Computed tomography (CT) has been the standard imaging modality for pre- and postsurgical evaluation of pediatric patients with craniofacial skeletal pathologies. The large difference in attenuation coefficients between bone and soft tissue allows for straightforward segmentation, enabling creation of 3D models at high spatial reso- lution. However, there is growing concern about potential adverse effects of repeated exposure to ionizing radia- tion during infancy and childhood. In fact, the FDA has issued warnings cautioning against unnecessarily expos- ing children to X-rays, stating that even though the risk is deemed relatively small, it could, nonetheless, contrib- ute to increased risk of cancer later in life. As a result, non-radiative alternatives to CT would be particularly valu- able to identify and manage these patients, who require multiple examinations over the course of their lifetime. Unfortunately, conventional MR is not suited for imaging bone, which appears with near background intensity due to very short T2 relaxation times and relatively low proton density, and thus is difficult to distinguish from air. The premise underlying this proposal is that a new solid-state (SS) subtraction imaging technique, making use of bone signal attenuation during and following excitation, collected in an interleaved fashion with an acquisition yielding maximal bone signal retention, allows for superior bone-selective imaging. Along with a new k-space data sharing approach and compressed sensing reconstruction, achieved by exploiting signal sparsity in the difference image domain, it is hypothesized that the method will achieve high-resolution whole-skull coverage, yielding 3D render- ings comparable to their CT analogs in less than three minutes of scan time. This hypothesis will be rigorously evaluated in three specific Aims, starting with full implementation of image data acquisition, reconstruction and processing (Aim 1). Subsequently, the methodology is tested in human cadaver skulls to compare its performance based on craniometric accuracy and agreement with the gold-standard CT data. The latter is quantified in terms of the Srenson-Dice (SD) coefficient of the binarized 3D-rendered images, and craniometric accuracy in the form of the concordance correlation coefficient (Aim 2). Finally, the SS-MRI method is evaluated in 30 children and ad- olescents who are clinically indicated for craniofacial surgery guided by high-resolution CT (Aim 3). We posit that the new method is superior to a competing gradient-echo imaging method that is unable to distinguish between bone and background (e.g. air in the sinuses) ? therefore denoted black-bone (BB) MRI. The specific hypothesis to be tested is that the SS MRI method is more accurate than its BB counterpart with respect to the reference CT- based renderings. Successful execution of this project should result in a clinically effective, radiation-free alterna- tive to CT for the pre- and post-surgical evaluation of pediatric patients with craniofacial abnormalities. Beyond the scope of the proposed pilot project, the rigorously evaluated method will provide a gateway toward establishment of a database of craniofacial anatomy during growth and development enabling recognition of abnormal pheno- types.
Craniofacial bone surgery in children currently relies on three-dimensional images of the skull obtained by X- ray computed tomography (CT). While these images provide an accurate picture of the surface of the skull and serve as a guide to the surgeon both before and after surgery, repeated exposure to radiation emitted by CT scans can put the patients at risk for developing cancer later in life. The proposed new method is based on magnetic resonance imaging (MRI), which is free of harmful radiation, and is projected to yield images comparable in quality to CT and, in future larger studies, should inform on normal and abnormal development of the skull.
