Image-based Metrics Research Articles

Rapid and Precise Computation of Fractional Flow Reserve from Routine Two-Dimensional Coronary Angiograms Based on Fluid Mechanics: The Pilot FFR2D Study.

Objective: To present a novel pipeline for rapid and precise computation of fractional flow reserve from an analysis of routine two-dimensional coronary angiograms based on fluid mechanics equations (FFR2D). Material and methods: This was a pilot analytical study that was designed to assess the diagnostic performance of FFR2D versus the gold standard of FFR (threshold ≤ 0.80) measured with a pressure wire for the physiological assessment of intermediate coronary artery stenoses. In a single academic center, consecutive patients referred for diagnostic coronary angiography and potential revascularization between 1 September 2020 and 1 September 2022 were screened for eligibility. Routine two-dimensional angiograms at optimal viewing angles with minimal overlap and/or foreshortening were segmented semi-automatically to derive the vascular geometry of intermediate coronary lesions, and nonlinear pressure-flow mathematical relationships were applied to compute FFR2D. Results: Some 88 consecutive patients with a single intermediate coronary artery lesion were analyzed (LAD n = 74, RCA n = 9 and LCX n = 5; percent diameter stenosis of 45.7 ± 11.0%). The computed FFR2D was on average 0.821 ± 0.048 and correlated well with invasive FFR (r = 0.68, p < 0.001). There was very good agreement between FFR2D and invasive-wire FFR with minimal measurement bias (mean difference: 0.000 ± 0.048). The overall accuracy of FFR2D for diagnosing a critical epicardial artery stenosis was 90.9% (80 cases classified correctly out of 88 in total). FFR2D identified 24 true positives, 56 true negatives, 4 false positives, and 4 false negatives and predicted FFR ≤ 0.80 with a sensitivity of 85.7%, specificity of 93.3%, positive likelihood ratio of 13.0, and negative likelihood ratio of 0.15. FFR2D had a significantly better discriminatory capacity (area under the ROC curve: 0.95 [95% CI: 0.91-0.99]) compared to 50%DS on 2D-QCA (area under the ROC curve: 0.70 [95% CI: 0.59-0.82]; p = 0.0001) in predicting wire FFR ≤ 0.80. The median time of image analysis was 2 min and the median time of computation of the FFR2D results was 0.1 s. Conclusion: FFR2D may rapidly derive a precise image-based metric of fractional flow reserve with high diagnostic accuracy based on a single two-dimensional coronary angiogram.

Neural-network based high-speed volumetric dynamic optical coherence tomography

We demonstrate deep-learning neural network (NN)-based dynamic optical coherence tomography (DOCT), which generates high-quality logarithmic-intensity-variance (LIV) DOCT images from only four OCT frames. The NN model is trained for tumor spheroid samples using a customized loss function: the weighted mean absolute error. This loss function enables highly accurate LIV image generation. The fidelity of the generated LIV images to the ground truth LIV images generated using 32 OCT frames is examined via subjective image observation and statistical analysis of image-based metrics. Fast volumetric DOCT imaging with an acquisition time of 6.55 s/volume is demonstrated using this NN-based method.

Assessment of visual landscape quality of urban green spaces using image-based metrics derived from perceived sensory dimensions

Urban green spaces (UGS) have been extensively studied for their psychological benefits to people. However, accurately assessing the quality of UGS, which is closely linked to human perception, remains challenging due to the complexity of this attribute. To address this gap, this study proposed a novel approach that uses comprehensive image-based metrics to evaluate multiple aspects of visual landscape quality (VLQ). These metrics were developed based on a theoretical framework of perceived sensory dimensions (PSD). We presented the meanings and calculation methods of metrics based on panoramic images. Additionally, we conducted an online survey to demonstrate the significant correlation between these metrics and the subjective perception of PSD. This study also proposed a new spatial mapping method to represent VLQ using these metrics, illustrated through a case study of the Singapore Botanic Gardens (SBG). We used analysis of variance to compare the distinctions between metrics naturalness, wildness, and GVI by selecting a few green spaces within the SBG. The approaches provide an innovative way to evaluate UGS visual quality using photographic imagery data, contributing to a more comprehensive understanding of the visual quality assessment of UGS.

