RESUMEN
BACKGROUND: Notwithstanding the encouraging results of previous studies reporting on the efficiency of deep learning (DL) in COVID-19 prognostication, clinical adoption of the developed methodology still needs to be improved. To overcome this limitation, we set out to predict the prognosis of a large multi-institutional cohort of patients with COVID-19 using a DL-based model. PURPOSE: This study aimed to evaluate the performance of deep privacy-preserving federated learning (DPFL) in predicting COVID-19 outcomes using chest CT images. METHODS: After applying inclusion and exclusion criteria, 3055 patients from 19 centers, including 1599 alive and 1456 deceased, were enrolled in this study. Data from all centers were split (randomly with stratification respective to each center and class) into a training/validation set (70%/10%) and a hold-out test set (20%). For the DL model, feature extraction was performed on 2D slices, and averaging was performed at the final layer to construct a 3D model for each scan. The DensNet model was used for feature extraction. The model was developed using centralized and FL approaches. For FL, we employed DPFL approaches. Membership inference attack was also evaluated in the FL strategy. For model evaluation, different metrics were reported in the hold-out test sets. In addition, models trained in two scenarios, centralized and FL, were compared using the DeLong test for statistical differences. RESULTS: The centralized model achieved an accuracy of 0.76, while the DPFL model had an accuracy of 0.75. Both the centralized and DPFL models achieved a specificity of 0.77. The centralized model achieved a sensitivity of 0.74, while the DPFL model had a sensitivity of 0.73. A mean AUC of 0.82 and 0.81 with 95% confidence intervals of (95% CI: 0.79-0.85) and (95% CI: 0.77-0.84) were achieved by the centralized model and the DPFL model, respectively. The DeLong test did not prove statistically significant differences between the two models (p-value = 0.98). The AUC values for the inference attacks fluctuate between 0.49 and 0.51, with an average of 0.50 ± 0.003 and 95% CI for the mean AUC of 0.500 to 0.501. CONCLUSION: The performance of the proposed model was comparable to centralized models while operating on large and heterogeneous multi-institutional datasets. In addition, the model was resistant to inference attacks, ensuring the privacy of shared data during the training process.
Asunto(s)
COVID-19 , Aprendizaje Profundo , Tomografía Computarizada por Rayos X , COVID-19/diagnóstico por imagen , Humanos , Pronóstico , Masculino , Femenino , Anciano , Persona de Mediana Edad , Privacidad , Radiografía Torácica , Conjuntos de Datos como AsuntoRESUMEN
Objective: To evaluate the frequency and significance of brain imaging findings in methanol poisoning patients and to propose a criterion for prioritizing brain imaging. Methods: We retrospectively reviewed the data of 306 patients (286 men and 34 women, mean age 32.10 ± 9.9 years) with confirmed methanol poisoning who were admitted to two hospitals in Iran during the COVID-19 pandemic. We analyzed their demographic, clinical, laboratory, and brain imaging data. Results: The main brain computed tomography (CT) scan findings were hypodensity in the putamen (11.1%), cerebellar nuclei (8.2%), diffuse cerebral edema (7.5%), and intracranial hemorrhage (ICH; 1.6%). These findings were associated with blood pH, Glasgow Coma Scale (GCS), renal failure, bicarbonate, oxygen, carbon dioxide, potassium, and glucose levels (p < 0.05). Poor prognosis was related to blindness, opium addiction, chronic alcohol use, hyperglycemia, and abnormal CT scans (p < 0.001 for all). The most predictive brain imaging findings for poor prognosis were hypodensity in the cerebellar nuclei, diffuse cerebral edema, and ICH. Conclusion: Brain imaging can provide valuable information for the diagnosis and management of methanol poisoning patients. We suggest that patients with severe acidosis, low GCS, low pH, low oxygen saturation, and high glucose levels should undergo brain CT scan as a priority.
RESUMEN
Abernethy malformation or congenital portosystemic shunt is a rare congenital vascular malformation and anomaly of the splanchnic venous system defined by diverting portal blood away from the liver. It is commonly associated with multiple congenital anomalies. Imaging modalities such as computed tomography or magnetic resonance have a crucial role in prompting diagnosis and determining the prognosis based on the type of malformation and associated anomalies. Misdiagnosis could be harmful and may lead to inappropriate treatment. We present a case of Abernethy malformation with a complete end-to-side shunt of portal venous flow into the systemic venous flow and complete bypass of the liver, which was initially misdiagnosed with portal venous thrombosis.
