Dataset Information

Machine Friendly Machine Learning: Interpretation of Computed Tomography Without Image Reconstruction.

ABSTRACT: Recent advancements in deep learning for automated image processing and classification have accelerated many new applications for medical image analysis. However, most deep learning algorithms have been developed using reconstructed, human-interpretable medical images. While image reconstruction from raw sensor data is required for the creation of medical images, the reconstruction process only uses a partial representation of all the data acquired. Here, we report the development of a system to directly process raw computed tomography (CT) data in sinogram-space, bypassing the intermediary step of image reconstruction. Two classification tasks were evaluated for their feasibility of sinogram-space machine learning: body region identification and intracranial hemorrhage (ICH) detection. Our proposed SinoNet, a convolutional neural network optimized for interpreting sinograms, performed favorably compared to conventional reconstructed image-space-based systems for both tasks, regardless of scanning geometries in terms of projections or detectors. Further, SinoNet performed significantly better when using sparsely sampled sinograms than conventional networks operating in image-space. As a result, sinogram-space algorithms could be used in field settings for triage (presence of ICH), especially where low radiation dose is desired. These findings also demonstrate another strength of deep learning where it can analyze and interpret sinograms that are virtually impossible for human experts.

SUBMITTER: Lee H

PROVIDER: S-EPMC6820559 | biostudies-literature | 2019 Oct

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Machine Friendly Machine Learning: Interpretation of Computed Tomography Without Image Reconstruction.

Lee Hyunkwang H Huang Chao C Yune Sehyo S Tajmir Shahein H SH Kim Myeongchan M Do Synho S

Scientific reports 20191029 1

Recent advancements in deep learning for automated image processing and classification have accelerated many new applications for medical image analysis. However, most deep learning algorithms have been developed using reconstructed, human-interpretable medical images. While image reconstruction from raw sensor data is required for the creation of medical images, the reconstruction process only uses a partial representation of all the data acquired. Here, we report the development of a system to ...[more]

PMID: 31664075

Similar Datasets

Project description:ObjectivesDeep-learning image reconstruction (DLIR) offers unique opportunities for reducing image noise without degrading image quality or diagnostic accuracy in coronary CT angiography (CCTA). The present study aimed at exploiting the capabilities of DLIR to reduce radiation dose and assess its impact on stenosis severity, plaque composition analysis, and plaque volume quantification.MethodsThis prospective study includes 50 patients who underwent two sequential CCTA scans at normal-dose (ND) and lower-dose (LD). ND scans were reconstructed with Adaptive Statistical Iterative Reconstruction-Veo (ASiR-V) 100%, and LD scans with DLIR. Image noise (in Hounsfield units, HU) and quantitative plaque volumes (in mm3) were assessed quantitatively. Stenosis severity was visually categorized into no stenosis (0%), stenosis (< 20%, 20-50%, 51-70%, 71-90%, 91-99%), and occlusion (100%). Plaque composition was classified as calcified, non-calcified, or mixed.ResultsReduction of radiation dose from ND scans with ASiR-V 100% to LD scans with DLIR at the highest level (DLIR-H; 1.4 mSv vs. 0.8 mSv, p < 0.001) had no impact on image noise (28 vs. 27 HU, p = 0.598). Reliability of stenosis severity and plaque composition was excellent between ND scans with ASiR-V 100% and LD scans with DLIR-H (intraclass correlation coefficients of 0.995 and 0.974, respectively). Comparison of plaque volumes using Bland-Altman analysis revealed a mean difference of - 0.8 mm3 (± 2.5 mm3) and limits of agreement between - 5.8 and + 4.1 mm3.ConclusionDLIR enables a reduction in radiation dose from CCTA by 43% without significant impact on image noise, stenosis severity, plaque composition, and quantitative plaque volume.Key points•Deep-learning image reconstruction (DLIR) enables radiation dose reduction by over 40% for coronary computed tomography angiography (CCTA). •Image noise remains unchanged between a normal-dose CCTA reconstructed by ASiR-V and a lower-dose CCTA reconstructed by DLIR. •There is no impact on the assessment of stenosis severity, plaque composition, and quantitative plaque volume between the two scans.

