Publications (14)1.07 Total impact
-
Article: Pleural nodule identification in low-dose and thin-slice lung computed tomography
[show abstract] [hide abstract]
ABSTRACT: A completely automated system for the identification of pleural nodules in low-dose and thin-slice computed tomography (CT) of the lung has been developed. The directional-gradient concentration method has been applied to the pleura surface and combined with a morphological opening -based procedure to generate a list of nodule candidates. Each nodule candidate is characterized by 12 morphological and textural features, which are analyzed by a rule-based filter and a neural classifier. This detection system has been developed and validated on a dataset of 42 annotated CT scans. The k -fold cross validation has been used to evaluate the neural classifier performance. The system performance variability due to different ground truth agreement levels is discussed. In particular, the poor 44% sensitivity obtained on the ground truth with agreement level 1 (nodules annotated by only one radiologist) with six FP per scan grows up to the 72% if the underlying ground truth is changed to the agreement level 2 (nodules annotated by two radiologists). -
Article: The CALMA project
[show abstract] [hide abstract]
ABSTRACT: The CALMA (Computer_Assisted Library for MAmmography) project was born as a collaboration between radiologists and physicists. Its goal is to collect a set of digital mammographic images and to work out a suitable Computed Assisted Diagnosis tool to be used in screening mammography. The project is an Italian collaboration among the Istituto Nazionale di Fisica Nucleare sections of Bologna, Pisa, Torino, Udine and Trento and the Hospitals of Bari, Bologna, Livorno, Udine and Trento. Some preliminary results obtained in the classification of breast disease are described here. -
Article: A Completely automated CAD system for mass detection in a large mammographic database
[show abstract] [hide abstract]
ABSTRACT: Mass localization plays a crucial role in computer-aided detection (CAD) systems for the classification of suspicious regions in mammograms. In this article we present a completely automated classification system for the detection of masses in digitized mammographic images. The tool system we discuss consists in three processing levels: (a) Image segmentation for the localization of regions of interest (ROIs). This step relies on an iterative dynamical threshold algorithm able to select iso-intensity closed contours around gray level maxima of the mammogram. (b) ROI characterization by means of textural features computed from the gray tone spatial dependence matrix (GTSDM), containing second-order spatial statistics information on the pixel gray level intensity. As the images under study were recorded in different centers and with different machine settings, eight GTSDM features were selected so as to be invariant under monotonic transformation. In this way, the images do not need to be normalized, as the adopted features depend on the texture only, rather than on the gray tone levels, too. (c) ROI classification by means of a neural network, with supervision provided by the radiologist’s diagnosis. The CAD system was evaluated on a large database of 3369 mammographic images [2307 negative, 1062 pathological (or positive), containing at least one confirmed mass, as diagnosed by an expert radiologist]. To assess the performance of the system, receiver operating characteristic (ROC) and free-response ROC analysis were employed. The area under the ROC curve was found to be A<sub>z</sub>=0.783±0.008 for the ROI-based classification. When evaluating the accuracy of the CAD against the radiologist-drawn boundaries, 4.23 false positives per image are found at 80% of mass sensitivity. -
Article: Mammogram segmentation by contour searching and mass lesions classification with neural network
[show abstract] [hide abstract]
ABSTRACT: The mammography is the most effective procedure for an early diagnosis of the breast cancer. In this paper, an algorithm for detecting masses in mammographic images will be presented. The database consists of 3762 digital images acquired in several hospitals belonging to the MAGIC-5 collaboration (Medical Applications on a Grid Infrastructure Connection). A reduction of the whole image's area under investigation is achieved through a segmentation process, by means of a ROI Hunter algorithm, without loss of meaningful information. In the following classification step, feature extraction plays a fundamental role: some features give geometrical information, other ones provide shape parameters. Once the features are computed for each ROI, they are used as inputs to a supervised neural network with momentum. The output neuron provides the probability that the ROI is pathological or not. Results are provided in terms of ROC and FROC curves: the area under the ROC curve was found to be A<sub>z</sub>=0.862plusmn0.007, and we get a 2.8 FP/Image at a sensitivity of 82%. This software is included in the CAD station actually working in the hospitals belonging to the MAGIC-5 Collaboration -
