ECHOCARDIOGRAPHIC TEXTURE ANALYSIS FOR RE COGNITION OF ACUTE CARDIAC REJECTION AFTER HETEROTOPIC HEART-TRANSPLANTATION

Experimental studies from our laboratory demonstrated that acute cardi ac rejections (AR) are associated with an increase in myocardial echo amplitudes which correlate with the histologic severity of AR. The aim of the present study was to test whether AR-induced myocardial struct ural changes would also alter the spatial distribution of the myocardi al echo amplitudes, assessed by echocardiographic texture analysis, an d whether this method would improve the sensitivity of AR diagnosis. E chocardiographic LV long axis views were daily recorded in 12 dogs aft er cervical heterotopic heart transplantation (mean survival time: 16. 1 days; immunosuppression: cyclosporine, azathioprine and prednisolone ; AR treatment: pulsed steroids) under strictly standardized condition s and digitized into a 256 x 256 x 8 pixel matrix. Texture analysis us ing Cooccurrence- and Run Length Matrix was performed in 45 x 12 pixel -sized regions of interest (ROI) within the septum (VS) and posterior wall (PW) in enddiastole. Results obtained without AR were then compar ed with those during biopsy-proven mild, moderate and severe AR. Sever e AR are associated with a significant increase in Cooccurrence Matrix parameters within the septum and posterior wall (Contrast: VS: 3.1 +/ - 0.7 vs. 3.9 +/- 1.1, p < 0.05. PW: 1.8 +/- 0.5 vs. 2.3 +/- 1.0, p < 0.05. Variance: VS: 19 +/- 15 vs. 36 +/- 18, p < 0.05. PW: 11 +/- 6 vs . 21 +/- 8, p < 0.01). During mild and moderate AR no significant chan ges in texture measurements were found. Testing the influence of ROI p lacement on myocardial echo amplitude, texture measurements showed a h igh variability in comparison to mean gray level (up to 30% vs. <5%). Only severe AR are associated with significant changes of texture anal ysis parameters, even under the ideal conditions of an animal experime nt. Therefore, echocardiographic texture analysis was not suitable to improve the sensitivity of AR diagnosis.