Meta-analysis Of Immune Checkpoint Inhibitor Doses And Immune-related Adverse Events

The rapid development of Internet technology has witnessed the"big data"as one of the most popular industries,which is catching more and more people's attentions,and the medical-related big data has a close relationship with people's health.Dealing with the big data can predict the occurrence of diseases,improve medical programs,reduce the risk of medical treatment,and lower medical costs,thereby improve the health of patients and promote people's living standards.Immune checkpoint inhibitors are currently one of the most effective means to treat cancer.Although the method can effectively improve the survival rate of cancer patients,it can make patients with immune-related adverse events,and seriously affect the physical conditions of the patients.Thus this issue should be paid special attention.There is no method to predict the association between different doses of immune checkpoint inhibitors and immune-related adverse events.Thus,this dissertation proposed a complete meta-analysis model by considering the effect size and 95%confidence interval as important indicators.The meta-analysis model,proposed in this dissertation,consisted of the heterogeneity test model,fixed effect model,random effect model and publication bias test model.In this dissertation,the heterogeneity test model consisted of the Q test model and the I–squared test model.The heterogeneity is judged by the p value of the Q test model and the I~2 value of the I-squared test model.If the p-value of the Q test model is greater than 0.1,or the indicator I of the I-squared test is less than 50%,the fixed-effect model is used;otherwise,the random effect model is used.The fixed effect model and the random effect model are the core of the meta-analysis model,through which the indicator of the effect size OR and the indicator of the 95%confidence interval CI can be used to judge the correlation between different doses of immune checkpoint inhibitors and the immune-related adverse events.If the indicator OR is greater than 1 and all the values in the indicator CI is greater than 1,patients receiving high-dose immune checkpoint inhibitors are more prone to immune-related adverse events;if the indicator OR is less than 1 and all the values in the indicator CI are less than 1,patients receiving high-dose immune checkpoint inhibitors are less prone to immune-related adverse events;otherwise,they are not related.The experimental results can be intuitively given by the forest plots.The publication bias test model consists of the Begg test model and the Egger test model.The indicator p values,obtained from the Begg test model and the Egger test model,are used to judge publication bias in the meta-analysis.If both the p values of the Begg test model and the Egger test model are less than 0.05,no publication bias is found in the meta-analysis;otherwise there exists the publication bias in the meta-analysis.Finally,this meta-analysis model was used to predict the association between different doses of immune checkpoint inhibitors and immune-related adverse events.The experimental results showed that high-dose(10 mg/kg)immune checkpoint inhibitors were more likely to elevate the immune adverse events of aspartate aminotransferases,colitis and rashes,compared with other low doses of immune checkpoint inhibitors.