Tim Trio EPI Stability Quality Assurance

EPI timeseries stability, temporal SNR, drift and ghost levels are monitored daily using a 1.9 Liter doped-water phantom provided by Siemens, with scan methods derived from Siemens own quality assurance routines. We monitor for and document:
- consistent single image SNR for improved functional MRI,
- high temporal SNR in EPI timeseries for fMRI,
- minimal signal drift in EPI timeseries for fMRI,
- minimal/consistent image ghost intensity on EPI for fMRI and diffusion weighted images.

Stability scans are multi-slice EPI scans measuring 16 slices with 500 time points. The scan is performed 5-6 times each day in a 1-hour period. This sequence, which includes automated analysis of the data at the scanner immediately after image reconstruction, was developed by our collaborators at the Martinos Center for Biomedical Imaging at Massachusetts General Hospital, and is available from MGH for use here via a C2P agreement involving MGH, Siemens and Harvard.

Details of the automated analysis are provided here. Some examples of plots generated at the scanner are shown below (screen-shots from the Siemens scanner).  The first figure shows signal intensity for a central ROI (top left), image SNR for that ROI (top right), and EPI ghost level (lower left) over the time-course of 500 measurements. Standard deviation and %Peak-to-Peak (P2P) for each measurement are given in the comments below the plots. A second signal intensity plot, detrended for drift, is shown in green if signal P2P is below 1%, or red if above 1%.  This analysis provides immediate feedback at the scanner on its performance within seconds of the scan completing. Signal P2P has been identified as the most important and informative indicator of suitable scanner stability as required for functional MRI scans.

The second figure (again - a screenshot from the Siemens scanner) shows calculated maps for each of the 16 slices after completion of the 500 time-point time series. Shown are a subtraction image (minimum - mean; top left); time-series mean (top right); time-series standard-deviation multiplied by 10 (lower left) and time-series (temporal) SNR (lower right). Standard deviation is very low in all slices, as are the subtraction maps. Temporal SNR is consistent in all slices and the pixel marker in this image shows tSNR > 400, as expected for a stable phantom.

In addition, the scanner automatically generates an email summarizing each stability scan. Signal intensity, Signal %P2P, image SNR and ghost levels are plotted and monitored on a routine basis. Plots from a recent month are shown below. Signal intensity varies based on position of the slice in the RF coil - this results in the "vertical-stripe" of data points on any given day. We aim to keep %P2P below 1% - one or two lone values between 1 and 1.5% indicate the phantom is not yet stable and at ambient temperature. Image SNR as determined by this algorithm varies more on a daily basis than SI or ghost level, as it is a calculated parameter. With our Siemens engineer, we investigate any consistent occurrences of %P2P, SI, SNR or ghost level outside the ranges shown.

Current plots of scanner stability measurement. Examples from a recent month are shown below.

Data from stability scans run since December 2011 (text format - subject lines of "Stability run on MRC35380").