Response Evaluation of Neoadjuvant Therapies in Sarcoma

Sarcoma is a complex and heterogeneous disease with a rapidly evolving treatment landscape. With a growing emphasis on neoadjuvant therapy as a way to improve surgical and oncologic outcomes, our approach to monitor treatment efficacy must also continue to evolve. This is paramount to both clinical trial design, where endpoints must accurately reflect disease outcomes, and individual patient, whose treatment response informs therapeutic decisions. In the era of personalized medicine, the response to neoadjuvant treatment in sarcoma remains most effectively gauged by pathologic review following surgical resection. Although measures of pathologic complete response most effectively predict outcome, the requisite surgical excision precludes their use in real-time monitoring of neoadjuvant treatment response. Current image-based metrics such as RECIST and PERCIST have been utilized in many trials; however, they are limited by their unilateral measurement approach. More effective tools are needed to better measure the response to therapy prior to neoadjuvant regimen completion, so that the medication or regimen may be best tailored to patient response in an ongoing fashion. Delta-radiomics and circulating tumor DNA (ctDNA) represent promising novel tools for real-time monitoring of treatment efficacy. These metrics have been shown to predict pathologic complete response and disease progression at a superior level to traditional CT-based guidelines. Delta-radiomics is currently being utilized in a clinical trial among soft tissue sarcoma patients in which radiation dosage is adjusted based on radiomic data. The ability of ctDNA to detect molecular residual disease is also under study in multiple clinical trials, although none in the field of sarcoma. Future directions in the field include the use of ctDNA and molecular residual disease testing among sarcoma patients, as well as increased utilization of delta-radiomics, to more effectively monitor neoadjuvant treatment response prior to surgical resection.

Physics-Informed Generative Adversarial Network-Based Modeling and Simulation of Linear Electric Machines

The demand for fast magnetic field approximation for the optimal design of electromagnetic devices is urgent nowadays. However, due to the lack of a publicly available dataset and the unclear definition of each parameter in the magnetic field dataset, the expansion of data-driven magnetic field approximation is severely limited. This study presents a physics-informed generative adversarial network (PIGAN), as well as a permanent magnet linear synchronous motor (PMLSM)-based magnetic field dataset, for fast magnetic field approximation. It includes the current density, material distribution, electromagnetic material properties, and other parameters of the electric machine. Physics-informed loss functions are utilized in the training process, making the output governed by Maxwell’s equation. Different slot-pole combinations of the PMLSM are involved in the dataset to extend the generalization of PIGAN. Some indicators for the further evaluation of magnetic approximation performance, including image-based metrics and calculation methods for the performance of electric motors, are presented in this study. Some challenges of magnetic field approximation using PIGAN are also discussed. The effectiveness of the physics-informed method is verified by comparing the magnetic field approximation results and the performance analysis results of the PMLSM with FEM, and the speed of PIGAN is approximately 40 times faster than that of FEM, while the accuracy is similar.

Pigment Penetration Characterization of Colored Boundaries in Powder-Based Color 3D Printing.

Color 3D printing has widely affected our daily lives; therefore, its precise control is essential for aesthetics and performance. In this study, four unique test plates were printed using powder-based full-color 3D printing as an example; moreover, the corresponding pigment-penetration depth, chromaticity value and image-based metrics were measured to investigate the lateral pigment penetration characteristics and relative surface-color reproduction of each color patch, and to perform an objective analysis with specific microscopic images. The results show that the lateral pigment-penetration depth correlates with the number of printed layers on the designed 3D test plates, and the qualitative analysis of microscopic images can explain the change in chromaticity well. Meanwhile, there is an obvious linear correlation between the mean structural similarity, color-image difference and color difference for current color samples. Thus, our proposed approach has a good practicality for powder-based color 3D printing, and can provide new insight into predicting the color-presentation efficiency of color 3D-printed substrates by the abovementioned objective metrics.

