Your browser doesn't support javascript.
loading
Mostrar: 20 | 50 | 100
Resultados 1 - 20 de 14.565
Filtrar
1.
Annu Int Conf IEEE Eng Med Biol Soc ; 2020: 1254-1257, 2020 07.
Artigo em Inglês | MEDLINE | ID: mdl-33018215

RESUMO

Computer-aided Diagnosis (CAD) systems have long aimed to be used in clinical practice to help doctors make decisions by providing a second opinion. However, most machine learning based CAD systems make predictions without explicitly showing how their predictions were generated. Since the cognitive process of the diagnostic imaging interpretation involves various visual characteristics of the region of interest, the explainability of the results should leverage those characteristics. We encode visual characteristics of the region of interest based on pairs of similar images rather than the image content by itself. Using a Siamese convolutional neural network (SCNN), we first learn the similarity among nodules, then encode image content using the SCNN similarity-based feature representation, and lastly, we apply the K-nearest neighbor (KNN) approach to make diagnostic characterizations using the Siamese-based image features. We demonstrate the feasibility of our approach on spiculation, a visual characteristic that radiologists consider when interpreting the degree of cancer malignancy, and the NIH/NCI Lung Image Database Consortium (LIDC) dataset that contains both spiculation and malignancy characteristics for lung nodules.Clinical Relevance - This establishes that spiculation can be quantified to automate the diagnostic characterization of lung nodules in Computed Tomography images.


Assuntos
Neoplasias Pulmonares , Interpretação de Imagem Radiográfica Assistida por Computador , Humanos , Pulmão , Neoplasias Pulmonares/diagnóstico por imagem , Redes Neurais de Computação , Tomografia Computadorizada por Raios X
2.
Annu Int Conf IEEE Eng Med Biol Soc ; 2020: 1331-1334, 2020 07.
Artigo em Inglês | MEDLINE | ID: mdl-33018234

RESUMO

Lung cancer is the most common form of cancer found worldwide with a high mortality rate. Early detection of pulmonary nodules by screening with a low-dose computed tomography (CT) scan is crucial for its effective clinical management. Nodules which are symptomatic of malignancy occupy about 0.0125 - 0.025% of volume in a CT scan of a patient. Manual screening of all slices is a tedious task and presents a high risk of human errors. To tackle this problem we propose a computationally efficient two stage framework. In the first stage, a convolutional neural network (CNN) trained adversarially using Turing test loss segments the lung region. In the second stage, patches sampled from the segmented region are then classified to detect the presence of nodules. The proposed method is experimentally validated on the LUNA16 challenge dataset with a dice coefficient of 0.984±0.0007 for 10-fold cross-validation.


Assuntos
Interpretação de Imagem Radiográfica Assistida por Computador , Tomografia Computadorizada por Raios X , Humanos , Pulmão/diagnóstico por imagem , Redes Neurais de Computação , Cintilografia
3.
Comput Math Methods Med ; 2020: 9756518, 2020.
Artigo em Inglês | MEDLINE | ID: mdl-33014121

RESUMO

The COVID-19 diagnostic approach is mainly divided into two broad categories, a laboratory-based and chest radiography approach. The last few months have witnessed a rapid increase in the number of studies use artificial intelligence (AI) techniques to diagnose COVID-19 with chest computed tomography (CT). In this study, we review the diagnosis of COVID-19 by using chest CT toward AI. We searched ArXiv, MedRxiv, and Google Scholar using the terms "deep learning", "neural networks", "COVID-19", and "chest CT". At the time of writing (August 24, 2020), there have been nearly 100 studies and 30 studies among them were selected for this review. We categorized the studies based on the classification tasks: COVID-19/normal, COVID-19/non-COVID-19, COVID-19/non-COVID-19 pneumonia, and severity. The sensitivity, specificity, precision, accuracy, area under the curve, and F1 score results were reported as high as 100%, 100%, 99.62, 99.87%, 100%, and 99.5%, respectively. However, the presented results should be carefully compared due to the different degrees of difficulty of different classification tasks.


