Deep learning based joint fusion approach to exploit anatomical and functional brain information in autism spectrum disorders.

BACKGROUND: The integration of the information encoded in multiparametric MRI images can enhance the performance of machine-learning classifiers. In this study, we investigate whether the combination of structural and functional MRI might improve the performances of a deep learning (DL) model trained to discriminate subjects with Autism Spectrum Disorders (ASD) with respect to typically developing controls (TD). MATERIAL AND METHODS: We analyzed both structural and functional MRI brain scans publicly available within the ABIDE I and II data collections. We considered 1383 male subjects with age between 5 and 40 years, including 680 subjects with ASD and 703 TD from 35 different acquisition sites. We extracted morphometric and functional brain features from MRI scans with the Freesurfer and the CPAC analysis packages, respectively. Then, due to the multisite nature of the dataset, we implemented a data harmonization protocol. The ASD vs. TD classification was carried out with a multiple-input DL model, consisting in a neural network which generates a fixed-length feature representation of the data of each modality (FR-NN), and a Dense Neural Network for classification (C-NN). Specifically, we implemented a joint fusion approach to multiple source data integration. The main advantage of the latter is that the loss is propagated back to the FR-NN during the training, thus creating informative feature representations for each data modality. Then, a C-NN, with a number of layers and neurons per layer to be optimized during the model training, performs the ASD-TD discrimination. The performance was evaluated by computing the Area under the Receiver Operating Characteristic curve within a nested 10-fold cross-validation. The brain features that drive the DL classification were identified by the SHAP explainability framework. RESULTS: The AUC values of 0.66±0.05 and of 0.76±0.04 were obtained in the ASD vs. TD discrimination when only structural or functional features are considered, respectively. The joint fusion approach led to an AUC of 0.78±0.04. The set of structural and functional connectivity features identified as the most important for the two-class discrimination supports the idea that brain changes tend to occur in individuals with ASD in regions belonging to the Default Mode Network and to the Social Brain. CONCLUSIONS: Our results demonstrate that the multimodal joint fusion approach outperforms the classification results obtained with data acquired by a single MRI modality as it efficiently exploits the complementarity of structural and functional brain information.

Effect of data harmonization of multicentric dataset in ASD/TD classification.

Machine Learning (ML) is nowadays an essential tool in the analysis of Magnetic Resonance Imaging (MRI) data, in particular in the identification of brain correlates in neurological and neurodevelopmental disorders. ML requires datasets of appropriate size for training, which in neuroimaging are typically obtained collecting data from multiple acquisition centers. However, analyzing large multicentric datasets can introduce bias due to differences between acquisition centers. ComBat harmonization is commonly used to address batch effects, but it can lead to data leakage when the entire dataset is used to estimate model parameters. In this study, structural and functional MRI data from the Autism Brain Imaging Data Exchange (ABIDE) collection were used to classify subjects with Autism Spectrum Disorders (ASD) compared to Typical Developing controls (TD). We compared the classical approach (external harmonization) in which harmonization is performed before train/test split, with an harmonization calculated only on the train set (internal harmonization), and with the dataset with no harmonization. The results showed that harmonization using the whole dataset achieved higher discrimination performance, while non-harmonized data and harmonization using only the train set showed similar results, for both structural and connectivity features. We also showed that the higher performances of the external harmonization are not due to larger size of the sample for the estimation of the model and hence these improved performance with the entire dataset may be ascribed to data leakage. In order to prevent this leakage, it is recommended to define the harmonization model solely using the train set.

A multicenter evaluation of a deep learning software (LungQuant) for lung parenchyma characterization in COVID-19 pneumonia.

