Experimental determination of depth-scaling factors and central axis depth dose for clinical electron beams

Application of Low-dose Stereoradiography in In Vivo Vertebral Morphologic Measurements: Comparison With Computed Tomography

BACKGROUND

Though computed tomography (CT) and 3 dimensional (3D) reconstruction of the spine and ribcage are powerful techniques for detailed monitoring of spinal growth and surgical planning of patients, drawbacks can arise. We explored the application of low-dose stereoradiography of the spine to calculate distinct morphologic parameters of the vertebral body in a juvenile patient population with early-onset scoliosis or congenital scoliosis. This study compares the 3D vertebral morphology measurements using low-dose stereoradiography with the currently accepted imaging modality for such measurements, CT scans.

METHODS

A total of 86 vertebrae of 6 patients with early-onset scoliosis and 3 patients with congenital scoliosis, age ranged between 7.8 and 12.5 years, who had both thoracic spine CT scan and low-dose stereoradiography of the spine were included. 3D reconstructions of CT and low-dose stereoradiography were generated. Using previously validated image processing techniques, vertebral anterior (A), posterior (P), left (L), and right (R) heights, superior and inferior endplates depth and width (S-D, I-D, S-W, I-W) were measured on the CTs' 3D reconstructions and were compared with the same parameters measured on low-dose stereoradiography reconstructions using a postprocessing custom code. The agreement between the 2 techniques in measurement of the vertebral morphology was assessed using the Bland-Altman plots.

RESULTS

No significant difference was observed in the A, P, L, R, S-W, and I-W between the stereoradiography and CT measurements (P>0.05). S-D and I-D were significantly greater in low-dose stereoradiography measurements P<0.05. Bland-Altman plots showed an agreement between the stereoradiography and CT techniques in vertebral height measurements (A, P, L, R); however, larger measurement bias and greater limits of agreement in S-D, I-D, S-W, and I-W measurements were shown.

CONCLUSIONS

In vivo measurements of the vertebral heights using low-dose stereoradiography 3D reconstructions were comparable with CT measurements except for the superior and inferior vertebral depths. The low-dose stereoradiography imaging modality and the postprocessing platform can be used for assessment of the vertebral heights and monitoring asymmetric growth in patients undergoing growth-sparing treatment except for the vertebral levels with congenital deformities.

LEVEL OF EVIDENCE

Level IV.

Characterization of the phantom material Virtual Water™ in high-energy photon and electron beams

The material Virtual Water has been characterized in photon and electron beams. Range-scaling factors and fluence correction factors were obtained, the latter with an uncertainty of around 0.2%. This level of uncertainty means that it may be possible to perform dosimetry in a solid phantom with an accuracy approaching that of measurements in water. Two formulations of Virtual Water were investigated with nominally the same elemental composition but differing densities. For photon beams neither formulation showed exact water equivalence-the water/Virtual Water dose ratio varied with the depth of measurement with a difference of over 1% at 10 cm depth. However, by using a density (range) scaling factor very good agreement (<0.2%) between water and Virtual Water at all depths was obtained. In the case of electron beams a range-scaling factor was also required to match the shapes of the depth dose curves in water and Virtual Water. However, there remained a difference in the measured fluence in the two phantoms after this scaling factor had been applied. For measurements around the peak of the depth-dose curve and the reference depth this difference showed some small energy dependence but was in the range 0.1%-0.4%. Perturbation measurements have indicated that small slabs of material upstream of a detector have a small (<0.1% effect) on the chamber reading but material behind the detector can have a larger effect. This has consequences for the design of experiments and in the comparison of measurements and Monte Carlo-derived values.