Assuntos
Betacoronavirus , Técnicas de Laboratório Clínico , Infecções por Coronavirus/diagnóstico por imagem , Pandemias , Pneumonia Viral/diagnóstico por imagem , Interpretação de Imagem Radiográfica Assistida por Computador/métodos , Tomografia Computadorizada por Raios X/estatística & dados numéricos , Inteligência Artificial , Infecções por Coronavirus/diagnóstico , Infecções por Coronavirus/epidemiologia , Aprendizado Profundo , Humanos , Redes Neurais de Computação , Pneumonia/classificação , Pneumonia/diagnóstico por imagem , Pneumonia Viral/epidemiologia , Interpretação de Imagem Radiográfica Assistida por Computador/estatística & dados numéricos , Radiografia Torácica/estatística & dados numéricos , Sensibilidade e Especificidade
4.
Annu Int Conf IEEE Eng Med Biol Soc ; 2020: 1132-1135, 2020 07.
Artigo em Inglês | MEDLINE | ID: mdl-33018186

RESUMO

CAD systems have shown good potential for improving breast cancer diagnosis and anomaly detection in mammograms. A basic enabling step for the utilization of CAD systems in mammographic analysis is the correct identification of the breast region. Therefore, several methods to segment the pectoral muscle in the medio-lateral oblique (MLO) mammographic view have been proposed in the literature. However, currently it is difficult to perform and objective comparison between different chest wall (CW) detection methods since they are often evaluated with different evaluation procedures, datasets and the implementations of the methods are not publicly available. For this reason, we propose a methodology to evaluate and compare the performance of CW detection methods using a publicly available dataset (INbreast). We also propose a new intensity-based method for automatic CW detection. We then utilize the proposed evaluation methodology to compare the performance of our CW detection algorithm with a state-of-the-art CW detection method. The performance was measured in terms of the Dice's coefficient similarity, the area error and mean contour distance. The proposed method achieves yielded the best results in all the performance measures.


Assuntos
Parede Torácica , Benchmarking , Humanos , Mamografia , Reconhecimento Automatizado de Padrão , Interpretação de Imagem Radiográfica Assistida por Computador , Parede Torácica/diagnóstico por imagem
5.
J Comput Assist Tomogr ; 44(5): 790-795, 2020.
Artigo em Inglês | MEDLINE | ID: mdl-32936580

RESUMO

BACKGROUND: The state-of-art motion correction algorithm is inadequate for correcting motion artifacts in coronary arteries in cardiovascular computed tomography angiography (CCTA) for children with high heart rates, and even less effective for heart structures beyond coronary arteries. PURPOSE: This study aimed to evaluate the effectiveness of a second-generation, whole-heart motion correction algorithm in improving the heart image quality of CCTA for children with high heart rates. MATERIALS AND METHODS: Forty-two consecutive symptomatic cardiac patients with high heart rates (122.6 ± 18.8 beats/min) were enrolled. All patients underwent CCTA on a 256-row CT using a prospective electrocardiogram-triggered single-beat protocol. Images were reconstructed using a standard algorithm (STD), state-of-the-art first-generation coronary artery motion correction algorithm (MC1), and second-generation, whole-heart motion correction algorithm (MC2). The image quality of the origin of left coronary, right coronary, aortic valve, pulmonary valve, mitral valve, tricuspid valve, aorta root, pulmonary artery root, ventricular septum (VS), and atrial septum (AS) was assessed by 2 experienced radiologists using a 4-point scale (1, nondiagnostic; 2, detectable; 3, measurable; and 4, excellent); nonparametric test was used to analyze and compare the differences among 3 groups; and post hoc multiple comparisons were used between different methods. RESULTS: There were group differences for cardiac structures except VS and AS, with MC2 having the best image quality and STD having the worst image quality. Post hoc multiple comparisons showed that MC2 was better than MC1 and STD in all structures except VS and AS where all 3 algorithms performed equally, whereas MC1 was better than STD only in the origin of left coronary, right coronary, and mitral valve. CONCLUSIONS: A second-generation, whole-heart motion correction algorithm further significantly improves cardiac image quality beyond the coronaries in CCTA for pediatric patients with high heart rates.


Assuntos
Angiografia por Tomografia Computadorizada/métodos , Angiografia Coronária/métodos , Vasos Coronários/diagnóstico por imagem , Vasos Coronários/fisiopatologia , Frequência Cardíaca/fisiologia , Interpretação de Imagem Radiográfica Assistida por Computador/métodos , Algoritmos , Criança , Pré-Escolar , Feminino , Humanos , Lactente , Masculino , Estudos Retrospectivos
6.
IEEE J Biomed Health Inform ; 24(10): 2806-2813, 2020 10.
Artigo em Inglês | MEDLINE | ID: mdl-32915751

