4.3 Implications
Inflammation is a pathological stage of many diseases. Inflammation manifests in a variety of disease forms, and the currently selected therapeutic drugs are mainly non-steroidal anti-inflammatory drugs, glucocorticoids and other drugs. However, the application of anti-inflammatory effects of plant monomers is relatively limited. In addition, animal experiments to study the anti-inflammatory effect of monomers lack the results of effective clinical transformation, and there is also a lack of corresponding preclinical data collection. This article comprehensively utilizes the research on the treatment of inflammation with icariin, conducts meta-analysis based on its outcome index data to explore the curative effect, machine learning was then used to mine for more effective doses and courses of Icariin and its derivatives for the treatment of inflammation, and the types of inflammatory diseases for which they have an efficacy advantage. The results showed that icariin and its derivatives could improve TNF-α, IL-1β, IL-6, IFN-γ, TGF-β1, IkB-α, NF-κB p65, NLRP3, PPARα, SOD, MDA, BCL -2 and caspase-1 levels. And icariin and its derivatives have more favorable effect of inhibiting inflammation in respiratory, urinary, neurological and digestive diseases and at a dose of 27.52 mg/kg/day or more and a course of treatment of 31.22 days or more.
Animal experiments have always been an integral part of basic medical research, but laboratory animal research has always been subjected to ethical tests(Meier & Stocker, 1989). With the increase in the number of animal experiments and the expansion of the scale of a single experiment, the ethical review of the use of animals in experiments has become more stringent(Gruber & Hartung, 2004). How to efficiently select appropriate experimental animal models, treatment regimens and dosages of reagents, and make better use of animal experimental results, has gradually become an important issue(Loeb, Hendee, Smith, & Schwartz, 1989). Meta-analysis of animals can integrate published experimental results, save experimental resources, and obtain higher-level and more accurate preclinical evidence. At the same time, machine learning can be used to screen out features that are more important for drug treatment of diseases, so as to provide a reference for the selection of sample features for more in-depth animal experiments.
According to research, many drugs have shown good curative effects in the laboratory for treating diseases, but their clinical transformation rate is still low(van der Worp & Sandercock, 2012). Although there are many common mechanisms discovered in the laboratory, there are still differences in the organ structure between animals and humans, which cannot be ignored. Therefore, it is necessary to use machine learning to mine key features and select an appropriate animal model thereby increasing the clinical translation rate.