Assessment of late gadolinium enhancement in nonischemic cardiomyopathy: comparison of a fast Phase-Sensitive Inversion Recovery Sequence (PSIR) and a conventional segmented 2D gradient echo recall (GRE) sequence--preliminary findings.

BACKGROUND

Reliable detection of myocardial scarring in nonischemic cardiomyopathy is time-consuming using techniques that require determination of optimal inversion time. Therefore we evaluated an inversion-time-insensitive approach using a fast phase-sensitive inversion recovery (PSIR) sequence to detect and quantify late gadolinium enhancement (LGE).

PATIENTS AND METHODS

Twenty patients (mean age 40 years, 9 females) with nonischemic cardiomyopathy and evidence of LGE were evaluated. After administration of 0.2 mmol/kg gadolinium diethylene triamine pentaacetic acid, a segmented 2D inversion recovery turbo fast low-angle shot gradient echo recall (GRE) sequence [echo time (TE) 4.3 milliseconds, repetition time (TR) 750 milliseconds, alpha 30 degrees , voxel size 1.7 x 1.3 x 8-10 mm] was obtained and served as the standard of reference. Second, a fast multislice single-shot 2D PSIR sequence (TE 1.1 millisecond, TR 700 milliseconds, alpha 40 degrees , voxel size 2.5 x 1.7 x 8-10 mm) was acquired in the same slice positions. The PSIR(IR) images were used to analyze LGE. Altogether 53 short-axis slices with LGE were evaluated. Contrast-to-noise ratio and area of LGE were calculated and compared by 2 experienced readers. Image quality and confidence level for identification of LGE were rated on 5-point scales. Interobserver variability was evaluated in 10 patients.

RESULTS

All images were interpretable. Imaging time was reduced from 385 +/- 127 seconds to 20 +/- 3 seconds (P < 0.001). Contrast-to-noise ratio was 8.29 for PSIRmag and 12.07 for the conventional GRE images (P < 0.001). The mean area of LGE was 1.01 +/- 0.62 cm(2) for the GRE sequence and 1.10 +/- 0.62 cm(2) for PSIR(IR) (P = NS). The general linear model showed no interaction between the results and no significant difference of the mean (r = 0.09, mean difference 0.09 cm(2)). The overall interobserver variability of PSIR(IR) and GRE was excellent, with Pearson's correlation coefficients of r = 0.96 for PSIR(IR) and r = 0.98 for GRE. PSIR(IR) and conventional GRE were comparable in terms of image quality and confidence level (image quality: 1.6 +/- 0.67 vs. 1.5 +/- 0.93, P = NS; confidence level: 1.4 +/- 0.84 vs. 1.3 +/- 0.5; P = NS).

CONCLUSIONS

Fast PSIR sequences enable accurate detection and quantification of LGE in nonischemic cardiomyopathies. The examination time can be significantly shortened using the single-shot approach of the PSIR technique.