机器学习模型用于妊娠糖尿病发病风险预测及干预对策分析

目的 采用新型机器学习模型用于预测妊娠糖尿病（GDM）发病风险，为采取相应干预措施提供依据。方法 纳入2019年10月至2021年11月南通大学附属海安市人民医院收治的孕妇412例，依据孕妇是否并发GDM予以分组，分成GDM组（n=115）与正常组（n=297）。随后，随机将整个患者队列分配至独立训练集与独立测试集。在训练集中，采用单变量分析与多变量Logistic回归分析影响孕妇并发GDM的危险因素，并采用随机森林算法建立机器学习模型和引入独立测试集，采用受试者操作特征（ROC）曲线、校准曲线、决策曲线分析交叉验证机器学习模型的准确度与稳健性。结果 机器学习模型具有较好的区分度，训练集中曲线下面积（AUC）为0.881，测试集中AUC为0.846。对模型的解释和分析表明，年龄、孕前体质量指数（BMI）、胎次、空腹血糖、糖化血红蛋白、B细胞比例、甘油三酯水平显著影响GDM的发生风险（P<0.05）。结论 机器学习模型具有较好的准确度和稳健性，能够预测孕妇并发GDM的风险，辅助临床医护人员开展针对性的干预，从而降低GDM的发生率并改善预后。     

基于增强CT门静脉期影像组学模型对局部进展期结直肠癌新辅助化疗疗效的预测研究

目的 探讨基于CT图像的影像组学模型预测局部进展期结直肠癌（LACRC）患者新辅助化疗（NAC）疗效的潜在价值。方法 回顾性分析。纳入2014年1月—; 2019年9月山西省肿瘤医院接受手术治疗、术前均行NAC的局部进展期结直肠腺癌患者181例,其中男96例、女85例,年龄23~85岁。181例患者按7∶; 3的比例随机分为训练集（127例）、验证集（54例）,依据肿瘤退缩分级（TRG）标准分为疗效反应良好组（TRG 0~1级,81例）、反应不良组（TRG 2~3级,100例）。所有患者在治疗前均行增强CT检查。提取门脉期CT图像的1 037个影像组学特征,通过以最小绝对收缩与选择算子算法（LASSO）为主的四步法进行特征降维,然后采用多因素logistic回归对筛选出的特征构建影像组学模型;在训练集中,通过单因素及多因素logistic回归筛选预测LACRC患者NAC疗效的临床病理独立危险因素并构建临床模型;联合临床病理独立危险因素及影像组学特征构建融合模型并绘制列线图。绘制受试者工作特征曲线（ROC曲线）、校正曲线和决策曲线分析（DCA）,评估各模型对LACRC患者NAC疗效的预测性能、校准性能及其临床效益。结果 验证集患者的年龄大于训练集,差异有统计学意义（Z=-3.47,P<; 0.05）;两组患者性别分布,肿瘤临床T分期、N分期、病理分化程度,以及TRG级别等基线资料比较,差异均无统计学意义（P值均>; 0.05）。训练集中,疗效反应良好组（57例）和反应不良组（70例）患者的性别以及肿瘤临床T分期、N分期、病理分化程度的差异均有统计学意义（P值均<; 0.05）;验证集中的反应良好组（24例）和反应不良组（30例）患者的肿瘤临床T分期、N分期的差异均有统计学意义（P值均<; 0.05）。基于门脉期CT图像降维选择后共得到4个关键影像组学特征（P值均<; 0.05）,用于构建影像组学模型。临床模型包括临床T分期和病理分化程度2个独立危险因素（P<; 0.05）。影像组学模型、临床模型和融合模型在训练集的ROC曲线下面积（AUC）分别为0.822、0.702、0.850,验证集对应的AUC分别为0.757、0.706、0.824。校正曲线分析显示,影像组学模型和融合模型均有良好的校准性能。DCA 曲线分析显示,3种预测模型均有一定的临床效益,其中融合模型净收益值最大。结论 基于增强CT图像的影像组学特征结合相关临床因素构建的融合模型在预测LACRC患者NAC疗效方面有一定的价值。     

