Radiomic-based nonlinear supervised learning classifiers on non-contrast CT to predict functional prognosis in patients with spontaneous intracerebral hematoma.

PURPOSE: To evaluate if nonlinear supervised learning classifiers based on non-contrast CT can predict functional prognosis at discharge in patients with spontaneous intracerebral hematoma. METHODS: Retrospective, single-center, observational analysis of patients with a diagnosis of spontaneous intracerebral hematoma confirmed by non-contrast CT between January 2016 and April 2018. Patients with HIE > 18 years and with TCCSC performed within the first 24 h of symptom onset were included. Patients with secondary spontaneous intracerebral hematoma and in whom radiomic variables were not available were excluded. Clinical, demographic and admission variables were collected. Patients were classified according to the Modified Rankin Scale (mRS) at discharge into good (mRS 0-2) and poor prognosis (mRS 3-6). After manual segmentation of each spontaneous intracerebral hematoma, the radiomics variables were obtained. The sample was divided into a training and testing cohort and a validation cohort (70-30% respectively). Different methods of variable selection and dimensionality reduction were used, and different algorithms were used for model construction. Stratified 10-fold cross-validation were performed on the training and testing cohort and the mean area under the curve (AUC) were calculated. Once the models were trained, the sensitivity of each was calculated to predict functional prognosis at discharge in the validation cohort. RESULTS: 105 patients with spontaneous intracerebral hematoma were analyzed. 105 radiomic variables were evaluated for each patient. P-SVM, KNN-E and RF-10 algorithms, in combination with the ANOVA variable selection method, were the best performing classifiers in the training and testing cohort (AUC 0.798, 0.752 and 0.742 respectively). The predictions of these models, in the validation cohort, had a sensitivity of 0.897 (0.778-1;95%CI), with a false-negative rate of 0% for predicting poor functional prognosis at discharge. CONCLUSION: The use of radiomics-based nonlinear supervised learning classifiers are a promising diagnostic tool for predicting functional outcome at discharge in HIE patients, with a low false negative rate, although larger and balanced samples are still needed to develop and improve their performance.

Clasificadores de aprendizaje supervisado no lineales basados en radiómica de la TC cerebral sin contraste para predecir el pronóstico funcional en pacientes con hematoma intracerebral espontáneo / Radiomic-based nonlinear supervised learning classifiers on non-contrast CT to predict functional prognosis in patients with spontaneous intracerebral hematoma

Objetivo: Evaluar si clasificadores de aprendizaje supervisado no lineales basados en radiómica de la TC cerebral sin contraste (TCCSC) pueden predecir el pronóstico funcional al alta en pacientes con hematoma intracerebral espontáneo (HIE). Material y método: Análisis observacional retrospectivo y unicéntrico de pacientes con diagnóstico de HIE confirmado por TCCSC entre enero de 2016 y abril de 2018. Se incluyeron pacientes con HIE >18años y con TCCSC realizado dentro de las primeras 24horas del inicio de los síntomas. Se excluyeron los HIE secundarios y en los que no se disponía de las variables de radiómica. Se recogieron datos clínicos, demográficos y variables al ingreso. Los pacientes se clasificaron según la Escala Modificada de Rankin (mRS) al alta en buen (mRS0-2) y mal pronóstico (mRS3-6). Tras la segmentación manual de la TCCSC de cada HIE se obtuvieron las variables de radiómica. La muestra se dividió en una cohorte de entrenamiento y prueba y otra cohorte de validación (70-30%, respectivamente). Se usaron diferentes métodos de selección de variables y reducción de dimensionalidad, así como diferentes algoritmos para la construcción del modelo. Se realizaron 10 iteraciones de validación cruzada estratificada en la cohorte de entrenamiento y prueba y se calculó la media de los valores de área bajo la curva (AUC). Una vez entrenados los modelos, se calculó la sensibilidad de cada uno para predecir el pronóstico funcional al alta en la cohorte de validación. Resultados: Se analizaron 105 pacientes con HIE. Se evaluaron 105 variables de radiómica de cada paciente. Los algoritmos P-SVM, KNN-E y RF-10, en combinación con el método de selección de variables ANOVA, fueron los clasificadores con mejor rendimiento en la cohorte de entrenamiento y prueba (AUC: 0,798, 0,752 y 0,742, respectivamente)...(AU)

Purpose: To evaluate if nonlinear supervised learning classifiers based on non-contrast cerebral CT can predict functional prognosis at discharge in patients with spontaneous intracerebral hematoma (HIE). Methods: Retrospective, single-center, observational analysis of patients with a diagnosis of spontaneous intracerebral hematoma confirmed by non-contrast CT between January 2016 and April 2018. Patients with HIE >18years and with non-contrast CT performed within the first 24hours of symptom onset were included. Patients with secondary spontaneous intracerebral hematoma and in whom radiomic variables were not available were excluded. Clinical, demographic and admission variables were collected. Patients were classified according to the Modified Rankin Scale (mRS) at discharge into good (mRS0-2) and poor prognosis (mRS3-6). After manual segmentation of each spontaneous intracerebral hematoma, the radiomics variables were obtained. The sample was divided into a training and testing cohort and a validation cohort (70-30%, respectively). Different methods of variable selection and dimensionality reduction were used, and different algorithms were used for model construction. Stratified 10-fold cross-validation were performed on the training and testing cohort and the mean area under the curve (AUC) were calculated. Once the models were trained, the sensitivity of each was calculated to predict functional prognosis at discharge in the validation cohort. Results: 105 patients with spontaneous intracerebral hematoma were analyzed. 105 radiomic variables were evaluated for each patient. P-SVM, KNN-E and RF-10 algorithms, in combination with the ANOVA variable selection method, were the best performing classifiers in the training and testing cohort (AUC: 0.798, 0.752 and 0.742, respectively)...(AU)