Project description:BackgroundMulti-energy computed tomography (CT) provides multiple channel-wise reconstructed images, and they can be used for material identification and k-edge imaging. Nonetheless, the projection datasets are frequently corrupted by various noises (e.g., electronic, Poisson) in the acquisition process, resulting in lower signal-noise-ratio (SNR) measurements. Multi-energy CT images have local sparsity, nonlocal self-similarity in spatial dimension, and correlation in spectral dimension.MethodsIn this paper, we propose an image-spectral decomposition extended-learning assisted by sparsity (IDEAS) method to fully exploit these intrinsic priors for multi-energy CT image reconstruction. Particularly, a nonlocal low-rank Tucker decomposition (TD) is employed to utilize the correlation and nonlocal self-similarity priors. Moreover, considering the advantages of multi-task tensor dictionary learning (TDL) in sparse representation, an adaptive spatial dictionary and an adaptive spectral dictionary are trained during the iterative reconstruction process. Furthermore, a weighted total variation (TV) regularization term is employed to encourage local sparsity.ResultsNumerical simulation, physical phantom, and preclinical mouse experiments are performed to validate the proposed IDEAS algorithm. Specifically, in the simulation experiments, the proposed IDEAS reconstructed high-quality images that are very close to the references. For example, the root mean square error (RMSE) of IDEAS image in energy bin 1 is as low as 0.0672, while the RMSE of other methods are higher than 0.0843. Besides, the structural similarity (SSIM) of IDEAS reconstructed image in energy bin 1 is greater than 0.98. For material decomposition, the RMSE of IDEAS bone component is as low as 0.0152, and other methods are higher than 0.0199. In addition, the computational cost of IDEAS is as low as 98.8 s for one iteration, and the competing tensor decomposition method is higher than 327 s.ConclusionsTo further improve the quality of the reconstructed multi-energy CT images, multiple prior regularizations are introduced to the multi-energy CT reconstructed model, leading to an IDEAS method. Both qualitative and quantitative evaluation of our results confirm the outstanding performance of the proposed algorithm compared to the state-of-the-arts.

Project description:In medical and industrial applications of computed tomography (CT) imaging, limited by the scanning environment and the risk of excessive X-ray radiation exposure imposed to the patients, reconstructing high quality CT images from limited projection data has become a hot topic. X-ray imaging in limited scanning angular range is an effective imaging modality to reduce the radiation dose to the patients. As the projection data available in this modality are incomplete, limited-angle CT image reconstruction is actually an ill-posed inverse problem. To solve the problem, image reconstructed by conventional filtered back projection (FBP) algorithm frequently results in conspicuous streak artifacts and gradual changed artifacts nearby edges. Image reconstruction based on total variation minimization (TVM) can significantly reduce streak artifacts in few-view CT, but it suffers from the gradual changed artifacts nearby edges in limited-angle CT. To suppress this kind of artifacts, we develop an image reconstruction algorithm based on ?0 gradient minimization for limited-angle CT in this paper. The ?0-norm of the image gradient is taken as the regularization function in the framework of developed reconstruction model. We transformed the optimization problem into a few optimization sub-problems and then, solved these sub-problems in the manner of alternating iteration. Numerical experiments are performed to validate the efficiency and the feasibility of the developed algorithm. From the statistical analysis results of the performance evaluations peak signal-to-noise ratio (PSNR) and normalized root mean square distance (NRMSD), it shows that there are significant statistical differences between different algorithms from different scanning angular ranges (p<0.0001). From the experimental results, it also indicates that the developed algorithm outperforms classical reconstruction algorithms in suppressing the streak artifacts and the gradual changed artifacts nearby edges simultaneously.

