RESUMO
Human interaction has become almost mandatory for an automated medical system wishing to be accepted by clinical regulatory agencies such as Food and Drug Administration. Since this interaction causes variability in the gathered data, the inter-observer and intra-observer variability must be analyzed in order to validate the accuracy of the system. This study focuses on the variability from different observers that interact with an automated lung delineation system that relies on human interaction in the form of delineation of the lung borders. The database consists of High Resolution Computed Tomography (HRCT): 15 normal and 81 diseased patients' images taken retrospectively at five levels per patient. Three observers manually delineated the lungs borders independently and using software called ImgTracer™ (AtheroPoint™, Roseville, CA, USA) to delineate the lung boundaries in all five levels of 3-D lung volume. The three observers consisted of Observer-1: lesser experienced novice tracer who is a resident in radiology under the guidance of radiologist, whereas Observer-2 and Observer-3 are lung image scientists trained by lung radiologist and biomedical imaging scientist and experts. The inter-observer variability can be shown by comparing each observer's tracings to the automated delineation and also by comparing each manual tracing of the observers with one another. The normality of the tracings was tested using D'Agostino-Pearson test and all observers tracings showed a normal P-value higher than 0.05. The analysis of variance (ANOVA) test between three observers and automated showed a P-value higher than 0.89 and 0.81 for the right lung (RL) and left lung (LL), respectively. The performance of the automated system was evaluated using Dice Similarity Coefficient (DSC), Jaccard Index (JI) and Hausdorff (HD) Distance measures. Although, Observer-1 has lesser experience compared to Obsever-2 and Obsever-3, the Observer Deterioration Factor (ODF) shows that Observer-1 has less than 10% difference compared to the other two, which is under acceptable range as per our analysis. To compare between observers, this study used regression plots, Bland-Altman plots, two tailed T-test, Mann-Whiney, Chi-Squared tests which showed the following P-values for RL and LL: (i) Observer-1 and Observer-3 were: 0.55, 0.48, 0.29 for RL and 0.55, 0.59, 0.29 for LL; (ii) Observer-1 and Observer-2 were: 0.57, 0.50, 0.29 for RL and 0.54, 0.59, 0.29 for LL; (iii) Observer-2 and Observer-3 were: 0.98, 0.99, 0.29 for RL and 0.99, 0.99, 0.29 for LL. Further, CC and R-squared coefficients were computed between observers which came out to be 0.9 for RL and LL. All three observers however manage to show the feature that diseased lungs are smaller than normal lungs in terms of area.
Assuntos
Interpretação de Imagem Assistida por Computador , Neoplasias Pulmonares/diagnóstico , Variações Dependentes do Observador , Tomografia Computadorizada por Raios X/métodos , Adolescente , Adulto , Idoso , Idoso de 80 Anos ou mais , Feminino , Humanos , Neoplasias Pulmonares/patologia , Masculino , Pessoa de Meia-Idade , Estados Unidos , Adulto JovemRESUMO
Interstitial Lung Disease (ILD) encompasses a wide array of diseases that share some common radiologic characteristics. When diagnosing such diseases, radiologists can be affected by heavy workload and fatigue thus decreasing diagnostic accuracy. Automatic segmentation is the first step in implementing a Computer Aided Diagnosis (CAD) that will help radiologists to improve diagnostic accuracy thereby reducing manual interpretation. Automatic segmentation proposed uses an initial thresholding and morphology based segmentation coupled with feedback that detects large deviations with a corrective segmentation. This feedback is analogous to a control system which allows detection of abnormal or severe lung disease and provides a feedback to an online segmentation improving the overall performance of the system. This feedback system encompasses a texture paradigm. In this study we studied 48 males and 48 female patients consisting of 15 normal and 81 abnormal patients. A senior radiologist chose the five levels needed for ILD diagnosis. The results of segmentation were displayed by showing the comparison of the automated and ground truth boundaries (courtesy of ImgTracer™ 1.0, AtheroPoint™ LLC, Roseville, CA, USA). The left lung's performance of segmentation was 96.52% for Jaccard Index and 98.21% for Dice Similarity, 0.61 mm for Polyline Distance Metric (PDM), -1.15% for Relative Area Error and 4.09% Area Overlap Error. The right lung's performance of segmentation was 97.24% for Jaccard Index, 98.58% for Dice Similarity, 0.61 mm for PDM, -0.03% for Relative Area Error and 3.53% for Area Overlap Error. The segmentation overall has an overall similarity of 98.4%. The segmentation proposed is an accurate and fully automated system.
