Room: AAPM ePoster Library
Purpose: many variables to adjust, the conventional manual forward planning for Gamma Knife (GK) radiosurgery makes plan quality heavily depend on planners’ skills, experiences and the amount of devoted efforts and hence vary significantly. This is a concern particularly for vestibular schwannoma cases, which usually have more challenging geometries. Quality control is hence highly desired to consistently provide high-quality care to each individual patient.
Methods: have built a database of high-quality GK plans by collecting 21 previous vestibular schwannoma cases treated with Leksell ICONTM GK unit in our institution, and replanning them using our recently developed multiresolution-level inverse planning (MRL-IP) approach, which has been validated to generate globally-optimal or approximately globally-optimal plans. Considering the mechanical design of GK unit, we quantify planning difficulty level using target volume, shape complexity (e.g., perimeter-to-area-ratio and convex-area-to-area-ratio of target contours), and target’s proximity to nearby organs (e.g., overlap-absolute-volume-histogram for brainstem and overlap-relative-volume-histogram for ipsilateral cochlea). For an under-assessment plan, a reference case with the closest planning difficulty level is identified to estimate expected plan quality (e.g., coverage, selectivity, intermediate dose spillage, brainstem D0.1cc dose, mean dose of ipsilateral cochlea, and beam-on time) to make quality control decision.
Results: cases (T1-3) were used to test our quality control method, and replanned using our MRL-IP approach for validation. T1’s plan quality was comparable to the expected quality, consistent with the negligible improvement obtained by replanning. T2’s plan quality was much worse than estimation, which was supported by the marked improvement gained by replanning. Due to the very limited cases in the current database, we cannot find a good reference case for T3.
Conclusion: preliminary study has demonstrated the feasibility of the proposed quality control method. We will keep adding high-quality plans of various geometries into our database to improve its versatility.
Funding Support, Disclosures, and Conflict of Interest: Supported by Winship Cancer Institute #IRG-17-181-06 from the American Cancer Society