Search

Abstract Pooling publicly-available MRI data from multiple sites allows to assemble extensive groups of subjects, increase statistical power, and promote data reuse with machine learning techniques. The harmonization of multicenter data is necessary to reduce the confounding effect associated with non-biological sources of variability in the data. However, when applied to the entire dataset before machine learning, the harmonization leads to data leakage, because information outside the training set may affect model building, and potentially falsely overestimate performance. We propose a 1) measurement of the efficacy of data harmonization; 2) harmonizer transformer, i.e., an implementation of the ComBat harmonization allowing its encapsulation among the preprocessing steps of a machine learning pipeline, avoiding data leakage by design. We tested these tools using brain T1-weighted MRI data from 1740 healthy subjects acquired at 36 sites. After harmonization, the site effect was removed or reduced, and we showed the data leakage effect in predicting individual age from MRI data, highlighting that introducing the harmonizer transformer into a machine learning pipeline allows for avoiding data leakage by design.

Abstract Radiomics is emerging as a promising and useful tool in cardiac magnetic resonance (CMR) imaging applications. Accordingly, the purpose of this study was to investigate, for the first time, the effect of image resampling/discretization and filtering on radiomic features estimation from quantitative CMR T1 and T2 mapping. Specifically, T1 and T2 maps of 26 patients with hypertrophic cardiomyopathy (HCM) were used to estimate 98 radiomic features for 7 different resampling voxel sizes (at fixed bin width), 9 different bin widths (at fixed resampling voxel size), and 7 different spatial filters (at fixed resampling voxel size/bin width). While we found a remarkable dependence of myocardial radiomic features from T1 and T2 mapping on image filters, many radiomic features showed a limited sensitivity to resampling voxel size/bin width, in terms of intraclass correlation coefficient (> 0.75) and coefficient of variation (

Abstract Background Radiofrequency ablation has been shown to be a safe and effective treatment for scar-related ventricular arrhythmias (VA). Recent preliminary studies have shown that real time integration of late gadolinium enhancement cardiac magnetic resonance (LGE-CMR) images with electroanatomical map (EAM) data may lead to increased procedure efficacy, efficiency, and safety. Methods VOYAGE is a prospective, randomized, multicenter controlled open label study designed to compare in terms of efficacy, efficiency, and safety a CMR aided/guided workflow to standard EAM-guided ventricular tachycardia (VT) ablation. Patients with an ICD or with ICD implantation expected within 1 month, with scar related VT, suitable for CMR and multidetector computed tomography (MDCT) will be randomized to a CMR-guided or CMR-aided approach, whereas subjects unsuitable for imaging or with image quality deemed not sufficient for postprocessing will be allocated to standard of care ablation. Primary endpoint is defined as VT recurrences (sustained or requiring appropriate ICD intervention) during 12 months follow-up, excluding the first month of blanking period. Secondary endpoints will include procedural efficiency, safety, impact on quality of life and comparison between CMR-guided and CMR-aided approaches. Patients will be evaluated at 1, 6 and 12 months. Discussion The clinical impact of real time CMR-guided/aided ablation approaches has not been thoroughly assessed yet. This study aims at defining whether such workflow results in more effective, efficient, and safer procedures. If proven to be of benefit, results from this study could be applied in large scale interventional practice. Trial registrationClinicalTrials.gov, NCT04694079, registered on January 1, 2021.

Abstract Background The objective of our study was to analyze the data of our biopsies, determine a detection rate (DR), compare it with the data in the literature and draw possible deductions, so as to offer the patient the possibility of not having other biopsies in the future. Methods We have enrolled 189 biopsy-naive patients in the period between September 2018 and December 2020. Each patient underwent multiparametric (mp)-MRI which was reviewed by our team of radiologists. In our center, each examination is examined by 4 radiologists separately with an overall final result. Through the t student test, any statistically significant differences between the DRs and the concordance rate between the positive cores and the suspected area on MRI were analyzed for each urologist who performed the procedure. Results The absolute (DR) was 69.3% (131/189 patients). The relative DR for each PIRADS score was 41% for PIRADS 3, 70.2% for PIRADS 4, 89.3% for PIRADS 5. We found a high percentage of agreement between the positive biopsy samples and the suspicious area identified on MRI: 90.8% (119/131 patients). There were no statistically significant differences between the DRs of the urologists who performed the procedure (p = 0.89), nor for the percentage of agreement between the positivity of the core and the suspected area on MRI (p = 0.92). Conclusions MRI in the future could become the gold standard for performing MRI fusion-guided biopsies to have a better diagnostic result and avoid rebiopsies. A team MRI reading allows greater accuracy in identifying the suspected lesion, which is demonstrated by a high rate of agreement with the positivity of the cores (90.8%). There is a cost problem due to the need to carry out the mpMRI but it could have less impact in the future. In addition, the MRI provides useful information on the extent of the disease (e.g., cT3a/b) which allows you to better plan the surgical strategy or other therapies.