Project description:BackgroundHead and neck computed tomography angiography (CTA) technology has become the noninvasive imaging method of choice for the diagnosis and long-term follow-up of vascular lesions of the head and neck. However, issues of radiation safety and contrast nephropathy associated with CTA examinations remain concerns. In recent years, deep learning image reconstruction (DLIR) algorithms have been increasingly used in clinical studies, demonstrating their potential for dose optimization. This study aimed to investigate the value of using a DLIR algorithm to reduce radiation and contrast doses in head and neck CTA.MethodsA total of 100 patients were prospectively enrolled and randomly divided into two groups. Group A (50 patients) consisted of those who underwent 70-kVp CTA with a low contrast volume and injection rate and who were classified according to the reconstruction algorithm into subgroups A1 [DLIR at high weighting (DLIR-H)], A2 [DLIR at low weighting (DLIR-L)], and A3 [volume-based adaptive statistical iterative reconstruction with 50% weighting (ASIR-V50%)]. Meanwhile, group B (50 patients) consisted of those who underwent standard radiation and contrast doses at 100 kVp with ASIR-V50% reconstruction. The computed tomography (CT) attenuation, background noise, signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), and subjective image quality score (SIQS) were statistically compared for several vessels among the four groups.ResultsGroup A showed significant reductions in contrast dosage, injection rate, and radiation dose of 36.09%, 20.88%, and 47.80%, respectively, compared to group B (all P<0.001). The four groups differed significantly in terms of background noise (all P<0.05) with group A1 having the lowest value. Group A1 also had significantly higher SNR and CNR values compared to group B in all vessels (all P<0.05) except the M1 of the middle cerebral artery for the SNR. Group A1 also had the highest SIQS, followed by the A2, B, and A3 groups. The SIQS showed good agreement between the two reviewers in all groups, with κ values between 0.88 and 1.ConclusionsCompared to the standard-dose protocol using 100 kVp and ASIR-V50%, a protocol of 70 kVp combined with DLIR-H significantly reduces the radiation dose, contrast dose, and injection rate in head and neck CTA while still significantly improving image quality for patients with a standard body size.

Project description:BackgroundComputed tomography perfusion (CTP) and computed tomography angiography (CTA) are valuable tools for diagnosing acute ischemic stroke (AIS). It is essential to obtain high-quality CTP and CTA images in short time. This study aimed to evaluate the image quality and diagnostic performance of brain CTP and CTA images generated from CTP reconstructed by a deep learning image reconstruction (DLIR) algorithm on patients with AIS.MethodsThe study prospectively enrolled 54 patients with suspected AIS undergoing non-contrast CT and CTP within 24 hours. CTP datasets were reconstructed with three levels of adaptive statistical iterative reconstruction-Veo algorithm [ASIR-V 0% with filtered back projection (FBP), ASIR-V 40%, and ASIR-V 80%] and three levels of DLIR, including low (DLIR-L), medium (DLIR-M), and high (DLIR-H). CTA images were generated using the CTP datasets at the peak arterial phase. Objective parameters including signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), and noise reduction rate. Subjective evaluation was assessed according to Abels scoring system. Perfusion parameters and detection accuracy for infarction core lesions were evaluated. The objective and subjective image quality of CTA images were also evaluated.ResultsAll reconstructions produced similar CT values (P>0.05). With the increase of ASIR-V and DLIR reconstruction strength, image noise decreased, while SNR and CNR increased for CTP images, especially in white matter. DLIR-H, DLIR-M, and ASIR-V80% yielded higher subjective scores than did ASIR-V40% and FBP. DLIR-H provided the highest noise reduction rate and detection accuracy. No significant difference was found in conventional parameters, the volume of infarct core, or ischemic penumbra among the 6 groups (P>0.05). The objective evaluation of reconstructed CTA images was comparable in DLIR-H, DLIR-M, and ASIR-V80% (P>0.05). The subjective scores of the DLIR-H and DLIR-M images were higher than those of the other groups, especially ASIR-V40% and FBP (P<0.05).ConclusionsCompared with FBP and ASIR-V40%, DLIR-H, DLIR-M, and ASIR-V80% improved the overall image quality of CTP and CTA images to varying degrees. Furthermore, DLIR-H and DLIR-M showed the best performance. DLIR-H is the best choice in diagnosing AIS with improved detection accuracy for cerebral infarction. Reconstructing CTA images using CTP datasets could reduce contrast agent and radiation dose.

Dataset Information

Machine Friendly Machine Learning: Interpretation of Computed Tomography Without Image Reconstruction.

Publications

Machine Friendly Machine Learning: Interpretation of Computed Tomography Without Image Reconstruction.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets