Left-ventricular non-compaction-comparison between different techniques of quantification of trabeculations: Should the diagnostic thresholds be modified?

Background.- Diagnosis of left ventricular non-compaction (LVNC) is challenging, and different imaging techniques propose different criteria. Aim. - To compare the value of two-dimensional transthoracic echocardiography (2D-TTE) and cardiac magnetic resonance (CMR) criteria in diagnosing LVNC, and to test a new trabecular quantification method obtained by 2D-TTE, exploring its relationship with CMR non-compacted mass quantification. Methods. - From a multicentre French study, we selected 48 patients with LVNC and 20 with dilated cardiomyopathy (DCM) who underwent 2D-TTE and CMR. Current 2D-TTE (Jenni et al.) and CMR criteria (Petersen et al., Jacquier et al.), were tested. A new 2D-TTE method of trabecular quantification (percentage of trabecular area) was also proposed, and compared with current criteria. Results. - The best cut-off values for the diagnosis of LVNC were a non-compacted/compacted ratio >= 2.3 (Petersen et al.), a trabeculated left ventricular mass >= 20% (Jacquier et al.) and a non-compacted/compacted ratio >= 1.8 (Jenni et al.). Lowering the threshold for the criterion of Jenni et at. from > 2 to >= 1.8 improved its sensitivity from 69% to 98%. The 2D-TTE percentage of trabecular area was 25.9 +/- 8% in the LVNC group vs. 9.9 +/- 4.4% in the DCM group (P< 0.05), and was well correlated with CMR non-compacted mass (r =0.65; P< 0.05). A 15.8% threshold value for 2D-TTE percentage of trabecular area predicted LVNC diagnosis with a specificity of 95% and a sensitivity of 92%; its sensitivity was better than that for the criteria of Jenni et al. (P <0.01) and Petersen et al. (P= 0.03). Conclusions. - Revision of the current threshold for the criterion of Jenni et al. from > 2 to >= 1.8 is necessary to improve LVNC diagnosis in patients with left ventricular dysfunction. A new 2D-TTE trabecular quantification method improves TTE diagnosis of LVNC. (C) 2020 Elsevier Masson SAS. All rights reserved.