Siamese-SR: A Siamese Super-Resolution Model for Boosting Resolution of Digital Rock Images for Improved Petrophysical Property Estimation

Digital Rock Physics leverages advances in digital image acquisition and analysis techniques to create 3D digital images of rock samples, which are used for computational modeling and simulations to predict petrophysical properties of interest. However, the accuracy of the predictions is crucially dependent on the quality of the digital images, which is currently limited by the resolution of the micro-CT scanning technology. We have proposed a novel Deep Learning based Super-Resolution model called Siamese-SR to digitally boost the resolution of Digital Rock images whilst retaining the texture and providing optimal de-noising. The Siamese-SR model consists of a generator which is adversarially trained with a relativistic and a siamese discriminator utilizing Materials In Context (MINC) loss estimator. This model has been demonstrated to improve the resolution of sandstone rock images acquired using micro-CT scanning by a factor of 2. Another key highlight of our work is that for the evaluation of the super-resolution performance, we propose to move away from image-based metrics such as Structural Similarity (SSIM) and Peak Signal to Noise Ratio (PSNR) because they do not correlate well with expert geological and petrophysical evaluations. Instead, we propose to subject the super-resolved images to the next step in the Digital Rock workflow to calculate a crucial petrophysical property of interest, viz. porosity and use it as a metric for evaluation of our proposed Siamese-SR model against several other existing super-resolution methods like SRGAN, ESRGAN, EDSR and SPSR. Furthermore, we also use Local Attribution Maps to show how our proposed Siamese-SR model focuses optimally on edge-semantics, which is what leads to improvement in the image-based porosity prediction, the permeability prediction from Multiple Relaxation Time Lattice Boltzmann Method (MRTLBM) flow simulations as well as the prediction of other petrophysical properties of interest derived from Mercury Injection Capillary Pressure (MICP) simulations.

Model for predicting perception of facial action unit activation using virtual humans

Blendshape facial rigs are used extensively in the industry for facial animation of virtual humans. However, storing and manipulating large numbers of facial meshes (blendshapes) is costly in terms of memory and computation for gaming applications. Blendshape rigs are comprised of sets of semantically-meaningful expressions, which govern how expressive the character will be, often based on Action Units from the Facial Action Coding System (FACS). However, the relative perceptual importance of blendshapes has not yet been investigated. Research in Psychology and Neuroscience has shown that our brains process faces differently than other objects so we postulate that the perception of facial expressions will be feature-dependent rather than based purely on the amount of movement required to make the expression. Therefore, we believe that perception of blendshape visibility will not be reliably predicted by numerical calculations of the difference between the expression and the neutral mesh. In this paper, we explore the noticeability of blendshapes under different activation levels, and present new perceptually-based models to predict perceptual importance of blendshapes. The models predict visibility based on commonly-used geometry and image-based metrics.

Advanced Surface Color Quality Assessment in Paper-Based Full-Color 3D Printing.

Color 3D printing allows for 3D-printed parts to represent 3D objects more realistically, but its surface color quality evaluation lacks comprehensive objective verification considering printing materials. In this study, a unique test model was designed and printed using eco-friendly and vivid paper-based full-color 3D printing as an example. By measuring the chromaticity, roughness, glossiness, and whiteness properties of 3D-printed surfaces and by acquiring images of their main viewing surfaces, this work skillfully explores the correlation between the color representation of a paper-based 3D-printed coloring layer and its attached underneath blank layer. Quantitative analysis was performed using ΔE*ab, feature similarity index measure of color image (FSIMc), and improved color-image-difference (iCID) values. The experimental results show that a color difference on color-printed surfaces exhibits a high linear correlation trend with its FSIMc metric and iCID metric. The qualitative analysis of microscopic imaging and the quantitative analysis of the above three surface properties corroborate the prediction of the linear correlation between color difference and image-based metrics. This study can provide inspiration for the development of computational coloring materials for additive manufacturing.

