Background:Hepatocellular carcinoma(HCC)is the most common cause of cancer-related death in Saudi Arabia.Our study aimed to investigate the patterns of HCC and the effect of TNM staging,Alfa-fetoprotein(AFP),and Child...Background:Hepatocellular carcinoma(HCC)is the most common cause of cancer-related death in Saudi Arabia.Our study aimed to investigate the patterns of HCC and the effect of TNM staging,Alfa-fetoprotein(AFP),and Child-Turcotte Pugh(CTP)on patients’overall survival(OS).Methods:A retrospective analysis was conducted on 43 HCC patients at a single oncology center in Saudi Arabia from 2015 to 2020.All patients had to fulfill one of the following criteria:(a)a liver lesion reported as definitive HCC on dynamic imaging and/or(b)a biopsy-confirmed diagnosis.Results:The mean patient age of all HCC cases was 66.8 with a male-to-female ratio of 3.3:1.All patients were stratified into two groups:viral HCC(n=22,51%)and non-viral HCC(n=21,49%).Among viral-HCC patients,55%were due to HBV and 45%due to HCV.Cirrhosis was diagnosed in 79%of cases.Age and sex did not significantly statistically differ in OS among viral and non-viral HCC patients(p-value>0.05).About 65%of patients had tumor size>5 cm during the diagnosis,with a significant statistical difference in OS(p-value=0.027).AFP was>400 ng/ml in 45%of the patients.There was a statistically significant difference in the OS in terms of AFP levels(p-value=0.021).A statistically significant difference was also observed between the CTP score and OS(p-value=0.02).CTP class B had the longest survival.BSC was the most common treatment provided to HCC patients followed by sorafenib therapy.There was a significant statistical difference in OS among viral and non-viral HCC patients(p-value=0.008).Conclusions:The most common predictors for OS were the underlying cause of HCC,AFP,and tumor size.Being having non-viral etiology,a tumor size>5 cm,an AFP>400 ng/mL,and a CTP score class C were all negatively associated with OS.展开更多
心电图(electrocardiogram,ECG)异常的自动检测是一个典型的多标签分类问题,训练分类器需要大量有高质量标签的样本.但心电数据集异常标签经常缺失或错误,如何清洗弱标签得到干净的心电数据集是一个亟待解决的问题.在一个标签完整且准...心电图(electrocardiogram,ECG)异常的自动检测是一个典型的多标签分类问题,训练分类器需要大量有高质量标签的样本.但心电数据集异常标签经常缺失或错误,如何清洗弱标签得到干净的心电数据集是一个亟待解决的问题.在一个标签完整且准确的示例数据集辅助下,提出一种基于异常特征模式(abnormality-feature pattern,AFP)的方法对弱标签心电数据进行标签清洗,以获取所有正确的异常标签.清洗分2个阶段,即基于聚类的规则构造和基于迭代的标签清洗.在第1阶段,通过狄利克雷过程混合模型(Dirichlet process mixture model,DPMM)聚类,识别每个异常标签对应的不同特征模式,进而构建异常发现规则、排除规则和1组二分类器.在第2阶段,根据发现和排除规则辨识初始相关标签集,然后根据二分类器迭代扩展相关标签并排除不相关标签.AFP方法捕捉了示例数据集和弱标签数据集的共享特征模式,既应用了人的知识,又充分利用了正确标记的标签;同时,渐进地去除错误标签和填补缺失标签,保证了标签清洗的可靠性.真实和模拟数据集上的实验证明了AFP方法的有效性.展开更多
文摘Background:Hepatocellular carcinoma(HCC)is the most common cause of cancer-related death in Saudi Arabia.Our study aimed to investigate the patterns of HCC and the effect of TNM staging,Alfa-fetoprotein(AFP),and Child-Turcotte Pugh(CTP)on patients’overall survival(OS).Methods:A retrospective analysis was conducted on 43 HCC patients at a single oncology center in Saudi Arabia from 2015 to 2020.All patients had to fulfill one of the following criteria:(a)a liver lesion reported as definitive HCC on dynamic imaging and/or(b)a biopsy-confirmed diagnosis.Results:The mean patient age of all HCC cases was 66.8 with a male-to-female ratio of 3.3:1.All patients were stratified into two groups:viral HCC(n=22,51%)and non-viral HCC(n=21,49%).Among viral-HCC patients,55%were due to HBV and 45%due to HCV.Cirrhosis was diagnosed in 79%of cases.Age and sex did not significantly statistically differ in OS among viral and non-viral HCC patients(p-value>0.05).About 65%of patients had tumor size>5 cm during the diagnosis,with a significant statistical difference in OS(p-value=0.027).AFP was>400 ng/ml in 45%of the patients.There was a statistically significant difference in the OS in terms of AFP levels(p-value=0.021).A statistically significant difference was also observed between the CTP score and OS(p-value=0.02).CTP class B had the longest survival.BSC was the most common treatment provided to HCC patients followed by sorafenib therapy.There was a significant statistical difference in OS among viral and non-viral HCC patients(p-value=0.008).Conclusions:The most common predictors for OS were the underlying cause of HCC,AFP,and tumor size.Being having non-viral etiology,a tumor size>5 cm,an AFP>400 ng/mL,and a CTP score class C were all negatively associated with OS.
文摘心电图(electrocardiogram,ECG)异常的自动检测是一个典型的多标签分类问题,训练分类器需要大量有高质量标签的样本.但心电数据集异常标签经常缺失或错误,如何清洗弱标签得到干净的心电数据集是一个亟待解决的问题.在一个标签完整且准确的示例数据集辅助下,提出一种基于异常特征模式(abnormality-feature pattern,AFP)的方法对弱标签心电数据进行标签清洗,以获取所有正确的异常标签.清洗分2个阶段,即基于聚类的规则构造和基于迭代的标签清洗.在第1阶段,通过狄利克雷过程混合模型(Dirichlet process mixture model,DPMM)聚类,识别每个异常标签对应的不同特征模式,进而构建异常发现规则、排除规则和1组二分类器.在第2阶段,根据发现和排除规则辨识初始相关标签集,然后根据二分类器迭代扩展相关标签并排除不相关标签.AFP方法捕捉了示例数据集和弱标签数据集的共享特征模式,既应用了人的知识,又充分利用了正确标记的标签;同时,渐进地去除错误标签和填补缺失标签,保证了标签清洗的可靠性.真实和模拟数据集上的实验证明了AFP方法的有效性.
