Accurate method for interpretation of heart images developed

VTT - Technical Research Centre of Finland has developed an internationally significant method for interpreting MRI images of the heart.

Myocardial diseases are among the most common causes of death in Western countries. Magnetic resonance imaging (MRI) of the heart is the most accurate imaging method that assesses the function of the cardiac muscle. As MRI produces hundreds of images, the analyzing process is very time-consuming.

The new method facilitates more accurate diagnoses, based on the interpretation of atrial and ventricular volumes from the images. Automated, or even semi-automated methods have not been available previously for these four cavities. Thanks to the new method, heart diseases can now be diagnosed with more accuracy at an early stage, thus speeding up therapy and reducing costs.

Automated analysis of heart images is a hot topic in medical discussions, since automated, accurate methods for studying the volumetry of atria and ventricles are not available. One MRI scan usually involves some 250 images of the heart. Determining the key indices for patient diagnosis - such as atrial and ventricular volumes - often requires a great deal of human effort, yet time is usually limited in clinical work. The new method allows the specialist to focus on other tasks.

The method developed at VTT facilitates the simultaneous use of both short and long-axis images. Accurate determination of interfaces between ventricles and atria, for example, from the commonly used short-axis images is difficult, while they are visible in long-axis images. The developed image segmentation (partitioning the image into anatomical parts) method makes it possible to use the information obtained from both images simultaneously.

The development of the automated method is continuing in an ongoing project. The new method may become available for clinical use in hospitals already in 2005.

New methods are being developed by VTT in co-operation with the Helsinki University of Technology and the Hospital District of Helsinki and Uusimaa (HUS). The business community was represented by GE Medical Systems, Elekta-Neuromag Oy, Nexstim Oy and CSC. The project also involved co-operation with the image processing laboratories of CREATIS (INSA of Lyon) in France and Leiden University Medical Center in the Netherlands.