Multiscale In Vivo Imaging of Collective Cell Migration in Drosophila Embryos.

Coordinated cell movements drive embryonic development and tissue repair, and can also spread disease. Time-lapse microscopy is an integral part in the study of the cell biology of collective cell movements. Advances in imaging techniques enable monitoring dynamic cellular and molecular events in real time within living animals. Here, we demonstrate the use of spinning disk confocal microscopy to investigate coordinated cell movements and epithelial-to-mesenchymal-like transitions during embryonic wound closure in Drosophila. We describe image-based metrics to quantify the efficiency of collective cell migration. Finally, we show the application of super-resolution radial fluctuation microscopy to obtain multidimensional, super-resolution images of protrusive activity in collectively moving cells in vivo. Together, the methods presented here constitute a toolkit for the modern analysis of collective cell migration in living animals.

Discriminant analysis for lodging severity classification in wheat using RADARSAT-2 and Sentinel-1 data

Crop lodging - the bending of crop stems from their upright position or the failure of root-soil anchorage systems - is a major yield-reducing factor in wheat and causes deterioration of grain quality. The severity of lodging can be measured by a lodging score (LS)- an index calculated from the crop angle of inclination (CAI) and crop lodged area (LA). LS is difficult and time consuming to measure manually meaning that information on lodging occurrence and severity is limited and sparse. Remote sensing-based estimates of LS can provide more timely, synoptic and reliable information on crop lodging across vast areas. This information could improve estimates of crop yield losses, inform insurance loss adjusters and influence management decisions for subsequent seasons.This research - conducted in the 600 ha wheat sown area in the Bonifiche Ferraresi farm, located in Jolanda di Savoia, Ferrara, Italy - evaluated the performance of RADARSAT-2 and Sentinel-1 data to discriminate and classify lodging severity based on field measured LS. We measured temporal crop status characteristics related to lodging (e.g. lodged area, CAI, crop height) and collected relevant meteorological data (wind speed and rainfall) throughout May-June 2018. These field measurements were used to distinguish healthy (He) wheat from lodged wheat with different degrees of lodging severity (moderate, severe and very severe). We acquired multi-incidence angle (FQ8-27° and FQ21-41°) RADARSAT-2 and Sentinel-1 (40°) images and derived multiple metrics from them to discriminate and classify lodging severity. As a part of our data exploration, we performed a correlation analysis between the image-based metrics and LS. Next, a multi-temporal discriminant analysis approach, including a partial least squares (PLS-DA) method, was developed to classify lodging severities. We used the area under the curve-receiver operating characteristics (AUC-ROC) and confusion matrices to evaluate the accuracy of the PLS-DA classification models.Results show that (1) volume scattering components were highly correlated with LS at low incidence angles while double and surface scattering was more prevalent at high incidence angles; (2) lodging severity was best classified using low incidence angle R-FQ8 data (overall accuracy 72%) and (3) the Sentinel-1 data-based classification model was able to correctly identify 60% of the lodging severity cases in the study site. The results from this first study on classifying lodging severity using satellite-based SAR platforms suggests that SAR-based metrics can capture a substantial proportion of the observed variation in lodging severity, which is important in the context of operational crop lodging assessment in particular, and sustainable agriculture in general.

Boundaries Extend and Contract in Scene Memory Depending on Image Properties.