RESUMO

The pandemic of coronavirus disease 2019 (COVID-19) has lead to a global public health crisis spreading hundreds of countries. With the continuous growth of new infections, developing automated tools for COVID-19 identification with CT image is highly desired to assist the clinical diagnosis and reduce the tedious workload of image interpretation. To enlarge the datasets for developing machine learning methods, it is essentially helpful to aggregate the cases from different medical systems for learning robust and generalizable models. This paper proposes a novel joint learning framework to perform accurate COVID-19 identification by effectively learning with heterogeneous datasets with distribution discrepancy. We build a powerful backbone by redesigning the recently proposed COVID-Net in aspects of network architecture and learning strategy to improve the prediction accuracy and learning efficiency. On top of our improved backbone, we further explicitly tackle the cross-site domain shift by conducting separate feature normalization in latent space. Moreover, we propose to use a contrastive training objective to enhance the domain invariance of semantic embeddings for boosting the classification performance on each dataset. We develop and evaluate our method with two public large-scale COVID-19 diagnosis datasets made up of CT images. Extensive experiments show that our approach consistently improves the performanceson both datasets, outperforming the original COVID-Net trained on each dataset by 12.16% and 14.23% in AUC respectively, also exceeding existing state-of-the-art multi-site learning methods.


Assuntos
Betacoronavirus , Técnicas de Laboratório Clínico/estatística & dados numéricos , Infecções por Coronavirus/diagnóstico por imagem , Infecções por Coronavirus/diagnóstico , Aprendizado Profundo , Pandemias , Pneumonia Viral/diagnóstico por imagem , Pneumonia Viral/diagnóstico , Tomografia Computadorizada por Raios X/estatística & dados numéricos , Biologia Computacional , Sistemas Computacionais , Infecções por Coronavirus/classificação , Bases de Dados Factuais/estatística & dados numéricos , Humanos , Aprendizado de Máquina , Pandemias/classificação , Pneumonia Viral/classificação , Interpretação de Imagem Radiográfica Assistida por Computador/estatística & dados numéricos
7.
PLoS One ; 15(9): e0238427, 2020.
Artigo em Inglês | MEDLINE | ID: mdl-32936816

RESUMO

In Graeco-Roman times in the Lower-Egyptian Fayoum region, a painted portrait was traditionally placed over the face of a deceased individual. These mummy portraits show considerable inter-individual diversity. This suggests that those portraits were created separately for each individual. In the present study, we investigated a completely wrapped young infant mummy with a typical mummy portrait by whole body CT analysis. This was used to obtain physical information on the infant and provided the basis for a virtual face reconstruction in order to compare it to the mummy portrait. We identified the mummy as a 3-4 years old male infant that had been prepared according to the typical ancient Egyptian mummification rites. It most probably suffered from a right-sided pulmonary infection which may also be the cause of death. The reconstructed face showed considerable similarities to the portrait, confirming the portrait's specificity to this individual. However, there are some differences between portrait and face. The portrait seems to show a slightly older individual which may be due to artistic conventions of that period.


Assuntos
Face/diagnóstico por imagem , Múmias/diagnóstico por imagem , Retratos como Assunto/história , Arte , Pré-Escolar , Egito , Face/anatomia & histologia , História Antiga , Humanos , Imageamento Tridimensional , Lactente , Masculino , Múmias/história , Interpretação de Imagem Radiográfica Assistida por Computador , Tomografia Computadorizada por Raios X , Imagem Corporal Total
8.
PLoS One ; 15(9): e0238926, 2020.
Artigo em Inglês | MEDLINE | ID: mdl-32925940

RESUMO

Fractures of the wrist are common in Emergency Departments, where some patients are treated with a procedure called Manipulation under Anaesthesia. In some cases, this procedure is unsuccessful and patients need to revisit the hospital where they undergo surgery to treat the fracture. This work describes a geometric semi-automatic image analysis algorithm to analyse and compare the x-rays of healthy controls and patients with dorsally displaced wrist fractures (Colles' fractures) who were treated with Manipulation under Anaesthesia. A series of 161 posterior-anterior radiographs from healthy controls and patients with Colles' fractures were acquired and analysed. The patients' group was further subdivided according to the outcome of the procedure (successful/unsuccessful) and pre- or post-intervention creating five groups in total (healthy, pre-successful, pre-unsuccessful, post-successful, post-unsuccessful). The semi-automatic analysis consisted of manual location of three landmarks (finger, lunate and radial styloid) and automatic processing to generate 32 geometric and texture measurements, which may be related to conditions such as osteoporosis and swelling of the wrist. Statistical differences were found between patients and controls, as well as between pre- and post-intervention, but not between the procedures. The most distinct measurements were those of texture. Although the study includes a relatively low number of cases and measurements, the statistical differences are encouraging.