Article: Distributed medical images analysis on a Grid infrastructure
[show abstract] [hide abstract]
ABSTRACT: In this paper medical applications on a Grid infrastructure, the MAGIC-5 Project, are presented and discussed. MAGIC-5 aims at developing Computer Aided Detection (CADe) software for the analysis of medical images on distributed databases by means of GRID Services. The use of automated systems for analyzing medical images improves radiologists’ performance; in addition, it could be of paramount importance in screening programs, due to the huge amount of data to check and the cost of related manpower. The need for acquiring and analyzing data stored in different locations requires the use of Grid Services for the management of distributed computing resources and data. Grid technologies allow remote image analysis and interactive online diagnosis, with a relevant reduction of the delays presently associated with the diagnosis in the screening programs. The MAGIC-5 project develops algorithms for the analysis of mammographies for breast cancer detection, Computed- Tomography (CT) images for lung cancer detection and Positron Emission Tomography (PET) images for the early diagnosis of Alzheimer Disease (AD). A Virtual Organization (VO) has been deployed, so that authorized users can share data and resources and implement the following use cases: screening, tele-training and tele-diagnosis for mammograms and lung CT scans, statistical diagnosis by comparison of candidates to a distributed data-set of negative PET scans for the diagnosis of the AD. A small-scale prototype of the required Grid functionality was already implemented for the analysis of digitized mammograms. -
Article: FLUXEN portable equipment for direct x-ray spectra measurements
[show abstract] [hide abstract]
ABSTRACT: The proper use of imaging equipment in radiological units is based on an appropriate knowledge of the physical characteristics of the X-ray beam used. The FLUXEN PROJECT is working on a portable apparatus which, together with dedicated software, is able to perform an exact spectral reconstruction of the radiation produced in diagnostic X-ray tubes. The apparatus characterizes the energy spectrum of radiological tubes and also provides a measurement of the emitted flux. The acquisition system is based on a commercial CZT detector (3×3×2 mm<sup>3</sup>), produced by AMPTEK, cooled by a Peltier cell, with a high efficiency in the diagnostic X-ray energy range and modified in the shaping electronics so as to obtain a faster response. The acquiring section lies on a NuDAQ I/O card with a sampling frequency of up to 20 MHz. The signal produced by the X-ray tube is wholly acquired and an off-line analysis is made so as to make possible an accurate recognition of pile-up events and a reconstruction of the emitted spectra. The reconstructed spectra of a General Electric Senographe DMR mammographic X-ray tube are shown. -
Article: Direct analysis of molybdenum target generated x-ray spectra with a portable device
[show abstract] [hide abstract]
ABSTRACT: In routine applications, information about the photon flux of x-ray tubes is obtained from exposure measurements and cataloged spectra. This approach relies mainly on the assumption that the real spectrum is correctly approximated by the cataloged one, once the main characteristics of the tube such as voltage, target material, anode angle, and filters are taken account of. In practice, all this information is not always available. Moreover, x-ray tubes with the same characteristics may have different spectra. We describe an apparatus that should be useful for quality control in hospitals and for characterizing new radiographic systems. The apparatus analyzes the spectrum generated by an x-ray mammographic unit. It is based on a commercial CZT produced by AMPTEK Inc. and a set of tungsten collimator disks. The electronics of the CZT are modified so as to obtain a faster response. The signal is digitized using an analog to digital converter with a sampling frequency of up to 20 MHz. The whole signal produced by the x-ray tube is acquired and analyzed off-line in order to accurately recognize pile-up events and reconstruct the emitted spectrum. The energy resolution has been determined using a calibrated x-ray source. Spectra were validated by comparison of the HVL measured using an ionization chamber. -
