Diagnostic Performance of CT-like Images for Lumbar Pedicle Screw Planning and Spinal Canal Area Measurement: A Comparative Study with Conventional CT and MRI.

Background

Although magnetic resonance imaging (MRI) provides excellent soft-tissue contrast for most spinal evaluations, computed tomography (CT) is still always required for preoperative planning to assess osseous anatomy and determine surgical device size, increasing the radiation exposure and workflow complexity. CT-like images enable visualization of precise bone morphology without ionizing radiation. In addition, these images often provide CT myelography-like contrasts, allowing the simultaneous depiction of the spinal canal area (SCA). This study aimed to evaluate whether CT-like images provide measurement accuracy equivalent to conventional CT and MRI for pedicle screw planning and spinal canal area assessment.

Methods

We retrospectively analyzed paired lumbar CT and MRI datasets obtained within ≤1 month in 51 patients. Pedicle width and length were measured on CT and CT-like images, whereas SCA was measured on T2 weighed-images and CT-like images. A total of 224 vertebrae were analyzed. Annotated images were independently evaluated by two readers in a randomized order. Inter-modality agreement was assessed using intraclass correlation coefficients (ICCs) and a Bland-Altman analysis.

Results

CT-like images demonstrated an excellent agreement with CT for pedicle measurements (ICCs: 0.968-0.985 for width; 0.922-0.966 for length). Mean differences were ≤0.1 mm for pedicle width and approximately 1 mm for pedicle length, which are unlikely to affect screw selection. The agreement with T2WI for SCA was good to excellent (ICCs: 0.766-0.945).

Conclusions

CT-like images provide comparable performance for quantitative pedicle assessment and show high agreement for SCA evaluation, supporting comprehensive preoperative assessment with a single MRI examination.