Assuntos
Fratura de Colles/diagnóstico por imagem , Fratura de Colles/terapia , Interpretação de Imagem Radiográfica Assistida por Computador/métodos , Adulto , Idoso , Idoso de 80 Anos ou mais , Estudos de Casos e Controles , Feminino , Humanos , Masculino , Manipulação Ortopédica , Pessoa de Meia-Idade , Radiografia , Resultado do Tratamento
9.
PLoS One ; 15(9): e0238582, 2020.
Artigo em Inglês | MEDLINE | ID: mdl-32966278

RESUMO

OBJECTIVE: To compare "virtual" unenhanced (VUE) computed tomography (CT) images, reconstructed from rapid kVp-switching dual-energy computed tomography (DECT), to "true" unenhanced CT images (TUE), in clinical abdominal imaging. The ability to replace TUE with VUE images would have many clinical and operational advantages. METHODS: VUE and TUE images of 60 DECT datasets acquired for standard-of-care CT of pancreatic cancer were retrospectively reviewed and compared, both quantitatively and qualitatively. Comparisons included quantitative evaluation of CT numbers (Hounsfield Units, HU) measured in 8 different tissues, and 6 qualitative image characteristics relevant to abdominal imaging, rated by 3 experienced radiologists. The observed quantitative and qualitative VUE and TUE differences were compared against boundaries of clinically relevant equivalent thresholds to assess their equivalency, using modified paired t-tests and Bayesian hierarchical modeling. RESULTS: Quantitatively, in tissues containing high concentrations of calcium or iodine, CT numbers measured in VUE images were significantly different from those in TUE images. CT numbers in VUE images were significantly lower than TUE images when calcium was present (e.g. in the spine, 73.1 HU lower, p < 0.0001); and significantly higher when iodine was present (e.g. in renal cortex, 12.9 HU higher, p < 0.0001). Qualitatively, VUE image ratings showed significantly inferior depiction of liver parenchyma compared to TUE images, and significantly more cortico-medullary differentiation in the kidney. CONCLUSIONS: Significant differences in VUE images compared to TUE images may limit their application and ability to replace TUE images in diagnostic abdominal CT imaging.


Assuntos
Abdome/diagnóstico por imagem , Neoplasias Pancreáticas/diagnóstico por imagem , Imagem Radiográfica a Partir de Emissão de Duplo Fóton/métodos , Tomografia Computadorizada por Raios X/métodos , Feminino , Humanos , Masculino , Interpretação de Imagem Radiográfica Assistida por Computador/métodos , Estudos Retrospectivos
10.
PLoS One ; 15(9): e0239459, 2020.
Artigo em Inglês | MEDLINE | ID: mdl-32946530

RESUMO

This study aimed to evaluate the visualization of peripheral bronchioles in normal lungs via quarter-detector computed tomography (QDCT). Visualization of bronchioles within 10 mm from the pleura is considered a sign of bronchiectasis. However, it is not known peripheral bronchioles how close to the pleura in normal lungs can be tracked using QDCT. This study included 228 parts in 76 lungs from 38 consecutive patients who underwent QDCT. Reconstruction was performed with different thicknesses, increments, and matrix sizes: 0.5-mm thickness and increment with 512 and 1024 matrixes (Group5 and Group10, respectively) and 0.25-mm thickness and increment with 1024 matrix (Group10Thin). The distance between the most peripheral bronchiole visible and the pleura was determined in the three groups. The distance between the peripheral bronchial duct ends and the nearest pleural surface were significantly shorter in the order of Group10Thin, Group10, and Group5, and the mean distances from the pleura in Group10Thin and Group10 were shorter than 10 mm. These findings suggest the visualization of peripheral bronchioles in QDCT was better with a 1024 axial matrix than with a 512 matrix, and with a 0.25-mm slice thickness/increment than with a 0.5-mm slice thickness/increment. Our study also indicates bronchioles within 10 mm of the pleura do not necessarily indicate pathology.


Assuntos
Bronquíolos/diagnóstico por imagem , Interpretação de Imagem Radiográfica Assistida por Computador/métodos , Tomografia Computadorizada por Raios X/métodos , Adulto , Idoso , Idoso de 80 Anos ou mais , Bronquiectasia/diagnóstico por imagem , Feminino , Humanos , Pulmão/diagnóstico por imagem , Masculino , Pessoa de Meia-Idade , Pleura/diagnóstico por imagem , Estudos Retrospectivos , Tomografia Computadorizada por Raios X/estatística & dados numéricos
11.
J Cancer Res Ther ; 16(4): 780-787, 2020.
Artigo em Inglês | MEDLINE | ID: mdl-32930118

