Evaluation of two treatment planning systems for single isocenter multiple metastases stereotactic radiosurgery treatment planning and delivery

Abstract

Plan quality and low dose spillage, in the context of single isocenter treatment of multiple metastases, are evaluated for two commercially available treatment planning systems (TPSs): Brainlab Elements Multiple Mets Software v.3.0 and v.4.0 (MBM) and RayStation v.11A (RS) representing a TPS specialized for this application and a more generalized TPS, respectively. Twenty-one patients, originally planned for treatment with MBM, were selected for analysis with 66 (2 to 5 per patient) total lesions with average of 3.1 lesions per isocenter. Organs-at-risk were auto-contoured in Brainlab MBM and reviewed by physicist or physician prior to treatment planning. Gross tumor volumes (GTVs) and planning target volumes (PTVs) were contoured in MBM. All structures, CT, MRI, and registration files were exported to RS. Structures were resampled upon export. Treatment plans were developed in RS. Plans in MBM used dynamic conformal arcs (DCA) and RS plans utilized VMAT, both implementing non-coplanar trajectories. Plan evaluation metrics include planning target volume (PTV) coverage (D98 %), conformity index (CI), gradient index (GI) normal brain V12Gy (cc) and V4Gy (cc). PTV D98 % coverage was comparable between MBM and RS plans demonstrating an average difference of 0.87 Gy (± 1.07 Gy) (with RS having higher average D98 %). The greatest difference between MBM and RS was the dose to the normal brain. Normal brain V12Gy was on average 4.54cc (± 3.02cc) vs 5.47cc (± 3.00cc) and V4Gy was 33.23cc (± 22.08cc) vs 43.35cc (± 26.24cc) for MBM vs RS, respectively. The conformity index (CI) improved in MBM vs RS (1.20 vs 1.56 average). MBM dynamic conformal arcs (DCA) and RS VMAT approaches produce comparable plans, however we found MBM treatment plans to be superior due to better target conformality and subsequent lower normal brain V12Gy and V4Gy.