Purpose: Sequence parameters for most clinical protocols have traditionally been designed for subjective image interpretation by a radiologist. Various types of sequences and imaging parameters are typically selected based on compromises in contrast, resolution, scan time, and motion robustness. However, while perceived differences in contrast exist for different sequences, validation and normalization of contrast behavior between these sequences cannot be performed with standard QA procedures. In addition, followup imaging is very often acquired with different parameters or sequence types, negating any radiological confidence in subtle observed contrast changes. In this study we used a system phantom to evaluate and normalize contrast of various T1 weighted sequences used for brain imaging in our institution.
Methods: An automated segmentation tool was developed to delineate the T1, T2, and PD compartments of the ISMRM-NIST system phantom. The phantom was acquired with T1 weighted 2D conventional spin echo, fast spin echo, and spoiled gradient-recalled (SPGR) echo sequences, and was also acquired with 3D SPGR and inversion recovery (IR) prepared SPGR sequences. Acquisitions were performed on 1.5T and 3T scanners using a standard head and neck coil. Signal intensity of the T1-sensitive compartments was measured for evaluation of relative contrast, as defined by change in signal intensity as a function of T1. Parameter adjustments were performed on the sequences that exhibited significantly less contrast than conventional spin echo.
Results: IR prepared sequences generally exhibited the most contrast, while 2D SPGR exhibited the least contrast. By adjusting TR or TI, and adding IR preparation to 2D SPGR, contrast curves could be made similar to that of single spin echo.
Conclusion: A standardized system phantom enables the measurement and normalization of T1 contrast between different types of sequences. The proposed process forms a foundation for QA of image contrast in clinical protocols.