RESUMO

Background: There are "blind spots" on chest computed tomography (CT) where pulmonary nodules can easily be overlooked. The number of missed pulmonary nodules can be minimized by instituting a training program with particular focus on the depiction of nodules at blind spots. Purpose: The purpose of this study was to assess the variation in lung nodule detection in chest CT based on location, attenuation characteristics, and reader experience. Materials and Methods: We selected 18 noncalcified lung nodules (6-8 mm) suspicious of primary and metastatic lung cancer with solid (n = 7), pure ground-glass (6), and part-solid ground-glass (5) attenuation from 12 chest CT scans. These nodules were randomly inserted in chest CT of 34 patients in lung hila, 1st costochondral junction, branching vessels, paramediastinal lungs, lung apices, juxta-diaphragm, and middle and outer thirds of the lungs. Two residents and two chest imaging clinical fellows evaluated the CT images twice, over a 4-month interval. Before the second reading session, the readers were trained and made aware of the potential blind spots. Chi-square test was used to assess statistical significance. Results: Pretraining session: Fellows detected significantly more part-solid ground-glass nodules compared to residents (P = 0.008). A substantial number of nodules adjacent to branching vessels and posterior mediastinum were missed. Posttraining session: There was a significant increase in detectability independent of attenuation and location of nodules for all readers (P < 0.0008). Conclusion: Dedicated chest CT training improves detection of lung nodules, especially the part-solid ground-glass nodules. Detection of nodules adjacent to branching vessels and the posterior mediastinal lungs is difficult even for fellowship-trained radiologists.


Assuntos
Neoplasias Pulmonares/diagnóstico , Nódulos Pulmonares Múltiplos/diagnóstico , Radiologia/educação , Treinamento por Simulação/métodos , Nódulo Pulmonar Solitário/diagnóstico , Adulto , Idoso , Idoso de 80 Anos ou mais , Algoritmos , Detecção Precoce de Câncer/métodos , Feminino , Humanos , Neoplasias Pulmonares/diagnóstico por imagem , Neoplasias Pulmonares/patologia , Masculino , Mediastino/diagnóstico por imagem , Mediastino/patologia , Pessoa de Meia-Idade , Nódulos Pulmonares Múltiplos/diagnóstico por imagem , Nódulos Pulmonares Múltiplos/patologia , Lesões Pré-Cancerosas/diagnóstico por imagem , Lesões Pré-Cancerosas/patologia , Interpretação de Imagem Radiográfica Assistida por Computador/métodos , Radiologia/métodos , Software , Nódulo Pulmonar Solitário/diagnóstico por imagem , Tomografia Computadorizada por Raios X/métodos
12.
PLoS One ; 15(9): e0239562, 2020.
Artigo em Inglês | MEDLINE | ID: mdl-32966330

RESUMO

Reproducible and unbiased methods to quantify alveolar structure are important for research on many lung diseases. However, manually estimating alveolar structure through stereology is time consuming and inter-observer variability is high. The objective of this work was to develop and validate a fast, reproducible and accurate (semi-)automatic alternative. A FIJI-macro was designed that automatically segments lung images to binary masks, and counts the number of test points falling on tissue and the number of intersections of the air-tissue interface with a set of test lines. Manual selection remains necessary for the recognition of non-parenchymal tissue and alveolar exudates. Volume density of alveolar septa ([Formula: see text]) and mean linear intercept of the airspaces (Lm) as measured by the macro were compared to theoretical values for 11 artificial test images and to manually counted values for 17 lungs slides using linear regression and Bland-Altman plots. Inter-observer agreement between 3 observers, measuring 8 lungs both manually and automatically, was assessed using intraclass correlation coefficients (ICC). [Formula: see text] and Lm measured by the macro closely approached theoretical values for artificial test images (R2 of 0.9750 and 0.9573 and bias of 0.34% and 8.7%). The macro data in lungs were slightly higher for [Formula: see text] and slightly lower for Lm in comparison to manually counted values (R2 of 0.8262 and 0.8288 and bias of -6.0% and 12.1%). Visually, semi-automatic segmentation was accurate. Most importantly, manually counted [Formula: see text] and Lm had only moderate to good inter-observer agreement (ICC 0.859 and 0.643), but agreements were excellent for semi-automatically counted values (ICC 0.956 and 0.900). This semi-automatic method provides accurate and highly reproducible alveolar morphometry results. Future efforts should focus on refining methods for automatic detection of non-parenchymal tissue or exudates, and for assessment of lung structure on 3D reconstructions of lungs scanned with microCT.


