Feasibility and limitation of constructive interference in steady-state (CISS) MR imaging in neonates with lumbosacral myeloschisis

The aim of this study was to evaluate three-dimensional Fourier transformation-constructive interference in steady-state (CISS) imaging as a preoperative anatomical evaluation of the relationship between the placode, spinal nerve roots, CSF space, and the myelomeningocele sac in neonates with lumbosacral myeloschisis. Five consecutive patients with lumbosacral myeloschisis were included in this study. Magnetic resonance (MR) CISS, conventional T1-weighted (T1-W) and T2-weighted (T2-W) images were acquired on the day of birth to compare the anatomical findings with each sequence. We also performed curvilinear reconstruction of the CISS images, which can be reconstructed along the curved spinal cord and neural placode. Neural placodes were demonstrated in two patients on T1-W images and in three patients on T2-W images. T2-W images revealed a small number of nerve roots in two patients, while no nerve roots were demonstrated on T1-W images. In contrast, CISS images clearly demonstrated neural placodes and spinal nerve roots in four patients. These findings were in accordance with intraoperative findings. Curvilinear CISS images demonstrated the neuroanatomy around the myeloschisis in one slice. The resulting images were degraded by a band artifact that obstructed fine anatomical analysis of the nerve roots in the ventral CSF space. The placode and nerve roots could not be visualized in one patient in whom the CSF space was narrow due to the collapse of the myelomeningocele sac. MR CISS imaging is superior to T1-W and T2-W imaging for demonstrating the neural placode and nerve roots, although problems remain in terms of artifacts. (orig.)