This study evaluates the efficacy of deep learning models in identifying infarct tissue on computed tomography perfusion (CTP) scans from patients with acute ischemic stroke due to large vessel occlusion, specifically addressing the potential influence of varying noise reduction techniques implemented by different vendors. We analyzed CTP scans from 60 patients who underwent mechanical thrombectomy achieving a modified thrombolysis in cerebral infarction (mTICI) score of 2c or 3, ensuring minimal changes in the infarct core between the initial CTP and follow-up MR imaging. Noise reduction techniques, including principal component analysis (PCA), wavelet, non-local means (NLM), and a no denoising approach, were employed to create hemodynamic parameter maps. Infarct regions identified on follow-up diffusion-weighted imaging (DWI) within 48 hours were co-registered with initial CTP scans and refined with ADC maps to serve as ground truth for training a data-augmented U-Net model. The performance of this convolutional neural network (CNN) was assessed using Dice coefficients across different denoising methods and infarct sizes, visualized through box plots for each parameter map. Our findings show no significant differences in model accuracy between PCA and other denoising methods, with minimal variation in Dice scores across techniques. This study confirms that CNNs are adaptable and capable of handling diverse processing schemas, indicating their potential to streamline diagnostic processes and effectively manage CTP input data quality variations.

MR methods have been proposed to estimate blood oxygen saturation in cerebral tissues by exploiting the magnetic susceptibility difference between oxy and deoxyhemoglobin. These models neglect the contribution of extravascular susceptibility sources leading to inaccurate measurements. Recent approaches combined quantitative BOLD and QSM methods, but these models neglected the impact of the geometrical arrangements of extravascular sources on signal decay. Consequently, we propose a new model to unravel the magnitude and phase of MRI gradient echo sequence signal contributions of intravascular and extravascular compartments to produce blood-oxygen saturation and non-vascular susceptibility maps. This new model and previous MRI approaches are implemented in rat data acquired in healthy and stroke settings. The results are compared across anatomical regions of interest characterized by different extravascular magnetic susceptibilities. Results obtained with the new model improved estimations in myelin-rich white matter regions and delineated oxygen saturation decrease in stroke lesions after 60 min of middle cerebral artery occlusion. Characterization of extravascular susceptibility sources improved the overall results, highlighting its critical role in the translation toward pathological applications.

The Instantaneous Signal Loss Simulation (InSiL) model is a promising alternative to the classical mono-exponential fitting of the Modified Look-Locker Inversion-recovery (MOLLI) sequence in cardiac T 1 mapping applications, which achieves better accuracy and is less sensitive to heart rate (HR) variations. Classical nonlinear least squares (NLLS) estimation methods require some parameters of the model to be fixed a priori in order to give reliable T 1 estimations and avoid outliers. This introduces further bias in the estimation, reducing the advantages provided by the InSiL model. In this paper, a novel Bayesian estimation method using a hierarchical model is proposed to fit the parameters of the InSiL model. The hierarchical Bayesian modeling has a shrinkage effect that works as a regularizer for the estimated values, by pulling spurious estimated values toward the group-mean, hence reducing greatly the number of outliers. Simulations, physical phantoms, and in-vivo human cardiac data have been used to show that this approach estimates accurately all the InSiL parameters, and achieve high precision estimation of the T 1 compared to the classical MOLLI model and NLLS InSiL estimation.