Abstract In recent years, 2D convolutional neural networks (CNNs) have been extensively used to diagnose neurological diseases from magnetic resonance imaging (MRI) data due to their potential to discern subtle and intricate patterns. Despite the high performances reported in numerous studies, developing CNN models with good generalization abilities is still a challenging task due to possible data leakage introduced during cross-validation (CV). In this study, we quantitatively assessed the effect of a data leakage caused by 3D MRI data splitting based on a 2D slice-level using three 2D CNN models to classify patients with Alzheimer’s disease (AD) and Parkinson’s disease (PD). Our experiments showed that slice-level CV erroneously boosted the average slice level accuracy on the test set by 30% on Open Access Series of Imaging Studies (OASIS), 29% on Alzheimer’s Disease Neuroimaging Initiative (ADNI), 48% on Parkinson’s Progression Markers Initiative (PPMI) and 55% on a local de-novo PD Versilia dataset. Further tests on a randomly labeled OASIS-derived dataset produced about 96% of (erroneous) accuracy (slice-level split) and 50% accuracy (subject-level split), as expected from a randomized experiment. Overall, the extent of the effect of an erroneous slice-based CV is severe, especially for small datasets.

The human brain is a highly complex structure, which can be only partially described by conventional metrics derived from magnetic resonance imaging (MRI), such as volume, cortical thickness, and gyrification index. In the last years, the fractal dimension (FD) – a useful quantitative index of fractal geometry – has proven to well express the morphological complexity of the cerebral cortex. However, this complexity is likely higher than that we can observe using MRI scanners with 1.5 T or 3 T field strength. Ultrahigh-field MRI (UHF-MRI) improves imaging of smaller anatomical brain structures by exploring down to a submillimetric spatial resolution with higher signal-to-noise and contrast-to-noise ratios. Accordingly, we hypothesized that UHF-MRI might reveal a higher level of the structural complexity of the cerebral cortex. In this study, using an improved box-counting algorithm, we estimated the FD of the cerebral cortex in six public or private T1-weighted MRI datasets of young healthy subjects (for a total of 87 subjects), acquired at different field strengths (1.5 T, 3 T, and 7 T). Our results showed, for the first time, that MRI-derived FD values of the cerebral cortex imaged at 7 T were significantly higher than those observed at lower field strengths. UHF-MRI provides an anatomical definition not achievable at lower field strengths and can improve unveiling the real structural complexity of the human brain.

Effective anticancer treatments have dramatically improved the outcome of patients with cancer, but cardiac toxicity reduces their clinical efficacy in a non-negligible percentage of patients. Sacubitril/valsartan is a new paradigm in the treatment of chronic heart failure, with a reduced ejection fraction due to the enhancement of natriuretic peptides' properties when coupled with a blocking effect on the angiotensin II type 1 (AT1) receptors. As with other clinical conditions of heart failure with potentially reversible declines in cardiac function, a wearable cardioverter defibrillator (WCD) is a valid tool for protection against sudden death until recovery occurs. We report a case series of four patients with chemotherapy-related acute cardiac failure with severely reduced cardiac function. They were successfully treated with sacubitril/valsartan while being protected from malignant arrhythmias using a wearable cardioverter defibrillator until the recovery of cardiac function. Sacubitril/valsartan was confirmed to be effective in anthracycline-related cardiac toxicity and the wearable cardioverter defibrillator should be considered as a support tool even in the oncology patient.

Radiomics and artificial intelligence have the potential to become a valuable tool in clinical applications. Frequently, radiomic analyses through machine learning methods present issues caused by high dimensionality and multicollinearity, and redundant radiomic features are usually removed based on correlation analysis. We assessed the effect of preprocessing—in terms of voxel size resampling, discretization, and filtering—on correlation-based dimensionality reduction in radiomic features from cardiac T1 and T2 maps of patients with hypertrophic cardiomyopathy. For different combinations of preprocessing parameters, we performed a dimensionality reduction of radiomic features based on either Pearson’s or Spearman’s correlation coefficient, followed by the computation of the stability index. With varying resampling voxel size and discretization bin width, for both T1 and T2 maps, Pearson’s and Spearman’s dimensionality reduction produced a slightly different percentage of remaining radiomic features, with a relatively high stability index. For different filters, the remaining features’ stability was instead relatively low. Overall, the percentage of eliminated radiomic features through correlation-based dimensionality reduction was more dependent on resampling voxel size and discretization bin width for textural features than for shape or first-order features. Notably, correlation-based dimensionality reduction was less sensitive to preprocessing when considering radiomic features from T2 compared with T1 maps.