Assuntos
Displasia Broncopulmonar/patologia , Interpretação de Imagem Assistida por Computador/métodos , Alvéolos Pulmonares/patologia , Animais , Displasia Broncopulmonar/diagnóstico por imagem , Modelos Animais de Doenças , Feminino , Técnicas Histológicas/estatística & dados numéricos , Variações Dependentes do Observador , Gravidez , Alvéolos Pulmonares/diagnóstico por imagem , Coelhos , Interpretação de Imagem Radiográfica Assistida por Computador/métodos , Microtomografia por Raio-X/estatística & dados numéricos
14.
BMC Med Imaging ; 20(1): 92, 2020 08 05.
Artigo em Inglês | MEDLINE | ID: mdl-32758155

RESUMO

BACKGROUND: To investigate the CT changes of different clinical types of COVID-19 pneumonia. METHODS: This retrospective study included 50 patients with COVID-19 from 16 January 2020 to 25 February 2020. We analyzed the clinical characteristics, CT characteristics and the pneumonia involvement of the patients between the moderate group and the severe and critical group, and the dynamic changes of severity with the CT follow-up time. RESULTS: There were differences in the CT severity score of the right lung in the initial CT, and total CT severity score in the initial and follow-up CT between the moderate group and the severe and critical group (all p < 0.05). There was a quadratic relationship between total CT severity score and CT follow-up time in the severe and critical group (r2 = 0.137, p = 0.008), the total CT severity score peaked at the second follow-up CT. There was no correlation between total CT severity score and CT follow-up time in the moderate group (p > 0.05). There were no differences in the occurrence rate of CT characteristics in the initial CT between the two groups (all p > 0.05). There were differences in the occurrence rate of ground-glass opacity and crazy-paving pattern in the second follow-up CT, and pleural thickening or adhesion in the third follow-up CT between the two groups (all p < 0.05). CONCLUSIONS: The CT changes of COVID-19 pneumonia with different severity were different, and the extent of pneumonia involvement by CT can help to assess the severity of COVID-19 pneumonia rather than the initial CT characteristics.


Assuntos
Infecções por Coronavirus/diagnóstico por imagem , Pneumonia Viral/diagnóstico por imagem , Pneumonia/virologia , Interpretação de Imagem Radiográfica Assistida por Computador/métodos , Adulto , Idoso , Infecções por Coronavirus/patologia , Feminino , Humanos , Masculino , Pessoa de Meia-Idade , Pandemias , Pneumonia/diagnóstico por imagem , Pneumonia/patologia , Pneumonia Viral/patologia , Estudos Retrospectivos , Índice de Gravidade de Doença , Tomografia Computadorizada por Raios X , Adulto Jovem
15.
IEEE Pulse ; 11(4): 2-7, 2020.
Artigo em Inglês | MEDLINE | ID: mdl-32804639

RESUMO

Qualitative interpretation is a good thing when it comes to reading lung images in the fight against coronavirus 2019 disease (COVID-19), but quantitative analysis makes radiology reporting much more comprehensive. To that end, several research groups have begun looking to artificial intelligence (AI) as a tool for reading and analyzing X-rays and computed tomography (CT) scans, and helping to diagnose and monitor COVID-19.


Assuntos
Inteligência Artificial , Infecções por Coronavirus/diagnóstico por imagem , Pulmão/diagnóstico por imagem , Pneumonia Viral/diagnóstico por imagem , Interpretação de Imagem Radiográfica Assistida por Computador , Tomografia Computadorizada por Raios X , Algoritmos , Betacoronavirus , Humanos , Pandemias
16.
Phys Med ; 77: 36-42, 2020 Sep.
Artigo em Inglês | MEDLINE | ID: mdl-32771702

RESUMO

PURPOSE: To assess the impact of iterative reconstructions on image quality and detectability of focal liver lesions in low-energy monochromatic images from a Fast kV-Switching Dual Energy CT (KVSCT) platform. METHODS: Acquisitions on an image-quality phantom were performed using a KVSCT for three dose levels (CTDIvol:12.72/10.76/8.79 mGy). Raw data were reconstructed for five energy levels (40/50/60/70/80 keV) using Filtered Back Projection (FBP) and four levels of ASIR (ASIR30/ASIR50/ASIR70/ASIR100). Noise power spectrum (NPS) and task-based transfer function (TTF) were measured before computing a Detectability index (d') to model the detection task of liver metastasis (LM) and hepatocellular carcinoma (HCC) as function of keV. RESULTS: From 40 to 70 keV, noise-magnitude was reduced on average by -68% ± 1% with FBP; -61% ± 3% with ASIR50 and -52% ± 6% with ASIR100. The mean spatial frequency of the NPS decreased when the energy level decreased and the iterative level increased. TTF values at 50% decreased as the energy level increased and as the percentage of ASIR increased. The detectability of both lesions increased with increasing dose level and percentage of ASIR. For the LM, d' peaked at 70 keV for all reconstruction types, except for ASIR70 at 12.72 mGy and ASIR100, where d' peaked at 50 keV. For HCC, d' peaked at 60 keV for FBP and ASIR30 but peaked at 50 keV for ASIR50, ASIR70 and ASIR100. CONCLUSIONS: Using percentage of ASIR above 50% at low-energy monochromatic images could limit the increase of noise-magnitude, benefit from spatial resolution improvement and hence enhance detectability of subtle low contrast focal liver lesions such as HCC.


