Myocardial Fibrosis in Dilated Cardiomyopathy: Comparison Between Magnetic Resonance and Histological Findings

Background: Myocardial fibrosis is known to be present in dilated cardiomyopathy and it predicts the occurrence of sudden death and congestive heart failure. Late gadolinium enhancement (LGE) cardiovascular magnetic resonance (CMR) allows identification of fibrosis with a typical distribution pattern. The pattern and distribution of late enhancement has not been correlated with pathological substrates in non-ischemic cardiomyopathies. Therefore its association with functional and arrhythmogenic prognosis deserves better understanding. Aim: Aim of this study was to correlate in vivo CMR findings with the ex vivo histological substrates in patients with non-ischemic cardiomyopathy who underwent cardiac transplantation. Methods: We performed in vivo CMR in five pts with dilated non-ischemic cardiomyopathy on the waiting list of cardiac transplantation. On explanted hearts we identified at histology the presence, location and pattern of fibrosis and we correlated them with the results of CMR according to a 48 segment model applied to both techniques. We identify three different patterns of LGE (striae, spot, and diffuse respectively) and we assessed their presence in each histological segments, where diffuse was divided into minimal interstitial net (lousy net) and thin interstitial net (thin net). Results: At CMR 178(74%) of the 240 sections studied showed a different patterns of fibrosis and 62 sections were negative (26%). In the 178 positive sections LGE-CMR showed diffuse patter in 43,8%, spot pattern in 5,4% and mid wall striae pattern in 25,8%. At LGE-CMR the normal pattern correspond to absence of fibrosis in 19,3% at histology. At LGE-CMR the mid wall striae pattern was negative for fibrosis in 6,4% at histology. Specificity and sensitivity of to detection fibrosis was 69% and 81 % respectively. Conclusions: Assessment of fibrosis by LGE-CMR showed low specificity suggesting that the presence of fibrosis is often underestimated by LGECMR. The resolution power of CMR can account for this discrepancy and quantitative histological studies need to be performed in order to established the cut- off values of fibrosis detectable by CMR. These values will set new clinico- pathological correlates with the development of heart failure and arrhythmias