RESUMEN
This review discusses pooled experience of creation, implementation and effectiveness of machine learning technologies in CT-based diagnosis of intracranial hemorrhages. The authors analyzed 21 original articles between 2015 and 2022 using the following keywords: «intracranial hemorrhage¼, «machine learning¼, «deep learning¼, «artificial intelligence¼. The review contains general data on basic concepts of machine learning and also considers in more detail such aspects as technical characteristics of data sets used for creation of AI algorithms for certain type of clinical task, their possible impact on effectiveness and clinical experience.
Asunto(s)
Inteligencia Artificial , Aprendizaje Automático , Humanos , Algoritmos , Hemorragias Intracraneales/diagnóstico por imagen , Tomografía Computarizada por Rayos XRESUMEN
BACKGROUND: Pathological low-energy (LE) vertebral compression fractures (VFs) are common complications of osteoporosis and predictors of subsequent LE fractures. In 84% of cases, VFs are not reported on chest CT (CCT), which calls for the development of an artificial intelligence-based (AI) assistant that would help radiology specialists to improve the diagnosis of osteoporosis complications and prevent new LE fractures. AIMS: To develop an AI model for automated diagnosis of compression fractures of the thoracic spine based on chest CT images. MATERIALS AND METHODS: Between September 2019 and May 2020 the authors performed a retrospective sampling study of ССТ images. The 160 of results were selected and anonymized. The data was labeled by seven readers. Using the morphometric analysis, the investigators received the following metric data: ventral, medial and dorsal dimensions. This was followed by a semiquantitative assessment of VFs degree. The data was used to develop the Comprise-G AI mode based on CNN, which subsequently measured the size of the vertebral bodies and then calculates the compression degree. The model was evaluated with the ROC curve analysis and by calculating sensitivity and specificity values. RESULTS: Formed data consist of 160 patients (a training group - 100 patients; a test group - 60 patients). The total of 2,066 vertebrae was annotated. When detecting Grade 2 and 3 maximum VFs in patients the Comprise-G model demonstrated sensitivity - 90,7%, specificity - 90,7%, AUC ROC - 0.974 on the 5-FOLD cross-validation data of the training dataset; on the test data - sensitivity - 83,2%, specificity - 90,0%, AUC ROC - 0.956; in vertebrae demonstrated sensitivity - 91,5%, specificity - 95,2%, AUC ROC - 0.981 on the cross-validation data; for the test data sensitivity - 79,3%, specificity - 98,7%, AUC ROC - 0.978. CONCLUSIONS: The Comprise-G model demonstrated high diagnostic capabilities in detecting the VFs on CCT images and can be recommended for further validation.
Asunto(s)
Fracturas por Compresión , Fracturas de la Columna Vertebral , Inteligencia Artificial , Fracturas por Compresión/diagnóstico , Humanos , Redes Neurales de la Computación , Estudios Retrospectivos , Fracturas de la Columna Vertebral/diagnósticoRESUMEN
Direct visualization of rapid cerebrospinal fluid movements is a topical task of neurosurgery, which has applications such as evaluating hydrocephalus and the effectiveness of 3rd ventriculostomy. PURPOSE: The study purpose was to evaluate the capabilities of a modified Time-SLIP pulse MRI sequence for visualization of fluid (CSF) movements in the phantom, healthy subject, and patient. MATERIAL AND METHODS: The study was performed in a phantom simulating pulsed CSF movements, healthy volunteers (9 people), and patients without impaired CSF dynamics (12 people), whose data were used to determine mean CSF flow parameters, as well as in 1 patient after 3rd ventriculostomy. A 1.5 T MRI instrument was used. The Time-SLIP parameters were as follows: TR = 8,500 ms; TEeff = 80 ms; Thk = 5.0 mm; tag spacing = 30 mm; NEX 7; inversion time (BBTI) = 2,000/3,000 ms; no cardiosynchronization. Scanning time was 2:16 min. The estimated parameter was the length of motion (LOM) of CSF. RESULTS: According to a study on a phantom simulating various conditions of oscillatory fluid motion, the mean LOM determination error in the modified Time-SLIP mode was 20%. This technique provided the following LOM data for the cerebral aqueduct (median, 25-75% quartiles): 13.0 (9.5-16.0) mm for BBTI of 2,000ms and 30.2 (23.7-35.3) mm for BBTI of 3,000 ms, i.e. 2.3-fold higher. This difference may be explained by an intense turbulent current leading to rapid CSF exchange between the 3rd and 4th ventricles and prolonged CSF movement during several heart contractions. Quantitative parameters of CSF movement at the C1-C2 level were determined. Additionally, Time-SLIP was used to evaluate performance of a third ventricle fistula. CONCLUSION: We have proposed a modified Time-SLIP pulse sequence that does not require cardiosynchronization. The mean relative error in determining the CSF movement distance was 20%. The mean quantitative parameters of CSF movement in the cerebral aqueduct and at the C1-C2 level were obtained. Turbulent CSF flow is found in the cerebral aqueduct, which leads to rapid exchange between the 3rd and 4th ventricles.