RESUMEN
OBJECTIVE: The primary objective was to assess the diagnostic accuracy of a deep learning-based artificial intelligence model for the detection of acute appendicular fractures in pediatric patients presenting with a recent history of trauma to the emergency department. The secondary goal was to examine the effect of assistive support on the emergency doctor's ability to detect fractures. METHODS: The dataset was 5,150 radiographs of which 850 showed fractures, while 4,300 radiographs did not show any fractures. The process utilized 4,532 (88%) radiographs, inclusive of both fractured and non-fractured radiographs, in the training phase. Subsequently, 412 (8%) radiographs were appraised during validation, and 206 (4%) were set apart for the testing phase. With and without artificial intelligence assistance, the emergency doctor reviewed another set of 2,000 radiographs (400 fractures and 600 non-fractures each) for labeling in the second test. RESULTS: The artificial intelligence model showed a mean average precision 50 of 89%, a specificity of 92%, a sensitivity of 90%, and an F1 score of 90%. The confusion matrix revealed that the model trained with artificial intelligence achieved accuracies of 93 and 95% in detecting fractures, respectively. Artificial intelligence assistance improved the reading sensitivity from 93.7% (without assistance) to 97.0% (with assistance) and the reading accuracy from 88% (without assistance) to 94.9% (with assistance). CONCLUSION: A deep learning-based artificial intelligence model has proven to be highly effective in detecting fractures in pediatric patients, enhancing the diagnostic capabilities of emergency doctors through assistive support.
Asunto(s)
Inteligencia Artificial , Fracturas Óseas , Humanos , Fracturas Óseas/diagnóstico por imagen , Niño , Preescolar , Sensibilidad y Especificidad , Femenino , Aprendizaje Profundo , Servicio de Urgencia en Hospital , Masculino , Reproducibilidad de los Resultados , Radiografía/métodos , Adolescente , LactanteRESUMEN
OBJECTIVE: The objective of this study was to investigate whether the measurement of mean optic nerve sheath diameter in patients with transient ischemic attack could be used to distinguish between control groups, the acute ischemic stroke group, and subgroups within the acute ischemic stroke category. METHODS: Retrospectively, the mean optic nerve sheath diameters of patients aged 18 years and older belonging to control, transient ischemic attack, acute ischemic stroke, and subgroups within the acute ischemic stroke category were measured with initial computed tomography conducted in the emergency department. RESULTS: Out of the 773 patients included in the study, 318 (41.1%) were in the control group, 77 (10%) had transient ischemic attack, and 378 (49%) were categorized as stroke patients. The average mean optic nerve sheath diameter was significantly higher in both the stroke and transient ischemic attack groups compared with the control group (p<0.001 for both comparisons). Furthermore, the mean optic nerve sheath diameter in the stroke subgroups was significantly higher than in both the transient ischemic attack and control groups (p<0.001 for all comparisons). In transient ischemic attack patients, the mean optic nerve sheath diameter showed a significant ability to predict transient ischemic attack (AUC=0.913, p<0.001), with a calculated optimal cutoff value of 4.72, sensitivity of 94.8%, and specificity of 73.9%. CONCLUSION: The mean optic nerve sheath diameter of patients in the transient ischemic attack group was lower compared with those in the stroke subgroups but higher compared with the control group.
Asunto(s)
Ataque Isquémico Transitorio , Accidente Cerebrovascular Isquémico , Accidente Cerebrovascular , Humanos , Ataque Isquémico Transitorio/diagnóstico por imagen , Estudios Retrospectivos , Accidente Cerebrovascular/diagnóstico por imagen , Tomografía Computarizada por Rayos X , Nervio Óptico/diagnóstico por imagenRESUMEN
SUMMARY OBJECTIVE: The objective of this study was to investigate whether the measurement of mean optic nerve sheath diameter in patients with transient ischemic attack could be used to distinguish between control groups, the acute ischemic stroke group, and subgroups within the acute ischemic stroke category. METHODS: Retrospectively, the mean optic nerve sheath diameters of patients aged 18 years and older belonging to control, transient ischemic attack, acute ischemic stroke, and subgroups within the acute ischemic stroke category were measured with initial computed tomography conducted in the emergency department. RESULTS: Out of the 773 patients included in the study, 318 (41.1%) were in the control group, 77 (10%) had transient ischemic attack, and 378 (49%) were categorized as stroke patients. The average mean optic nerve sheath diameter was significantly higher in both the stroke and transient ischemic attack groups compared with the control group (p<0.001 for both comparisons). Furthermore, the mean optic nerve sheath diameter in the stroke subgroups was significantly higher than in both the transient ischemic attack and control groups (p<0.001 for all comparisons). In transient ischemic attack patients, the mean optic nerve sheath diameter showed a significant ability to predict transient ischemic attack (AUC=0.913, p<0.001), with a calculated optimal cutoff value of 4.72, sensitivity of 94.8%, and specificity of 73.9%. CONCLUSION: The mean optic nerve sheath diameter of patients in the transient ischemic attack group was lower compared with those in the stroke subgroups but higher compared with the control group.
RESUMEN
SUMMARY OBJECTIVE: The primary objective was to assess the diagnostic accuracy of a deep learning-based artificial intelligence model for the detection of acute appendicular fractures in pediatric patients presenting with a recent history of trauma to the emergency department. The secondary goal was to examine the effect of assistive support on the emergency doctor's ability to detect fractures. METHODS: The dataset was 5,150 radiographs of which 850 showed fractures, while 4,300 radiographs did not show any fractures. The process utilized 4,532 (88%) radiographs, inclusive of both fractured and non-fractured radiographs, in the training phase. Subsequently, 412 (8%) radiographs were appraised during validation, and 206 (4%) were set apart for the testing phase. With and without artificial intelligence assistance, the emergency doctor reviewed another set of 2,000 radiographs (400 fractures and 600 non-fractures each) for labeling in the second test. RESULTS: The artificial intelligence model showed a mean average precision 50 of 89%, a specificity of 92%, a sensitivity of 90%, and an F1 score of 90%. The confusion matrix revealed that the model trained with artificial intelligence achieved accuracies of 93 and 95% in detecting fractures, respectively. Artificial intelligence assistance improved the reading sensitivity from 93.7% (without assistance) to 97.0% (with assistance) and the reading accuracy from 88% (without assistance) to 94.9% (with assistance). CONCLUSION: A deep learning-based artificial intelligence model has proven to be highly effective in detecting fractures in pediatric patients, enhancing the diagnostic capabilities of emergency doctors through assistive support.