The evaluation of large sets of parameters is a common requirement when new MR sequences are being optimized. For dynamic studies, it is often desirable to adjust parameters in “real time” depending on decisions made by user-defined algorithms. By performing a modification to the software tool MatMRI^[1], we developed a method that allows changing multiple MR sequence parameters, dynamically. We present results of this new method when applied to the modification of sequence parameters for MR-Acoustic Radiation Force Imaging (MR-ARFI)^[2]. This work of is of interest to researchers and clinicians interested in MRI-guided interventions, dynamic studies and dynamic control of MR sequence parameters.

MatMRI is an open access tool that encapsulates the XTC interface^[3] with Philips MRI scanners to make it accessible in a Matlab environment. XTC includes a mechanism to update the geometrical location of imaging slices in a dynamic study. We adapted the existing process of geometry updates to allow for modification of other MR sequence parameters. In MatMRI, a communication message is used to update the geometry of an MRI acquisition. This message consists of a series of N double-precision 4x4-matrices, where N is the number of imaging stacks being used in a dynamic study. In each matrix, six values store cosine vectors and are used to specify the orientation information. Since a valid cosine vector requires values between -1.0 and +1.0, we use this requirement to create matrices to be detectable for purposes other than changing the geometry; by example, by assigning a value of 2.0 in the first cosine value. Once the message is intercepted, then all the remaining values in the matrices can be used to update other MR sequence parameters. In principle, hundreds of double-precision values could be used for user-defined purposes.

We applied this method to update parameters controlling the MR-ARFI sequence^[2]. Experiments were conducted using a 3T MRI and a Sonalleve MR-HIFU system (Philips Healthcare, Best, The Netherlands). MR images were acquired every 0.25s using an EPI gradient echo sequence (EPI factor 9) with an echo time of 30 ms and a repetition time of 44 ms. The image field of view was 45×45 mm² with a voxel size of 0.7×0.7×4 mm³. HIFU pulses with a length of 0.5ms and acoustic power of 200W were used to produce the radiation force displacement. The image was positioned perpendicular to the acoustic beam and centered at the focus of the ultrasound beam. As shown in Figure 1, we adapted the MR sequence to update MRI parameters on demand, including the Trigger Delay (T_Dy), Gradient Delay (G_Dy), Gradient Duration (G_Du), and Gradient Strength (G_St). An user interface written in Matlab, which ran in a computer connected to the main MRI console via a 1 Gb Ethernet link, was used to modify the MR parameters dynamically and collect MR data in “real time”. The toolbox MatHIFU^[1] was used to control the HIFU system.

Figure 2 shows a section of the user interface detailing all the MR parameters (12 in total) that were available to be modified dynamically. An experiment was performed where the Trigger Delay (T_Dy) was modified over time. As shown in Figure 1, T_Dy controls the timing between the beginning of the HIFU pulse and the motion sensing gradient of the ARFI sequence. In this experiment, T_Dy was varied from 0 to 2 ms in steps of 0.2 ms every time a group of 8 images was collected. Figure 3 shows the displacement measurement over time at a voxel centered at the focused beam. To verify baseline conditions, the test included the collection of displacement images before the first HIFU pulse and after the last HIFU pulse. Changes of the measured displacement were clearly observable as the change in T_Dy was applied after a group of 8 images were collected. In each group of 8 images, the last 5 images were used to calculate the mean and standard deviation of the displacement. Figure 4 shows the measured displacement as function of T_Dy. In total, 12 changes to the MR parameters were applied and the whole test lasted only 38 seconds.In this study, we demonstrated that it is feasible to reprogram MR sequence parameters dynamically using existing technology for the control of scanners. This new feature in the MatMRI toolbox facilitates the testing of multiple parameters in a very efficient manner. The capability of changing MR parameters “on demand” will be greatly beneficial in the future to optimize MRI techniques such as MR-ARFI and other interventional procedures.