BACKGROUND: The role of computed tomography (CT) in the diagnosis and characterization of coronavirus disease 2019 (COVID-19) pneumonia has been widely recognized. We evaluated the performance of a software for quantitative analysis of chest CT, the LungQuant system, by comparing its results with independent visual evaluations by a group of 14 clinical experts. The aim of this work is to evaluate the ability of the automated tool to extract quantitative information from lung CT, relevant for the design of a diagnosis support model. METHODS: LungQuant segments both the lungs and lesions associated with COVID-19 pneumonia (ground-glass opacities and consolidations) and computes derived quantities corresponding to qualitative characteristics used to clinically assess COVID-19 lesions. The comparison was carried out on 120 publicly available CT scans of patients affected by COVID-19 pneumonia. Scans were scored for four qualitative metrics: percentage of lung involvement, type of lesion, and two disease distribution scores. We evaluated the agreement between the LungQuant output and the visual assessments through receiver operating characteristics area under the curve (AUC) analysis and by fitting a nonlinear regression model. RESULTS: Despite the rather large heterogeneity in the qualitative labels assigned by the clinical experts for each metric, we found good agreement on the metrics compared to the LungQuant output. The AUC values obtained for the four qualitative metrics were 0.98, 0.85, 0.90, and 0.81. CONCLUSIONS: Visual clinical evaluation could be complemented and supported by computer-aided quantification, whose values match the average evaluation of several independent clinical experts. KEY POINTS: We conducted a multicenter evaluation of the deep learning-based LungQuant automated software. We translated qualitative assessments into quantifiable metrics to characterize coronavirus disease 2019 (COVID-19) pneumonia lesions. Comparing the software output to the clinical evaluations, results were satisfactory despite heterogeneity of the clinical evaluations. An automatic quantification tool may contribute to improve the clinical workflow of COVID-19 pneumonia.

Quantification of pulmonary involvement in COVID-19 pneumonia: an upgrade of the LungQuant software for lung CT segmentation.

Computed tomography (CT) scans are used to evaluate the severity of lung involvement in patients affected by COVID-19 pneumonia. Here, we present an improved version of the LungQuant automatic segmentation software (LungQuant v2), which implements a cascade of three deep neural networks (DNNs) to segment the lungs and the lung lesions associated with COVID-19 pneumonia. The first network (BB-net) defines a bounding box enclosing the lungs, the second one (U-net 1 ) outputs the mask of the lungs, and the final one (U-net 2 ) generates the mask of the COVID-19 lesions. With respect to the previous version (LungQuant v1), three main improvements are introduced: the BB-net, a new term in the loss function in the U-net for lesion segmentation and a post-processing procedure to separate the right and left lungs. The three DNNs were optimized, trained and tested on publicly available CT scans. We evaluated the system segmentation capability on an independent test set consisting of ten fully annotated CT scans, the COVID-19-CT-Seg benchmark dataset. The test performances are reported by means of the volumetric dice similarity coefficient (vDSC) and the surface dice similarity coefficient (sDSC) between the reference and the segmented objects. LungQuant v2 achieves a vDSC (sDSC) equal to 0.96 ± 0.01 (0.97 ± 0.01) and 0.69 ± 0.08 (0.83 ± 0.07) for the lung and lesion segmentations, respectively. The output of the segmentation software was then used to assess the percentage of infected lungs, obtaining a Mean Absolute Error (MAE) equal to 2%.

Multi-site harmonization of MRI data uncovers machine-learning discrimination capability in barely separable populations: An example from the ABIDE dataset.

Machine Learning (ML) techniques have been widely used in Neuroimaging studies of Autism Spectrum Disorders (ASD) both to identify possible brain alterations related to this condition and to evaluate the predictive power of brain imaging modalities. The collection and public sharing of large imaging samples has favored an even greater diffusion of the use of ML-based analyses. However, multi-center data collections may suffer the batch effect, which, especially in case of Magnetic Resonance Imaging (MRI) studies, should be curated to avoid confounding effects for ML classifiers and masking biases. This is particularly important in the study of barely separable populations according to MRI data, such as subjects with ASD compared to controls with typical development (TD). Here, we show how the implementation of a harmo- nization protocol on brain structural features unlocks the case-control ML separation capability in the analysis of a multi-center MRI dataset. This effect is demonstrated on the ABIDE data collection, involving subjects encompassing a wide age range. After data harmonization, the overall ASD vs. TD discrimination capability by a Random Forest (RF) classifier improves from a very low performance (AUC = 0.58 ± 0.04) to a still low, but reasonably significant AUC = 0.67 ± 0.03. The performances of the RF classifier have been evaluated also in the age-specific subgroups of children, adolescents and adults, obtaining AUC = 0.62 ± 0.02, AUC = 0.65 ± 0.03 and AUC = 0.69 ± 0.06, respectively. Specific and consistent patterns of anatomical differences related to the ASD condition have been identified for the three different age subgroups.