Assuntos
Diagnóstico por Computador/métodos , Doenças Pulmonares Intersticiais/diagnóstico , Doenças Pulmonares Intersticiais/patologia , Reconhecimento Automatizado de Padrão/métodos , Interpretação de Imagem Radiográfica Assistida por Computador/métodos , Algoritmos , Feminino , Humanos , Doenças Pulmonares Intersticiais/diagnóstico por imagem , Masculino , Reprodutibilidade dos Testes , Tomografia Computadorizada por Raios XRESUMO
Lung disease risk stratification is important for both diagnosis and treatment planning, particularly in biopsies and radiation therapy. Manual lung disease risk stratification is challenging because of: (a) large lung data sizes, (b) inter- and intra-observer variability of the lung delineation and (c) lack of feature amalgamation during machine learning paradigm. This paper presents a two stage CADx cascaded system consisting of: (a) semi-automated lung delineation subsystem (LDS) for lung region extraction in CT slices followed by (b) morphology-based lung tissue characterization, thereby addressing the above shortcomings. LDS primarily uses entropy-based region extraction while ML-based lung characterization is mainly based on an amalgamation of directional transforms such as Riesz and Gabor along with texture-based features comprising of 100 greyscale features using the K-fold cross-validation protocol (K = 2, 3, 5 and 10). The lung database consisted of 96 patients: 15 normal and 81 diseased. We use five high resolution Computed Tomography (HRCT) levels representing different anatomy landmarks where disease is commonly seen. We demonstrate the amalgamated ML stratification accuracy of 99.53%, an increase of 2% against the conventional non-amalgamation ML system that uses alone Riesz-based feature embedded with feature selection based on feature strength. The robustness of the system was determined based on the reliability and stability that showed a reliability index of 0.99 and the deviation in risk stratification accuracies less than 5%. Our CADx system shows 10% better performance when compared against the mean of five other prominent studies available in the current literature covering over one decade.
Assuntos
Bases de Dados Factuais , Processamento de Imagem Assistida por Computador/métodos , Pneumopatias/diagnóstico por imagem , Aprendizado de Máquina , Tomografia Computadorizada por Raios X , Feminino , Humanos , Masculino , Medição de Risco/métodosRESUMO
BACKGROUND: The extent of calcium volume in the carotid arteries of contrast-based computer tomography (CT) is a valuable indicator of stroke risk. This study presents an automated, simple and fast calcium volume computation system. Since the high contrast agent can sometimes obscure the presence of calcium in the CT slices, it is therefore necessary to identify these slices before the corrected volume can be estimated. METHODS: The system typically consists of segmenting the calcium region from the CT scan into slices based on Hounsfield Unit-based threshold, and subsequently computing the summation of the calcium areas in each slice. However, when the carotid volume has intermittently higher concentration of contrast agent, a dependable approach is adapted to correct the calcium region using the neighboring slices, thereby estimating the correct volume. Furthermore, mitigation is provided following the regulatory constraints by changing the system to semi-automated criteria as a fall back solution. We evaluate the automated and semi-automated techniques using completely manual calcium volumes computed based on the manual tracings by the neuroradiologist. RESULTS: A total of 64 patients with calcified plaque in the internal carotid artery were analyzed. Using the above algorithm, our automated and semi-automated system yields correlation coefficients (CC) of 0.89 and 0.96 against first manual readings and 0.90 and 0.96 against second manual readings, respectively. Using the t-test, there was no significant difference between the automated and semi-automated methods against manual. The intra-observer reliability was excellent with CC 0.98. CONCLUSIONS: Compared to automated method, the semi-automated method for calcium volume is acceptable and closer to manual strategy for calcium volume. Further work evaluating and confirming the performance of our semi-automated protocol is now warranted.