基于多参数MR的影像组学融合模型术前预测宫颈鳞癌脉管浸润的应用价值

目的 探讨基于多参数MR的影像组学融合模型术前预测宫颈鳞癌脉管间隙浸润（LVSI）的应用价值。方法 回顾性研究。纳入2016年6月—2019年3月山西省肿瘤医院宫颈鳞癌患者168例。患者年龄22~76（52.0±10.1）岁,临床分期为国际妇产联盟（FIGO）ⅠB期127例、ⅡA期41例。所有患者术前行多参数盆腔MR扫描,均接受根治性子宫切除术联合盆腔淋巴结清扫术治疗。收集其临床病理资料和多参数MRI数据,以7∶3的比例按照随机抽样法分为训练集117例和验证集51例。在T₂加权像（T₂WI）、表观弥散系数[ADC,由2个b值的弥散加权成像数据自动生成]及增强T₁加权像（cT₁WI）3个序列的MRI上,对病灶进行手动分割勾画肿瘤轮廓感兴趣区（ROI）,得到三维感兴趣区（VOI）并提取特征,通过以最大相关最小冗余和最小绝对收缩与选择算子回归为主的三步降维法筛选特征并构建影像组学模型。多因素logistic回归分析筛选临床特征并联合影像组学模型建立融合模型,制作列线图。受试者操作特征曲线（ROC 曲线）、校正曲线、决策分析曲线评估列线图的效能及临床效益。结果 术后病理检查确诊LVSI阳性42例,阴性126例。训练集与验证集患者的年龄、FIGO分期、肿瘤最大径、肿瘤分化程度、LVSI状态等临床病理特征比较,差异均无统计学意义（P值均>0.05）。基于T₂WI、ADC及cT₁WI多参数MRI提取的影像组学特征,经特征筛选后得到7个关键特征,均与宫颈癌LVSI相关（P值均<0.05）,并构建影像组学模型。训练集T₂WI、ADC及cT₁WI 3个序列独立构建的影像组学模型预测宫颈癌LVSI的ROC曲线下面积（AUC）分别为0.630[95%可信区间（CI）0.557~0.698]、0.686（95%CI 0.563~0.694）、0.761（95%CI 0.702~0.818）,3个序列共同构建的联合影像组学模型对应的AUC为0.887（95%CI 0.842~0.925）,诊断效能最优,并在验证集中得到验证。联合影像组学模型与肿瘤分化程度构建的融合模型列线图预测宫颈癌LVSI,在训练集与验证集中的AUC分别为0.893（95%CI 0.851~0.929）、0.854（95%CI 0.749~0.943）,校正曲线显示出列线图有良好的校正性能;决策曲线表明当风险阈值概率范围在0.50~0.96时,采用影像组学融合模型预测宫颈癌LVSI的净收益优于“将所有患者视为宫颈癌LVSI阳性或阴性”。结论 基于多参数MRI影像组学特征与临床特征的融合模型对宫颈癌LVSI状态有良好的预测价值。     

幕上高血压脑出血预后影响因素的多重线性回归分析

目的探讨影响幕上高血压脑出血（SICH）患者预后的相关因素,以指导临床治疗和评估预后。方法回顾性分析符合本研究纳入标准的幕上高血压脑出血324例完整病历。以一般资料、起病症状、入院查体、影像学资料、治疗方式、并发症等43项为自变量,以发病后1个月后功能独立性评定评分（functional independence measure,FIM）为因变量,建立多重线性回归模型,筛选出对预后有影响的因素,并比较各因素的影响大小。结果经统计学处理发现血肿体积、入院时收缩期血压、GCS评分、脑室是否积血、血肿体积扩大、是否并发肺部感染和应激性溃疡等7项对预后有显著性的影响。结论血肿体积、入院GCS评分和脑室是否积血对预后影响有重要意义,可作为SICH患者预后的关键性指标。     