Microvascular dysfunction and no-reflow are considered a major cause of secondary damage despite revascularization in acute ischemic stroke (AIS), ultimately affecting patient outcomes. We used quantitative PET-MRI imaging to characterize early microvascular damages in a preclinical non-human primate model mimicking endovascular mechanical thrombectomy (EVT). During occlusion, PET perfusion and MRI diffusion were used to measure ischemic and lesion core volumes respectively. Following revascularization, multiparametric PET-MRI included perfusion, diffusion, blood-brain barrier (BBB) permeability MRI, and 15O-oxygen metabolism PET. Lesion growth on MRI was evaluated at one week, and the neurological score was assessed daily; a poor outcome was defined as a score>6 (0-normal, 60-death) after one week. Early after recanalization, the gold-standard PET ischemic threshold (<0.2 ​mL/min/g) identified post-EVT hypoperfusion in 67 ​% of the cases (14/21) located in the occlusion acute lesion. Acquired 110 ​min post-EVT, the area of MRI Tmax hypoperfusion was larger and even more frequent (18/20) and was also located within the acute lesion. Eight of the total cases (38 ​%) had a poor outcome, and all of them had no-reflow (7/8 MRI no-reflow and 6/8 ​PET no-reflow). Diffusion ADC alterations and post-EVT oxygen extraction fraction (OEF) values were significantly different in PET no-reflow cases compared to those without no-reflow, exhibiting an inverse correlation. Independently of no-reflow, long perfusion Tmax and post-EVT high BBB Ktrans in the lesion core were the hallmarks of poor outcome and infarct growth. This early quantitative imaging signature may predict infarct growth and poor outcome and help to identify neuroprotection targets.

Objectives: The non-linear relationship between gadolinium concentration and the signal in perfusion cardiac magnetic resonance (CMR) poses a significant challenge for accurate quantification of pharmacokinetic parameters. This phenomenon primarily impacts the arterial input function (AIF), causing it to appear saturated in comparison to the temporal concentration profile. This study aims to leverage a blind deconvolution strategy through a deep-learning approach to address the saturation in the AIF. Methods: We propose the utilization of a convolutional neural network (CNN) architecture with the saturated AIF and a set of myocardial tissue signals as inputs, generating the corrected AIF as the output. To train the network, a dataset comprising over 3×10^6 simulated AIFs with associated signals from five simulated tissues response for each instance was employed. To assess the effectiveness of the approach, the trained network was evaluated using a dual-saturation sequence to compare the corrected AIF with the unsaturated version. The clinical dataset encompassed scans from 43 patients. Results: The mean square error (MSE) for the testing subset of the simulated database was 0.69% of the peak. In the in vivo dataset, the coefficient of determination R2 was 0.26 and 0.86 for the saturated and corrected AIF, respectively, in comparison to the unsaturated AIF. Conclusion: The proposed network successfully corrects the acquisition-induced effects on the AIF. Moreover, the extensive simulated database, featuring diverse acquisition parameters, facilitates the robust generalization of the network's application.

To investigate the relationship between collaterals and blood–brain barrier (BBB) permeability on pre-treatment MRI in a cohort of acute ischemic stroke (AIS) patients treated with thrombectomy. We conducted a retrospective analysis of the HIBISCUS-STROKE cohort, a single-center observational study that enrolled patients treated with thrombectomy from 2016 to 2022. Dynamic-susceptibility MRIs were post-processed to generate K2 maps with arrival-time correction, which were co-registered with apparent diffusion coefficient (ADC) maps. The 90th percentile of K2 was extracted from the infarct core—defined by an ADC ≤ 620 × 10−6 mm2/s with manual adjustments—and expressed as a percentage change compared to the contralateral white matter. Collaterals were assessed using pre-thrombectomy digital subtraction arteriography with an ASITN/SIR score < 3 defining poor collaterals. Out of 249 enrolled, 101 (40.6