Quantification of pulmonary involvement in COVID-19 pneumonia by means of a cascade of two U-nets: training and assessment on multiple datasets using different annotation criteria.

PURPOSE: This study aims at exploiting artificial intelligence (AI) for the identification, segmentation and quantification of COVID-19 pulmonary lesions. The limited data availability and the annotation quality are relevant factors in training AI-methods. We investigated the effects of using multiple datasets, heterogeneously populated and annotated according to different criteria. METHODS: We developed an automated analysis pipeline, the LungQuant system, based on a cascade of two U-nets. The first one (U-net[Formula: see text]) is devoted to the identification of the lung parenchyma; the second one (U-net[Formula: see text]) acts on a bounding box enclosing the segmented lungs to identify the areas affected by COVID-19 lesions. Different public datasets were used to train the U-nets and to evaluate their segmentation performances, which have been quantified in terms of the Dice Similarity Coefficients. The accuracy in predicting the CT-Severity Score (CT-SS) of the LungQuant system has been also evaluated. RESULTS: Both the volumetric DSC (vDSC) and the accuracy showed a dependency on the annotation quality of the released data samples. On an independent dataset (COVID-19-CT-Seg), both the vDSC and the surface DSC (sDSC) were measured between the masks predicted by LungQuant system and the reference ones. The vDSC (sDSC) values of 0.95±0.01 and 0.66±0.13 (0.95±0.02 and 0.76±0.18, with 5 mm tolerance) were obtained for the segmentation of lungs and COVID-19 lesions, respectively. The system achieved an accuracy of 90% in CT-SS identification on this benchmark dataset. CONCLUSION: We analysed the impact of using data samples with different annotation criteria in training an AI-based quantification system for pulmonary involvement in COVID-19 pneumonia. In terms of vDSC measures, the U-net segmentation strongly depends on the quality of the lesion annotations. Nevertheless, the CT-SS can be accurately predicted on independent test sets, demonstrating the satisfactory generalization ability of the LungQuant.

Experimental optimization of the energy for breast-CT with synchrotron radiation.

Breast Computed Tomography (bCT) is a three-dimensional imaging technique that is raising interest among radiologists as a viable alternative to mammographic planar imaging. In X-rays imaging it would be desirable to maximize the capability of discriminating different tissues, described by the Contrast to Noise Ratio (CNR), while minimizing the dose (i.e. the radiological risk). Both dose and CNR are functions of the X-ray energy. This work aims at experimentally investigating the optimal energy that, at fixed dose, maximizes the CNR between glandular and adipose tissues. Acquisitions of both tissue-equivalent phantoms and actual breast specimens have been performed with the bCT system implemented within the Syrma-3D collaboration at the Syrmep beamline of the Elettra synchrotron (Trieste). The experimental data have been also compared with analytical simulations and the results are in agreement. The CNR is maximized at energies around 26-28 keV. These results are in line with the outcomes of a previously presented simulation study which determined an optimal energy of 28 keV for a large set of breast phantoms with different diameters and glandular fractions. Finally, a study on photon starvation has been carried out to investigate how far the dose can be reduced still having suitable images for diagnostics.

Dealing with confounders and outliers in classification medical studies: The Autism Spectrum Disorders case study.