T1WI增强纹理参数联合临床病理学特征对脑胶质瘤患者术后1年内复发的预测价值研究

目的 探讨T₁WI增强纹理参数联合临床病理学特征对脑胶质瘤患者术后1年内复发的预测价值。方法 回顾性队列研究。纳入2018年1月—2021年2月川北医学院附属医院及重庆市人民医院经术后病理确诊的脑胶质瘤患者90例。其中男60例、女30例,年龄17~76（47±14）岁,术后1年内复发46例、未复发44例。按照2∶1比例将患者随机分为训练集与验证集。采用MaZda软件于术前MR T₁WI增强图像上提取胶质瘤纹理参数。在训练集中对纹理参数进行降维筛选并利用最小绝对收缩和选择算子（LASSO）算法分析建立预后纹理积分模型。同时在训练集中对术后1年内复发与未复发患者的临床病理学特征及纹理积分进行影响因素分析,构建可视化联合预测模型。观察项目：（1）分析影响脑胶质瘤术后1年内复发的危险因素,并建立联合预测模型。（2）在验证集中,对纹理积分模型及联合预测模型进行验证。采用受试者工作特征（ROC）曲线、校准曲线及决策曲线分析法评价模型的诊断效能与临床净获益。结果 单因素分析显示,肿瘤WHO分级、异柠檬酸脱氢酶-1突变、术后放化疗、纹理积分是脑胶质瘤术后1年内复发的影响因素。多因素分析结果显示,WHO分级、异柠檬酸脱氢酶-1突变、术后放化疗及纹理积分均为脑胶质瘤术后1年内复发的独立危险因素[风险比（OR）=6.527、0.160、0.052、6.300,95％可信区间（CI）=1.201~35.485、0.031~0.827、0.004~0.708、1.905~20.841,P=0.030、0.029、0.026、0.003]。联合预测模型预测脑胶质瘤术后1年内复发的效能高于纹理积分模型,其训练集与验证集的曲线下面积（AUC）分别为0.92（95％CI 0.86~0.99）和0.86（95％CI 0.74~0.99）,灵敏度分别为0.90、0.88,特异度分别为0.83、0.79;纹理积分模型在训练集与验证集中的预测AUC分别为0.85（95％CI 0.75~0.95）和0.82（95％CI 0.67~0.97）,灵敏度分别为0.73、0.63,特异度分别为0.90、0.99。在两种模型中,预测概率与实际概率间的一致性及临床净获益均较高。结论 基于术前T₁WI增强图像的纹理参数联合临床病理学特征建立的联合预测模型对脑胶质瘤全切术后1年内的复发具有一定的预测价值。     

SICH量表对自发性脑出血短期临床结局的预测作用

目的：探讨影响脑出血短期预后的相关危险因素,制订一个基于量表的自发性脑出血危重程度评价模式,并评估其对自发性脑出血短期临床结局的预测效能。方法：回顾性分析我院从2011年6月至2014年12月收治的自发性脑出血患者的临床资料。将可能影响短期预后的危险因素做单因素分析及Logistic回归分析,根据相关危险因素制定SICH量表,得出其预测自发性脑出血患者短期临床结局的预测效能。结果：年龄、GCS评分、血肿位置、血肿体积、收缩压范围、血肿破入脑室、合并症为自发性脑出血预后不良的独立预后危险因素,根据以上因素制订SICH量表,评分范围从0~8分。SICH量表诊断不良结局的最佳诊断界点是≥4,由此得出的Youden指数分别是0.61,ROC曲线下面积为0.88(95% CI：0.85~0.91);诊断较好结局的最佳诊断界点是≤2,由此得出的Youden指数分别是0.55,ROC曲线下面积为0.85(95% CI：0.81~0.89)。结论：SICH量表可以对ICH患者进行危重程度的评估,并对短期临床结局的预测诊断效能较好。     