Article: The CALMA system: an artificial neural network method for detecting masses and microcalcifications in digitized mammograms
[show abstract] [hide abstract]
ABSTRACT: The CALMA (Computer Assisted Library for MAmmography) project is a five years plan developed in a physics research frame in collaboration between INFN (Istituto Nazionale di Fisica Nucleare) and many Italian hospitals. At present a large database of digitized mammographic images (more than 6000) was collected and a software based on neural network algorithms for the search of suspicious breast lesions was developed. Two tools are available: a microcalcification clusters hunter, based on supervised and unsupervised feedforward neural network, and a massive lesions searcher, based on a hibrid approach. Both the algorithms analyzed preprocessed digitized images by high frequency filters. Clinical tests were performed to evaluate sensitivity and specificity of the system, considering the system as alone and as secon reader. Results show that the system is ready to be implemented by medical industry. The CALMA project, just ended, has its natural development in the GPCALMA (Grid Platform for CALMA) project, where distributed users join common resources (images, tools, statistical analysis). -
Article: Classifiers trained on dissimilarity representation of medical pattern: a comparative study
[show abstract] [hide abstract]
ABSTRACT: In this paper we investigate the feasibility of some typical techniques of pattern recognition for the classification of medical examples. The learning of the classifiers is not made in the traditional features space but it can be made by constructing decision rules on dissimilarity (distance) representations. In such a recognition process a new object is described by its distances to (a subset of) the training samples. Purpose of this work is the development of an automatic classification system which could be useful for radiologists in the investigation of breast cancer. The software has been designed in the framework of the MAGIC-5 collaboration. In the automatic classification system the suspicious regions with high probability to include a lesion are extracted from the image as regions of interest (ROIs). Each ROI is characterized by some features extracted from co-occurrence matrix containing spatial statistics information on ROI pixel gray tones. A dissimilarity representation of these features is made before the classification. A Feed-Forward Neural Network (FF-NN), a K-Nearest Neighbour (K-NN) and a Linear Discriminant Analysis (LDA) are employed to distinguish pathological records from not-pathological ones by the new features. The results obtained in terms of sensitivity (percentage of pathological ROIs correctly classified) and specificity (percentage of healthy ROIs correctly classified) will be comparatively presented. The K-NN classifier gives slightly better results than FF-NN and LDA accuracy (percentage of cases correctly classified) on two-classes problem (pathologic or healthy patients). -
Article: Comparative study of feature classification methods for mass lesion recognition in digitized mammograms
[show abstract] [hide abstract]
ABSTRACT: In this work a comparison of different classification methods for the identification of mass lesions in digitized mammograms is performed. These methods, used in order to develop Computer Aided Detection (CAD) systems, have been implemented in the framework of the MAGIC-5 Collaboration. The system for identification of mass lesions is based on a three-step procedure: a) preprocessing and segmentation, b) region of interest (ROI) searching, c) feature extraction and classification. It was tested on a very large mammographic database (3369 mammographic images from 967 patients). Each ROI is characterized by eight features extracted from a co-occurrence matrix containing spatial statistics information on the ROI pixel grey tones. -
Article: Measurements of spectral and position resolution on a 16x16 pixel CZT imaging hard x-ray detector
[show abstract] [hide abstract]
ABSTRACT: Cadmium zinc telluride (CZT) pixel detectors show very good spectral and spatial resolution and are suitable for use in compact hard X-ray sensors operated without cryogenics. One of the more interesting astrophysical application is their use as focal plane detectors for multilayer hard X-ray telescopes operating in the 15-70 keV energy band. Here we report on results obtained using a 16 x 16 CZT pixel detector (10 x 10 x 1 mm<sup>3</sup> single crystal) with 500 µm pixels operated at room temperature using standard commercial electronics. The results clearly show that the use of small pixels is effective in reducing one of the major drawbacks of CZT planar detectors i.e. the considerable amount of charge loss, due to hole trapping, which gives rise to a reduced energy resolution and a low energy tail in the pulse-height spectra. -
