Science and Technology

Technique allows more precise and integrated analysis of brain lesions

Method developed at the State University of Campinas assist in the diagnosis of lupus, multiple sclerosis and stroke

A new technique developed at the State University of Campinas (Unicamp) allows us to analyze in an integrated and more precisely the different types of brain injuries. The method, automated, assist doctors and specialists in the diagnosis of white matter changes in the brain that may be associated with multiple sclerosis, lupus and cerebrovascular accident (CVA).

The study, preliminary, was conducted by engineer teleinformatics Mariana Pinheiro Benedict. Some results have been presented at this year's Congress of the International Society for Optics and Photonics (SPIE, its acronym in English).

The method may have advantages, mainly for investigating lesions have not yet been characterized, for example, those caused by lupus, autoimmune disease rare. The technique, explains the researcher Mariana Bento, you can collaborate both in diagnosis and in the monitoring of patients.

Even we can not know the future only if the lesions increased or decreased, but as they increased or decreased and how it happened. This becomes very important to monitor the patient over time, reinforces the engineer.

Still according to her, another differential compared to conventional methods is the ability to analyze comparatively injuries caused by different diseases. There is software on the market that make the analysis of certain disease and injury. We are not aware of methods for integrated analysis of different types of disease, as is our proposal, in contrast Mariana Benedict.

In her research, she was guided by the teacher Roberto de Alencar Lotufo, Department of Computer Engineering of FEEC. The same department, Professor Leticia Rittner coorientou work. The study also relied on the collaboration of the Hospital de Clinicas (HC) Simone Appenzeller, a professor in the Faculty of Medical Sciences (FCM).

The Unicamp provided medical magnetic resonance images of patients HC used to subsidize research informs Mariana Benedict. The lesions in the white matter of the brain are often found in neurological diagnosis, whether symptomatic or asymptomatic. Monitoring and analysis are very important for diagnosis. Currently, magnetic resonance imaging of the brain is the kind of examination that conveys more information to the physician. Only the analyzes are made based only on visual observations, ponders.

Our methodology proposes automatic analysis of a region to tell whether or not it is an injury and also what type of injury if positive

The researcher explains that the conventional method, an expert opens the magnetic resonance imaging in certain software and lesions observed visually, frame by frame. Differentiate, visually, the types of injuries is very difficult for doctors because changes in the white matter are very similar. Our methodology proposes automatic analysis of a region to tell whether or not it is an injury and also what type of injury and if so, compare.

The goal, she explains, is not to replace the doctor, but help you. It indicates which region has a higher likelihood of injury and point to what would be its origin. It is a more robust analysis, based on attributes and other information than just visual ones, says the scholar.

The techniques will propose measures of longitudinal patient. For example, it begins to be treated and four months and information attributes are compared to these lesions over time, as the thickness and volume changes. Thus, the doctor can more accurately examine if the treatment is progressing or not complete.

<i> information Unicamp </ i>