Search

Introduction. The Ph.D. thesis addresses the development of innovative techniques to create 3D-printed anatomical brain phantoms, which can be used for quantitative technical assessments on morpho-functional imaging devices, providing simulation accuracy not obtainable with currently available phantoms. 3D printing (3DP) technology is paving the way for advanced anatomical modelling in biomedical applications. Despite the potential already expressed by 3DP in this field, it is still little used for the realization of anthropomorphic phantoms of human organs with complex internal structures. Making an anthropomorphic phantom is very different from making a simple anatomical model and 3DP is still far from being plug-and-print. Hence, the need to develop ad-hoc techniques providing innovative solutions for the realization of anatomical phantoms with unique characteristics, and greater ease-of-use. Aim. The thesis explores the entire workflow (brain MRI images segmentation, 3D modelling and materialization) developed to prototype a new complex anthropomorphic brain phantom, which can simulate three brain compartments simultaneously: grey matter (GM), white matter (WM) and striatum (caudate nucleus and putamen, known to show a high uptake in nuclear medicine studies). The three separate chambers of the phantom will be filled with tissue-appropriate solutions characterized by different concentrations of radioisotope for PET/SPECT, para-/ferro-magnetic metals for MRI, and iodine for CT imaging. Methods. First, to design a 3D model of the brain phantom, it is necessary to segment MRI images and to extract an error-less STL (Standard Tessellation Language) description. Then, it is possible to materialize the prototype and test its functionality. - Image segmentation. Segmentation is one of the most critical steps in modelling. To this end, after demonstrating the proof-of-concept, a multi-parametric segmentation approach based on brain relaxometry was proposed. It includes a pre-processing step to estimate relaxation parameter maps (R1 = longitudinal relaxation rate, R2 = transverse relaxation rate, PD = proton density) from the signal intensities provided by MRI sequences of routine clinical protocols (3D-GrE T1-weighted, FLAIR and fast-T2-weighted sequences with ≤ 3 mm slice thickness). In the past, maps of R1, R2, and PD were obtained from Conventional Spin Echo (CSE) sequences, which are no longer suitable for clinical practice due to long acquisition times. Rehabilitating the multi-parametric segmentation based on relaxometry, the estimation of pseudo-relaxation maps allowed developing an innovative method for the simultaneous automatic segmentation of most of the brain structures (GM, WM, cerebrospinal fluid, thalamus, caudate nucleus, putamen, pallidus, nigra, red nucleus and dentate). This method allows the segmentation of higher resolution brain images for future brain phantom enhancements. - STL extraction. After segmentation, the 3D model of phantom is described in STL format, which represents the shapes through the approximation in manifold mesh (i.e., collection of triangles, which is continuous, without holes and with a positive – not zero – volume). For this purpose, we developed an automatic procedure to extract a single voxelized surface, tracing the anatomical interface between the phantom's compartments directly on the segmented images. Two tubes were designed for each compartment (one for filling and the other to facilitate the escape of air). The procedure automatically checks the continuity of the surface, ensuring that the 3D model could be exported in STL format, without errors, using a common image-to-STL conversion software. Threaded junctions were added to the phantom (for the hermetic closure) using a mesh processing software. The phantom's 3D model resulted correct and ready for 3DP. Prototyping. Finally, the most suitable 3DP technology is identified for the materialization. We investigated the material extrusion technology, named Fused Deposition Modeling (FDM), and the material jetting technology, named PolyJet. FDM resulted the best candidate for our purposes. It allowed materializing the phantom's hollow compartments in a single print, without having to print them in several parts to be reassembled later. FDM soluble internal support structures were completely removable after the materialization, unlike PolyJet supports. A critical aspect, which required a considerable effort to optimize the printing parameters, was the submillimetre thickness of the phantom walls, necessary to avoid distorting the imaging simulation. However, 3D printer manufacturers recommend maintaining a uniform wall thickness of at least 1 mm. The optimization of printing path made it possible to obtain strong, but not completely waterproof walls, approximately 0.5 mm thick. A sophisticated technique, based on the use of a polyvinyl-acetate solution, was developed to waterproof the internal and external phantom walls (necessary requirement for filling). A filling system was also designed to minimize the residual air bubbles, which could result in unwanted hypo-intensity (dark) areas in phantom-based imaging simulation. Discussions and conclusions. The phantom prototype was scanned trough CT and PET/CT to evaluate the realism of the brain simulation. None of the state-of-the-art brain phantoms allow such anatomical rendering of three brain compartments. Some represent only GM and WM, others only the striatum. Moreover, they typically have a poor anatomical yield, showing a reduced depth of the sulci and a not very faithful reproduction of the cerebral convolutions. The ability to simulate the three brain compartments simultaneously with greater accuracy, as well as the possibility of carrying out multimodality studies (PET/CT, PET/MRI), which represent the frontier of diagnostic imaging, give this device cutting-edge prospective characteristics. The effort to further customize 3DP technology for these applications is expected to increase significantly in the coming years.

