基于虚拟现实技术的放疗CT模拟定位远程培训系统研发

目的 研发一种基于虚拟现实技术的放疗CT模拟定位的远程培训系统,探索一种医学培训的新方法。方法 使用3DMax与Maya进行3D建模,Unity3D引擎开发3D虚拟操作及交互系统;Java的SpringMvc架构作为系统后台服务,MySQL作为后台数据库系统;并将用户分为教师和学员两种角色,模式分为教学与考核模式。结果 系统功能涵盖CT模拟定位全过程,主要包括患者信息管理、CT模拟定位机认知、体位固定技术、CT定位扫描、处理突发事件等模块。自2018年投入使用以来,运行稳定,系统浏览量达14 920人次,培训通过率为86.66%。与传统培训相比,培训效率明显提升,并获得一致好评。结论 远程培训系统能有效提升学员的临床实践能力、人文关怀能力,具有良好的自主性、共享性、创新性。目前系统已上线且推广性较强,应用前景广阔。     

基于勾画内容检索方法的鼻咽癌危及器官结构标准化命名的研究

目的 根据AAPM TG-263,提出基于内容的放疗结构命名标准化(CBSN),探索其用于鼻咽癌危及器官(OAR)标准化验证的可行性。方法 回顾性收集2017—2019年经调强放疗的855例鼻咽癌患者的放疗结构文件(其中15例含临床异常结构)。通过Matlab自编程软件获取医师勾画OAR的图像位置、几何特征、一阶灰度直方图和基于灰度共生矩阵的纹理特征等,建立CBSN位置验证模型和CBSN知识库。采用Fisher判别分析建立CBSN的OAR分类模型,利用自身验证、交互验证和外部验证对模型评估。结果 99%(69/70)的模拟异常结构在CBSN知识库90%参考值范围之外,不同OAR之间的特征参数不同(P<0.001)。CBSN的OAR分类模型的自身验证、交互验证和外部验证的准确率分别为92.1%、92.0%和91.8%。14例临床异常结构成功被CBSN检测,准确率为93%(14/15)。模拟测试中左右位置异常结构验证的准确率为100%,如检测到命名为Len_L的右晶体。结论 CBSN可用于鼻咽癌OAR验证,为鼻咽癌多中心合作及标准化的放疗提供参考依据。     

肺癌插植放疗中插植针排列规则对肺受量影响研究

目的 分析肺部肿瘤的近距离组织间插植放疗插植针的排列规则对肺受照剂量影响。方法 针对15例近距离组织间插植放疗肺癌患者病例,设计不受肋骨限制的插植针规则排列的放疗计划,命名为虚拟计划并与原实际计划进行比较。分别统计当两种计划方案的处方剂量为 10、30、60、120 Gy时肺V5、V20、V30、MLD。采用Wilcoxon符号秩检验。结果 虚拟计划方案的肺V5、V20、V30、MLD均小于实际计划方案,差异均有统计学意义(P<0.05)。结论 插植针的不规则排列限制了近距离组织间插植放疗在肺受量上进一步降低;所以在插植手术中应尽可能使得插植针规则排列。     

利用多目标优化技术的KBP模型精炼方法研究

目的 研究利用多目标优化（MCO）技术提高基于先验知识自动计划（KBP）模型的计划设计质量的可行性。方法 选择55例已完成放疗的鼻咽癌患者,每个病例都采用固定野调强放疗技术。随机选择40个病例的调强放疗计划作为训练集1,通过多目标优化技术对训练集1中的放疗计划进行预处理,构建新的训练集2。将初始训练集1和处理后的训练集2作为样本,分别训练得到传统的KBP模型和多目标优化技术精炼后的MCO‐KBP模型。在剩余15个病例中随机选择5例作为验证集,其余10例作为测试集。验证通过后用测试集对初始的人工计划以及传统KBP模型和MCO‐KBP模型生成自动计划的计划质量进行统计学分析。结果 传统KBP模型和MCO‐KBP模型生成自动计划的靶区剂量（D_95%）均能满足临床要求,适形性指数与均匀性指数基本一致（P>0.05）,且MCO‐KBP模型生成的自动计划中重要危及器官的剂量比传统KBP模型更低（脑干D_max平均偏低2.13 Gy,左、右腮腺D_mean分别平均偏低1.39、1.59 Gy,左、右视神经及垂体的D_max分别平均偏低1.42、1.16、1.88 Gy,差异均有统计学意义）。结论 与传统的KBP模型相比,精修后的MCO‐KBP模型设计的调强放疗计划对危及器官的保护具有明显优势,利用MCO技术提高KBP模型的计划设计质量是可行的。     