BACKGROUND AND OBJECTIVES:The aim of this study was to investigate the relationship between baseline blood-brain barrier (BBB) permeability and the kinetics of circulating inflammatory markers in a cohort of acute ischemic stroke (AIS) patients treated with mechanical thrombectomy. METHODS:The CoHort of Patients to Identify Biological and Imaging markerS of CardiovascUlar Outcomes in Stroke includes AIS patients treated with mechanical thrombectomy after admission MRI and undergoing a sequential assessment of circulating inflammatory markers. Baseline dynamic susceptibility perfusion MRI was postprocessed with arrival time correction to provide K2 maps reflecting BBB permeability. After coregistration of apparent diffusion coefficient and K2 maps, the 90th percentile of K2 value was extracted within baseline ischemic core and expressed as a percentage change compared with contralateral normal-appearing white matter. Population was dichotomized according to the median K2 value. Univariable and multiple variable logistic regression analyses were performed to investigate factors associated with increased pretreatment BBB permeability in the whole population and in patients with symptom onset <6 hours. RESULTS:In the whole population (n = 105 patients, median K2 = 1.59), patients with an increased BBB permeability had higher serum levels of matrix metalloproteinase (MMP)-9 at H48 (p = 0.02), a higher C-reactive protein (CRP) serum level at H48 (p = 0.01), poorer collateral status (p = 0.01), and a larger baseline ischemic core (p < 0.001). They were more likely to have hemorrhagic transformation (p = 0.008), larger final lesion volume (p = 0.02), and worst neurologic outcome at 3 months (p = 0.04). The multiple variable logistic regression indicated that an increased BBB permeability was associated only with ischemic core volume (odds ratio [OR] 1.04, 95% CI 1.01-1.06, p < 0.0001). Restricting analysis to patients with symptom onset <6 hours (n = 72, median K2 = 1.27), participants with an increased BBB permeability had higher serum levels of MMP-9 at H0 (p = 0.005), H6 (p = 0.004), H24 (p = 0.02), and H48 (p = 0.01), higher CRP levels at H48 (p = 0.02), and a larger baseline ischemic core (p < 0.0001). The multiple variable logistic analysis showed that increased BBB permeability was independently associated with higher H0 MMP-9 levels (OR 1.33, 95% CI 1.12-1.65, p = 0.01) and a larger ischemic core (OR 1.27, 95% CI 1.08-1.59, p = 0.04). DISCUSSION:In AIS patients, increased BBB permeability is associated with a larger ischemic core. In the subgroup of patients with symptom onset <6 hours, increased BBB permeability is independently associated with higher H0 MMP-9 levels and a larger ischemic core.

Abstract Non-human primate (NHP) studies are unique in translational research, especially in neurosciences and neuroimaging approaches are a preferred method for scaling cross-species comparative neurosciences. In this regard, neuroimaging database development and sharing are encouraged to increase the number of subjects available to the community while limiting the number of animals used in research. We present here a simultaneous PET/MR dataset of 20 Macaca Fascicularis structured according to the Brain Imaging Data Structure (BIDS) standards. This database contains multiple MRI sequences (anatomical, diffusion and perfusion imaging notably), as well as PET perfusion and inflammation using respectively [ 15 O]H 2 O and [ 11 C]PK11195 radiotracers. We describe the pipeline method to assemble baseline data from various cohorts and qualitatively assessed all the data using signal-to-noise and contrast-to-noise ratios as well as the median of intensity. The database is stored and available through the PRIME-DE consortium repository.

Non-human primate studies are unique in translational research, especially in neurosciences where neuroimaging approaches are the preferred methods used for cross-species comparative neurosciences. In this regard, neuroimaging database development and sharing are encouraged to increase the number of subjects available to the community, while limiting the number of animals used in research. Here we present a simultaneous positron emission tomography (PET)/magnetic resonance (MR) dataset of 20 Macaca fascicularis images structured according to the Brain Imaging Data Structure standards. This database contains multiple MR imaging sequences (anatomical, diffusion and perfusion imaging notably), as well as PET perfusion and inflammation imaging using respectively [ 15 O]H 2 O and [ 11 C]PK11195 radiotracers. We describe the pipeline method to assemble baseline data from various cohorts and qualitatively assess all the data using signal-to-noise and contrast-to-noise ratios as well as the median of intensity and the pseudo-noise-equivalent-count rate (dynamic and at maximum) for PET data. Our study provides a detailed example for quality control integration in preclinical and translational PET/MR studies with the aim of increasing reproducibility. The PREMISE database is stored and available through the PRIME-DE consortium repository.

This work provides a new method for fast post-processing of MRI data acquired using the WASAB1 sequence for simultaneous B 0 and B 1 mapping, used in CEST imaging for field inhomogeneity corrections. We are proposing a new processing method with outstanding acceleration of the parameter estimation procedure, without compromising the stability of the estimation. The stability of the method is demonstrated on phantom data and in vivo 3 Tesla clinical data.