Relaxation parameter maps (RPMs) calculated from spin-echo data have provided a basis for the segmentation of normal brain tissues and white matter lesions in multiple sclerosis (MS) MRI studies. However, Conventional Spin-Echo (CSE) sequences, once the core of clinical MRI studies, have been largely replaced by faster ones, which do not allow the calculation a-posteriori of RPMs from clinical studies. Aim of the study was to develop and validate a method to estimate RPMs (pseudo-RPMs) from routine clinical MRI protocols (including 3D-Gradient Echo T1w, FLAIR and fast-T2w sequences), suitable for fully automatic multiparametric segmentation of normal-appearing and pathological brain tissues in MS.The proposed method processes spatially normalized clinical MRI studies through a multistep pipeline, to collect a set of data points of matched signal intensities (from MRI studies) and relaxation parameters (from a CSE-derived digital template and an MS lesion database), which are then fitted by a multiple and multivariate 4-th degree polynomial regression, providing pseudo-RPMs. The method was applied to a dataset of 59 clinical MRI studies providing pseudo-RPMs that were segmented through a method originally developed for the CSE-derived RPMs. Results of the segmentation in 12 studies were used to iteratively optimize method parameters. Accuracy of segmentation of normal-appearing brain tissues from the pseudo-RPMs was assessed by comparing their age-related changes, as measured in 47 clinical studies, against those measured acquired using CSE sequences in a comparable dataset of 47 patients. Lesion segmentation was validated against manual segmentation carried out by three neuroradiologists.Age-related changes of normal-appearing brain tissue volumes measured using the pseudo-RPMs substantially overlapped those measured using the RPMs obtained from CSE sequences, and segmentation of MS lesions showed a moderate-high spatial overlap with manual segmentation, comparable to that achieved by the widely used Lesion Segmentation Tool on FLAIR images, with a greater volumetric agreement.The proposed approach allows calculation from clinical studies of pseudo-RPMs, which are equivalent to those obtainable from CSE sequences, avoiding the need for the acquisition of additional, dedicated sequences for segmentation purposes.

Two structurally connected brain regions are more likely to interact, with the lengths of the structural bundles, their widths, myelination, and the topology of the structural connectome influencing the timing of the interactions. We introduce anin vivoapproach for measuring functional delays across the whole brain in humans (of either sex) using magneto/electroencephalography (MEG/EEG) and integrating them with the structural bundles. The resulting topochronic map of the functional delays/velocities shows that larger bundles have faster velocities. We estimated the topochronic map in multiple sclerosis patients, who have damaged myelin sheaths, and controls, demonstrating greater delays in patients across the network and that structurally lesioned tracts were slowed down more than unaffected ones. We provide a novel framework for estimating functional transmission delaysin vivoat the single-subject and single-tract level.SIGNIFICANCE STATEMENTThis article provides a straightforward way to estimate patient-specific delays and conduction velocities in the CNS, at the individual level, in healthy and diseased subjects. To do so, it uses a principled way to merge magnetoencephalography (MEG)/electroencephalography (EEG) and tractography.

Comparison of derivative operators via quantitative performance analysis is rarely addressed in medical imaging. Indeed, the main application of such operators is the extraction of edges and, since there is no unequivocal definition of edges, the common trend is to identify the best performing operator based on a qualitative match between the extracted edges and the fickle human perception of object boundaries. This study presents an objective comparison of four first-order derivative operators through quantitative analysis of results yielded in a specific task, i.e. a spine kinematics application. Such application is based on a template matching method, which estimates common kinematic parameters of intervertebral segments from an X-ray fluoroscopy sequence of spine motion, by operating on the image derivatives of each frame. Therefore, differences in image derivatives, computed via different derivative operators, may lead to differences in estimated parameters of intervertebral kinematics. The comparison presented in this study focused on the trajectory of the instantaneous center of rotation (ICR) of an intervertebral segment, as it is particularly sensitive even to very small differences in displacements and velocities. Therefore, a quantitative analysis of the discrepancies between the ICR trajectories, obtained with each of the four considered derivative operators, was carried out by defining quantitative measures. The results showed detectable differences in the obtained ICR trajectories, thus highlighting the need for quantitative analysis of derivative operator performances in applications aimed at providing quantitative results. However, the significance level of such differences for clinical applications should be further assessed, but, currently, it is not possible, as there is no consensus and sufficient data on kinematic parameters features associated with specific spinal pathologies.

Background Tasting is a complex process involving chemosensory perception and cognitive evaluation. Different experimental designs and solution delivery approaches may in part explain the variability reported in literature. These technical aspects certainly limit the development of taste-related brain computer interface devices. New Method We propose a novel modular, scalable and low-cost device for rapid injection of small volumes of taste solutions during fMRI experiments that gathers the possibility to flexibly increase the number of channels, allowing complex multi-dimensional taste experiments. We provide the full description of the hardware and software architecture and illustrate the application of the working prototype in single-subject event-related fMRI experiments by showing the BOLD responses to basic taste qualities and to five intensities of tastes during the course of perception. Results The device is shown to be effective in activating multiple clusters within the gustatory pathway and a precise time-resolved event-related analysis is shown to be possible by the impulsive nature of the induced perception. Comparison with Existing Method(s) This gustometer represents the first implementation of a low-cost, easily replicable and portable device that is suitable for all kinds of fMRI taste experiments. Its scalability will boost the experimental design of more complex multi-dimensional fMRI studies of the human taste pathway. Conclusions The gustometer represents a valid open-architecture alternative to other available devices and its spread and development may contribute to an increased standardization of experimental designs in human fMRI studies of taste perception and pave the way to the development of novel taste-related BCIs.