放射治療碰撞自動預測與射束角度最佳化

Collision Prediction and Beam Angle Optimization for External-Beam Radiation Therapy

王裕仁(Yu-Jen Wang)

指導教授 ： 賴飛羆

共同指導教授 ： 成佳憲(Jason Chia-Hsien Cheng)

國立臺灣大學/電機資訊學院/生醫電子與資訊學研究所/博士(2021年)

https://doi.org/10.6342/NTU202100867

摘要

目的：
放射線治療的角度最適化為執行現代放射治療時的重要挑戰。不論是採用強度調控或是弧形放射治療，皆缺乏便利與智慧的角度選取方式。治療採用非共面角度，具有劑量上的優勢，可以減少正常組織所受到的放射劑量，然而，可能會增加物理上碰撞的風險。尤其是在病人體型比較寬大、治療輔具比較突出，或是外加使用如生理監視器等設備時。本研究研發一套軟體能精準預測病人接受放射線治療時，會產生碰撞的角度，並自動生成治療最適化角度的組合。
材料與方法:
我們建構了兩組不同廠牌的直線加速器模型，以模擬不同機器不同環境下可能的碰撞組合。加速器的外形與體積是根據加速器手冊與實際量測的數值而來。病人與輔具的外部輪廓來自模擬定位時，電腦斷層所取得的影像。我們使用擬人假體搭配真空固定墊裝置進行現場實測軟體預測的準確程度。並回溯過往實際治療病人因為遭遇可能碰撞而更改治療計畫者，使用本軟體進行預測。同時，也利用過往病人的定位影像，進行最佳化治療角度的選取與運算。
結果：
我們所研發出的軟體，就模擬預測與實際現場測量的比較來看，兩台直線加速器與軟體模擬預測的誤差皆在5度以內。我們使用混淆矩陣來評估預測成果，顯示在單純只有直線加速器的情況下，正確率為98.7%與97.3%。真陽性率為97.7%與96.9%，而真陰性率為99.8%與97.9%。而使用擬人假體來執行的預測則顯示，正確率各為96.8%和97.3%。在真實臨床情境測試上，本軟體亦成功預測過往實際治療病人有因為遭遇可能碰撞進而更改治療計畫者的情形。另外，本研究也成功研發角度選取的模組，可以直接快速選出一組最適化的角度。

結論：
本軟體能成功預測體外放射線治療時，可能產生之碰撞情況，能對採用更多非共面放射線治療角度有所幫助。角度選取器能幫助快速選出非碰撞可用的最適化治療角度組合。

關鍵字

放射線治療 ； 碰撞 ； 非共面 ； 治療角度 ； 軟體

並列摘要

Purpose:
Beam angle optimization is a critical issue and is a challenging task for modern radiation therapy (RT). Until now, it still lacks a convenient strategy to select beams wisely. Noncoplanar RT techniques may have dosimetric advantages but increase mechanical collision risk, especially for large body sizes, large immobilization equipment or with physiological monitor during RT. We propose a software solution to accurately predict colliding/noncolliding configurations for coplanar and noncoplanar beams and noncolliding optimized beam angle sets for the dosimetric plan.
Materials and Methods:
We built the models of two different linear accelerators to simulate noncolliding gantry orientations for phantom/patient subjects. The sizes and shapes of the accelerators and the relative position between the couch and the gantry were delineated based on their manuals and the on-site measurements. The subjects’ external surfaces including the body and immobilization device, were automatically extracted based on computed tomography (CT) simulations. An Alderson Radiation Therapy phantom with vacuum bag was used to predict spatial collision prediction accuracy by the software. We used the simulation with one patient encountering a gantry collision problem during the initial setup to test the software’s validity. Patients with hepatocellular carcinoma (HCC) previously treated with RT were used to estimate the optimized beam sets of intensity-modulated radiation therapy (IMRT).
Results:
The difference between software estimates and on-site measurements demonstrated the noncoplanar collision angles all predicted accurately within a 5-degree difference in gantry position. The confusion matrix was calculated for each of the two empty accelerator models, and the accuracies were 98.7% and 97.3%, respectively. The true positive rates were 97.7% and 96.9%, while the true negative rates were 99.8% and 97.9%, respectively. For the phantom study, the accuracies were 96.8% and 97.3%, respectively. The collision problem encountered of the breast cancer patient in the initial setup position was identified successfully by the software. Moreover, the software provides the function to help physicians choose the beam angles with the optimized dosimetry.
Conclusion:
The developed software effectively and accurately predicted the collisions for the accelerator, phantom, and patient setups. This software may help prevent collisions and expand the spatial range of applicable beam angles. Beam angle selectors can help t choose the noncolliding and optimized beam sets efficiently.

並列關鍵字

Radiotherapy ； Collision ； Noncoplanar ； Beam angle ； Software

參考文獻

[1] M. K. Bucci, A. Bevan, and M. Roach III, "Advances in radiation therapy: conventional to 3D, to IMRT, to 4D, and beyond," CA: a cancer journal for clinicians, vol. 55, no. 2, pp. 117-134, 2005.

[2] K. Otto, "Volumetric modulated arc therapy: IMRT in a single gantry arc," Medical physics, vol. 35, no. 1, pp. 310-317, 2008.

[3] I. M. R. T. C. W. Group, "Intensity-modulated radiotherapy: current status and issues of interest," International Journal of Radiation Oncology* Biology* Physics, vol. 51, no. 4, pp. 880-914, 2001.

[4] H. Rocha, J. M. Dias, T. Ventura, B. d. C. Ferreira, and M. d. C. Lopes, "Beam angle optimization in IMRT: are we really optimizing what matters?," International Transactions in Operational Research, vol. 26, no. 3, pp. 908-928, 2019.

[5] L. Yuan et al., "Lung IMRT planning with automatic determination of beam angle configurations," Physics in Medicine Biology, vol. 63, no. 13, p. 135024, 2018.