Article
Value of virtually calculated low tube voltage monoenergetic datasets on a 3rd generation dual-source CT for the assesment of cervical spinal stenosis
Search Medline for
Authors
-
Mirko Arp
- Neurochirurgische Klinik, Universitätsmedizin Mannheim, Medizinische Fakultät, Universität Heidelberg, Mannheim, Deutschland
-
Gregory Ehrlich
- Neurochirurgische Klinik, Universitätsmedizin Mannheim, Medizinische Fakultät, Universität Heidelberg, Mannheim, Deutschland
-
Joshua Gawlitza
- Institut für Klinische Radiologie und Nuklearmedizin, Universitätsmedizin Mannheim, Medizinische Fakultät, Universität Heidelberg, Mannheim, Deutschland
-
Thomas Henzler
- Institut für Klinische Radiologie und Nuklearmedizin, Universitätsmedizin Mannheim, Medizinische Fakultät, Universität Heidelberg, Mannheim, Deutschland
-
Daniel Hänggi
- Neurochirurgische Klinik, Universitätsmedizin Mannheim, Medizinische Fakultät, Universität Heidelberg, Mannheim, Deutschland
-
Jason Michael Perrin
- Neurochirurgische Klinik, Universitätsmedizin Mannheim, Medizinische Fakultät, Universität Heidelberg, Mannheim, Deutschland
| Published: | June 9, 2017 |
|---|
Outline
Text
Objective: Assessment of the cervical spinal cord and CSF spaces with CT remains difficult due to the low spinal cord to CSF attenuation differences on CT images. Thus, patients with suspected spinal stenosis frequently undergo MRI as well as cervical CT prior to surgery. CT is of particular importance for patients with contraindications for MRI or with metal implants producing artifacts. In dual-energy CT (DECT), two CT datasets are acquired with different x-ray spectra, which allow the specific differentiation of anatomical structures or the reduction of metal artifacts. Virtually calculated low tube voltage monoenergetic datasets calculated from DECT have the potential to increase the spinal-cord-to-CSF ratio (SC/CSFratio) due to increased spinal cord attenuation at lower keV levels. Thus, the aim of this study was to prospectively evaluate the diagnostic value of low tube voltage monoenergetic datasets in patients with spinal stenosis.
Methods: 55 patients with suspected spinal stenosis prospectively underwent DECT of the cervical spine on a 3rd generation dual-source CT. Low tube voltage monoenergetic datasets were calculated in 10 keV intervals from 40-190 keV intervals. The datasets were compared to standard 120 kVp CT datasets that were also calculated from the DECT raw data. For objective image quality evaluation, the CT attenuation difference between SC and CSF was calculated. Subjective diagnostic image quality was evaluated by an experienced and independent radiologist and neurosurgeon by choosing one subjectively best image quality in all acquired datasets.
Results: The attenuation difference between SC and CSF as objective marker for the spinal cord delineation continuously increased with decreasing keV levels from 24,5 HU (190 keV) to 39,2 HU (40 keV). From all monoenergetic datasets 80 keV images showed the highest subjective image quality. In 69% the 80 KeV dataset was chosen by the examining neurosurgeon and in 75 % by the examining radiologist [range 70-90KeV]. Compared to the standard 120 kVp datasets image quality of the 80 keV datasets was rated superior. The difference between objective and subjective image evaluation is explained by the higher image noise at 40 keV when compared to 80 keV.
Conclusion: Low keV datasets calculated from DECT increase objective and subjective diagnostic image quality of the cervical spinal cord when compared to standard CT. This is of particular importance evaluating the indication for surgery in patients with clinical symptoms suggestive for cervical spinal stenosis and contraindications for MRI or with metal implants causing artifacts.