Boundary extension, a memory distortion in which observers consistently recall a scene with visual information beyond its boundaries, is widely accepted across the psychological sciences as a phenomenon revealing fundamental insight into memory representations [1-3], robust across paradigms [1, 4] and age groups [5-7]. This phenomenon has been taken to suggest that the mental representation of a scene consists of an intermingling of sensory information and a schema that extrapolates the views of a presented scene [8], and it has been used to provide evidence for the role of the neocortex [9] and hippocampus [10, 11] in the schematization of scenes during memory. However, the study of boundary extension has typically focused on object-oriented images that are not representative of our visuospatial world. Here, using a broad set of 1,000 images tested on 2,000 participants in a rapid recognition task, we discover "boundary contraction" as an equally robust phenomenon. Further, image composition largely drives whether extension or contraction is observed-although object-oriented images cause more boundary extension, scene-oriented images cause more boundary contraction. Finally, these effects also occur during drawing tasks, including a task with minimal memory load-when participants copy an image during viewing. Collectively, these results show that boundary extension is not a universal phenomenon and put into question the assumption that scene memory automatically combines visual information with additional context derived from internal schema. Instead, our memory for a scene may be largely driven by its visual composition, with a tendency to extend or contract the boundaries equally likely.

Robust Multiple Sclerosis Lesion Inpainting with Edge Prior.

Inpainting lesions is an important preprocessing task for algorithms analyzing brain MRIs of multiple sclerosis (MS) patients, such as tissue segmentation and cortical surface reconstruction. We propose a new deep learning approach for this task. Unlike existing inpainting approaches which ignore the lesion areas of the input image, we leverage the edge information around the lesions as a prior to help the inpainting process. Thus, the input of this network includes the T1-w image, lesion mask and the edge map computed from the T1-w image, and the output is the lesion-free image. The introduction of the edge prior is based on our observation that the edge detection results of the MRI scans will usually contain the contour of white matter (WM) and grey matter (GM), even though some undesired edges appear near the lesions. Instead of losing all the information around the neighborhood of lesions, our approach preserves the local tissue shape (brain/WM/GM) with the guidance of the input edges. The qualitative results show that our pipeline inpaints the lesion areas in a realistic and shape-consistent way. Our quantitative evaluation shows that our approach outperforms the existing state-of-the-art inpainting methods in both image-based metrics and in FreeSurfer segmentation accuracy. Furthermore, our approach demonstrates robustness to inaccurate lesion mask inputs. This is important for practical usability, because it allows for a generous over-segmentation of lesions instead of requiring precise boundaries, while still yielding accurate results.

Spatial Frequency and the Performance of Image-Based Visual Complexity Metrics

There is a wide range of visual and spatial complexity measurement methods that aim to quantify perceived image complexity. While image-based calculation methods (edge detection, image compression, contrast) characterize a digital image, visual perception studies focus on fundamental visual mechanisms, such as contrast sensitivity and visual task performance. Despite the evidence from several vision studies, spatial frequency information has not been widely utilized to assess image complexity. Previous studies suggest that image-based performance metrics are limited in explaining perceived complexity due to confounding factors, such as context, memory, familiarity, and expectation. Here, a visual experiment is conducted to assess the performance of image-based metrics and spatial frequency information using 16 abstract and natural images. Anew image complexity metric(R <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">spt</sub> ), based on detectability suprathreshold, was proposed to benchmark the performance of existing measures. Forty-four naïve participants used a 5-point Likert-type scale to judge the visual complexity of the images displayed on a tablet. Results indicate that root-mean-square error (RMSE) and R <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">spt</sub> correlate statistically significantly with subjective evaluations. Biological sex did not affect perceived spatial complexity. While RMSE and R <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">spt</sub> can potentially be used to estimate the spatial complexity of display images, the performance of spatial frequency information and image assessment measures in immersive viewing conditions require further research.