Utilidad de la angio-TC para la caracterización de malformaciones arteriovenosas cerebrales con presentación hemorrágica comparada con la angiografía por sustracción digital / Usefulness of CT angiography for characterizing cerebral arteriovenous malformations presenting as hemorrhage: comparison with digital subtraction angiography

OBJETIVO: El objetivo es determinar la utilidad de la angio-TC cerebral en la caracterización de las malformaciones arteriovenosas (MAV) cerebrales con presentación hemorrágica comparada con la angiografía por sustracción digital (DSA) como patrón de referencia. MATERIAL Y MÉTODOS: Se realizó un análisis retrospectivo de una base de datos prospectiva de pacientes con sangrado intracraneal debido a una MAV cerebral desde enero de 2007 hasta diciembre de 2012. Se revisaron variables radiológicas, como las características de la malformación (tamaño, localización, presencia de drenaje venoso profundo), afectación de un área elocuente y presencia de aneurismas relacionados. Dos neurorradiólogos ciegos a cualquier información clínico-radiológica analizaron por consenso las imágenes de tomografía computarizada y DSA. RESULTADOS: Veintidós pacientes fueron incluidos en el estudio. La angio-TC clasificó correctamente 15 de los 16 casos de MAV menores de 3cm, con una sensibilidad del 93,75%. Todos los casos con drenaje venoso profundo y localizados en un área elocuente fueron correctamente detectados (sensibilidad 100%). La presencia de cualquier tipo de aneurisma relacionado con la MAV fue detectada en 13 de 15 pacientes (sensibilidad 86,6%); 7 de 9 en los intranidales (sensibilidad 77,78%) y 6 de 9 de los aneurismas de flujo (sensibilidad 66,67%). CONCLUSIÓN: La angio-TC tiene una alta sensibilidad en la caracterización de MAV cerebrales en cuanto al tamaño menor de 3cm, localización en área elocuente, presencia de drenaje venoso profundo y la detección de cualquier aneurisma relacionado con la MAV. Sin embargo, la angio-TC tiene una menor sensibilidad en la detección de aneurismas intranidales y de flujo relacionados con la MAV

OBJECTIVE: To compare the usefulness of CT angiography against the gold standard, digital subtraction angiography (DSA), in the characterization of cerebral arteriovenous malformations (AVM) that present with bleeding. MATERIAL AND METHODS: We retrospectively analyzed patients with intracranial bleeding due to an AVM who were included in a prospective database in the period comprising January 2007 through December 2012. We reviewed radiologic variables such as the characteristics of the AVM (size, location, presence of deep venous drainage), involvement of eloquent areas, and the presence of associated aneurysms. Two neuroradiologists blinded to clinical and radiological information analyzed the CT and DSA in consensus. RESULTS: A total of 22 patients were included in the study. CT angiography correctly classified 15 of the 16 cases of AVM measuring less than 3cm (93.75% sensitivity). All cases of deep venous drainage and all those located in eloquent areas were correctly detected (100% sensitivity). The presence of any type of aneurysm related with the AVM was detected in 13 of 15 cases (86.6% sensitivity); 7 of 9 of the intranidal aneurysms were detected (77.78% sensitivity), as were 6 of the 9 flow aneurysms (66.67% sensitivity). CONCLUSION: CT angiography is highly sensitive in the characterization of cerebral AVMs measuring less than 3cm, of those located in eloquent areas, and of those with deep venous drainage; it is also highly sensitive in detecting aneurysms related with AVMs. However, CT angiography is less sensitive in detecting intranidal and flow aneurysms related with AVMs

Usefulness of CT angiography for characterizing cerebral arteriovenous malformations presenting as hemorrhage: comparison with digital subtraction angiography. / Utilidad de la angio-TC para la caracterización de malformaciones arteriovenosas cerebrales con presentación hemorrágica comparada con la angiografía por sustracción digital.

OBJECTIVE: To compare the usefulness of CT angiography against the gold standard, digital subtraction angiography (DSA), in the characterization of cerebral arteriovenous malformations (AVM) that present with bleeding. MATERIAL AND METHODS: We retrospectively analyzed patients with intracranial bleeding due to an AVM who were included in a prospective database in the period comprising January 2007 through December 2012. We reviewed radiologic variables such as the characteristics of the AVM (size, location, presence of deep venous drainage), involvement of eloquent areas, and the presence of associated aneurysms. Two neuroradiologists blinded to clinical and radiological information analyzed the CT and DSA in consensus. RESULTS: A total of 22 patients were included in the study. CT angiography correctly classified 15 of the 16 cases of AVM measuring less than 3cm (93.75% sensitivity). All cases of deep venous drainage and all those located in eloquent areas were correctly detected (100% sensitivity). The presence of any type of aneurysm related with the AVM was detected in 13 of 15 cases (86.6% sensitivity); 7 of 9 of the intranidal aneurysms were detected (77.78% sensitivity), as were 6 of the 9 flow aneurysms (66.67% sensitivity). CONCLUSION: CT angiography is highly sensitive in the characterization of cerebral AVMs measuring less than 3cm, of those located in eloquent areas, and of those with deep venous drainage; it is also highly sensitive in detecting aneurysms related with AVMs. However, CT angiography is less sensitive in detecting intranidal and flow aneurysms related with AVMs.