Improved post-processing strategy for MOLLI based tissue characterization allows application in patients with dyspnoe and impaired left ventricular function

Abstract

Contrast and non-contrast MRI based characterization of myocardium by T₁-mapping will be of paramount importance to obtain biomarkers, e.g. fibrosis, which determines the risk of heart failure patients. T₁-mapping by the standard post-processing of the modified look-locker inversion recovery (MOLLI) lacks of accuracy when trying to reduce its duration, which on the other hand, is highly desirable in patients with heart failure. The recently suggested inversion group fitting (IGF) technique, which considers more parameters for fitting, has a superior accuracy for long T₁ times despite a shorter duration. However, for short T₁ values, the standard method has a superior precision. A conditional fitting routine is proposed which ideally takes advantage of both algorithms.

Materials and methods: All measurements were performed on a 1.5T clinical scanner (ACHIEVA, Philips Healthcare, The Netherlands) using a MOLLI 5(n)3(n)3 prototype with n(heart beats) being a variable waiting time between inversion experiments. Phantom experiments covered a broad range of T₁ times, waiting times and heart rates. A saturation recovery experiment served as a gold standard for T₁ measurement. All data were analyzed with the standard MOLLI, the IGF fit and the conditional fitting routine and the obtained T₁ values were compared with the gold standard. In vivo measurements were performed in a healthy volunteer and a total of 34 patients with normal findings, dilative cardiomyopathy and amyloidosis.

Results: Theoretical analysis and phantom experiments provided a threshold value for an apparent IGF T₁^* determining processing with IGF post processing for values above, or switching to the standard technique for values below. This was validated in phantoms and patients measurements. A reduction of the waiting time to 1 instead of 3 heart beats between the inversion experiments showed reliable results. The acquisition time was reduced from 17 to 13 heart beats. The in vivo measurements showed ECV values between 25% (18-33%; SD 0.03) in the healthy, 30% (22-40%; SD 0.04) in patients with DCM and 45% (30-60%; SD 0.9) in patients with amyloidosis.

Conclusion: The adopted post-processing algorithm determines long T₁ values with high accuracy and short T₁ values while maintaining a high precision. Based on reduction of waiting time, and independence of heart rate, it shortens breath hold duration and allows fast T₁-mapping, which is frequently a prerequisite in patients with cardiac diseases.

Keywords: ECV; MOLLI; Post-processing; T(1)-mapping.