T. Ketsiri1, S. P. Srivastava2, R. Dortch1, and Z. Li1; 1Barrow Neurological Institute, Phoenix, AZ, 2Dignity Health Cancer Institute, Phoenix, AZ
Purpose/Objective(s): MRI is increasingly used in radiation therapy. Standard MRI has been used in tumor localization, while diffusion MRI (dMRI) has shown potential as a biomarker for radiotherapy response. CT is widely used for planning and delivering radiotherapy. A common challenge for spine imaging post-surgery is the presence of metallic devices that generate geometric distortion artifacts in MRI, preventing accurate delineation of the target volumes and organs at risk. For CT, metal artifacts also affect the accurate calculation of treatment doses. There have been numerous efforts to reduce the artifacts in both CT and MRI, especially dMRI. Several dMRI imaging methods have been proposed, including multi-shot echo planar imaging (msEPI), MultiVane (MV), and single-shot turbo spin echo (ssTSE). In this work, we develop and investigate ssTSE dMRI with multispectral imaging (MSI) for improved spine dMRI near metal. Materials/
Methods: MRI study was tested with phantom and evaluated on a patient with a cervical implant and an existing clinical CT spine scan, approved by IRB. The CT scan was acquired with a standard clinical cervical spine protocol. For MRI, a standard T2-weighted scan was acquired for localization along with three types of dMRI images in the sagittal plane – standard of care (SOC) msEPI, MV, and ssTSE MSI. The scans were performed on a 3T MRI scanner and a total-spine coil. Reference CT data were registered to the MRI images. The artifacts in the spinal cord were qualitatively assessed. The apparent diffusion coefficient (ADC) within the spinal cord dMRI was quantitatively analyzed. Results: Signal loss and geometric distortion artifacts were observed near the implant in MRI images. The artifacts became less visible further away from the implant and were minimized with MV and ssTSE MSI as compared to msEPI. The CT and MRI images were relatively well registered. While there was less distortion artifact of the implant in the CT image compared to MRI, streaking artifacts and signal nonuniformity were observed in the surrounding area in the CT images, especially within the spinal cord area near the implant. The results are consistent for phantom and patient. The ADC maps from the patient’s ssTSE exhibited increased uniformity for diffusivity within the spinal cord near the implant, while the increased standard deviation (SD) was observed with msEPI near the implant likely attributed to the distortion artifacts (Table 1). More patients will be recruited for validation. Conclusion: We demonstrated the feasibility of dMRI with ssTSE MRI in spine imaging near metals. This technique could potentially be useful for radiotherapy by providing more accurate spine tumor visualization for treatment planning and a biomarker for monitoring responses. Abstract 2324 – Table 1: Mean ± SD of ADC (x10-3 mm2/s) within the spinal cord for different dMRI
