Giorgio De Nunzio

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.

The Channeler Ant Model: Object segmentation with virtual ant colonies

3-D object segmentation is an important and challenging topic in computer vision that could be tackled with artificial life models. A Channeler Ant Model (CAM), based on the natural ant capabilities of dealing with 3-D environments through self-organization and emergent behaviours, is proposed. Ant colonies, defined in terms of moving, pheromone laying, reproduction, death and deviating behaviours rules, is able to segment artificially generated objects of different shape, intensity, background. The model depends on few parameters and provides an elegant solution for the segmentation of 3-D structures in noisy environments with unknown range of image intensities: even when there is a partial overlap between the intensity and noise range, it provides a complete segmentation with negligible contamination (i.e., fraction of segmented voxels that do not belong to the object).

Modeling the microscopic electrical properties of thrombin binding aptamer (TBA) for label-free biosensors

Aptamers are chemically produced oligonucleotides, able to bind a variety of targets such as drugs, proteins and pathogens with high sensitivity and selectivity. Therefore, aptamers are largely employed for producing label-free biosensors (aptasensors), with significant applications in diagnostics and drug delivery. In particular, the anti-thrombin aptamers are biomolecules of high interest for clinical use, because of their ability to recognize and bind the thrombin enzyme. Among them, the DNA 15-mer aptamer (TBA), has been widely explored around the possibility of using it in aptasensors. This paper proposes a microscopic model of the electrical properties of TBA and of the aptamer-thrombin complex, combining information from both structure and function, following the issues addressed in an emerging branch of electronics known as proteotronics. The theoretical results are compared and validated with measurements reported in the literature. Finally, the model suggests resistance measurements as a novel tool for testing aptamer-target affinity.

An innovative lung segmentation algorithm in CT images with accurate delimitation of the hilus pulmonis

This paper proposes a new segmentation method for the delimitation of the lung parenchyma in thorax Computed-Tomography (CT) datasets, which will be used as pre-processing step in the CAD (Computer Assisted Detection) system for lung nodule detection that is being developed by the MAGIC-5 (Medical Applications in a Grid Infrastructure Connection) Collaboration. Once finished, the CAD software will run in an integrated “grid” environment, where the potentiality of distributed resources for both data and computation will be exploited. The algorithm is fully automated and three-dimensional (3D). Its most innovative part - to the best of our knowledge - is the segmentation of the external airways (trachea and bronchi), obtained by 3D region growing with wavefront simulation and suitable stop conditions. Another original element is the technique used to check and solve the problem of the apparent ‘fusion’ between the lungs, caused by partial volume effects. A general overview of the algorithm is given, with some details of the innovative parts. The results of its application to a database of about 130 high-resolution low-dose images are discussed.

Automatic segmentation and therapy follow-up of cerebral glioma in diffusion-tensor images

Gliomas are the most common primary brain tumors, with a typical infiltrative growth pattern along white matter (WM) fibers. Diffusion Tensor Imaging (DTI) is sensitive to the directional diffusion of water along WM tracts, which allows the identification of subtle peritumoral glioma infiltration that are not apparent on conventional Magnetic Resonance imaging. The aim of this study was to characterize pathological and healthy tissue in DTI datasets by statistical texture analysis, developing a Computer Assisted Detection (CAD) technique for cerebral glioma. This system, coupled to voxel-based tumor evolution analysis, could allow objective tumor identification and qualitative and quantitative measurements in the follow-up of patients during chemotherapy. In this paper, preliminary results of tumor segmentation and evolution analysis are shown.

Thrombin Aptamer-Based Biosensors: A Model of the Electrical Response

Aptamers are target specific single stranded DNA, RNA or peptide sequences having the ability to bind a variety of proteins, molecules and also ions. Aptasensors, sensors based on aptamers, are at the frontier of sensing technology, mainly in diagnosis and therapy. They appear to be competitive with traditional sensors due to the possibility of detecting and measuring very low concentrations of many different ligands, whose detection and quantification are usually complex, expensive and time-consuming. In this paper we report about a thrombin aptasensor, able to resolve concentrations in a range of 6 orders of magnitude, and provide the microscopic interpretation of its electrical response on the basis of a single macromolecule approach. This investigation has been performed in the framework of an emerging branch of electronics devoted to proteins and living matter, also known as proteotronics.

