessibility
If you are unable to use this file in its current format, please select the Contact Us link and we can modify it to make it more accessible to you.