利用多目标优化技术的KBP模型精炼方法研究

目的 研究利用多目标优化（MCO）技术提高基于先验知识自动计划（KBP）模型的计划设计质量的可行性。方法 选择55例已完成放疗的鼻咽癌患者,每个病例都采用固定野调强放疗技术。随机选择40个病例的调强放疗计划作为训练集1,通过多目标优化技术对训练集1中的放疗计划进行预处理,构建新的训练集2。将初始训练集1和处理后的训练集2作为样本,分别训练得到传统的KBP模型和多目标优化技术精炼后的MCO‐KBP模型。在剩余15个病例中随机选择5例作为验证集,其余10例作为测试集。验证通过后用测试集对初始的人工计划以及传统KBP模型和MCO‐KBP模型生成自动计划的计划质量进行统计学分析。结果 传统KBP模型和MCO‐KBP模型生成自动计划的靶区剂量（D_95%）均能满足临床要求,适形性指数与均匀性指数基本一致（P>0.05）,且MCO‐KBP模型生成的自动计划中重要危及器官的剂量比传统KBP模型更低（脑干D_max平均偏低2.13 Gy,左、右腮腺D_mean分别平均偏低1.39、1.59 Gy,左、右视神经及垂体的D_max分别平均偏低1.42、1.16、1.88 Gy,差异均有统计学意义）。结论 与传统的KBP模型相比,精修后的MCO‐KBP模型设计的调强放疗计划对危及器官的保护具有明显优势,利用MCO技术提高KBP模型的计划设计质量是可行的。     

食管癌根治性调强放疗大体肿瘤体积因素与靶区内放疗失败的关系

目的 回顾性分析GTV-T因素与肿瘤局部失败的关系。方法 整理并计算133例行根治性调强放疗的食管癌患者GTV-T因素。包括:GTV-T体积,GTV-T最大管壁厚度(GTV-T区域中食管平面最大厚度),GTV-T体积/长度 (GTV-T体积除以GTV-T层面数计算病变长度)。结果 GTV-T区域治疗失败率与GTV-T体积呈正相关线性趋势关系。GTV-T肿瘤体积为36 cm³,最大管壁厚度为2.5 cm,以及体积/长度为5.3 cm²,可作为预估食管癌放疗失败风险的临界值。结论 根治性调强放疗大体肿瘤体积(GTV-T)因素能在一定程度上预示食管癌根治性放疗后局部控制情况,预估食管癌的放疗失败风险。     

BOLD-fMRI在脑胶质瘤术后IMRT中视觉功能区保护价值

目的 评价血氧水平依赖功能MRI (BOLD-fMRI)在枕叶脑胶质瘤术后放疗视觉功能区保护中的临床应用价值。方法 选取2014—2016年间枕叶区脑胶质瘤术后拟行放疗病例28例。所有患者放疗前行定位CT、常规MRI及BOLD-fMRI检查,获得在3DT1解剖图像上显示的视皮质激活图,并以此为指导在定位CT图像上标记视皮质区,采用IMRT技术分别制定视觉功能区保护性放疗计划和未考虑视觉功能区的常规放疗计划。通过比较PTV的CI、HI、Dmax、Dmean以及视皮质区所受Dmax、Dmean,评价视觉功能区保护性方案的可行性。结果 常规放疗计划及视觉功能区保护性放疗计划在确保靶区有效治疗剂量及常规OAR剂量限制在安全可接受范围前提下,PTV的CI及HI差异无统计学意义(P=0.874、0.602)。视觉功能区保护性放疗计划患侧视皮质Dmax、Dmean分别降低8.40%、9.25%,健侧13.26%、14.77%。结论 利用BOLD-fMRI指导制定视觉功能区保护性放疗计划,在保证靶区处方剂量前提下,相比常规放疗计划,降低了视觉功能区受量,进而可以实现相应的功能区保护。     