Assuntos
Carcinoma Hepatocelular/diagnóstico por imagem , Neoplasias Hepáticas/diagnóstico por imagem , Fígado/diagnóstico por imagem , Tomografia Computadorizada por Raios X/métodos , Humanos , Imagens de Fantasmas , Doses de Radiação , Interpretação de Imagem Radiográfica Assistida por Computador/métodos , Tomografia Computadorizada por Raios X/instrumentação
17.
Medicine (Baltimore) ; 99(34): e21886, 2020 Aug 21.
Artigo em Inglês | MEDLINE | ID: mdl-32846848

RESUMO

BACKGROUND: As the gold standard for imaging sinus disease, the main disadvantage of computed tomography (CT) of the pediatric paranasal sinus is radiation exposure. Because of this, 1 protocol for CT should reduce radiation dose while maintaining image quality. The aim of this study is to evaluate the image quality of dose-reduced paranasal sinus computed tomography (CT) using an ultralow tube voltage (70 kVp) combined with iterative reconstruction (IR) in children. METHODS: CT scans of the paranasal sinus were performed using different protocols [70 kVp protocols with IR, Group A, n = 80; 80 kVp protocols with a filtered back projection algorithm, Group B, n = 80] in 160 pediatric patients. Then, the volume-weighted CT dose index, dose-length product, and effective dose were estimated. Image noise, the signal-to-noise ratio and the diagnostic image quality were also evaluated. RESULTS: For the radiation dose, the volume-weighted CT dose index, dose-length product and effective dose values were significantly lower for the 70 kVp protocols than for the 80 kVp protocols (P < .001). Compared with the 80 kVp protocols, the 70 kVp protocols had significantly higher levels of image noise (P = .001) and a lower signal-to-noise ratio (P = .002). No significant difference in the overall subjective image quality grades was observed between these 2 groups (P = .098). CONCLUSION: The ultralow tube voltage (70 kVp) technique combined with IR enabled a significant dose reduction in CT examinations performed in the pediatric paranasal sinus while maintaining diagnostic image quality with clinically acceptable image noise.


Assuntos
Aumento da Imagem/instrumentação , Doenças dos Seios Paranasais/diagnóstico por imagem , Exposição à Radiação/efeitos adversos , Interpretação de Imagem Radiográfica Assistida por Computador/métodos , Tomografia Computadorizada por Raios X/instrumentação , Adolescente , Algoritmos , Criança , Pré-Escolar , China/epidemiologia , Estudos de Viabilidade , Feminino , Humanos , Masculino , Estudos Prospectivos , Doses de Radiação , Exposição à Radiação/estatística & dados numéricos , Razão Sinal-Ruído
18.
Nat Commun ; 11(1): 4080, 2020 08 14.
Artigo em Inglês | MEDLINE | ID: mdl-32796848

RESUMO

Chest CT is emerging as a valuable diagnostic tool for clinical management of COVID-19 associated lung disease. Artificial intelligence (AI) has the potential to aid in rapid evaluation of CT scans for differentiation of COVID-19 findings from other clinical entities. Here we show that a series of deep learning algorithms, trained in a diverse multinational cohort of 1280 patients to localize parietal pleura/lung parenchyma followed by classification of COVID-19 pneumonia, can achieve up to 90.8% accuracy, with 84% sensitivity and 93% specificity, as evaluated in an independent test set (not included in training and validation) of 1337 patients. Normal controls included chest CTs from oncology, emergency, and pneumonia-related indications. The false positive rate in 140 patients with laboratory confirmed other (non COVID-19) pneumonias was 10%. AI-based algorithms can readily identify CT scans with COVID-19 associated pneumonia, as well as distinguish non-COVID related pneumonias with high specificity in diverse patient populations.