基于一般线性模型的机器学习方法在BOLD-fMRI脑胶质瘤患者个体化运动功能区定位中的应用价值

目的 探讨基于一般线性模型（GLM）的机器学习方法在血氧水平依赖功能磁共振成像（BOLD-fMRI）定位脑胶质瘤患者个体化运动功能中的应用价值。方法 前瞻性研究。纳入2017年11月—2021年11月西安交通大学第一附属医院神经外科确诊为脑胶质瘤且病灶位于大脑运动功能区的38例患者作为机器学习模型的验证集（男25例、女13例,年龄24~69岁）,同期招募健康志愿者50例作为模型的训练集（男26例、女24例,年龄22~68岁）。采用独立成分分析法（ICA）,随机提取98例人类连接组计划（HCP）受试者的静息态功能核磁共振（rs-fMRI）特征。依据健康志愿者的rs-fMRI和基于任务的功能磁共振（tb-fMRI）的相关性,训练基于GLM的机器学习模型。观察项目：（1）采用Pearson相关系数（CC）分析比较GLM预测的激活与实际激活的相似度。（2）采用Dice系数（DC）作为模型预测效能的定量指标,比较GLM与ICA方法的预测效能。结果 （1）胶质瘤患者基于GLM的机器学习方法所预测的激活与实际tb-fMRI的功能激活相似度高[（89.47% （34/38）的患者CC值>0.30）]。（2）胶质瘤患者GLM预测任务态运动功能激活的效能,DC为0.34（0.27,0.42）,优于ICA方法的效能DC 0.26（0.16,0.30）,差异有统计学意义（Z=-3.88,P<0.001）;GLM在肿瘤半球的预测效能优于ICA方法,DC分别为0.36（0.17,0.48）和0.34（0.04,0.45）,差异有统计学意义（Z=-2.43,P=0.015）;2种方法在非肿瘤半球的预测效果均显著高于肿瘤半球（Z=-4.33、-3.59,P值均<0.001）。结论 基于GLM的机器学习方法能够很好地在术前利用rs-fMRI数据预测出胶质瘤患者的tb-fMRI运动功能激活,且其预测效果好于ICA方法。     

基于3D深度残差网络和多模态MRI的脑胶质瘤自动分级

目的:利用3D深度残差网络和多模态MRI实现对脑胶质瘤的自动分级。方法:利用BraTS2020公共数据集的293例高级别胶质瘤（HGG）和76例低级别胶质瘤（LGG）的多模态MRI数据训练和测试3D深度残差卷积网络模型。多模态MRI图像经过3D剪裁、重采样和归一化的预处理,随机分组为训练（64%）、验证（16%）和测试（20%）样本,将预处理后的多模态MRI图像和分级标注输入到网络模型进行训练、验证和测试。利用准确率（ACC）和受试者工作特征（ROC）曲线下面积（AUC）评价分级结果。结果:在59例（48例HGG和11例LGG）验证数据集上,ACC和AUC分别为0.93和0.97,在75例（62例HGG和13例LGG）测试数据集上,ACC和AUC分别为0.89和0.93。结论:3D深度残差网络在多模态MRI数据集上获得了较好的脑胶质瘤自动分级结果,可以为确定治疗方案和预测预后方面提供重要参考。     

联合免疫分型构建肺腺癌铁死亡相关长链非编码RNA风险评分模型

目的：依据肺腺癌（LUAD）患者铁死亡与长链非编码RNA(lncRNA)之间的相关性，结合免疫分型构建新型风险评分模型以评估LUAD患者的预后。方法：基于生物信息学技术下载TCGA数据库中LUAD样本转录组数据和临床数据，获取来自FerrDb数据库铁死亡相关的基因，通过“caret”包将整理后的504例LUAD样本随机分为50%训练集与50%验证集，采用Pearson相关性分析、单因素Cox回归得到与LUAD预后有关的铁死亡相关lncRNA，利用R软件的“ConensusClusterPlus”包和CIBERSORT软件进行免疫相关分析，LASSO回归分析建立铁死亡相关lncRNA模型，受试者工作特征（ROC）曲线及曲线下面积（AUC）评估预后模型的效能，并通过验证集进行验证。结果：单因素Cox及LASSO回归分析筛选出9个铁死亡相关lncRNA构成的风险评分模型。单因素以及多因素Cox回归分析均显示，本预后模型可以作为一个独立的预后因素（P<0.001），该模型在训练集、内部验证集及外部验证集中均具有较好的预测效能。结论：本研究构建的LUAD患者风险评分模型可作为一种新的独立预...     

基于密度分布特征的深度神经网络模型诊断COVID-19的价值

目的:评价基于密度分布特征（CDD）的深度神经网络（DNN）模型对新型冠状病毒肺炎（COVID-19）的诊断价值。方法:收集42例COVID-19病例和43例非COVID-19肺炎病例。将所有患者的211份胸部CT图像分为训练集（n=128）和验证集（n=83）。参考北美放射学会发布的COVID-19相关性肺炎的CT结构化报告，构建基于CT影像特征的DNN模型（DNN-CTIF）。根据胸部CT图像上肺炎CDD建立DNN-CDD模型。采用ROC曲线分析和决策曲线分析对两种模型进行评价。结果:DNN-CTIF模型的AUC在训练集为0.927，在验证集为0.829。DNN-CDD模型的AUC在训练集为0.965，在验证集为0.929。DNN-CDD模型在验证集的AUC高于DNN-CTIF模型（P=0.047）。决策曲线分析表明在0.04~1.00概率阈值范围内，DNN-CDD模型相比DNN-CTIF模型使患者的净获益更高。结论:DNN-CTIF和DNN-CDD模型对COVID-19均具有较好的诊断性能，其中DNN-CDD模型优于DNN-CTIF模型。     