Machine learning (ML) approaches have been widely applied to medical data in order to find reliable classifiers to improve diagnosis and detect candidate biomarkers of a disease. However, as a powerful, multivariate, data-driven approach, ML can be misled by biases and outliers in the training set, finding sample-dependent classification patterns. This phenomenon often occurs in biomedical applications in which, due to the scarcity of the data, combined with their heterogeneous nature and complex acquisition process, outliers and biases are very common. In this work we present a new workflow for biomedical research based on ML approaches, that maximizes the generalizability of the classification. This workflow is based on the adoption of two data selection tools: an autoencoder to identify the outliers and the Confounding Index, to understand which characteristics of the sample can mislead classification. As a study-case we adopt the controversial research about extracting brain structural biomarkers of Autism Spectrum Disorders (ASD) from magnetic resonance images. A classifier trained on a dataset composed by 86 subjects, selected using this framework, obtained an area under the receiver operating characteristic curve of 0.79. The feature pattern identified by this classifier is still able to capture the mean differences between the ASD and Typically Developing Control classes on 1460 new subjects in the same age range of the training set, thus providing new insights on the brain characteristics of ASD. In this work, we show that the proposed workflow allows to find generalizable patterns even if the dataset is limited, while skipping the two mentioned steps and using a larger but not well designed training set would have produced a sample-dependent classifier.

Radiochromic film dosimetry in synchrotron radiation breast computed tomography: a phantom study.

This study relates to the INFN project SYRMA-3D for in vivo phase-contrast breast computed tomography using the SYRMEP synchrotron radiation beamline at the ELETTRA facility in Trieste, Italy. This peculiar imaging technique uses a novel dosimetric approach with respect to the standard clinical procedure. In this study, optimization of the acquisition procedure was evaluated in terms of dose delivered to the breast. An offline dose monitoring method was also investigated using radiochromic film dosimetry. Various irradiation geometries have been investigated for scanning the prone patient's pendant breast, simulated by a 14 cm-diameter polymethylmethacrylate cylindrical phantom containing pieces of calibrated radiochromic film type XR-QA2. Films were inserted mid-plane in the phantom, as well as wrapped around its external surface, and irradiated at 38 keV, with an air kerma value that would produce an estimated mean glandular dose of 5 mGy for a 14 cm-diameter 50% glandular breast. Axial scans were performed over a full rotation or over 180°. The results point out that a scheme adopting a stepped rotation irradiation represents the best geometry to optimize the dose distribution to the breast. The feasibility of using a piece of calibrated radiochromic film wrapped around a suitable holder around the breast to monitor the scan dose offline is demonstrated.

Image quality comparison between a phase-contrast synchrotron radiation breast CT and a clinical breast CT: a phantom based study.

In this study we compared the image quality of a synchrotron radiation (SR) breast computed tomography (BCT) system with a clinical BCT in terms of contrast-to-noise ratio (CNR), signal-to-noise ratio (SNR), noise power spectrum (NPS), spatial resolution and detail visibility. A breast phantom consisting of several slabs of breast-adipose equivalent material with different embedded targets (i.e., masses, fibers and calcifications) was used. Phantom images were acquired using a dedicated BCT system installed at the Radboud University Medical Center (Nijmegen, The Netherlands) and the SR BCT system at the SYRMEP beamline of Elettra SR facility (Trieste, Italy) based on a photon-counting detector. Images with the SR setup were acquired mimicking the clinical BCT conditions (i.e., energy of 30 keV and radiation dose of 6.5 mGy). Images were reconstructed with an isotropic cubic voxel of 273 µm for the clinical BCT, while for the SR setup two phase-retrieval (PhR) kernels (referred to as "smooth" and "sharp") were alternatively applied to each projection before tomographic reconstruction, with voxel size of 57 × 57 × 50 µm3. The CNR for the clinical BCT system can be up to 2-times higher than SR system, while the SNR can be 3-times lower than SR system, when the smooth PhR is used. The peak frequency of the NPS for the SR BCT is 2 to 4-times higher (0.9 mm-1 and 1.4 mm-1 with smooth and sharp PhR, respectively) than the clinical BCT (0.4 mm-1). The spatial resolution (MTF10%) was estimated to be 1.3 lp/mm for the clinical BCT, and 5.0 lp/mm and 6.7 lp/mm for the SR BCT with the smooth and sharp PhR, respectively. The smallest fiber visible in the SR BCT has a diameter of 0.15 mm, while for the clinical BCT is 0.41 mm. Calcification clusters with diameter of 0.13 mm are visible in the SR BCT, while the smallest diameter for the clinical BCT is 0.29 mm. As expected, the image quality of the SR BCT outperforms the clinical BCT system, providing images with higher spatial resolution and SNR, and with finer granularity. Nevertheless, this study assesses the image quality gap quantitatively, giving indications on the benefits associated with SR BCT and providing a benchmarking basis for its clinical implementation. In addition, SR-based studies can provide a gold-standard in terms of achievable image quality, constituting an upper-limit to the potential clinical development of a given technique.