基于Xenapp虚拟系统在放疗流程中的应用

目的 放疗流程中牵涉到许多软件应用系统和复杂的硬件设备,借助基于Xenapp虚拟系统整合放疗流程中的各个应用系统。方法 采用windows2008r2操作系统,Citrix Xenapp 6.5 sp1企业版,sqlserverr2008 express 数据库。安装Xenapp服务器,将放疗科常用的软件应用在服务器上安装并发布,在放疗科局域网内的电脑上安装CitrixReceiver,并使用网页浏览器登录Xenapp服务器站点,即可使用站点发布的所有应用。结果 使用了Citrix Xenapp的虚拟化系统,大大减少了计算机硬件的开销,使得计算机硬件的开销减少50%以上,减少了技术人员的维护成本,使得放疗科的工作做到无缝衔接,提高了放疗科工作人员的办公效率。结论 基于Xenapp的虚拟化系统在放疗流程中可以起到节约开销,提高效率的作用,值得在临床工作中大力推广。     

直肠癌术前/术后适形/调强放疗靶区勾画共识与图谱

目的 在三维放疗时代,为直肠癌术前/术后放疗制定靶区勾画的中国专家共识,并辅以高分辨率图谱。方法 收集直肠癌放疗靶区勾画的国际指南,结合国内肿瘤中心的经验,起草“直肠癌术前/术后适形调强放疗靶区勾画共识与图谱”,在“中国医师协会结直肠肿瘤专委会放疗专委会”集中会议上进行讨论及投票,由专门的执行小组对最终的结果进行归纳。结果 主要针对直肠癌放疗的定位技术、术前放疗靶区定义、根治术后放疗靶区定义、OAR定义、剂量和分割等方面,根据证据的强弱和专家意见一致的程度,给予分级推荐;同时在高分辨率CT上提供了CTV的图示。结论为直肠癌放疗提供中国专家共识,规范直肠癌放疗流程,减少不同单位或不同医生之间勾画的差异。     

基于实时超声图像引导技术和线性判别模型分析前列腺癌放疗分次内运动模式

目的 基于经会阴超声(TPUS)实时扫描技术和线性判别分析(LDA)法,定性分析并自动判别前列腺癌放疗分次内运动模式,为个体化精确放疗奠定基础。方法 应用TPUS技术记录了61例前列腺癌患者共1265个分次近百万个实时监测数据,划分为稳定型、波浪型、小偏执型、银叉型、回归型、大偏执型和稽留型运动模式。对运动轨迹量化并提取特征参数,通过LDA法建立判别方程式,评估训练集和测试集的判别效果。结果 平均每位患者存在4种不同的运动模式,不稳定型占(35.00±21.49)%。随着治疗次数的增加,运动轨迹并未表现出越来越稳定的趋势,不同模式的出现极不规则。构建的线性判别模型对训练集和测试集的判别准确率分别为90.4%和89.5%,敏感性和特异性分别为84.9%和91.1%。结论 前列腺癌患者分次内运动模式多样且随机,具有不可预测的特点。LDA法可以有效地对分次内运动模式进行判别,同时在治疗过程中利用判别方程和中心坐标实现对未知样本的自动鉴别。     

Research and development of remote training system for radiotherapy CT simulation based on virtual reality technology

Objective To develop a remote training system for CT simulation positioning of radiotherapy using virtual reality technology, and to explore a new method of medical training. Methods The 3DMax and Maya were employed to establish the 3D model. The unity3D engine was adopted to develop 3D virtual operation and interaction system. Java spring MVC architecture was utilized as the system background service. MySQL was used as the background database system. The users were assigned into two roles:teacher and student, and the modes were divided into teaching and assessment modes. Results The function of the system covered the whole process of CT simulation positioning, mainly including modules of patient information management, CT simulation positioning machine cognition, body position fixation technology, CT positioning scanning, and emergency handling, etc. Since it was put into use in 2018, the system has been running stably, with 14920 pages views and an 86.66% pass rate. Compared with the traditional training, the training efficiency has been significantly improved and has received unanimous recognition. Conclusions The remote training system can effectively improve the clinical practice ability and humanistic care ability of the trainees, which has good autonomy, sharing, and innovation. At present, the system has been put online and has strong popularization with prospects for broad application.     