Assuntos
Aterosclerose/diagnóstico por imagem , Cálcio/análise , Artéria Carótida Interna/diagnóstico por imagem , Angiografia por Tomografia Computadorizada , Placa Aterosclerótica/diagnóstico por imagem , Idoso , Algoritmos , Meios de Contraste/administração & dosagem , Feminino , Humanos , Masculino , Curva ROC , Reprodutibilidade dos Testes , Estudos Retrospectivos , Software , Estados UnidosRESUMO
The knowledge of teeth positions on the maxillary arch is useful in the rehabilitation of the edentulous patient. A combination of angular (θ), and linear (l) variables representing position of four teeth were initially proposed as the shape descriptor of the maxillary dental arch. Three categories of shape were established, each having a multivariate normal distribution. It may be argued that 4 selected teeth on the standardized digital images of the dental casts could be considered as insufficient with respect to representing shape. However, increasing the number of points would create problems with dimensions and proof of existence of the multivariate normal distribution is extremely difficult. This study investigates the ability of Fourier descriptors (FD) using all maxillary teeth to find alternative shape models. Eight FD terms were sufficient to represent 21 points on the arch. Using these 8 FD terms as an alternative shape descriptor, three categories of shape were verified, each category having the complex normal distribution.
Assuntos
Arco Dental/patologia , Arcada Edêntula/patologia , Maxila/patologia , Algoritmos , Análise de Fourier , Humanos , Interpretação de Imagem Assistida por Computador , Análise MultivariadaRESUMO
Selected landmarks from each of 47 maxillary dental casts were used to define a Cartesian-coordinate system from which the positions of selected teeth were determined on standardized digital images. The position of the i-th tooth was defined by a line of length (l(i)) joining the tooth to the origin, and the angle (θ(i)) of this line to the horizontal Cartesian axis. Four teeth, the central incisor, lateral incisor, canine and first molar were selected and their position were collectively used to represent the shape of the dental arch. A pilot study using clustering and principal component analysis strongly suggest the existence of 3 groups of arch shape. In this study, the homogeneity of the 3 groups was further investigated and confirmed by the Dunn and Davies-Bouldein validity indices. This is followed by an investigation of the probability distribution of these 3 groups. The main result of this study suggests 3 groups of multivariate (MV) normal distribution. The MV normal probability distribution of these groups may be used in further studies to investigate the issues of variation of arch shape, which is fundamental to the practice of prosthodontics and orthodontics.
Assuntos
Arco Dental/anatomia & histologia , Processamento de Imagem Assistida por Computador , Maxila/anatomia & histologia , Probabilidade , Análise por Conglomerados , Humanos , Projetos Piloto , Análise de Componente PrincipalRESUMO
Standardized digital images of maxillary dental casts of 47 subjects were analyzed using MATLAB software whereby the two hamular notches and the incisive papilla defines the Cartesian vertical and horizontal axes, as well as the origin. The angle and length of the midpoints of the anterior teeth, mesiobuccal and distobuccal cusp of the posterior teeth were measured from the origin and denoted as θ(1), , θ(18) and l(1), , l(18) respectively. These values were collectively used to represent the shape of each dental cast. Clustering and principal component analyses were employed to find possible groups of dental arches using the above measure of shape. The main result of this study is that the 3 groups of dental arch shape may be represented by the novel feature vector v(k) = (θ(k)(1), l(k)(1), θ(k)(3), l(k)(3), θ(k)(5), l(k)(5), θ(k)(13), l(k)(13)), k = 1, 2, 3. Knowledge of v(k) implies three impression trays should be sufficient in a particular prosthetic dentistry application for Malaysian patients. Further, given that v(k) are accurately measured they may be potential candidates as evidence in specific application of forensic dentistry.