Blood Stasis Imaging Predicts Cerebral Microembolism during Acute Myocardial Infarction

Cardioembolic stroke is a major source of mortality and disability worldwide. The authors hypothesized that quantitative characterization of intracardiac blood stasis may be useful to determine cardioembolic risk in order to personalize anticoagulation therapy. The aim of this study was to assess the relationship between image-based metrics of blood stasis in the left ventricle and brain microembolism, a surrogate marker of cardiac embolism, in a controlled animal experimental model of acute myocardial infarction (AMI). Intraventricular blood stasis maps were derived from conventional color Doppler echocardiography in 10 pigs during anterior AMI induced by sequential ligation of the mid and proximal left anterior descending coronary artery (AMI-1 and AMI-2 phases). From these maps, indices of global and local blood stasis were calculated, such as the average residence time and the size and ratio of contact with the endocardium of blood regions with long residence times. The incidence of brain microemboli (high-intensity transient signals [HITS]) was monitored using carotid Doppler ultrasound. HITS were detected in 0%, 50%, and 90% of the animals at baseline and during AMI-1 and AMI-2 phases, respectively. The average residence time of blood in the left ventricle increased in parallel. The residence time performed well to predict microemboli (C-index=0.89, 95% CI, 0.75-1.00) and closely correlated with the number of HITS (R=0.87, P<.001). Multivariate and mediation analyses demonstrated that the number of HITS during AMI phases was best explained by stasis. Among conventional echocardiographic variables, only apical wall motion score weakly correlated with the number of HITS (R=0.3, P=.04). Mural thrombosis in the left ventricle was ruled out in all animals. The degree of stasis of blood in the left ventricle caused by AMI is closely related to the incidence of brain microembolism. Therefore, stasis imaging is a promising tool for a patient-specific assessment of cardioembolic risk.

Optimization of the electrode drive pattern for imaging fascicular compound action potentials in peripheral nerve with fast neural electrical impedance tomography

Objective: The main objective of this study was to investigate which injection pattern led to the best imaging of fascicular compound activity in fast neural EIT of peripheral nerve using an external cylindrical 2 × 14-electrodes cuff. Specifically, the study addressed the identification of the optimal injection pattern and of the optimal region of the reconstructed volume to image fascicles. Approach: The effect of three different measurement protocol features (transversal/longitudinal injection, drive electrode spacing, referencing configuration) over imaging was investigated in simulation with the use of realistic impedance changes and noise levels. Image-based metrics were employed to evaluate the quality of the reconstructions over the reconstruction domain. The optimal electrode addressing protocol suggested by the simulations was validated in vivo on the tibial and peroneal fascicles of rat sciatic peripheral nerves (N = 3) against MicroCT reference images. Main results: Injecting current transversally, with spacing of ⩾4 electrodes apart (⩾100°) and single-ring referencing of measurements, led to the best overall localization when reconstructing on the edge of the electrode array closest to the reference. Longitudinal injection protocols led to a higher SNR of the reconstructed image but poorer localization. All in vivo EIT recordings had statistically significant impedance variations (p < 0.05). Overall, fascicle center-of-mass (CoM) localization error was estimated at 141 ± 56 µm (−26 ± 94 µm and 5 ± 29° in radial coordinates). Significant difference was found (p < 0.05) between mean angular location of the tibial and peroneal CoMs. Significance: This study gives the reader recommendations for performing fast neural EIT of fascicular compound activity using the most effective protocol features.

Visual Quality Assessment of 3D Models

Geometric modifications of three-dimensional (3D) digital models are commonplace for the purpose of efficient rendering or compact storage. Modifications imply visual distortions that are hard to measure numerically. They depend not only on the model itself but also on how the model is visualized. We hypothesize that the model’s light environment and the way it reflects incoming light strongly influences perceived quality. Hence, we conduct a perceptual study demonstrating that the same modifications can be masked, or conversely highlighted, by different light-matter interactions. Additionally, we propose a new metric that predicts the perceived distortion of 3D modifications for a known interaction. It operates in the space of 3D meshes with the object’s appearance, that is, the light emitted by its surface in any direction given a known incoming light. Despite its simplicity, this metric outperforms 3D mesh metrics and competes with sophisticated perceptual image-based metrics in terms of correlation to subjective measurements. Unlike image-based methods, it has the advantage of being computable prior to the costly rendering steps of image projection and rasterization of the scene for given camera parameters.