Training in surgical oncology - the role of VR simulation

There have been dramatic changes in surgical training over the past two decades which have resulted in a number of concerns for the development of future surgeons. Changes in the structure of cancer services, working hour restrictions and a commitment to patient safety has led to a reduction in training opportunities that are available to the surgeon in training. Simulation and in particular virtual reality (VR) simulation has been heralded as an effective adjunct to surgical training. Advances in VR simulation has allowed trainees to practice realistic full length procedures in a safe and controlled environment, where mistakes are permitted and can be used as learning points. There is considerable evidence to demonstrate that the VR simulation can be used to enhance technical skills and improve operating room performance. Future work should focus on the cost effectiveness and predictive validity of VR simulation, which in turn would increase the uptake of simulation and enhance surgical training.     

Virtual reality in radiation therapy training

Integration of virtual reality (VR) in clinical training programs is a novel tool in radiotherapy. This paper presents a review of the experience with VR and Immersive visualization in 3D perspective for planning and delivery of external radiotherapy.Planning and delivering radiation therapy is a complex process involving physicians, physicists, radiographers and radiation therapists/nurses (RTT’s). The specialists must be able to understand spatial relationships in the patient anatomy.Although still in its infancy, VR tools have become available for radiotherapy training, enabling students to simulate and train clinical situations without interfering with the clinical workflow, and without the risk of making errors. Immersive tools like a 3D linear accelerator and 3D display of dose distributions have been integrated into training, together with IT-labs with clinical software. Training in a VR environment seems to be cost-effective for the clinic. Initial reports suggest that 3D display of dose distributions may improve treatment planning and decision making. Whether VR training qualifies the students better than conventional training is still unsettled, but the first results are encouraging.     

Target Volume Delineation Training in Radiation Oncology in India : A Survey Evaluating Its Status,the Need for Educational Programs and the Utility of Virtual Teaching

Introduction: The complexity of modern-day radiotherapy techniques calls for greater accuracy in target volume delineation which requires intensive training and expert guidance. In this study, we intend to evaluate the status of training in target delineation across radiation oncology curriculums in India and utility of webinars in teaching it. Methodology: We organized daily webinars on the topics of radiological anatomy and target volume delineation in common sites of cancer over a period of 2 weeks. At the end of the program, a 35 item survey questionnaire was shared with the participants. The responses were analysed and are reported here. Results: Out of the 797 participants that registered for the course, 356 radiation oncologists responded to the survey questionnaire. Majority (96%) of our respondents believe that there is a need for additional training in target volume delineation. Ninety percent of the participants felt that radiation oncology curriculum requires a formal radiology training but only 6.7% reported that their training consisted of a dedicated rotation and regular lectures in radiological anatomy. Majority (97%) responded that they were  likely to incorporate the points learnt from the webinar sessions in their daily practice. Forty eight percent of respondents opted for attending a webinar to an in-person event in the future while 34% would like to have the option to choose between the two. Ninety four percent agreed that online webinars should be conducted routinely even after the pandemic ends. Conclusion: Our survey results suggest that target volume delineation training in India is deficient and virtual teaching programs can be a practical and effective way to improve it.     

MR引导的放疗系统临床应用

目的:探讨MR引导的Unity放疗系统在临床应用中的可行性。方法:临床试验入选24例患者,按治疗部位分为头颈组、胸腹组、盆腔组、脊柱组和肢体组,总共384个治疗分次。在无外置激光的情况下进行摆位,统计在线治疗流程中不同阶段所需时间和每个分次的配准误差。每周利用MR几何形变模体检测MR图像几何形变大小。并利用ArcChe...     