Assuntos
Inteligência Artificial , Técnicas de Laboratório Clínico/métodos , Infecções por Coronavirus/diagnóstico por imagem , Pneumonia Viral/diagnóstico por imagem , Tomografia Computadorizada por Raios X/métodos , Adolescente , Adulto , Idoso , Idoso de 80 Anos ou mais , Algoritmos , Betacoronavirus/isolamento & purificação , Criança , Pré-Escolar , Infecções por Coronavirus/diagnóstico , Infecções por Coronavirus/virologia , Aprendizado Profundo , Feminino , Humanos , Imageamento Tridimensional/métodos , Pulmão/diagnóstico por imagem , Masculino , Pessoa de Meia-Idade , Pandemias , Pneumonia Viral/virologia , Interpretação de Imagem Radiográfica Assistida por Computador/métodos , Adulto Jovem
19.
PLoS One ; 15(8): e0235672, 2020.
Artigo em Inglês | MEDLINE | ID: mdl-32845877

RESUMO

A new computer-aided detection scheme is proposed, the 3D U-Net convolutional neural network, based on multiscale features of transfer learning to automatically detect pulmonary nodules from the thoracic region containing background and noise. The test results can be used as reference information for doctors to assist in the detection of early lung cancer. The proposed scheme is composed of three major steps: First, the pulmonary parenchyma area is segmented by various methods. Then, the 3D U-Net convolutional neural network model with a multiscale feature structure is built. The network model structure is subsequently fine-tuned by the transfer learning method based on weight, and the optimal parameters are selected in the network model. Finally, datasets are extracted to train the fine-tuned 3D U-Net network model to detect pulmonary nodules. The five-fold cross-validation method is used to obtain the experimental results for the LUNA16 and TIANCHI17 datasets. The experimental results show that the scheme not only has obvious advantages in the detection of medium and large-sized nodules but also has an accuracy rate of more than 70% for the detection of small-sized nodules. The scheme provides automatic and accurate detection of pulmonary nodules that reduces the overfitting rate and training time and improves the efficiency of the algorithm. It can assist doctors in the diagnosis of lung cancer and can be extended to other medical image detection and recognition fields.


Assuntos
Imageamento Tridimensional/métodos , Neoplasias Pulmonares/diagnóstico por imagem , Pulmão/diagnóstico por imagem , Redes Neurais de Computação , Nódulo Pulmonar Solitário/diagnóstico por imagem , Algoritmos , Detecção Precoce de Câncer/métodos , Humanos , Aprendizado de Máquina , Interpretação de Imagem Radiográfica Assistida por Computador/métodos , Tomografia Computadorizada por Raios X/métodos
20.
IEEE J Biomed Health Inform ; 24(10): 2776-2786, 2020 10.
Artigo em Inglês | MEDLINE | ID: mdl-32750973

RESUMO

Fast and accurate diagnosis is essential for the efficient and effective control of the COVID-19 pandemic that is currently disrupting the whole world. Despite the prevalence of the COVID-19 outbreak, relatively few diagnostic images are openly available to develop automatic diagnosis algorithms. Traditional deep learning methods often struggle when data is highly unbalanced with many cases in one class and only a few cases in another; new methods must be developed to overcome this challenge. We propose a novel activation function based on the generalized extreme value (GEV) distribution from extreme value theory, which improves performance over the traditional sigmoid activation function when one class significantly outweighs the other. We demonstrate the proposed activation function on a publicly available dataset and externally validate on a dataset consisting of 1,909 healthy chest X-rays and 84 COVID-19 X-rays. The proposed method achieves an improved area under the receiver operating characteristic (DeLong's p-value < 0.05) compared to the sigmoid activation. Our method is also demonstrated on a dataset of healthy and pneumonia vs. COVID-19 X-rays and a set of computerized tomography images, achieving improved sensitivity. The proposed GEV activation function significantly improves upon the previously used sigmoid activation for binary classification. This new paradigm is expected to play a significant role in the fight against COVID-19 and other diseases, with relatively few training cases available.


Assuntos
Algoritmos , Betacoronavirus , Técnicas de Laboratório Clínico/métodos , Infecções por Coronavirus/diagnóstico , Pandemias , Pneumonia Viral/diagnóstico , Teorema de Bayes , Técnicas de Laboratório Clínico/estatística & dados numéricos , Biologia Computacional , Infecções por Coronavirus/diagnóstico por imagem , Infecções por Coronavirus/epidemiologia , Bases de Dados Factuais/estatística & dados numéricos , Aprendizado Profundo , Humanos , Redes Neurais de Computação , Pneumonia Viral/diagnóstico por imagem , Pneumonia Viral/epidemiologia , Interpretação de Imagem Radiográfica Assistida por Computador/métodos , Tomografia Computadorizada por Raios X/estatística & dados numéricos
SELEÇÃO DE REFERÊNCIAS
DETALHE DA PESQUISA