A prediction model for asthma exacerbations after stopping asthma biologics

BackgroundLittle is known regarding the prediction of the risks of asthma exacerbation after stopping asthma biologics.ObjectiveTo develop and validate a predictive model for the risk of asthma exacerbations after stopping asthma biologics using machine learning models.MethodsWe identified 3057 people with asthma who stopped asthma biologics in the OptumLabs Database Warehouse and considered a wide range of demographic and clinical risk factors to predict subsequent outcomes. The primary outcome used to assess success after stopping was having no exacerbations in the 6 months after stopping the biologic. Elastic-net logistic regression (GLMnet), random forest, and gradient boosting machine models were used with 10-fold cross-validation within a development (80%) cohort and validation cohort (20%).ResultsThe mean age of the total cohort was 47.1 (SD, 17.1) years, 1859 (60.8%) were women, 2261 (74.0%) were White, and 1475 (48.3%) were in the Southern region of the United States. The elastic-net logistic regression model yielded an area under the curve (AUC) of 0.75 (95% confidence interval [CI], 0.71-0.78) in the development and an AUC of 0.72 in the validation cohort. The random forest model yielded an AUC of 0.75 (95% CI, 0.68-0.79) in the development cohort and an AUC of 0.72 in the validation cohort. The gradient boosting machine model yielded an AUC of 0.76 (95% CI, 0.72-0.80) in the development cohort and an AUC of 0.74 in the validation cohort.ConclusionOutcomes after stopping asthma biologics can be predicted with moderate accuracy using machine learning methods.     

Predicting the Availability of Hematopoietic Stem Cell Donors Using Machine Learning

Hematopoietic stem cell transplantation (HSCT) is firmly established as an important curative therapy for patients with hematologic malignancies and other blood disorders. Apart from finding HLA-matched donors during the HSCT process, donor availability remains a key consideration as the time taken from diagnosis to transplant is recognized to adversely affect patient outcome. In this study, we aimed to develop and validate a machine learning approach to predict the availability of stem cell donors. We retrospectively collected a data set containing 10,258 verification typing requests made during the HSCT process in the British Bone Marrow Registry (BBMR) between January 1, 2013, and December 31, 2018. Three machine learning algorithms were implemented and compared, including boosted decision trees (BDTs), logistic regression, and support vector machines. Area under the receiver operating characteristic curve (AUC) was primarily used to assess the algorithms. The experimental results showed that BDTs performed better in predicting the availability of BBMR donors. The overall predictive power of the model, using AUC on the test cohort of 2052 records, was found to be 0.826. Our findings show that machine learning can predict the availability of donors with a high degree of accuracy. We propose the use of the BDT machine learning approach to predict the availability of BBMR donors and use the predictive scores during the HSCT process to ensure patients with blood cancers or disorders receive a transplant at the optimum time.     

Machine learning of clinical performance in a pancreatic cancer database

ObjectiveWe consider predictive models for clinical performance of pancreatic cancer patients based on machine learning techniques. The predictive performance of machine learning is compared with that of the linear and logistic regression techniques that dominate the medical oncology literature.Methods and materialsWe construct predictive models over a clinical database that we have developed for the University of Massachusetts Memorial Hospital in Worcester, Massachusetts, USA. The database contains retrospective records of 91 patient treatments for pancreatic tumors. Classification and regression targets include patient survival time, Eastern Cooperative Oncology Group (ECOG) quality of life scores, surgical outcomes, and tumor characteristics. The predictive performance of several techniques is described, and specific models are presented.ResultsWe show that machine learning techniques attain a predictive performance that is as good as, or better than, that of linear and logistic regression, for target attributes that include tumor N and T stage, survival time, and ECOG quality of life scores. Bayesian techniques are found to provide the best performance overall. For tumor size as the target attribute, however, logistic regression (respectively linear regression in the case of a numerical as opposed to discrete target) performs best. Preprocessing in the form of attribute selection and supervised attribute discretization improves predictive performance for most of the predictive techniques and target attributes considered.ConclusionMachine learning provides techniques for improved prediction of clinical performance. These techniques therefore merit consideration as valuable alternatives to traditional multivariate regression techniques in clinical medical studies.     