Evaluation of Altered Functional Connections in Male Children With Autism Spectrum Disorders on Multiple-Site Data Optimized With Machine Learning.

No univocal and reliable brain-based biomarkers have been detected to date in Autism Spectrum Disorders (ASD). Neuroimaging studies have consistently revealed alterations in brain structure and function of individuals with ASD; however, it remains difficult to ascertain the extent and localization of affected brain networks. In this context, the application of Machine Learning (ML) classification methods to neuroimaging data has the potential to contribute to a better distinction between subjects with ASD and typical development controls (TD). This study is focused on the analysis of resting-state fMRI data of individuals with ASD and matched TD, available within the ABIDE collection. To reduce the multiple sources of heterogeneity that impact on understanding the neural underpinnings of autistic condition, we selected a subgroup of 190 subjects (102 with ASD and 88 TD) according to the following criteria: male children (age range: 6.5-13 years); rs-fMRI data acquired with open eyes; data from the University sites that provided the largest number of scans (KKI, NYU, UCLA, UM). Connectivity values were evaluated as the linear correlation between pairs of time series of brain areas; then, a Linear kernel Support Vector Machine (L-SVM) classification, with an inter-site cross-validation scheme, was carried out. A permutation test was conducted to identify over-connectivity and under-connectivity alterations in the ASD group. The mean L-SVM classification performance, in terms of the area under the ROC curve (AUC), was 0.75 ± 0.05. The highest performance was obtained using data from KKI, NYU and UCLA sites in training and data from UM as testing set (AUC = 0.83). Specifically, stronger functional connectivity (FC) in ASD with respect to TD involve (p < 0.001) the angular gyrus with the precuneus in the right (R) hemisphere, and the R frontal operculum cortex with the pars opercularis of the left (L) inferior frontal gyrus. Weaker connections in ASD group with respect to TD are the intra-hemispheric R temporal fusiform cortex with the R hippocampus, and the L supramarginal gyrus with L planum polare. The results indicate that both under- and over-FC occurred in a selected cohort of ASD children relative to TD controls, and that these functional alterations are spread in different brain networks.

Optimization of the energy for Breast monochromatic absorption X-ray Computed Tomography.

The limits of mammography have led to an increasing interest on possible alternatives such as the breast Computed Tomography (bCT). The common goal of all X-ray imaging techniques is to achieve the optimal contrast resolution, measured through the Contrast to Noise Ratio (CNR), while minimizing the radiological risks, quantified by the dose. Both dose and CNR depend on the energy and the intensity of the X-rays employed for the specific imaging technique. Some attempts to determine an optimal energy for bCT have suggested the range 22 keV-34 keV, some others instead suggested the range 50 keV-60 keV depending on the parameters considered in the study. Recent experimental works, based on the use of monochromatic radiation and breast specimens, show that energies around 32 keV give better image quality respect to setups based on higher energies. In this paper we report a systematic study aiming at defining the range of energies that maximizes the CNR at fixed dose in bCT. The study evaluates several compositions and diameters of the breast and includes various reconstruction algorithms as well as different dose levels. The results show that a good compromise between CNR and dose is obtained using energies around 28 keV.

