An Evidence-Based Evaluation of Lesion Size and Distance Criteria for Rotational Error Mitigation In Single-Isocenter Multi-Target (SIMT) Radiosurgery
Abstract
Purpose
In comparison to treating a single lesion at isocenter, the use of SIMT to treat multiple lesions is more sensitive to rotational errors, which can affect the delivered dose distribution. The aim of this study is to retrospectively assess the robustness of SIMT plans treated at our institution to rotational errors. Additionally, our study aims to provide guidance for planning future cases as to whether a lesion can be robustly treated with SIMT.
Methods
6 Patients with 51 brain metastases were retrospectively included in the study. SIMT plans were developed at varying dose levels of 18Gyx1, 8Gyx3, 9Gyx3, and 10Gyx3 prescribed to the GTV. V100%, V95%, D99%, conformity index (CI), gradient index (GI), and Brain V12Gy were collected for each GTV. A 1° pitch, roll, and yaw, error was introduced for each GTV using the Eclipse Plan Uncertainty Parameter function. A Logistic regression model was used to fit lesion size (LS) and distance-to-isocenter (DTI) to predict V100% being less than 95% (V100%<95) after a simulated rotation.
Results
The patient cohort consisted of a median of 4 lesions per plan (3-12), with a median DTI of 2.83 cm (0.80– 6.63cm) and a median LS of 0.17cc (0.01 - 13.63cc). The lesion cohort was divided using an 80/20 split for training and testing. LS, DTI, original V100% and the Paddick CI were found to be significantly correlated with V100%<95% after introduced rotations with Pearson coefficients of -0.40, 0.31, -0.38, -0.31, respectively. The model was able to achieve an AUC of 0.96 on the testing cohort when predicting V100%<95% using LS and DTI.
Conclusion
The results of this abstract demonstrate a simple and effective method, contained mostly within the Eclipse environment, to determine the LS and DTI limits when aiming to achieve a robust SIMT plan using institutional data.