Study design: Retrospective.
Objective: To evaluate the image quality of low-radiation-dose computed tomography (LD-CT) of the thoracolumbar spine, using model-based iterative reconstruction (MBIR) for measuring pedicle diameter.
Summary of background data: MBIR can drastically reduce radiation dose but its utility in spine surgery planning is unknown.
Methods: We identified patients (mean age, 70.5 ± 13.3 yrs) who incidentally underwent both standard-radiation-dose CT (SD-CT) with hybrid iterative reconstruction and LD-CT with MBIR of the thoracolumbar spine within 2 years. We compared radiation dose, subjective image sharpness, signal-to-noise ratio, and contrast-to-noise ratio for the two tests. Additionally, inner pedicle diameters were measured on SD-CT (DSD) and LD-CT (DLD), and statistically compared.
Results: We included 24 CT and 84 pedicles for each CT group. The radiation dose of LD-CT estimated by volume CT dose index was 1.21 ± 0.42 mGy, one-sixth the dose of SD-CT. The effective dose of LD-CT was 0.58 ± 0.31 mSv, equivalent to or less than that of a one-time lumbar X-ray in a previous report. LD-CT was significantly inferior in subjective image sharpness for the contour of vertebrae and trabecular structure, but superior for signal-to-noise ratio and contrast-to-noise ratio. The intra-rater reliability (intra-RR) and inter-RR for DLD were 0.985 and 0.892, respectively, comparable to those of DSD. DLD was consistently 0.30 mm smaller than DSD when compared within the same pedicle, regardless of pedicle diameter.
Conclusion: LD-CT with MBIR produced a radiation dose equivalent to a one-time lumbar X-ray and provided excellent images for measuring pedicle diameter. LD-CT can be a substitute for SD-CT when planning spine surgery if the relationship between DSD and DLD is sufficiently understood.
Level of evidence: 3.