基于影像基因组学的完全性肺静脉异位引流的精准分型研究

肺静脉狭窄严重影响完全性肺静脉异位引流（TAPVC）手术预后，与TAPVC解剖异质性强和缺乏精准分型有关。近年，影像基因组学逐渐应用于心血管疾病的智能诊治。我们前期基于灰度Hessian-传递函数和Graphcut算法实现肺静脉左心房的分割和重建；对部分TAPVC患儿的高通量测序获得325353个单核苷酸多态性变化（SNP），并通过斑马鱼模型发现SNAI1、HMGA2和VAV2突变导致静脉内皮出芽功能障碍。由此假设，基于影像基因组学可挖掘生物学解释性强的特异影像组学标签，精准分型TAPVC。为证实假设，本项目拟（1）基于影像组学挖掘表征TAPVC解剖形态异质性的定量形态特征；（2）通过全基因组测序，获得SNP等相关基因组学特征；（3）融合影像和基因组学特征，通过聚类分析和机器学习等方法以精准分型TAPVC。本项目为TAPVC的智能诊断提供理论基础，对复杂先心病的定量、精准化诊疗具有一定意义

Owing to a wide spectrum of morphological variation of total anomalous pulmonary venous connection (TAPVC) and lack of precise evaluation of anatomic classification, pulmonary vein stenosis remains a vexing problem after surgical repair of TAPVC, which is associated with late morbidity and mortality. In recent years, intelligent diagnosis based on radiogenomics has become popular in cardiovascular disease. We have segmented the pulmonary vein (PV) as well as left atrium (LA) based on grayscale hessian-transfer function. Subsequently, we reconstruct PV and LA in a 3D visualization using Graphcut calculation. Meanwhile, we have finished whole genome sequencing and gained 32535 single nucleotide polymorphism (SNP). And we find SNAI1, HMGA2 and VAV2 mutation which are associated with abnormal pulmonary vein development in the zebrafish model. Therefore, we hypothesize that radiogenomics can help make a precise classification of the morphology of TAPVC. To verify the hypothesis, this study is intended to (1) quantify the specific anatomic morphology based on radiomics, (2) attain the genomic features such as SNP based on whole genome sequencing, and (3) obtain the fusion of the radiomic and genomic features using cluster analysis and machine learning technique to achieve the specific classification of TAPVC variation. This research provides theoretical foundation and application perspective in intelligent diagnosis of TAPVC. Moreover, it provides a new sight in the clinical application of radiogenomics, which helps to achieve precision medicine in the complicated congenital heart disease.