In order to assess possible influences of occlusion on motor performance, we studied by functional magnetic resonance imaging (fMRI) the changes in the blood oxygenation level dependent (BOLD) signal induced at brain level by a finger to thumb motor task in a population of subjects characterized by an asymmetric activation of jaw muscles during clenching (malocclusion). In these subjects, appropriate occlusal correction by an oral orthotic (bite) reduced the masticatory asymmetry. The finger to thumb task was performed while the subject’s dental arches were touching, in two conditions: (a) with the teeth in direct contact (Bite OFF) and (b) with the bite interposed between the arches (Bite ON). Both conditions required only a very slight activation of masticatory muscles. Maps of the BOLD signal recorded during the movement were contrasted with the resting condition (activation maps). Between conditions comparison of the activation maps (Bite OFF/Bite ON) showed that, in Bite OFF, the BOLD signal was significantly higher in the trigeminal sensorimotor region, the premotor cortex, the cerebellum, the inferior temporal and occipital cortex, the calcarine cortex, the precuneus on both sides, as well as in the right posterior cingulate cortex. These data are consistent with the hypothesis that malocclusion makes movement performance more difficult, leading to a stronger activation of (a) sensorimotor areas not dealing with the control of the involved body part, (b) regions planning the motor sequence, and (c) the cerebellum, which is essential in motor coordination. Moreover, the findings of a higher activation of temporo-occipital cortex and precuneus/cingulus, respectively, suggest that, during malocclusion, the movement occurs with an increased visual imagery activity, and requires a stronger attentive effort.

Task- and stimulus-based neuroimaging studies have begun to unveil the central autonomic network which modulates autonomic nervous system activity. In the present study, we aimed to evaluate the central autonomic network without the bias constituted by the use of a task. Additionally, we assessed whether this circuitry presents signs of dysregulation in the early stages of Parkinson's disease (PD), a condition which may be associated with dysautonomia. We combined heart-rate-variability based methods for time-varying assessments of the autonomic nervous system outflow with resting-state fMRI in 14 healthy controls and 14 de novo PD patients, evaluating the correlations between fMRI time-series and the instantaneous high-frequency component of the heart-rate-variability power spectrum, a marker of parasympathetic outflow. In control subjects, the high-frequency component of the heart-rate-variability power spectrum was significantly anti-correlated with fMRI time-series in several cortical, subcortical and brainstem regions. This complex central network was not detectable in PD patients. In between-group analysis, we found that in healthy controls the brain activation related to the high-frequency component of the heart-rate-variability power spectrum was significantly less than in PD patients in the mid and anterior cingulum, sensorimotor cortex and supplementary motor area, insula and temporal lobe, prefrontal cortex, hippocampus and in a region encompassing posterior cingulum, precuneus and parieto-occipital cortex. Our results indicate that the complex central network which modulates parasympathetic outflow in the resting state is impaired in the early clinical stages of PD.

PURPOSE:To assess the potential of histogram metrics of diffusion-tensor imaging (DTI)-derived indices in revealing neurodegeneration and its progression in spinocerebellar ataxia type 2 (SCA2). MATERIALS AND METHODS:Nine SCA2 patients and 16 age-matched healthy controls, were examined twice (SCA2 patients 3.6±0.7 years and controls 3.3±1.0 years apart) on the same 1.5T scanner by acquiring T1-weighted and diffusion-weighted (b-value = 1000 s/mm2) images. Cerebrum and brainstem-cerebellum regions were segmented using FreeSurfer suite. Histogram analysis of DTI-derived indices, including mean diffusivity (MD), fractional anisotropy (FA), axial (AD) / radial (RD) diffusivity and mode of anisotropy (MO), was performed. RESULTS:At baseline, significant differences between SCA2 patients and controls were confined to brainstem-cerebellum. Median values of MD/AD/RD and FA/MO were significantly (p

Spinocerebellar ataxia type 2 (SCA2) is the second most frequent autosomal dominant inherited ataxia worldwide. We investigated the capability of magnetic resonance imaging (MRI) to track in vivo progression of brain atrophy in SCA2 by examining twice 10 SCA2 patients (mean interval 3.6 years) and 16 age- and gender-matched healthy controls (mean interval 3.3 years) on the same 1.5 T MRI scanner. We used T1-weighted images and tensor-based morphometry (TBM) to investigate volume changes and the Inherited Ataxia Clinical Rating Scale to assess the clinical deficit. With respect to controls, SCA2 patients showed significant higher atrophy rates in the midbrain, including substantia nigra, basis pontis, middle cerebellar peduncles and posterior medulla corresponding to the gracilis and cuneatus tracts and nuclei, cerebellar white matter (WM) and cortical gray matter (GM) in the inferior portions of the cerebellar hemisphers. No differences in WM or GM volume loss were observed in the supratentorial compartment. TBM findings did not correlate with modifications of the neurological deficit. In conclusion, MRI volumetry using TBM is capable of demonstrating the progression of pontocerebellar atrophy in SCA2, supporting a possible role of MRI as biomarker in future trials.