Osteoporosis Risk Prediction for Bone Mineral Density Assessment of Postmenopausal Women Using Machine Learning

Purpose

A number of clinical decision tools for osteoporosis risk assessment have been developed to select postmenopausal women for the measurement of bone mineral density. We developed and validated machine learning models with the aim of more accurately identifying the risk of osteoporosis in postmenopausal women compared to the ability of conventional clinical decision tools.

Materials and Methods

We collected medical records from Korean postmenopausal women based on the Korea National Health and Nutrition Examination Surveys. The training data set was used to construct models based on popular machine learning algorithms such as support vector machines (SVM), random forests, artificial neural networks (ANN), and logistic regression (LR) based on simple surveys. The machine learning models were compared to four conventional clinical decision tools: osteoporosis self-assessment tool (OST), osteoporosis risk assessment instrument (ORAI), simple calculated osteoporosis risk estimation (SCORE), and osteoporosis index of risk (OSIRIS).

Results

SVM had significantly better area under the curve (AUC) of the receiver operating characteristic than ANN, LR, OST, ORAI, SCORE, and OSIRIS for the training set. SVM predicted osteoporosis risk with an AUC of 0.827, accuracy of 76.7%, sensitivity of 77.8%, and specificity of 76.0% at total hip, femoral neck, or lumbar spine for the testing set. The significant factors selected by SVM were age, height, weight, body mass index, duration of menopause, duration of breast feeding, estrogen therapy, hyperlipidemia, hypertension, osteoarthritis, and diabetes mellitus.

Conclusion

Considering various predictors associated with low bone density, the machine learning methods may be effective tools for identifying postmenopausal women at high risk for osteoporosis.     

基于神经网络的川崎病并发冠状动脉病变预测模型

建立川崎病并发冠状动脉病变预测模型。从电子病历数据库中收集1 000例(343例患冠状动脉病变)川崎病患儿的人口学资料、实验室检验数据、超声心动图数据,对数据进行预处理后用关联规则筛选川崎病并发冠状动脉病变的危险指标,划分训练集和测试集分别为总样本集的70%和30%,分别建立神经网络模型和Logistic回归模型,并用灵敏度及特异性等指标对模型的预测效果予以评估。结果显示,神经网络模型的灵敏度=0.718,特异性=0.746,准确率=0.737及AUC(ROC曲线下面积)=0.796,优于Logistic回归模型[灵敏度=0.175,特异性=0.893,准确率为0.647及AUC=0.624]。研究结果表明,神经网络模型对川崎病并发冠状动脉病变的预测效果优于Logistic回归模型。     

基于多模态磁共振放射组学与临床指标的前列腺癌智能检测及风险预测模型建立

目的：利用多模态磁共振放射组学开发前列腺癌自动检测模型,并使用列线图构建多因素回归模型,将前列腺MRI放射组学特征与临床多个检测指标进行整合,从而对患前列腺癌风险性进行预测。方法：回顾性研究于2019年2月～2021年10月病理证实为前列腺癌和其他前列腺良性肿瘤的患者133例。所有病例均行前列腺直肠指检(DRE)、前列腺特异性抗原(PSA)、游离前列腺特异性抗原(F-PSA)、FPSA/PSA检测。治疗前多模态前列腺MRI图像(DWI+DCE+T₂WI)用于提取放射特征,最大相关最小冗余(m RMR)算法用于消除混杂变量,使用最小绝对收缩和选择算子(LASSO)逻辑回归进行放射特征选择。通过曲线下面积(AUC)、准确性、特异性、敏感性评估放射特征的诊断性能;通过多元logistic回归选择临床指标和放射组学特征模型来制定放射组学列线图,并使用校准曲线和Hosmer-lemeshow试验验证其可靠性。结果：两名观察者测量的所有数据ICC均在0.80以上。所有前列腺MRI图像随机分为训练组和验证组(7:3)。在训练组中,DWI、DCE和T₂WI的...     

