Purpose To develop and validate a semi-quantification method (time-delayed ratio, TDr) applied to amyloid PET scans, based on tracer kinetics information. Methods The TDr method requires two static scans per subject: one early (~0-10 min after the injection) and one late (typically 50-70 min or 90-100 min after the injection, depending on the tracer). High perfusion regions are delineated on the early scan and applied onto the late scan. A SUVr-like ratio is calculated between the average intensities in the high perfusion regions and the late scan hotspot. TDr was applied to a naturalistic multicenter dataset of 143 subjects acquired with [ 18 F]florbetapir. TDr values are compared to visual evaluation, cortical-cerebellar SUVr, and to the geometrical semi-quantification method ELBA. All three methods are gauged versus the heterogeneity of the dataset.Results TDr shows excellent agreement with respect to the binary visual assessment (AUC = 0.99) and significantly correlates with both validated semi-quantification methods, reaching a Pearson correlation coefficient of 0.86 with respect to ELBA. Conclusions TDr is an alternative approach to previously validated ones (SUVr and ELBA). It requires minimal image processing; it is independent on predefined regions of interest and does not require MR registration. Besides, it takes advantage on the availability of early scans which are becoming common practice while imposing a negligible added patient discomfort.

The large amount of data produced by the ATLAS experiment needs new computing paradigms for data processing and analysis, which involve many computing centres spread around the world. The computing workload is managed by regional federations, called "clouds". The Italian cloud consists of a main (Tier-1) center, located in Bologna, four secondary (Tier-2) centers, and a few smaller (Tier-3) sites. In this contribution we describe the Italian cloud facilities and the activities of data processing, analysis, simulation and software development performed within the cloud, and we discuss the tests of the new computing technologies contributing to evolution of the ATLAS Computing Model.

Abstract:Research in Alzheimer's disease (AD) has seen a tremendous growth of candidate biomarkers in the last decade. The role of such established or putative biomarkers is to allow an accurate diagnosis of AD, to infer about its prognosis, to monitor disease progression and evaluate changes induced by disease-modifying drugs. An ideal biomarker should detect a specific pathophysiological feature of AD, not present in the healthy condition, in other primary dementias, or in confounding conditions. Besides being reliable, a biomarker should be detectable by means of procedures which must be relatively non-invasive, simple to perform, widely available and not too expensive. At present, no candidate meets these requirements representing the high standards aimed at by researchers. Among others, various morphological brain measures performed by means of magnetic resonance imaging (MRI), ranging from the total brain volume to some restricted regions such as the hippocampal volume, have been proposed. Nowadays the efforts are directed toward finding an automated, unsupervised method of evaluating atrophy in some specific brain region, such as the medial temporal lobe (MTL). In this work we provide an extensive review of the state of the art on the automatic and semi-automatic image processing techniques for the early assessment of patients at risk of developing AD. Our main focus is the relevance of the morphological analysis of MTL, and in particular of the hippocampal formation, in making the diagnosis of AD and in distinguishing it from other dementias.

No abstract

No abstract

No abstract

No abstract

We present the development of a web-based interface (proAD) for the automatic analysis of structural T1-weighted magnetic resonance images (MRI). This web-based tool is meant to be a high level interface to a sophisticated analysis environment aimed at the early diagnosis of Alzheimer's disease patients. The analysis procedure relies on clinical diagnosis and follow up data. It extracts several portions of the temporal lobe, mainly around the hippocampal area, and processes them with no interactive input from the user. The output of the MRI analysis is a classification index which is able to distinguish among: (a) patients with Alzheimer's disease (AD), (b) patients with amnestic mild cognitive impairment (aMCI), (c) patients with amnestic mild cognitive impairment who are likely to develop AD in a time frame of 2 years (aMCIconv), and (c) normal cognitive subjects. The interface is still under development and has been built as easy and user-friendly as possible. Robustness has been considered as essential part for the development. Results are presented to users in a very simple manner, thereby serving as a potential diagnostic tool for neurologists.A working testbed of proAD (still in alpha stage) can be reached at http://magic5gui.ge.infn.it/tesi