A neural response-based model to predict discomfort glare from luminance image

It is well known that one of the problems of the current method for discomfort glare evaluation, called the unified glare rating, is the non-uniform luminance of the glare source. This paper addresses this issue by considering the spatial contrast of luminance as a measure of non-uniformity. An image-based metric is proposed to evaluate discomfort glare by modeling the neural response of human vision. The model takes an absolute luminance image as input and predicts visual discomfort based on the spatial distribution of the luminance of the stimulus and the background. The developed model was tested to predict subjective glare ratings based on an experiment conducted using non-uniform LED sources with symmetric and asymmetric patterns of LEDs, and its performance was compared with the unified glare rating. As expected, the unified glare rating predictions correlated well with the subjective glare evaluations of luminaires with symmetric patterns of LEDs (as they appear less non-uniform) but not for those with asymmetric patterns. Results showed that the developed model, named the Neural Response-based Glare Model, gave similar performance to unified glare rating for symmetric patterns but outperformed UGR for asymmetric patterns of LEDs.

Diagnosis of pulmonary hypertension from magnetic resonance imaging-based computational models and decision tree analysis.

Accurately identifying patients with pulmonary hypertension (PH) using noninvasive methods is challenging, and right heart catheterization (RHC) is the gold standard. Magnetic resonance imaging (MRI) has been proposed as an alternative to echocardiography and RHC in the assessment of cardiac function and pulmonary hemodynamics in patients with suspected PH. The aim of this study was to assess whether machine learning using computational modeling techniques and image-based metrics of PH can improve the diagnostic accuracy of MRI in PH. Seventy-two patients with suspected PH attending a referral center underwent RHC and MRI within 48 hours. Fifty-seven patients were diagnosed with PH, and 15 had no PH. A number of functional and structural cardiac and cardiovascular markers derived from 2 mathematical models and also solely from MRI of the main pulmonary artery and heart were integrated into a classification algorithm to investigate the diagnostic utility of the combination of the individual markers. A physiological marker based on the quantification of wave reflection in the pulmonary artery was shown to perform best individually, but optimal diagnostic performance was found by the combination of several image-based markers. Classifier results, validated using leave-one-out cross validation, demonstrated that combining computation-derived metrics reflecting hemodynamic changes in the pulmonary vasculature with measurement of right ventricular morphology and function, in a decision support algorithm, provides a method to noninvasively diagnose PH with high accuracy (92%). The high diagnostic accuracy of these MRI-based model parameters may reduce the need for RHC in patients with suspected PH.

Exploring the use of shape and texture descriptors of positron emission tomography tracer distribution in imaging studies of neurodegenerative disease

Positron emission tomography (PET) data related to neurodegeneration are most often quantified using methods based on tracer kinetic modeling. In contrast, here we investigate the ability of geometry and texture-based metrics that are independent of kinetic modeling to convey useful information on disease state. The study was performed using data from Parkinson’s disease subjects imaged with 11C-dihydrotetrabenazine and 11C-raclopride. The pattern of the radiotracer distribution in the striatum was quantified using image-based metrics evaluated over multiple regions of interest that were defined on co-registered PET and MRI images. Regression analysis showed a significant degree of correlation between several investigated metrics and clinical evaluations of the disease (p < 0.01). The best results were obtained with the first-order moment invariant of the radioactivity concentration values estimated over the full structural extent of the region as defined by MRI (R2 = 0.94). These results demonstrate that there is clinically relevant quantitative information in the tracer distribution pattern that can be captured using geometric and texture descriptors. Such metrics may provide an alternate and complementary data analysis approach to traditional kinetic modeling.