Advancements towards the implementation of clinical phase-contrast breast computed tomography at Elettra.

Breast computed tomography (BCT) is an emerging application of X-ray tomography in radiological practice. A few clinical prototypes are under evaluation in hospitals and new systems are under development aiming at improving spatial and contrast resolution and reducing delivered dose. At the same time, synchrotron-radiation phase-contrast mammography has been demonstrated to offer substantial advantages when compared with conventional mammography. At Elettra, the Italian synchrotron radiation facility, a clinical program of phase-contrast BCT based on the free-space propagation approach is under development. In this paper, full-volume breast samples imaged with a beam energy of 32 keV delivering a mean glandular dose of 5 mGy are presented. The whole acquisition setup mimics a clinical study in order to evaluate its feasibility in terms of acquisition time and image quality. Acquisitions are performed using a high-resolution CdTe photon-counting detector and the projection data are processed via a phase-retrieval algorithm. Tomographic reconstructions are compared with conventional mammographic images acquired prior to surgery and with histologic examinations. Results indicate that BCT with monochromatic beam and free-space propagation phase-contrast imaging provide relevant three-dimensional insights of breast morphology at clinically acceptable doses and scan times.

Quantitative characterization of breast tissues with dedicated CT imaging.

A quantitative characterization of the soft tissues composing the human breast is achieved by means of a monochromatic CT phase-contrast imaging system, through accurate measurements of their attenuation coefficients within the energy range of interest for breast CT clinical examinations. Quantitative measurements of linear attenuation coefficients are performed on tomographic reconstructions of surgical samples, using monochromatic x-ray beams from a synchrotron source and a free space propagation setup. An online calibration is performed on the obtained reconstructions, in order to reassess the validity of the standard calibration procedure of the CT scanner. Three types of healthy tissues (adipose, glandular, and skin) and malignant tumors, when present, are considered from each sample. The measured attenuation coefficients are in very good agreement with the outcomes of similar studies available in the literature, although they span an energy range that was mostly neglected in the previous studies. No globally significant differences are observed between healthy and malignant dense tissues, although the number of considered samples does not appear sufficient to address the issue of a quantitative differentiation of tumors. The study assesses the viability of the proposed methodology for the measurement of linear attenuation coefficients, and provides a denser sampling of attenuation data in the energy range useful to breast CT.

Monochromatic breast computed tomography with synchrotron radiation: phase-contrast and phase-retrieved image comparison and full-volume reconstruction.

A program devoted to performing the first in vivo synchrotron radiation (SR) breast computed tomography (BCT) is ongoing at the Elettra facility. Using the high spatial coherence of SR, phase-contrast (PhC) imaging techniques can be used. The latest high-resolution BCT acquisitions of breast specimens, obtained with the propagation-based PhC approach, are herein presented as part of the SYRMA-3D collaboration effort toward the clinical exam. Images are acquired with a 60 - µ m pixel dead-time-free single-photon-counting CdTe detector. The samples are imaged at 32 and 38 keV in a continuous rotating mode, delivering 5 to 20 mGy of mean glandular dose. Contrast-to-noise ratio (CNR) and spatial resolution performances are evaluated for both PhC and phase-retrieved images, showing that by applying the phase-retrieval algorithm a five-time CNR increase can be obtained with a minor loss in spatial resolution across soft tissue interfaces. It is shown that, despite having a poorer CNR, PhC images can provide a sharper visualization of microcalcifications, thus being complementary to phase-retrieved images. Furthermore, the first full-volume scan of a mastectomy sample ( 9 × 9 × 3 cm 3 ) is reported. This investigation into surgical specimens indicates that SR BCT in terms of CNR, spatial resolution, scan duration, and scan volume is feasible.

Phase-contrast breast CT: the effect of propagation distance.

