An MRI sequence independent convolutional neural network for synthetic head CT generation in proton therapy

A magnetic resonance imaging (MRI) sequence independent deep learning technique was developed and validated to generate synthetic computed tomography (sCT) scans for MR guided proton therapy. 47 meningioma patients previously undergoing proton therapy based on pencil beam scanning were divided into training (33), validation (6), and test (8) cohorts. T₁, T₂, and contrast enhanced T₁ (T1CM) MRI sequences were used in combination with the planning CT (pCT) data to train a 3D U-Net architecture with ResNet-Blocks. A hyperparameter search was performed including two loss functions, two group sizes of normalisation, and depth of the network. Training outcome was compared between models trained for each individual MRI sequence and for all sequences combined. The performance was evaluated based on a metric and dosimetric analysis as well as spot difference maps. Furthermore, the influence of immobilisation masks that are not visible on MRIs was investigated. Based on the hyperparameter search, the final model was trained with fixed features per group for the group normalisation, six down-convolution steps, an input size of 128 × 192 × 192, and feature loss. For the test dataset for body/bone the mean absolute error (MAE) values were on average 79.8/216.3 Houndsfield unit (HU) when trained using T1 images, 71.1/186.1 HU for T2, and 82.9/236.4 HU for T1CM. The structural similarity metric (SSIM) ranged from 0.95 to 0.98 for all sequences. The investigated dose parameters of the target structures agreed within 1% between original proton treatment plans and plans recalculated on sCTs. The spot difference maps had peaks at ±0.2 cm and for 98% of all spots the difference was less than 1 cm. A novel MRI sequence independent sCT generator was developed, which suggests that the training phase of neural networks can be disengaged from specific MRI acquisition protocols. In contrast to previous studies, the patient cohort consisted exclusively of actual proton therapy patients (i.e. “real-world data”).     

DEep learning-based rapid Spiral Image REconstruction (DESIRE) for high-resolution spiral first-pass myocardial perfusion imaging

The objective of the current study was to develop and evaluate a DEep learning-based rapid Spiral Image REconstruction (DESIRE) technique for high-resolution spiral first-pass myocardial perfusion imaging with whole-heart coverage, to provide fast and accurate image reconstruction for both single-slice (SS) and simultaneous multislice (SMS) acquisitions. Three-dimensional U-Net–based image enhancement architectures were evaluated for high-resolution spiral perfusion imaging at 3 T. The SS and SMS MB = 2 networks were trained on SS perfusion images from 156 slices from 20 subjects. Structural similarity index (SSIM), peak signal-to-noise ratio (PSNR), and normalized root mean square error (NRMSE) were assessed, and prospective images were blindly graded by two experienced cardiologists (5: excellent; 1: poor). Excellent performance was demonstrated for the proposed technique. For SS, SSIM, PSNR, and NRMSE were 0.977 [0.972, 0.982], 42.113 [40.174, 43.493] dB, and 0.102 [0.080, 0.125], respectively, for the best network. For SMS MB = 2 retrospective data, SSIM, PSNR, and NRMSE were 0.961 [0.950, 0.969], 40.834 [39.619, 42.004] dB, and 0.107 [0.086, 0.133], respectively, for the best network. The image quality scores were 4.5 [4.1, 4.8], 4.5 [4.3, 4.6], 3.5 [3.3, 4], and 3.5 [3.3, 3.8] for SS DESIRE, SS L1-SPIRiT, MB = 2 DESIRE, and MB = 2 SMS-slice-L1-SPIRiT, respectively, showing no statistically significant difference (p = 1 and p = 1 for SS and SMS, respectively) between L1-SPIRiT and the proposed DESIRE technique. The network inference time was ~100 ms per dynamic perfusion series with DESIRE, while the reconstruction time of L1-SPIRiT with GPU acceleration was ~ 30 min. It was concluded that DESIRE enabled fast and high-quality image reconstruction for both SS and SMS MB = 2 whole-heart high-resolution spiral perfusion imaging.     

Assessment of a Large-Scale Peer Learning Program’s Value by Manual Review of Case Submissions

ObjectivePrior studies used submission numbers or report addendum rates to measure peer learning programs’ (PLP) impact. We assessed the educational value of a PLP by manually reviewing cases submitted to identify factors correlating with meaningful learning opportunities (MLOs).MethodsThis institutional review board–exempted, retrospective study was performed in a large academic radiology department generating >800,000 reports annually. A PLP facilitating radiologist-to-radiologist feedback was implemented May 1, 2017, with subsequent pay-for-performance initiatives encouraging increasing submissions, >18,000 by 2019. Two radiologists blinded to submitter and receiver identity categorized 336 randomly selected submissions as a MLO, not meaningful, or equivocal, resolving disagreements in consensus review. Primary outcome was proportion of MLOs. Secondary outcomes included percent engagement by subspecialty clinical division and comparing MLO and report addendum rates via Fisher’s exact tests. We assessed association between peer learning category, pay-for-performance interventions, and subspecialty division with MLOs using logistic regression.ResultsOf 336 PLP submissions, 65.2% (219 of 336) were categorized as meaningful, 27.4% (92 of 336) not meaningful, and 7.4% (25 of 336) equivocal, with substantial reviewer agreement (86.0% [289 of 336], κ = 0.71, 95% confidence interval 0.64-0.78). MLO rate (65.2% [219 of 336]) was five times higher than addendum rate (12.9% [43 of 333]) for the cohort. MLO proportion (adjusted odds ratios 0.05-1.09) and percent engagement (0.5%-3.6%) varied between subspecialty divisions, some submitting significantly fewer MLOs (P < .01). MLO proportion did not vary between peer learning categories.ConclusionEducational value of a large-scale PLP, estimated through manual review of case submissions, is likely a more accurate measure of program impact. Incentives to enhance PLP use did not diminish the program’s educational value.     