On-demand lung CT analysis with the M5L-CAD via the WIDEN front-end web interface and an OpenNebula-based cloud back-end

The development of algorithms for the analysis of medical images has been progressively growing over the past two decades. The most common approach is the deployment of standalone workstations, equipped with provider-dependent Graphic User Interfaces (GUI) from which the algorithm execution is triggered interactively. There are, however, several drawbacks: among them, the GUI development cost, the GUI learning curve for the users, the high fixed cost of the software licenses, the difficulty in upgrading the software release. For a few years, the hypothesis of using Grid Services has been explored by several research groups. It turned out that there were other drawbacks: the high costs and security risks of integrating computing resources of medical centers into a Grid Computing Infrastructure. The emerging of Cloud computing, accessible via secure Web protocols, solves most - if not all - the problems. In the specific case of lung Computer Assisted Detection, a further important reason favors the SaaS (Software as a Service) approach: it was demonstrated by several works that combining CAD algorithms improves the overall performance. The system we present is composed by three main building blocks: WIDEN (Web-based Image and Diagnosis Exchange Network) handles the workflow, the image upload and the CAD result notification; the OpenNebula-based cloud IaaS (Infrastructure as a Service) batch farm allocates virtual computing and storage resources; the M5L CAD provides the nodule detection functionality. Our proposed implementation securely handles sensitive patient data, since images are transferred with the HTTPS protocol and the underlying virtual batch farm is isolated. Moreover it is efficient since it dynamically backend.

Lung uniformization for juxta-pleural nodule detection

We propose a method for automatic lung juxta-pleural nodule detection in thorax CT images, to be used as a Computer Assisted Detection (CAD) tool by radiologists. It is based on the calculation and automatic analysis of local curvature on the lung surface as extracted from high-resolution CT scans, and exploits uniformization to a sphere (e.g. through conformal mapping) to allow a global view of the lung surface, with marking of high curvature regions which can be suspected of being pleural nodules. Schematically, the tool works as follows. First, lung binary masks are extracted from the image by 3D segmentation of the CT scan. On these masks, pleural nodules appear as small surface concavities of the mask surface. After patching the entrance of vessels into the parenchyma in the hilus pulmonis, the lung frontier Σ is a smooth genus-0 surface. This surface is triangulated and is then uniformized to a sphere Σ′. In this parameterization a suitable function Ψ of the mean and the Gaussian curvatures can be calculated over Σ. Function Ψ is displayed as a colour variation onto both Σ and Σ′, so marking regions that represent high-curvature concavities. A threshold on Ψ is then applied and regions of interest (ROIs), containing little concavities with a low radius of curvature (such as pleural nodules), are detected. ROIs are then examined and classified; techniques such as spherical wavelets are available on the sphere, which will be used to distinguish between false and true positives, helping in diagnosing pleural nodules.

Integrated model for the analysis of two-dimensional electrophore is gel image

Proteomics is the science that studies the proteome, that is the proteic expression of the genome. Cell proteome is extremely complex, and is composed of several thousand proteins. Twodimensional polyacrylamide gel electrophoresis (2DPAGE) is widely used as a standard method to separate and display proteins in a tissue or compound with a theoretical resolution of 104 proteins simultaneously. This technique combines the resolution power of isoelectrofocalization (IEF), which distinguishes proteins by their isoelectric point (pI), with SDSPAGE (sodium dodecyl sulphate PAGE), in which proteins are separated according to their weight and molecular size. Our group is developing software algorithms for the automatic analysis of images obtained by 2DPAGE gel optical scanning; our aim is the reduction of human intervention in the analysis process, currently quite slow, operatordependent, and prone to errors. In our approach, image noise is first reduced, in order to limit false positives and protein missing. Proteins appear as dark spots on a light background, so the next step is local minima search. Then, the watershed transform is applied, which partitions the gel image into basins: each basin contains a single (recognized) minimum, but can possibly include more than one protein spot if lessdeep minima are masked by the main one. At this point, we perform a registration between the work image and an atlas image (already analyzed by a biologist), and map the atlas spot positions to the work image. Each basin is then used as a region of interest (ROI) in which the shape of the spot (or spots) is fit to a model through a χ2-minimization procedure. The coordinates of the transformed spots are used as the fit initialization parameters.

MRDTI: a semi-automated algorithm to identify damaged brain areas from fractional anisotropy maps

Aim of this study was to analyse diffusion tensor imaging (DTI) datasets in order to identify damaged areas or disorders of the brain in a semi-automatic way. For this purpose, a software tool has been developed: it takes in input the fractional anisotropy (FA) map of a (damaged) brain and, after several steps involving the comparison between the two brain hemispheres, it gives back, as output, a binary mask with a ROI (Region of Interest) that shows the probably damaged area. In the same way, starting from the MR image without diffusion weighting (b0), we find another ROI that we compare with the one previously detected from the FA map. Then we overlay these ROIs onto both the FA map and the image without diffusion weighting, trying to quantify how well the ROIs cover the pathological tissue. This procedure was repeated on a few patients (healthy and pathological ones). The algorithm worked well, showing as a preliminary result that FA maps allow a neater detection of the pathological tissue if compared to MR images without diffusion weighting.