Page 158 - 南京医科大学自然版
P. 158

第45卷第12期
               ·1840 ·                           南 京    医 科 大 学 学         报                        2025年12月


             (续表2)
                    Algorithm                                Modeling indicator                       Reference
                Ten⁃by⁃tenfold nested Rituximab⁃related 40 genes:SHC3,XCR1,TCN1,DLX4,PLEKHG,etc.       [60]
                cross⁃validation  Tocilizumab⁃related 39 genes:SHC3,XCR1,DLX4,MYH6,TCN1,etc.
                Cross⁃validation  Clinical baseline data for monitoring the efficacy of drug treatment  [61]
                                                                        +
                XGBoost,RF        DNA methylation sequencing of peripheral blood CD4 T cells after drug treatment  [63]
                SVM               Clinical baseline data on the effectiveness of leflunomide treatment  [64]
                                  The efficacy of leflunomide treatment on peripheral blood differential DNA methylation sites:
                                  cg17330251,cg19814518,cg20124410,cg21109666,cg22572476,cg23403192,and
                                  cg24432675
                XGBoost           Whether RA will undergo surgery:patient’s clinical baseline data     [66]
                 RF:random forest;CNN:convolutional neural network;DCNN:deep convolutional neural network;ANN:artificial neural network;Bi⁃LSTM:bidi⁃
              rectional long short term memory;AM:additive model;SVM:support vector machine;SVM⁃RFE:support vector machine recursive feature elimination;
              SVM⁃RBF:support vector machine radial basis function;LASSO:least absolute shrinkage and selection operator;LSTM:long short term memory;
              MSResNet:multiscale residual network;GBT:gradient boosting decision tree;SSGB:stability selection gradient boosting;LGBM:light gradient boosting
              machine;XGBoost:eXtreme gradient boosting;glmBoost:generalized linear model boost;GMM:gaussian mixture model;ANFIS:adaptive neuron fuzzy
              inference system;RC:ridge regression;ENR:elastic net regression;KNN:K nearest neighbor;PLS:partial least squares regression;CC:consensus clus⁃
              ter plus;DEC:deep embedded cluster;LR:logistic regression;MA:multivariate analysis;MLR:multiple linear regression;SFS:sequential forward
              selection.







                                                Clinical baseline
                                                    data




                                                  Imaging
                                                                                     Diagnosis and
                                                  diagnosis
                                                                                       treatment
                                  RA
                                                                      ML
                                                  Molecular
                                                  detection
                                                                                      Drug efficacy
                                                                                       monitoring


                                                 Cell markers
                                            图1 ML辅助RA诊疗流程图(Office Plus绘制)
                            Figure 1  ML assisted RA diagnosis and treatment process diagram(drawn by Office Plus)


              练过程中融入广泛且多样化的现实世界数据,如                             开发,大幅提高了对不同数据的整合能力,并呈现
              病历、特定领域知识和多轮对话咨询,通过医学                             数据处理过程的详细解析。同时这种诊断模型是
              知识问答、医疗检查、患者咨询和病历诊断分析                             整个医院甚至整个社会公用的,这也就需要全社会
              等进行多维度验证,其诊断能力均显著优于其他                             建立公认算法,并加强对医护人员的持续培训。最
              模型  [71] 。虽然前述模型并未特别针对 RA 诊疗,                     后,加强多中心交互实践,汇集多家医院、诊疗中心
              但模型构建机制对于开发 RA 诊疗模型有很高的                           以及实验室的检测数据进行整合分析,建立可推
              借鉴价值。                                             广、普适性的诊疗模型,同期构建诊疗模型监管网
                  现阶段还需突破ML对于不同维度数据的整合                          络,杜绝人为篡改或擅用医疗数据,以便真正助力
              分析能力,伴随 AI,例如 DeepSeek 等相关大模型的                    诊疗,服务患者。
   153   154   155   156   157   158   159   160   161   162   163