基于直线加速器虚拟源模型的蒙特卡洛剂量计算及在IMRT独立验算中的初步应用

目的：研究临床放疗蒙特卡洛剂量计算方法中虚拟源模型的可行性。方法通过蒙特卡洛方法模拟得到记录医用直线加速器机头出射粒子物理特性的相空间文件,分析提取相空间文件中粒子的种类、能谱及位置分布,建立半经验虚拟双光子源抽样模型。结合并行剂量计算引擎GMC,得到3 cm×3 cm、5 cm×5 cm、10 cm×10 cm、20 cm×20 cm和30 cm×30 cm射野及2例临床调强计划的三维水模剂量分布的蒙特卡洛模拟结果,将其与水箱测量结果或医科达Monaco计划系统结果比较,以验证基于虚拟源的蒙特卡洛剂量计算的准确性。结果对5个射野下的水箱中心轴的百分深度剂量曲线以及不同深度的离轴剂量曲线,蒙特卡洛模拟结果与测量结果相差在1％以内。对2例临床调强计划, Monaco计算结果与蒙特卡洛模拟结果的三维通过率分别为98．9％和99．4％（3％／3 mm）,95．1％和95．4％（2％／2 mm）。结论基于虚拟源模型的蒙特卡洛模拟能得到准确的放疗剂量计算结果。     

Radiotherapy quality insurance by individualized in vivo dosimetry: State of the art

The quality insurance in radiotherapy in the frame of highly complex technical process as Intensity modulated radiotherapy (IMRT) needs independent control of the delivered dose to the patient. Actually, up to now, most of the radiotherapy treatments rely only on computed dosimetry through a rather complicated series of linked simulation tool. This dosimetry approach requires also qualified treatment means based on cautious quality insurance procedures. However, erroneous parameters could be difficult to detect and systematical errors could happen leading to radiotherapy accidents. In this context, in vivo dosimetry has a critical role of final control of the delivered dose. As many beam incidences and ports are used for any photontherapy treatment, external control could be very tedious and time consuming. Therefore, innovations are needed for in vivo dosimetry to provide ergonomic and efficient tools for these controls. This paper presents a review of technologies and products that can be used for in vivo dosimetry. It proposes also a reflection on the concepts to develop future devices suitable for this purpose. The technical means with their physical principles are reviewed, the clinical experiences demonstrating the feasibility of new techniques are then summarized and finally, the early clinical use and its impact on clinical practice is review.     

内镜下钛夹标记在早期食管癌精确放疗中应用

目的 介绍内镜下钛夹标记病变上、下界位置在早期食管癌精确放疗中的应用。方法 17例早期食管癌患者,通过内镜下钛夹标记病变上、下界,再行CT模拟定位勾画大体肿瘤体积(GTV),评估该项技术在精确放疗中的应用价值。结果 所有患者(包括13例初治、4例经内镜下切除后需放疗者)在CT和常规钡餐造影片上肿瘤显示不明显,无法准确确定GTV范围,内镜下在病变上下界均成功置入至少一个钛夹,标记肿瘤位置,在模拟定位CT上都能成功显示,并以此为依据准确勾画靶区,且未发生出血穿孔等3-4级不良反应。结论 早期食管癌在内镜下置入钛夹定位简便易行,对放疗靶区的确定有重要的临床价值。     

Virtual Reality Simulation in Ear Microsurgery: A Pilot Study

Classically, teaching of otological micro procedures is achieved by cadaveric dissection of the temporal bones, achieved by attending expensive temporal bone dissection courses. The difficulty in acquiring cadaveric specimens and the cost of courses has led to the development of alternative techniques. The aim of this study is to evaluate the efficacy of using virtual reality temporal bone simulation in training otolaryngology residents for mastoidectomy procedures. Four senior residents were evaluated during six canal wall down mastoidectomy procedures. Three of these procedures were done conventionally. The other three procedures were preceded by virtual reality simulation 1 day before the operation. In simulated cases, the residents scored higher on the global rating scale and task-based checklist, and were faster and more confident. Further, fewer instructions were required during the actual operation. This outcome is encouraging for the utilization of virtual reality simulation in otolaryngology residency training. Further studies should be performed on a larger number of simulators and surgical procedures to validate our results.