RESUMEN
The thoracic kidney is the rarest form of renal ectopia. Furthermore, it is usually asymptomatic and discovered incidentally. It is seen as a mass in the posterior mediastinum or juxta-diaphragmatic on chest radiography. A computed tomography scan or magnetic resonance imaging is usually needed for a definitive diagnosis. The thoracic kidney typically exits the retroperitoneal space through the foramen of Bochdalek.
RESUMEN
BACKGROUND: We aimed to analyze the prognostic power of CT-based radiomics models using data of 14,339 COVID-19 patients. METHODS: Whole lung segmentations were performed automatically using a deep learning-based model to extract 107 intensity and texture radiomics features. We used four feature selection algorithms and seven classifiers. We evaluated the models using ten different splitting and cross-validation strategies, including non-harmonized and ComBat-harmonized datasets. The sensitivity, specificity, and area under the receiver operating characteristic curve (AUC) were reported. RESULTS: In the test dataset (4,301) consisting of CT and/or RT-PCR positive cases, AUC, sensitivity, and specificity of 0.83 ± 0.01 (CI95%: 0.81-0.85), 0.81, and 0.72, respectively, were obtained by ANOVA feature selector + Random Forest (RF) classifier. Similar results were achieved in RT-PCR-only positive test sets (3,644). In ComBat harmonized dataset, Relief feature selector + RF classifier resulted in the highest performance of AUC, reaching 0.83 ± 0.01 (CI95%: 0.81-0.85), with a sensitivity and specificity of 0.77 and 0.74, respectively. ComBat harmonization did not depict statistically significant improvement compared to a non-harmonized dataset. In leave-one-center-out, the combination of ANOVA feature selector and RF classifier resulted in the highest performance. CONCLUSION: Lung CT radiomics features can be used for robust prognostic modeling of COVID-19. The predictive power of the proposed CT radiomics model is more reliable when using a large multicentric heterogeneous dataset, and may be used prospectively in clinical setting to manage COVID-19 patients.
Asunto(s)
COVID-19 , Neoplasias Pulmonares , Algoritmos , COVID-19/diagnóstico por imagen , Humanos , Aprendizaje Automático , Pronóstico , Estudios Retrospectivos , Tomografía Computarizada por Rayos X/métodosRESUMEN
Inferior epigastric artery (IEA) pseudoaneurysm is a rare complication following abdominal wall procedures near the artery. This is a case of inferior epigastric artery pseudoaneurysm after therapeutic paracentesis for large-volume ascites caused by chronic kidney failure. The patient was operated on, and the artery was ligated.
RESUMEN
BACKGROUND: This study was performed with the intention of comparing the clinical, laboratory, and chest computed tomography (CT) findings between severe and non-severe patients as well as between different age groups composed of pediatric patients with confirmed COVID-19. METHOD: This study was carried out on a total of 53 confirmed COVID-19 pediatric patients who were hospitalized in Namazi and Ali Asghar Hospitals, Shiraz, Iran. The patients were divided into two severe (n = 27) and non-severe (n = 28) groups as well as into other three groups in terms of their age: aged less than two years, aged 3-12 years and 13-17 years. It should be noted that CT scans, laboratory, and clinical features were taken from all patients at the admission time. Abnormal chest CT in COVID-19 pneumonia was found to show one of the following findings: ground-glass opacities (GGO), bilateral involvement, peripheral and diffuse distribution. RESULT: Fever (79.2%) and dry cough (75.5%) were the most common clinical symptoms. Severe COVID-19 patients showed lymphocytosis, while the non-severe ones did not (P = 0.03). C-reactive protein (CRP) was shown to be significantly lower in patients aged less than two years than those aged 3-12 and 13-17 years (P = 0.01). It was shown also that O2 saturation experienced a significant increase as did patients' age (P = 0.01). Severe patients had significantly higher CT abnormalities than non-severe patients (48.0% compared to 17.9%, respectively) (P = 0.02). CONCLUSION: Lymphocytosis and abnormal CT findings are among the factors most associated with COVID-19 severity. It was, moreover, showed that the severity of COVID-19, O2 saturation, and respiratory distress were improved as the age of confirmed COVID-19 pediatric patients increased.