Deformable registration for image-guided radiotherapy

项目描述（由申请人提供）：本项目旨在通过在诊断和治疗计划图像中准确、自动地注册可变形组织，提高图像引导放射治疗中剂量靶向的准确性和精度。医学成像和放射治疗的最新进展有可能通过非侵入性地确定癌症的位置和范围，并允许医生在不影响周围健康组织的情况下对癌变病变进行升级和靶向辐射剂量，从而改善患者护理。为了弥补由于成像要求（如直肠内探头、患者体位差异、体重变化和其他因素）在诊断和治疗阶段之间发生的重大变化，将开发并验证一种新的可变形图像配准计算工具。给定分割的参考图像和治疗过程中获得的图像，该算法将生成组织的3D表示，使用有限元模型估计组织位移和变形，考虑未知模型参数造成的不确定性，并输出图像之间的映射。低计算时间是临床可行性的关键。新方法将应用于将前列腺CT或MRI/MRSI参考图像与使用megvoltage Cone-Beam GT （MV CBCT）获得的治疗图像进行注册，MV CBCT是一种正在开发的3D成像方式，用于对治疗台上的患者进行调强放疗（IMRT）成像。该方法的验证将通过骨轮廓和测量标记种子的注册误差来进行，这些标记种子已经植入了一小部分患者的前列腺内。针对前列腺癌治疗的新软件工具，旨在通过使临床医生在放射治疗期间提高剂量与组织类型的一致性，显著改善患者护理。这些改进可以通过降低癌症患者的恢复时间、复发率和治疗费用，显著改善公众健康。

DESCRIPTION (provided by applicant): The objective of this project is to improve the accuracy and precision of dose targeting in image-guided radiation treatment by accurately and automatically registering deformable tissues in diagnostic and treatment planning images. Recent advances in medical imaging and radiation therapy have the potential to improve patient care by noninvasively identifying the location and extent of cancer and by allowing physicians to escalate and target radiation dose to cancerous lesions while sparing surrounding healthy tissues. To compensate for significant changes that occur between diagnostic and treatment phases due to imaging requirements such as endorectal probes, patient position differences, weight change, and other factors, a new computational tool for deformable image registration will be developed and validated. Given a segmented reference image and an image obtained during treatment, the algorithm will generate a 3D representation of the tissues, estimate tissue displacements and deformations using a finite element model, account for uncertainty due to unknown model parameters, and output a mapping between the images. Low computation time is critical for clinical viability. The new method will be applied to register prostate CT or MRI/MRSI reference images to treatment images obtained using Megavoltage Cone-Beam GT (MV CBCT), a 3D imaging modality being developed to image the patient on the treatment table for Intensity-Modulated Radiation Therapy (IMRT). Validation of the method will be performed using bone outlines and by measuring registration errors for marker seeds that have been implanted inside the prostate in a subset of patients. The new software tool, targeted at prostate cancer treatment, aims to significantly improve patient care by enabling clinicians to improve conformality of dose to tissue types during radiation therapy. These improvements could significantly enhance public health by lowering recovery times, recurrence rates, and treatment costs for cancer patients.