Evaluación morfométrica de neuroimágenes como factor pronóstico de supervivencia en pacientes con hemorragia cerebral

Abstract

Objective

To describe a strategy that allows to evaluate the morphometry effects of images of computed tomography (CT) as a prognostic factor for the survival of patients with cerebral hemorrhages.

Materials and methods

To explain and/or predict a patient clinic response in relation to morphometric variables of CT images, a latent variable approach is proposed that uses the following statistical tools: logistic regression, multilevel modeling and principal component analysis. To illustrate the methodology, data from the medical records of 39 patients with cerebral hemorrhage were used. Five morphometric indicators were measured in each of the 140 collected CT images.

Results

The application of the strategy allowed to make inferences at patient level, combining in a single predictive model of latent variables, the morphometric information of the CT images and characteristics of the patients such as age and sex. The results reveal that the latent variables represent in a synthetic way, the morphometric differences of the images between the patients and that these can be considered an important predictor of the survival in certain groups of individuals.

Conclusions

The proposed strategy provides an appropriate tool to evaluate the effect of morphometry of CT images on some clinical response of the patient. On the basis of a real example, the utility of the strategy in the development of a prognostic model of the survival of patients after a stroke was demonstrated.