Dosimetric Evaluation Of An Automatic Treatment Planning Option For Multiple Brain Metastases Stereotactic Radiosurgery (Srs)

Multiple brain metastases stereotactic radiosurgery treatment planning is challenging due to its geometry complexity and small target size. This study dosimetrically evaluated a manual and an automatic treatment planning approach in terms of target coverage and normal tissue sparing. Five patients with 3-10 lesions, previously planned in RayStation (v8.0A, RaySearch, Stockholm, Sweden), were re-planned in a stereotactic radiosurgery planning software. CT and MR images along with structures were imported into Elements from RayStation to eliminate variables. PTV size ranged from 0.08 to 4.43cc in volume and 0.11 to 2.84cm in diameter. All cases were planned on a 6FFFwith HDMLC and optimized to meet institutional planning standards. Manual plans were created with VMAT by an experienced planner with 1-3 iso centers based on best clinical judgement. Automatic plans were generated in Elements with dynamic conformal arcs and a single iso center. Specifically, automatic plans used optimization objectives of 100% prescription dose covering at least 99.5% of each PTV while keeping D0.5% within 125.0%. Initial beam arrangements for automatic plans were selected from 3 predefined templates with 4 couch angles and 4 or 5 partial arcs. Prescription dose ranged from 16 to 20Gy in 1 fraction. PTV coverage, maximum dose (Dmax), Conformity Index (CI), Homogeneity Index (HI), Volume of normal brain receiving 12Gy (V12Gy) and 3Gy (V3Gy), MU per fraction, and Dmax to Organs at Risk (OAR) were recorded and analyzed with two-tailed paired student t-test. The PTV coverage was at least 99.0% and 99.5%, with Dmax less than 133.2% and 122.5% for manual and automatic plans respectively. CI and HI were 1.38 ± 0.08 and 1.28 ± 0.02 for automatic plans, which were reduced by 4.82 ± 2.96 and 0.18 ± 0.07 compared to manual plans. V12Gy and V3Gy for automatic plans were 9.06 ± 2.11cc and 201.11 ± 95.88cc respectively, which were 4.42 ± 2.07cc and 63.60 ± 153.74cc smaller than manual plans. Automatic plans utilized 8034 ± 2173 MU per fraction compared to 15560 ± 3444 MU for manual plans, resulting in beam-on time reduction of 6.27 ± 1.06 mins. Both manual and automatic plans met OAR dose constraints, while automatic plans reduced dose to brainstem, cochleae, lenses, optic chiasm, and optic nerves by 0.39 ± 4.18Gy, 0.88 ± 1.23Gy, 0.18 ± 0.88Gy, 1.38 ± 2.62Gy, and 1.09 ± 2.43Gy respectively. On average, automatic plans finished within 10 mins while manual plans took about 3 hours for each iso center. Both manual and automatic planning can provide clinically acceptable plans. For the same PTV coverage, automatic plans provide less dose to OARs and reduced treatment time significantly. Automatic planning requires less efforts and is not dependent on planner’s experience. Despite only 5 patients have been treated and analyzed in our clinic, it revealed that automatic planning is of great importance to standardize and improve complicated multiple brain metastases plans.