Myocardial fibrosis assessment by semiquantitative, point-counting and computer-based methods in patients with heart muscle disease: a comparative study.

Abstract

No study has directly compared different histomorphometric methods of quantification of myocardial fibrosis. Therefore we compared the results of semiquantitative, point-counting and computer-based methods in the assessement of myocardial fibrosis in a consecutive series of endomyocardial biopsy samples from patients with heart muscle disease. Histological samples (at least three per patient) were obtained by endomyocardial biopsy from 11 patients with focal myocarditis and from 24 ambulatory patients with idiopathic dilated cardiomyopathy, or during surgery in 10 patients who underwent partial left ventriculectomy. Samples were cut and stained with Masson-trichrome for better contrast. From each sample, a representative field was digitized, and the amount of fibrosis was assessed by semiquantitative scoring, by point-counting, and by computer-based software. Semiquantitative scoring correlated with both point-counting (Spearman's r = 0.69, P < 0.0001) and computer-based (Spearman's r = 0.83, P < 0.0001) methods. There was also a good correlation between point-counting and computer-based methods (r = 0.71, P < 0.0001). However, when compared with the point-counting method, the computer-based method overestimated percent fibrosis by 3.0 +/- 6.7% (P = 0.004). This overestimation correlated with the mean percent fibrosis (r = 0.38, P = 0.014). Our data show good correlations between the three methods of myocardial fibrosis assessment. However, systematic differences between them emphasize that this should be taken into consideration when comparing results of the studies using different methods of fibrosis assessment.