Purpose:The size and location of infarct and penumbra are key to decision-making for acute ischemic stroke (AIS) management. CT perfusion (CTP) software estimate infarct and penumbra volume using contralateral hemisphere relative thresholding. This approach is not robust and widely contested by the scientific community. In this study, we investigate the use of deep learning-based algorithms to efficiently locate infarct and penumbra tissue on CTP hemodynamic maps.Approach:CTP scans were retrospectively collected for 60 and 59 patients in the infarct only and infarct + penumbra substudies respectively. Commercial CTP software was used to generate cerebral blood flow, cerebral blood volume, mean transit time, time to peak, and delay time maps. U-Net-shaped architectures were trained to segment infarct or infarct + penumbra. Test-time-augmentation, ensembling, and watershed segmentation were used as postprocessing techniques. Segmentation performance was evaluated using Dice coefficients (DC) and mean absolute volume errors (MAVE).Results:The algorithm segmented infarct tissue resulted in DC of 0.64 ± 0.03 (0.63, 0.65), and MAVE of 4.91 ± 0.94 (4.5, 5.32) mL. In comparison, the commercial software predicted infarct with a DC of 0.31 ± 0.17 (0.26, 0.36) and MAVE of 9.77 ± 8.35 (7.12, 12.42) mL. The algorithm was able to segment infarct + penumbra with a DC of 0.61 ± 0.04 (0.6, 0.63), and MAVE of 6.51 ± 1.37 (5.91, 7.11) mL. In comparison, the commercial software predicted infarct + penumbra with a DC of 0.3 ± 0.19 (0.25, 0.35) and MAVE of 9.18 ± 7.55 (7.25, 11.11) mL.Conclusions:Use of deep learning algorithms to assess severity of AIS in terms of infarct and penumbra volume is precise and outperforms current relative thresholding methods. Such an algorithm would enhance the reliability of CTP in guiding treatment decisions.

This work provides a new denoising methodology for multi-acquisition magnetic resonance images (MRI) based on principal components analysis (PCA). We are proposing a new principal component selection criterion that identifies spatial information in the extracted component coefficients, leading to a better preservation of anatomical structures and pathological information. In addition, our adaptive filtering step allows us to further denoise the MRI data, rejecting persistent spatial noise from the extracted component coefficients. In our investigations the proposed method outperformed the eigenvalue based selection criteria on Amide Proton Transfer weighted CEST data.

CEST MRI methods, such as APT and NOE imaging reveal biomarkers with significant diagnostic potential due to their ability to access molecular tissue information. Regardless of the technique used, CEST MRI data are affected by static magnetic B0 and radiofrequency B1 field inhomogeneities that degrade their contrast. For this reason, the correction of B0 field-induced artefacts is essential, whereas accounting for B1 field inhomogeneities have shown significant improvements in image readability. In a previous work, an MRI protocol called WASABI was presented, which can map simultaneously B0 and B1 field inhomogeneities, while maintaining the same sequence and readout types as used for CEST MRI. Despite the highly satisfactory quality of B0 and B1 maps computed from the WASABI data, the post-processing method is based on an exhaustive search of a four-parameter space and an additional four-parameter non-linear model fitting step. This leads to long post-processing times that are prohibitive in clinical practice. This work provides a new method for fast post-processing of WASABI data with outstanding acceleration of the parameter estimation procedure and without compromising its stability. The resulting computational acceleration makes the WASABI technique suitable for clinical use. The stability of the method is demonstrated on phantom data and clinical 3 Tesla in vivo data.