Quantitative Detection of Cirrhosis: Towards the Development of Computer-Assisted Detection Method

There are distinct morphologic features of cirrhosis on CT examinations; however, such impressions may be subtle or subjective. The purpose of this study is to build a computer-aided diagnosis (CAD) method to help radiologists with this diagnosis. One hundred sixty-seven abdominal CT examinations were randomly divided into training (n = 88) and validation (n = 79) sets. Livers were analyzed for morphological markers of cirrhosis and logistic regression models were created. Using the area under curve (AUC) for model performance, the best model had 0.89 for the training set and 0.85 for the validation set. For radiology reports, sensitivity of reporting cirrhosis was 0.45 and specificity 0.99. Using the predictive model adjunctively, radiologists’ sensitivity increased to 0.63 and specificity slightly decreased to 0.97. This study demonstrates that quantifying morphological features in livers may be utilized for diagnosing cirrhosis and for developing a CAD method for it.     

基于18F-FDG PET/CT影像组学的非小细胞肺癌病理亚型分类

目的：旨在建立一种基于¹⁸F-FDG PET/CT的临床—影像组学相结合的综合模型用于区分非小细胞肺癌中的腺癌和鳞癌。方法：回顾性收集上海交通大学附属胸科医院120例经病理学验证为腺癌(65例)和鳞癌(55例)的患者,从预处理的CT图像和PET图像中分别提取1218、108个影像组学特征,并纳入10个临床特征因素;卡方检验和Wilcoxon检验用于对临床特征进行筛选,并使用Relief算法和最小绝对收缩和选择算子(LASSO)对影像组学特征进行筛选;通过6种机器学习分类器分别建立临床、影像组学、综合模型。通过受试者工作特征(ROC)曲线及曲线下面积(AUC)来评价模型的分类能力。结果：综合模型在训练集和测试集中均表现出最高的AUC值和准确率,其中随机森林(RF)和Bagging分类器表现出的分类效果最佳。经五折交叉验证后,训练集中RF和Bagging的AUC值和准确率分别为0.92±0.03、0.86±0.06和0.92±0.02、0.83±0.02;测试集中RF和Bagging的AUC值和准确率分别为0.92、0.81和0.91、0.86。结论：结合^1...     

Predicting three-year kidney graft survival in recipients with systemic lupus erythematosus

Predicting the outcome of kidney transplantation is important in optimizing transplantation parameters and modifying factors related to the recipient, donor, and transplant procedure. As patients with end-stage renal disease (ESRD) secondary to lupus nephropathy are generally younger than the typical ESRD patients and also seem to have inferior transplant outcome, developing an outcome prediction model in this patient category has high clinical relevance. The goal of this study was to compare methods of building prediction models of kidney transplant outcome that potentially can be useful for clinical decision support. We applied three well-known data mining methods (classification trees, logistic regression, and artificial neural networks) to the data describing recipients with systemic lupus erythematosus (SLE) in the US Renal Data System (USRDS) database. The 95% confidence interval (CI) of the area under the receiver-operator characteristic curves (AUC) was used to measure the discrimination ability of the prediction models. Two groups of predictors were selected to build the prediction models. Using input variables based on Weka (a open source machine learning software) supplemented with additional variables of known clinical relevance (38 total predictors), the logistic regression performed the best overall (AUC: 0.74, 95% CI: 0.72-0.77)-significantly better (p < 0.05) than the classification trees (AUC: 0.70, 95% CI: 0.67-0.72) but not significantly better (p = 0.218) than the artificial neural networks (AUC: 0.71, 95% CI: 0.69-0.73). The performance of the artificial neural networks was not significantly better than that of the classification trees (p = 0.693). Using the more parsimonious subset of variables (six variables), the logistic regression (AUC: 0.73, 95% CI: 0.71-0.75) did not perform significantly better than either the classification tree (AUC: 0.70, 95% CI: 0.68-0.73) or the artificial neural network (AUC: 0.73, 95% CI: 0.70-0.75) models. We generated several models predicting 3-year allograft survival in kidney transplant recipients with SLE that potentially can be used in practice. The performance of logistic regression and classification tree was not inferior to more complex artificial neural network. Prediction models may be used in clinical practice to identify patients at risk.