Modeling the transport of photons from source to detector as a means of improving radiation therapy by enhanced utilization of kV and MV x-ray imaging

导言：本研究项目旨在开发新的算法，从现代放射治疗中获得的诊断和治疗医学图像中去除受污染的X射线散射。目前的放射治疗单位使用千伏(KV)X射线进行解剖成像(即。在提供治疗之前使用巨伏(MV)X射线束的锥形束计算机断层扫描。这些单元的治疗和诊断成像应用目前还处于初级阶段。*假设：MV和KV能量X射线的混合散射估计算法将提高图像重建和密度转换(CBCT)的精度，并改善剂量重建(MV)的精度，从而提高错误检测的灵敏度。*目标：1.发展一种结合确定性方法和新的快速蒙特卡罗方法的混合散射估计算法。2.用MV和KV两种X射线束验证和优化了该算法。3.结合交错的KV和MV成像，以便在治疗期间进行诊断成像。*方法：基于我们以前的经验，将开发一种混合算法来估计KV和MV能量X射线束对患者的X射线散射。快速的解析解将估计单次散射的X射线，而将每个体素视为散射源的新的蒙特卡罗方法将估计多次散射的X射线。该方法将在MV治疗性波束和KV(CBCT)诊断成像波束上进行验证，并根据重建准确性和改善的对比度(CBCT)以及患者在各种几何模型和拟人模型上的重建剂量精度(MV)进行定量评估。此外，将开发一种交错采集方法，通过使用XML控制软件在放射治疗提供期间同时获取KV和MV图像数据，将患者运动的影响降至最低。性能将通过图像质量结果与标准顺序图像采集的定量比较来评估。*意义：这项研究计划将提供关于放射治疗单位的质量显著提高的CBCT数据(即改善的肿瘤可见性、准确的密度转换)。对于MV成像，将提供更准确的患者剂量重建，从而提高检测交付错误的敏感度。交错图像采集将在治疗过程中提供更准确的患者解剖表示。未来将创造实时肿瘤跟踪和适应性放射治疗的研究机会。*结论：这项研究计划建立在我们开发散布估计算法及其应用的经验基础上，以提高放射治疗的准确性。最终，这些进步可能会在加拿大和世界各地用于改善所有接受放射治疗的患者的生活质量。

Introduction: This research program seeks to develop new algorithms to remove contaminating x-ray scatter from diagnostic and therapeutic medical images obtained in modern radiation therapy. Current radiation treatment units use kilovoltage (KV) x-rays for anatomical imaging (ie. cone-beam computed tomography, CBCT) just prior to therapy delivery using a megavoltage (MV) x-ray beam. Therapy and diagnostic imaging applications on these units are currently at a rudimentary level.***Hypothesis: A hybrid scatter estimation algorithm for both MV and KV energy x-rays will improve the accuracy of image reconstruction and density conversion (CBCT) and also improve the accuracy of dose reconstruction (MV) providing increased sensitivity for error detection. ***Objectives: 1. Develop a hybrid scatter estimation algorithm combining a deterministic approach with a novel, fast Monte Carlo method. 2. Validate and optimize this algorithm using both MV and KV x-ray beams. 3. Combine interleaved KV and MV imaging allowing diagnostic imaging during therapy delivery.***Methods: Building on our previous experience, a hybrid algorithm will be developed to estimate x-ray scatter from patients for both KV and MV energy x-ray beams. A fast analytical solution will estimate singly-scattered x-rays, while a novel Monte Carlo approach treating each voxel as a scatter source will estimate multiply-scattered x-rays. The method will be validated on both MV therapeutic beams and KV (CBCT) diagnostic imaging beams and quantitatively evaluated in terms of reconstruction accuracy and improved contrast (CBCT), and patient reconstructed dose accuracy (MV) on a variety of geometric and anthropomorphic phantoms. Furthermore, an interleaved acquisition approach will be developed to concurrently obtain KV and MV image data during radiation treatment delivery through the use of XML control software, minimizing the impact of patient motion. Performance will be assessed through quantitative comparison of image quality results to standard sequential image acquisition.***Significance: This research program will provide significantly improved quality CBCT data on radiation treatment units (i.e. improved tumour visibility, accurate density conversions). For MV imaging, more accurate patient dose reconstructions will be provided, improving sensitivity in detecting delivery errors. Interleaved image acquisition will provide a more accurate representation of patient anatomy during treatment. Future research opportunities for real-time tumour tracking and adaptive radiation therapy will be created.***Conclusion: This research program builds on our experience developing scatter estimation algorithms and their applications to improve the accuracy of radiation treatments. Eventually these advances may be used throughout Canada and the world to improve quality of life for all patients treated with radiation.**