Magnetic resonance imaging (MRI) plays an increasingly significant role in diagnosis and follow-up of breast lesions. A myriad of sequences are now available to provide various biomarkers for lesion characterization and grading to supplement mammographic findings. More recently, along with dynamic-contrast enhanced (DCE) and T2-weighted imaging, diffusion-weighted imaging (DWI) has emerged as a contrast-agent-free technique allowing to discriminate benign from malignant lesions. However, the lack of consensus regarding acquisition protocols requires significant efforts from medical and scientific communities toward standardization. The absence of guidelines leaves the vendor uncertain as to what characteristics in their software product would be the most applicable for the market. In the meanwhile, until such consensus is reached, vendors can only offer a large spectrum of metrics to allow multiparametric analysis. Olea Sphere combined with breastscape offers a large panel of advanced postprocessing for diffusion, permeability, relaxometry, and texture to facilitate breast lesion characterization and, hopefully, contribute to the standardization of clinical practice for the utmost benefit to patients. In this chapter, the authors describe the various DWI models implemented within Olea Sphere, with specific focus on the clinical interest of such advanced metrics, based on the proprietary Bayesian postprocessing method.

The ischemic penumbra is defined as the severely hypoperfused, functionally impaired, at-risk but not yet infarcted tissue that will be progressively recruited into the infarct core. Early reperfusion aims to save the ischemic penumbra by preventing infarct core expansion and is the mainstay of acute ischemic stroke therapy. Intravenous thrombolysis and mechanical thrombectomy for selected patients with large vessel occlusion has been shown to improve functional outcome. Given the varying speed of infarct core progression among individuals, a therapeutic window tailored to each patient has recently been proposed. Recent studies have demonstrated that reperfusion therapies are beneficial in patients with a persistent ischemic penumbra, beyond conventional time windows. As a result, mapping the penumbra has become crucial in emergency settings for guiding personalized therapy. The penumbra was first characterized as an area with a reduced cerebral blood flow, increased oxygen extraction fraction and preserved cerebral metabolic rate of oxygen using positron emission tomography (PET) with radiolabeled O2. Because this imaging method is not feasible in an acute clinical setting, the magnetic resonance imaging (MRI) mismatch between perfusion-weighted imaging and diffusion-weighted imaging, as well as computed tomography perfusion have been proposed as surrogate markers to identify the penumbra in acute ischemic stroke patients. Transversal studies comparing PET and MRI or using longitudinal assessment of a limited sample of patients have been used to define perfusion thresholds. However, in the era of mechanical thrombectomy, these thresholds are debatable. Using various MRI methods, the original penumbra definition has recently gained a significant interest. The aim of this review is to provide an overview of the evolution of the ischemic penumbra imaging methods, including their respective strengths and limitations, as well as to map the current intellectual structure of the field using bibliometric analysis and explore future directions.

Cardiac magnetic resonance myocardial perfusion imaging can detect coronary artery disease and is an alternative to single-photon emission computed tomography or positron emission tomography. However, the complex, non-linear MR signal and the lack of robust quantification of myocardial blood flow have hindered its widespread clinical application thus far. Recently, a new Bayesian approach was developed for brain imaging and evaluation of perfusion indexes (Kudo et al., 2014). In addition to providing accurate perfusion measurements, this probabilistic approach appears more robust than previous approaches, particularly due to its insensitivity to bolus arrival delays. We assessed the performance of this approach against a well-known and commonly deployed model-independent method based on the Fermi function for cardiac magnetic resonance myocardial perfusion imaging. The methods were first evaluated for accuracy and precision using a digital phantom to test them against the ground truth; next, they were applied in a group of coronary artery disease patients. The Bayesian method can be considered an appropriate model-independent method with which to estimate myocardial blood flow and delays. The digital phantom comprised a set of synthetic time-concentration curve combinations generated with a 2-compartment exchange model and a realistic combination of perfusion indexes, arterial input dynamics, noise and delays collected from the clinical dataset. The myocardial blood flow values estimated with the two methods showed an excellent correlation coefficient (r 2 > 0.9) under all noise and delay conditions. The Bayesian approach showed excellent robustness to bolus arrival delays, with a similar performance to Fermi modeling when delays were considered. Delays were better estimated with the Bayesian approach than with Fermi modeling. An in vivo analysis of coronary artery disease patients revealed that the Bayesian approach had an excellent ability to distinguish between abnormal and normal myocardium. The Bayesian approach was able to discriminate not only flows but also delays with increased sensitivity by offering a clearly enlarged range of distribution for the physiologic parameters.