Article: Mass lesion detection in mammographic images using Haralik textural features
[show abstract] [hide abstract]
ABSTRACT: In this article we present a classification system for an automatic detection of masses in digitized mammographic images. The systems consists in three main processing levels: a) image segmentation for the localization of regions of interest (ROIs); b) ROI characterization by means of textural features computed from the Gray Tone Spatial Dependence Matrix (GTSDM), containing second order spatial statistics information on the pixel grey level intensity; c) ROI classification by means of a neural network, with supervision provided by the radiologist’s diagnosis. The CAD system was developed and evaluated using a database of N<sub>I</sub> = 3369 mammographic images: the breakdown of the cases was N<sub>In</sub> = 2307 negative images, and N<sub>Ip</sub> = 1062 pathological (or positive) images, containing at least one confirmed mass, as diagnosed by an expert radiologist. To examine the performance of the overall CAD system, receiver operating characteristic (ROC) and free-response ROC (FROC) analysis were employed. The area under the ROC curve was found to be A<sub>z</sub> = 0.78 ± 0.008 for ROI-based classification. When evaluating the accuracy of the CAD against the radiologist-drawn boundaries, 4.23 false positive per image (FPpI) are found at 80% mass sensitivity. -
Article: GPCALMA: a grid-based tool for mammographic screening
[show abstract] [hide abstract]
ABSTRACT: The next generation of High Energy Physics (HEP) experiments requires a GRID approach to a distributed computing system and the associated data management: the key concept is the Virtual Organisation (VO), a group of distributed users with a common goal and the will to share their resources. A similar approach is being applied to a group of Hospitals which joined the GPCALMA project (Grid Platform for Computer Assisted Library for MAmmography), which will allow common screening programs for early diagnosis of breast and, in the future, lung cancer. HEP techniques come into play in writing the application code, which makes use of neural networks for the image analysis and proved to be useful in improving the radiologists' performances in the diagnosis. GRID technologies allow remote image analysis and interactive online diagnosis, with a potential for a relevant reduction of the delays presently associated to screening programs. A prototype of the system, based on AliEn GRID Services [1], is already available, with a central Server running common services [2] and several clients connecting to it. Mammograms can be acquired in any location; the related information required to select and access them at any time is stored in a common service called Data Catalogue , which can be queried by any client. The result of a query can be used as input for analysis algorithms, which are executed on nodes that are in general remote to the user (but always local to the input images) thanks to the PROOF facility [3], a set of C++ classes that provide the functionality required to configure several distributed nodes in a way that allows parallel analysis of similar data samples. The selected approach avoids data transfers for all the images with a negative diagnosis (about 95% of the sample) and allows an almost real time diagnosis for the 5% of images with high cancer probability. -
Article: A massive lesion detection algorithm in mammography.
[show abstract] [hide abstract]
ABSTRACT: A new algorithm for massive lesion detection in mammography is presented. The algorithm consists in three main steps: 1) reduction of the dimension of the image to be processed through the identification of regions of interest (roi) as candidates for massive lesions; 2) characterization of the RoI by means of suitable feature extraction; 3) pattern classification through supervised neural networks. Suspect regions are detected by searching for local maxima of the pixel grey level intensity. A ring of increasing radius, centered on a maximum, is considered until the mean intensity in the ring decreases to a defined fraction of the maximum. The ROIS thus obtained are described by average, variance, skewness and kurtosis of the intensity distributions at different fractions of the radius. A neural network approach is adopted to classify suspect pathological and healthy pattern. The software has been designed in the framework of the INFN (Istituto Nazionale Fisica Nucleare) research project GPCALMA (Grid Platform for Calma) which recruits physicists and radiologists from different Italian Research Institutions and hospitals to develop software for breast cancer detection.Physica Medica 21(1):23-30. · 1.07 Impact Factor
