Do we need a new CT scan for retreatment of intracranial SRS patients?

PURPOSE To determine if the treatment planning computed tomography scan (CT) from an initial intracranial stereotactic radiosurgery (SRS) treatment can be used for repeat courses of SRS. METHODS AND MATERIALS Twenty-five patients with 40 brain metastases that received multiple courses of SRS were retrospectively studied. Magnetic resonance scans from repeat SRS (rMR) courses were registered to CT scans from the initial SRS (iCT) and repeat SRS (rCT). The CT scans were then registered to find the displacement of the rMR between iCT and rCT registrations. The distance from each target to proximal skull surface was measured in 16 directions on each CT scan after registration. The mutual information (MI) coefficients from the registration process were used to evaluate image set similarity. Targets and plans from the rCTs were transferred to the iCTs, and doses were recalculated on the iCT for repeat plans. The two dose distributions were compared through 3D gamma analysis. RESULTS The magnitude of the mean linear translations from the MR registrations was 0.6 ± 0.3 mm. The mean differences in distance from target to skull on a per target basis were 0.3 ± 0.2 mm. The MI was 0.582 ± 0.042. Registration between a comparison group of 30 CT scans that had the same data resampled and 30 scans that were intercompared with different patients gave MI = 0.721 ± 0.055 and MI = 0.359 ± 0.031, respectively. The mean gamma passing rates were 0.997 ± 0.007 for 1 mm/1% criteria. CONCLUSIONS The rMR can be aligned to the iCT to accurately define targets. The skull shows minimal change between scans so the iCT can be used for set-up at repeat treatments. The dosimetry provided by the iCT dose calculation is adequate for repeat SRS. Treatment based on iCT is feasible.