动态PET-CT/动态增强MRI双模分子影像在鼻咽癌治疗监控及预后的应用研究

The research of molecular imaging is of help to select the imaging biomarkers which found the base of individual targeted therapy of cancer. It has been shown that dynamic PET-CT, compared to conventional whole-body static PET-CT scan, more accurately reflect the metabolism of fluorodeoxyglucose (FDG). Dynamic contrast-enhanced MRI (DCE-MRI) also shows more encouraging future in evaluating the molecular targeted treatment. In previous study the applicants have justified the technical feasibility of these two imaging modalities in pre-treatment characterization of nasopharyngeal carcinoma (NPC), however no publications have reported the roles in treatment monitoring and prognosis. In this project we plan to:: (1) recruit a NPC cohort, for whom dynamic FDG PET-CT and DCE-MRI will be performed before and after treatment; (2) from the dynamic PET-CT images to calculate the parameters K1, k2, k3, k4 and Ki to study the glucose transport and metabolism, and from the DCE-MRI images to evaluate the functional parameters (Ktrans, kep and ve) to study the blood perfusion and permeability in primary tumor and metastatic lymph nodes; (3) to evaluate the relationship between the glucose metabolism parameters and perfusion parameters, and the correlation with tumor stage and epstein-barr virus (EBV) data of these NPC patients; (4) to calculate the change of these parameters after treatment, and then evaluate if these parameters or changes have significant impact on the treatment outcome using overall survival/progression free survival (OS/PFS) as the end points. From this project we expect to: (1) establish the systematic methodology of applying the two imaging techniques in NPC studies; (2) to study the correlation between glucose transport/metabolism and perfusion/permeability in NPC by comparing dynamic PET-CT and DCE-MRI; and (3) to evaluate and select the biomarkers which may predict the treatment monitoring and prognosis of NPC. This project will have great impact on the treatment monitoring and prognosis of NPC patients.

分子影像研究有助于筛选影像生物标志，是个体化靶向治疗癌症的基础。其中，动态PET-CT扫描较静态扫描可更准确反映脱氧葡萄糖FDG代谢。动态增强MRI（DCE-MRI）同样较普通MRI更具评估分子靶向治疗的潜力。两者的技术可行性及在鼻咽癌疗前评估中的作用在申请者前期研究中已获证实，而于疗效监控及预后方面尚无报道。本研究拟：1、进行治疗前后动态PET-CT扫描并改进数学模型以计算FDG的灌注、输运及新陈代谢指标；2、进行治疗前后DCE-MRI扫描以量化血管灌注及渗透性；3、研究两种技术相关性；4、评估原发病灶及转移淋巴结内各参数及其改变与疗效及病人存活率的关系。本研究的成功将：1、改进动态PET-CT模型以促进其在鼻咽癌中的临床应用；2、了解鼻咽癌内FDG代谢与血管灌注及渗透性的相互关系；3、有助于筛选预先反映治疗反应及疗效的影像生物标志。本研究将对鼻咽癌的治疗监控和预后有重要的临床促进作用。

项目背景.分子影像研究有助于筛选影像生物标志，是个体化靶向治疗癌症的基础。其中，动态PET-CT扫描较静态扫描可更准确反映脱氧葡萄糖FDG代谢。动态增强MRI（DCE-MRI）同样较普通MRI更具评估分子靶向治疗的潜力。两者的技术可行性及在鼻咽癌（NPC）疗前评估中的作用在申请者前期研究中已获证实，而于疗效监控及预后方面尚无报道。..主要研究内容.1）改进数学模型，编写程序，计算NPC的FDG的灌注参数；.2）探讨并筛选出能早期预测病人治疗反应及生存率的最佳影像生物标志；.3）基于PET-CT和DCE-MRI图像的NPC病灶的的自动分割；.4）改进heterogeneity的计算方法。..重要成果.1）改进数学模型以同时计算FDG的灌注、输运及新陈代谢速率指标，并计算了血管通透性等；筛选能早期预测病人治疗反应及生存率的影像生物标志（如肿瘤代谢体积；代谢指标；血管通透性等）。.2）国际、国内学术会议分别发表论文及讨论1次及3次；发表2篇EI收录国内核心期刊论文；发表3篇SCI论文；申请软件著作权2个。另外， 2篇SCI论文在审，2篇SCI论文在撰写。.3）参加培养1名博士研究生，培养1名硕士研究生；培养了一个癌症影像研究团队。..关键数据和科学意义.1）肿瘤代谢体积可以有效预测NPC预后（P=0.024），联合其他代谢则准确度达100%。在初期治疗效果差的病人中，血管通透性参数较高（P在0.05左右）。这些影像生物标志有望为个性化治疗方案及结果预测提供依据。.2）实现了NPC的半自动分割，准确度在80%以上，且在动态PET-CT的结果更好。我们应用深度学习技术，已经取得很好结果（准确度85%以上）。这一结果有望在临床方面提供参考，如放疗靶区的勾画、疗后效果的评估等。.3）基于机器学习技术改进heterogeneity的计算方法。这方面的研究有望为临床实践提供参考。