X-ray phase imaging has the potential to dramatically improve soft tissue contrast sensitivity, which is a crucial requirement in many diagnostic applications such as breast imaging. In this context, a program devoted to perform in vivo phase-contrast synchrotron radiation breast computed tomography is ongoing at the Elettra facility (Trieste, Italy). The used phase-contrast technique is the propagation-based configuration, which requires a spatially coherent source and a sufficient object-to-detector distance. In this work the effect of this distance on image quality is quantitatively investigated scanning a large breast surgical specimen at three object-to-detector distances (1.6, 3, 9 m) and comparing the images both before and after applying the phase-retrieval procedure. The sample is imaged at 30 keV with a [Formula: see text] pixel pitch CdTe single-photon-counting detector, positioned at a fixed distance of 31.6 m from the source. The detector fluence is kept constant for all acquisitions. The study shows that, at the largest distance, a 20-fold SNR increase can be obtained by applying the phase-retrieval procedure. Moreover, it is shown that, for phase-retrieved images, changing the object-to-detector distance does not affect spatial resolution while boosting SNR (four-fold increase going from the shortest to the largest distance). The experimental results are supported by a theoretical model proposed by other authors, whose salient results are presented in this paper.

Large-area single-photon-counting CdTe detector for synchrotron radiation computed tomography: a dedicated pre-processing procedure.

Large-area CdTe single-photon-counting detectors are becoming more and more attractive in view of low-dose imaging applications due to their high efficiency, low intrinsic noise and absence of a scintillating screen which affects spatial resolution. At present, however, since the dimensions of a single sensor are small (typically a few cm2), multi-module architectures are needed to obtain a large field of view. This requires coping with inter-module gaps and with close-to-edge pixels, which generally show a non-optimal behavior. Moreover, high-Z detectors often show gain variations in time due to charge trapping: this effect is detrimental especially in computed tomography (CT) applications where a single tomographic image requires hundreds of projections continuously acquired in several seconds. This work has been carried out at the SYRMEP beamline of the Elettra synchrotron radiation facility (Trieste, Italy), in the framework of the SYRMA-3D project, which aims to perform the world's first breast-CT clinical study with synchrotron radiation. An ad hoc data pre-processing procedure has been developed for the PIXIRAD-8 CdTe single-photon-counting detector, comprising an array of eight 30.7 mm × 24.8 mm modules tiling a 246 mm × 25 mm sensitive area, which covers the full synchrotron radiation beam. The procedure consists of five building blocks, namely dynamic flat-fielding, gap seaming, dynamic ring removal, projection despeckling and around-gap equalization. Each block is discussed and compared, when existing, with conventional approaches. The effectiveness of the pre-processing is demonstrated for phase-contrast CT images of a human breast specimen. The dynamic nature of the proposed procedure, which provides corrections dependent upon the projection index, allows the effective removal of time-dependent artifacts, preserving the main image features including phase effects.

Gray Matter Alterations in Young Children with Autism Spectrum Disorders: Comparing Morphometry at the Voxel and Regional Level.

BACKGROUND AND PURPOSE: Sophisticated algorithms to infer disease diagnosis, pathology progression and patient outcome are increasingly being developed to analyze brain MRI data. They have been successfully implemented in a variety of diseases and are currently investigated in the field of neuropsychiatric disorders, including autism spectrum disorder (ASD). We aim to test the ability to predict ASD from subtle morphological changes in structural magnetic resonance imaging (sMRI). METHODS: The analysis of sMRI of a cohort of male ASD children and controls matched for age and nonverbal intelligence quotient (NVIQ) has been carried out with two widely used preprocessing software packages (SPM and Freesurfer) to extract brain morphometric information at different spatial scales. Then, support vector machines have been implemented to classify the brain features and to localize which brain regions contribute most to the ASD-control separation. RESULTS: The features extracted from the gray matter subregions provide the best classification performance, reaching an area under the receiver operating characteristic curve (AUC) of 74%. This value is enhanced to 80% when considering only subjects with NVIQ over 70. CONCLUSIONS: Despite the subtle impact of ASD on brain morphology and a limited cohort size, results from sMRI-based classifiers suggest a consistent network of altered brain regions.