Machine learning models for Parkinson’s disease detection and stage classification based on spatial-temporal gait parameters

BackgroundParkinson’s disease (PD) is a chronic and progressive neurodegenerative disease with no cure, presenting a challenging diagnosis and management. However, despite a significant number of criteria and guidelines have been proposed to improve the diagnosis of PD and to determine the PD stage, the gold standard for diagnosis and symptoms monitoring of PD is still mainly based on clinical evaluation, which includes several subjective factors. The use of machine learning (ML) algorithms in spatial-temporal gait parameters is an interesting advance with easy interpretation and objective factors that may assist in PD diagnostic and follow up.Research questionThis article studies ML algorithms for: i) distinguish people with PD vs. matched-healthy individuals; and ii) to discriminate PD stages, based on selected spatial-temporal parameters, including variability and asymmetry.MethodsGait data acquired from 63 people with PD with different levels of PD motor symptoms severity, and 63 matched-control group individuals, during self-selected walking speed, was study in the experiments.ResultsIn the PD diagnosis, a classification accuracy of 84.6 %, with a precision of 0.923 and a recall of 0.800, was achieved by the Naïve Bayes algorithm. We found four significant gait features in PD diagnosis: step length, velocity and width, and step width variability. As to the PD stage identification, the Random Forest outperformed the other studied ML algorithms, by reaching an Area Under the ROC curve of 0.786. We found two relevant gait features in identifying the PD stage: stride width variability and step double support time variability.SignificanceThe results showed that the studied ML algorithms have potential both to PD diagnosis and stage identification by analysing gait parameters.     

Frame Invariance and Scalability of Neural Operators for Partial Differential Equations

Partial differential equations (PDEs) play a dominant role in the mathematical modeling of many complex dynamical processes. Solving these PDEs often requires prohibitively high computational costs, especially when multiple evaluations must be made for different parameters or conditions. After training, neural operators can provide PDEs solutions significantly faster than traditional PDE solvers. In this work, invariance properties and computational complexity of two neural operators are examined for transport PDE of a scalar quantity. Neural operator based on graph kernel network (GKN) operates on graph-structured data to incorporate nonlocal dependencies. Here we propose a modified formulation of GKN to achieve frame invariance. Vector cloud neural network (VCNN) is an alternate neural operator with embedded frame invariance which operates on point cloud data. GKN-based neural operator demonstrates slightly better predictive performance compared to VCNN. However, GKN requires an excessively high computational cost that increases quadratically with the increasing number of discretized objects as compared to a linear increase for VCNN.     

Machine learning in clinical diagnosis,prognostication, and management of acute traumatic spinal cord injury (SCI): A systematic review

BackgroundMachine learning has been applied to improve diagnosis and prognostication of acute traumatic spinal cord injury. We investigate potential for clinical integration of machine learning in this patient population to navigate variability in injury and recovery.Materials and methodsWe performed a systematic review using PRISMA guidelines through PubMed database to identify studies that use machine learning algorithms for clinical application toward improvements in diagnosis, management, and predictive modeling.ResultsOf the 132 records identified, a total of 13 articles met inclusion criteria and were included in final analysis. Of the 13 articles, 5 focused on diagnostic accuracy and 8 were related to prognostication or management of traumatic spinal cord injury. Across studies, 1983 patients with spinal cord injury were evaluated with most classifying as ASIA C or D. Retrospective designs were used in 10 of 13 studies and 3 were prospective. Studies focused on MRI evaluation and segmentation for diagnostic accuracy and prognostication, investigation of mean arterial pressure in acute care and intraoperative settings, prediction of ambulatory and functional ability, chronic complication prevention, and psychological quality of life assessments. Decision tree, random forests (RF), support vector machines (SVM), hierarchical cluster tree analysis (HCTA), artificial neural networks (ANN), convolutional neural networks (CNN) machine learning subtypes were used.ConclusionsMachine learning represents a platform technology with clinical application in traumatic spinal cord injury diagnosis, prognostication, management, rehabilitation, and risk prevention of chronic complications and mental illness. SVM models showed improved accuracy when compared to other ML subtypes surveyed. Inherent variability across patients with SCI offers unique opportunity for ML and personalized medicine to drive desired outcomes and assess risks in this patient population.     

临床护理教师胜任力现状及与护生的评价比较

目的 调查临床护理教师胜任力水平及影响因素，为提高临床护理教学质量提供参考。方法 采用临床护理教师胜任力问卷、学习动机量表对重庆市11所三级医院的749名临床护理教师及458名实习护生进行调查。结果 临床护理教师胜任力自评均分为4.38±0.49,护生评价均分为4.12±0.65,护生评价总分及6个维度得分显著低于教师自评(均P<0.01);护龄、职务和学习动机是临床护理教师胜任力的影响因素(均P<0.01)。结论 临床护理教师胜任力处于较好水平，但护生评价低于教师自评；护理管理者可通过激发教师的学习动机进一步培养和提升临床护理教师的胜任力。