冠心病法医病理学诊断模型的构建

冠心病（coronary artery heart disease,CHD）是造成法医心源性猝死（(sudden cardiac death，SCD)）的主要原因。病理学诊断在实际工作中已经凸显一定局限性，虚拟尸检、微创尸检也存在误差和难以向基层推广，造成鉴定困难和司法争议。因此，如何利用非侵入性的手段获得尸体数据，并与计算机深度学习方法结合判断死亡原因是法医病理学今后努力的研究方向。最新研究发现，极限梯度增强（XGBoost）的决策系统在很多自动化诊断系统开发上凸显优势，可构建快速准确的诊断工具并建立疾病预测模型推测患者发生某种疾病的可能。本研究拟采用目前先进的利用机器学习中Xgboost算法，通过对法医病理尸检数据、蛋白数据库筛选的生物标记物、危险因素等数据进行训练，以病理学检测结果作为金标准，建立精准、高特异性、高分辨性的诊断模型，冠心病的死因鉴定、发病机制及预测有一定的实际意义。

Coronary artery heart disease (CHD) is the main cause of sudden cardiac death (SCD).Pathological diagnosis has highlighted certain limitations in practical work, virtual autopsy and minimally invasive autopsy also have errors and it is difficult to promote to the grass-roots level, resulting in identification difficulties and judicial disputes.Therefore, how to obtain cadaver data by non-invasive means and combine it with computer deep learning method to determine the cause of death is the future research direction of forensic pathology.Recent research has found that extreme gradient enhanced (XGBoost) decision systems have advantages in the development of many automated diagnostic systems, which can build rapid and accurate diagnostic tools and disease prediction models to predict the likelihood of patients developing certain diseases.This study intends to adopt the advanced Xgboost in the use of machine learning algorithm, based on the forensic pathological autopsy data, a database of protein screening biomarkers, risk factors, such as data for training, testing results of pathology as the gold standard, establish precise and diagnosis model of high specificity, high resolution, the cause of coronary heart disease death identification, pathogenesis and prediction has certain practical significance.