Abstract
Quantitative myocardial $T_2$ is a promising technique to assess myocardial inflammation and edema. Recent implementations have utilized $T_2$-prepared ($T_2$prep) SSFP sequences to acquire multiple $T_2$-weighted images at different echo times, generating $T_2$ maps based on a 2-parameter (2P-fit) model of $T_2$ decay. Recently, a 3-parameter fitting (3P-fit) model was found superior to the conventional 2P-fit model, as it compensates for $T_1$ relaxation effects, resulting in more accurate $T_2$ measurements. In this work, we sought to characterize the 3P-fit approach in terms of precision and reproducibility and to evaluate the influence of the number of employed $T_2$prep echo times on these two metrics.
Introduction
Quantitative $T_2$ mapping is a non-invasive MRI technique used to detect myocardial inflammation and edema. Traditionally, clinicians use a 2-parameter (2P-fit) model to calculate $T_2$ values from a series of images taken at different echo times. However, this standard approach does not account for $T_1$ relaxation effects during the imaging sequence, which can bias results. Furthermore, there has been no clear consensus on the optimal number of echo images required to balance scan efficiency with data quality. This study demonstrates that switching to a 3-parameter fitting model (3P-fit) significantly improves the reliability of myocardial $T_2$ mapping.
Methods
The researchers evaluated the 3P-fit approach using both numerical simulations and in-vivo testing:
- Numerical Simulations: Monte-Carlo simulations (1,000 repetitions) modeled signal intensities for theoretical tissue ($T_2 = 50\text{ms}$) across various Signal-to-Noise Ratio (SNR) levels to isolate the effect of echo count on mathematical stability.
- In-Vivo Testing: Ten healthy volunteers (average age $27 \pm 10$ years) were imaged using a 1.5T MRI scanner with a free-breathing, ECG-triggered $T_2$-prepared bSSFP sequence. To test reproducibility, each subject was scanned five times.
- Model Comparison: The conventional 2P-fit (simple exponential decay) was compared against the 3P-fit, which accounts for $T_1$ relaxation and other sequence-specific factors.
Key Insights and Results
The 2P-fit model was found to be "echo-dependent," meaning $T_2$ values shifted as the number of echoes changed. In contrast, the 3P-fit model provided consistent $T_2$ measurements regardless of echo count, indicating higher physical accuracy.
Precision improved as more echoes were added, but reached a point of diminishing returns. Both simulations and in-vivo data showed that precision "saturates" at approximately 10 echoes. Adding more than 10 echoes provides little benefit while increasing scan complexity.
The proposed optimized sequence (10 echoes + 3P-fit) outperformed the conventional sequence (3 echoes + 2P-fit) significantly:
- In-Vivo Precision: Improved from 11ms to 7ms.
- In-Vivo Reproducibility: Improved from 2.4ms to 1.2ms.
Visual analysis via "bull’s-eye" plots (segmental maps of the heart) confirmed much lower variability across myocardial regions when using the 10-echo 3P-fit approach.
Conclusion and Limitations
By accounting for $T_1$ relaxation and optimizing the number of echoes to 10, the 3P-fit model provides a more robust tool for diagnosing heart inflammation. While this study establishes a strong baseline in healthy volunteers, future work must validate these findings in patients with myocardial pathology (e.g., acute myocardial infarction) where $T_2$ values are elevated. Additionally, the trade-off between total exam time and image quality in a clinical workflow warrants further exploration.
Summary updated