Purpose: To evaluate the dosimetric qualities of plans generated using a novel automatic inverse planning system for Gamma Knife stereotactic radiosurgery (GKSRS).
Methods: Patients treated with GKSRS for pituitary adenoma (PA), vestibular schwannoma (VS), and post-operative brain metastases (pBM) at a single institute were randomly selected. The original clinical plans were manually generated and reviewed prior to treatment. Each case was re-planned by two independent planners using the inverse planning (IP) system in Gamma Knife Lightning. The planning goal was to achieve at least 99% prescription dose coverage to targets, while minimizing beam-on time and dose to organs at risk (OARs), if any. The beam-on time, scaled by plan dose rate, and dosimetric metrics were compared among three plans. Statistical analyses were performed using the ANOVA and Tukey tests.
Results: Thirty GKSRS patients were identified with 10 in each diagnosis type. At least 99% target coverages were achieved in the IPs, comparable to those in the clinical ones. There were no significant differences in the Paddick Conformity Index. Whereas the Gradient Indices were equivalent among the plans for PA and VS, the GIs in the IPs for pBM were significantly lower (p<0.01) due to the relatively large targets. Similar dosimetric metrics of OARs were also observed between the IPs and clinical plans. The average beam-on times in the IPs were reduced by 25 minutes for PA (p<0.01) and by 10 minutes for VS (p=0.60), but unchanged for pBM, However, the reductions in the IPs for PA and VS were offset by the increased numbers of shots used to control the dose to OARs, resulting in comparable total treatment durations.
Conclusion: Inverse planning in GK Lightning allows planners to generate GKSRS plans of equivalent plan qualities compared to manual planning. The automation accelerates GKSRS planning and therefore improves the clinical workflow.