[Pneumonia in the inmunocompromised host: perspective through images diagnosis and a Bayesian inference]

Abstract

Introduction: The population of immunocompromised patients has increased in recent decades. Many of these patients eventually present infectious complications including pneumonia, which is a diagnostic that must to be prompt and accurate.

Objective: To review the basis of the diagnosis of pneumonia in the immunocompromised patient. Sorted by the methodology of Bayesian inference, very relevant in the diagnostic attribution, we review the main basis of the diagnosis of pneumonia of immunocompromised patients: the epidemiology, the clinical history including the type of immunosuppression that weigh the likelihood of attribution a priori of an etiologic agent, and finally, the findings in the image (or likelihood function).

Conclusion: Although in general the findings are not pathognomonic and there is much overlap in the images, there are several features that orient in one direction or another. Proper assessment of the prior probability and the likelihood function is allowing ultimately a good diagnostic proposition.