Abstract
Quantitative myocardial T2 mapping is a promising technique for in-vivo assessment of inflammation and edema. Free breathing T2 mapping sequences increase the flexibility in the choice of the number of T2prep echoes times (TET2P), but should be combined with respiratory motion correction technique. In this study, we sought to evaluate the performance of the Adaptive Registration of varying Contrast-weighted images for improved TIssue Characterization (ARCTIC) algorithm for in-plane motion correction in T2 mapping data and its impact on in-vivo reproducibility and spatial variability of myocardial T2 estimates.
Introduction
Quantitative T2 mapping is a powerful MRI technique used to detect myocardial inflammation and edema. Traditionally, these scans require breath-holding to prevent respiratory motion artifacts. However, many patients with heart conditions find it difficult to hold their breath for the required duration. While "free-breathing" sequences are more patient-friendly and allow for more data points (echoes), they introduce respiratory motion. Even with navigators tracking the diaphragm, residual in-plane motion causes misalignment between images of varying contrast, leading to blurring and incorrect T2 values in the final maps.
Methods
The study involved seven healthy volunteers (mean age 30±17 years) imaged on a 1.5T scanner. Researchers compared several acquisition strategies focusing on the 2D short-axis slice of the left ventricle:
- Breath-held (BH): The standard approach.
- Free-breathing (FB): Without motion compensation.
- Free-breathing with Navigator (FB+NAV): Using hardware-based tracking.
- High-density sampling: A sequence with 20 echo times (T2P20TE) to test stability.
The core evaluation focused on the ARCTIC (Adaptive Registration of varying Contrast-weighted images for improved TIssue Characterization) algorithm. ARCTIC is specifically designed to register images where brightness and contrast change significantly due to T2 decay, a common challenge in parametric mapping.
Results
The application of ARCTIC yielded statistically significant improvements across all metrics. The DICE similarity coefficient improved from 0.79 to 0.91 in free-breathing data, and the Myocardial Boundary Error (MBE) was nearly halved from 1.16mm to 0.60mm.
A key finding was the benefit of high-density sampling. By using a free-breathing approach, researchers could collect 20 echo times instead of the usual 4. This over-sampling, combined with ARCTIC, significantly improved map homogeneity and reproducibility. For the 20-echo sequence, reproducibility error dropped from 4.3ms to 2.4ms, and spatial variability decreased from 10.6ms to 7.9ms, indicating a much smoother and more reliable map.
Discussion
The study demonstrates that software-based motion correction like ARCTIC is a necessary second step even when hardware navigators are used. ARCTIC’s ability to handle varying contrast is crucial for making free-breathing T2 mapping viable and robust, potentially matching or exceeding the quality of traditional breath-held techniques.
Limitations include the small, healthy cohort and the focus on in-plane motion. Future work should validate these findings in clinical populations with arrhythmias or edema and explore 3D mapping sequences to address through-plane motion.
Summary updated