Clinical Validation of BoneMRI in the Spine

Overview

BoneMRI is a quantitative 3D MRI technique that has been developed recently by MRIGuidance BV©, which is based on a multiple gradient-echo sequence and a machine learning processing pipeline. The BoneMRI technology is capable of generating CT-like, quantitative radiodensity bone MRI images to visualize cortical and trabecular bone, allowing to assess bone structure and morphology, in addition to regular clinical MRI images. The use of BoneMRI has been investigated and clinically validated in multiple musculoskeletal studies involving the cervical spine, hip and sacro-iliac joint. In order to clinically use BoneMRI in the entire spine, the BoneMRI technology needs to be validated in that area as well, focussing on geometrical and voxelwise accuracy of the radiodensity contrast to assure accurate visualization of the osseous structures. As robustness against expected data variability between hospitals is crucial for successful machine learning algorithms, multiple MR field strengths and scanner types from different manufacturers will be included in this study. If successful, BoneMRI will facilitate a better, easier and cheaper workflow by enabling diagnosis, treatment planning and surgical navigation using a single radiological examination, without the potential hazards of ionizing radiation.

Rationale: BoneMRI is a quantitative 3D MRI technique that has been developed recently by MRIGuidance BV©, which is based on a multiple gradient-echo sequence and a machine learning processing pipeline. The BoneMRI technology is capable of generating CT-like, quantitative radiodensity bone MRI images to visualize cortical and trabecular bone, allowing to assess bone structure and morphology, in addition to regular clinical MRI images. The use of BoneMRI has been investigated and clinically validated in multiple musculoskeletal studies involving the cervical spine, hip and sacro-iliac joint. In order to clinically use BoneMRI in the entire spine, the BoneMRI technology needs to be validated in that area as well, focussing on geometrical and voxelwise accuracy of the radiodensity contrast to assure accurate visualization of the osseous structures. As robustness against expected data variability between hospitals is crucial for successful machine learning algorithms, multiple MR field strengths and scanner types from different manufacturers will be included in this study. If successful, BoneMRI will facilitate a better, easier and cheaper workflow by enabling diagnosis, treatment planning and surgical navigation using a single radiological examination, without the potential hazards of ionizing radiation. Primary objective: The primary objective of this study is to investigate the performance of BoneMRI in terms of geometrical accurate visualization of the spinal osseous structures by radiodensity reconstruction when exposed to clinically relevant data variability. Study design: This study is a prospective multi-center clinical validation study, following a comparative design. Study population: Subjects referred to the radiology department for an MRI and CT scan of the spine having symptoms related to a spine disorder with suspected underlying involvement of osseous structures, will be asked to participate in this study. Duration of the study: Expectation is that it will take approximately 36-48 months to include 50 patients per center. Main study parameters/endpoints: Geometric accuracy in terms of visualization of the 3D osseous morphology of the spinal column. Nature and extent of the burden and risk associated with participation, benefit and group relatedness: The patient does not benefit from participating in this study and will receive routine care, which includes undergoing an MRI and CT scan. For research purposes an additional MRI sequence will be obtained for each patient. The CT scan is part of routine clinical care, so patients do not receive additional ionizing radiation compared to standard care. The subjects will in no way be exposed to BoneMRI as BoneMRI will not be installed at the investigation site and will not be part of the clinical workflow, nor the BoneMRI reconstructions will be part of the patient's file or decision making process of the healthcare professional. Therefore, there are no additional risks for the patients when participating in this study. This study may contribute to lower radiation doses in future patients when concluded that BoneMRI accurately visualizes the 3D morphology of the spinal osseous structures. This would render an additional conventional CT scan redundant.