Authors: Peter Anthony Chiarelli, MD, DPhil
Continued advancements in magnetic resonance imaging (MRI) technology have provided improved ways to acquire imaging data, and to more seamlessly incorporate MRI into modern pediatric surgical practice, including intracranial catheter placement. Conventionally, image guidance has required acquisition of CT or MRI images, associated with ionizing radiation or long scan/sedation, respectively. The delivery of ionizing radiation in patients with cerebral ventricular shunts is a well-understood concern that requires a tailored approach for prevention.Methods: Five established methods were implemented to accelerate the MRI acquisition, including (i) modification of k-space readout direction (i.e. spiral acquisition),20 (ii) partial k-space acquisition, (iii) sparse sampling of k-space (i.e. compressed sensing; CS)19,21, (iv) variation of in-plane spatial resolution, and (v) variation in slice thickness. 10 healthy volunteer participants were included in the current study. Target registration error (TRE), root mean square error (RMSE), and Bland-Altman analysis techniques were used to assess the accuracy and precision of imaging sequence registration.
This work investigates a new protocol design for fast magnetic resonance (MR) sequences in ventricular neuronavigation that provide scan duration comparable to CT acquisition, visualization of ventricular morphology, and accurate image registration by surface rendering. Systematic refinement of this protocol was accomplished through testing of 13 candidate sequences, and further evaluation of the optimized protocol in ten subjects. The sequence (Presurgical Imaging with Compressed sensing for time Optimization, PICO), is acquired in 25 seconds, visualizes ventricular morphology, and shows non-inferior accuracy of registration compared to a default MRI neuronavigation protocol.
We present PICO, an ultrafast brain MR sequence compatible with IGS, requiring acquisition less than 1/10th the time of current standard imaging. The sequence developed and tested in this work captures sufficient surface feature data to allow for accurate intraoperative image registration and has sufficient parenchymal contrast to provide visualization of ventricular anatomy.