Background Combining MRI techniques with machine learning methodology is rapidly gaining attention as a promising method for staging of brain gliomas. This study assesses the diagnostic value of such a framework applied to dynamic susceptibility contrast (DSC)-MRI in classifying treatment-naïve gliomas from a multi-center patients into WHO grades II-IV and across their isocitrate dehydrogenase (IDH) mutation status. Methods Three hundred thirty-three patients from 6 tertiary centres, diagnosed histologically and molecularly with primary gliomas (IDH-mutant = 151 or IDH-wildtype = 182) were retrospectively identified. Raw DSC-MRI data was post-processed for normalised leakage-corrected relative cerebral blood volume (rCBV) maps. Shape, intensity distribution (histogram) and rotational invariant Haralick texture features over the tumour mask were extracted. Differences in extracted features across glioma grades and mutation status were tested using the Wilcoxon two-sample test. A random-forest algorithm was employed (2-fold cross-validation, 250 repeats) to predict grades or mutation status using the extracted features. Results Shape, distribution and texture features showed significant differences across mutation status. WHO grade II-III differentiation was mostly driven by shape features while texture and intensity feature were more relevant for the III-IV separation. Increased number of features became significant when differentiating grades further apart from one another. Gliomas were correctly stratified by mutation status in 71% and by grade in 53% of the cases (87% of the gliomas grades predicted with distance less than 1). Conclusions Despite large heterogeneity in the multi-center dataset, machine learning assisted DSC-MRI radiomics hold potential to address the inherent variability and presents a promising approach for non-invasive glioma molecular subtyping and grading.

GlucoCEST is an MRI contrast enhancement technique sensitive to the concentration of sugar in the tissue. Because of a differencein metabolism, it is thought that tumors consume more sugar than normal tissue. However, glucose metabolism is complex and depends onmany processes, which are all important to understand the origin of the measured signal. To achieve this goal we apply here a process analysismethod to a deterministic system describing the metabolism of glucose in the tissue.

Objectives The aims of this study were to evaluate the agreement of computed tomography (CT)-perfusion parameter values of the normal renal cortex and various renal tumors, which were obtained by different mathematical models, and to evaluate their diagnostic accuracy. Materials and Methods Perfusion imaging was performed prospectively in 35 patients to analyze 144 regions of interest of the normal renal cortex and 144 regions of interest of renal tumors, including 21 clear-cell renal cell carcinomas (RCC), 6 papillary RCCs, 5 oncocytomas, 1 chromophobe RCC, 1 angiomyolipoma with minimal fat, and 1 tubulocystic RCC. Identical source data were postprocessed and analyzed on 2 commercial software applications with the following implemented mathematical models: maximum slope, Patlak plot, standard singular-value decomposition (SVD), block-circulant SVD, oscillation-limited block-circulant SVD, and Bayesian estimation technique. Results for blood flow (BF), blood volume (BV), and mean transit time (MTT) were recorded. Agreement and correlation between pairs of models and perfusion parameters were assessed. Diagnostic accuracy was evaluated by receiver operating characteristic (ROC) analysis. Results Significant differences and poor agreement of BF, BV, and MTT values were noted for most of model comparisons in both the normal renal cortex and different renal tumors. The correlations between most model pairs and perfusion parameters ranged between good and perfect (Spearman = 0.79-1.00), except for BV values obtained by Patlak method ( = 0.61-0.72). All mathematical models computed BF and BV values, which differed significantly between clear cell RCCs, papillary RCCs, and oncocytomas, which introduces them as useful diagnostic tests to differentiate between different histologic subgroups (areas under ROC curve, 0.83-0.99). The diagnostic accuracy to discriminate between clear-cell RCCs and the renal cortex was the lowest based on the Patlak plot model (area under ROC curve, 0.76); BF and BV values obtained by other algorithms did not differ significantly in their diagnostic accuracy. Conclusions Quantitative perfusion parameters obtained from different mathematical models cannot be used interchangeably. Based on BF and BV estimates, all models are a useful tool in the differential diagnosis of kidney tumors, with the Patlak plot model yielding a significantly lower diagnostic accuracy.