Automatic lung segmentation in CT images with accurate handling of the hilar region.

A fully automated and three-dimensional (3D) segmentation method for the identification of the pulmonary parenchyma in thorax X-ray computed tomography (CT) datasets is proposed. It is meant to be used as pre-processing step in the computer-assisted detection (CAD) system for malignant lung nodule detection that is being developed by the Medical Applications in a Grid Infrastructure Connection (MAGIC-5) Project. In this new approach the segmentation of the external airways (trachea and bronchi), is obtained by 3D region growing with wavefront simulation and suitable stop conditions, thus allowing an accurate handling of the hilar region, notoriously difficult to be segmented. Particular attention was also devoted to checking and solving the problem of the apparent 'fusion' between the lungs, caused by partial-volume effects, while 3D morphology operations ensure the accurate inclusion of all the nodules (internal, pleural, and vascular) in the segmented volume. The new algorithm was initially developed and tested on a dataset of 130 CT scans from the Italung-CT trial, and was then applied to the ANODE09-competition images (55 scans) and to the LIDC database (84 scans), giving very satisfactory results. In particular, the lung contour was adequately located in 96% of the CT scans, with incorrect segmentation of the external airways in the remaining cases. Segmentation metrics were calculated that quantitatively express the consistency between automatic and manual segmentations: the mean overlap degree of the segmentation masks is 0.96 ± 0.02, and the mean and the maximum distance between the mask borders (averaged on the whole dataset) are 0.74 ± 0.05 and 4.5 ± 1.5, respectively, which confirms that the automatic segmentations quite correctly reproduce the borders traced by the radiologist. Moreover, no tissue containing internal and pleural nodules was removed in the segmentation process, so that this method proved to be fit for the use in the framework of a CAD system. Finally, in the comparison with a two-dimensional segmentation procedure, inter-slice smoothness was calculated, showing that the masks created by the 3D algorithm are significantly smoother than those calculated by the 2D-only procedure.

A novel multithreshold method for nodule detection in lung CT.

Multislice computed tomography (MSCT) is a valuable tool for lung cancer detection, thanks to its ability to identify noncalcified nodules of small size (from about 3 mm). Due to the large number of images generated by MSCT, there is much interest in developing computer-aided detection (CAD) systems that could assist radiologists in the lung nodule detection task. A complete multistage CAD system, including lung boundary segmentation, regions of interest (ROIs) selection, feature extraction, and false positive reduction is presented. The selection of ROIs is based on a multithreshold surface-triangulation approach. Surface triangulation is performed at different threshold values, varying from a minimum to a maximum value in a wide range. At a given threshold value, a ROI is defined as the volume inside a connected component of the triangulated isosurface. The evolution of a ROI as a function of the threshold can be represented by a treelike structure. A multithreshold ROI is defined as a path on this tree, which starts from a terminal ROI and ends on the root ROI. For each ROI, the volume, surface area, roundness, density, and moments of inertia are computed as functions of the threshold and used as input to a classification system based on artificial neural networks. The method is suitable to detect different types of nodules, including juxta-pleural nodules and nodules connected to blood vessels. A training set of 109 low-dose MSCT scans made available by the Pisa center of the Italung-CT trial and annotated by expert radiologists was used for the algorithm design and optimization. The system performance was tested on an independent set of 23 low-dose MSCT scans coming from the Pisa Italung-CT center and on 83 scans made available by the Lung Image Database Consortium (LIDC) annotated by four expert radiologists. On the Italung-CT test set, for nodules having a diameter greater than or equal to 3 mm, the system achieved 84% and 71% sensitivity at false positive/scan rates of 10 and 4, respectively. For nodules having a diameter greater than or equal to 4 mm, the sensitivities were 97% and 80% at false positive/scan rates of 10 and 4, respectively. On the LIDC data set, the system achieved a 79% sensitivity at a false positive/scan rate of 4 in the detection of nodules with a diameter greater than or equal to 3 mm that have been annotated by all four radiologists.