Prognostic Risk Factors And Nomogram Model Construction Of Polymyositis/Dermatomyositis Associated Interstitial Lung Disease

BackgroundPolymyositis(PM)and dermatomyositis(DM)is a group of immune-mediated heterogeneous diseases,which may involve multiple organs.Interstitial lung disease(ILD)is the most prevalent in lung diseases.ILD contributes to major mortality in PM/DM patients.However,the prognosis and treatment response are highly variable among PM/DM-ILD.Therefore,it is essential to identify the risk factors and establish a prognosis model.Currently,several different risk prediction models concerning PM/DM-associated ILD had been proposed.They are the ILD-GAP model,FLIAR model and MCK model.However,the above models have their limitations.ILD-GAP model was less effective in myositis associated ILD.FLIAR model scores were intricate to calculate,which may require professional imaging experts to evaluate.Besides,further research ought to be carried out to test the effectiveness of MCK model.ObjectiveTo verify the MCK model,explore the prognostic risk factors of PM/DM-ILD and establish a prognostic model of nomogram.MethodsPatients who were diagnosed with PM/DM-ILD in the First Affiliated Hospital of Zhengzhou University during August 2017 to June 2020 were taken as the research objects.Taking time as cut-off point,patients who were hospitalized from August 2017 to September 2019 were included in the derivative group,and patients who were hospitalized from October 2019 to June 2020 were included in the validation group.Derivation group and validation group were compared with baseline characteristics and prognosis.In derivation group,patients were scored and grouped according to MCK model.Kaplan Meier survival curve and log-rank test were used to verify the effectiveness of MCK model.Univariate and treatment adjusted multivariate Cox regression analysis was used to screen the independent prognostic risk factors of PM/DM-ILD patients.Cut-off values of the above independent risk factors were calculated to convert them into binary variables.Kaplan Meier survival curve of each variable were generated to compare their survival rates.Then the above independent risk factors were incorporate into nomogram model.Patients in derivative group and validation group were scored and grouped according to nomogram model.Decision curve analysis,Kaplan Meier survival curve,Logrank test and C-index were carried out to verify the effectiveness of nomogram model internally and externally.Results①In derivation group,predicted forced vital capacity(FVC%),predicted forced expiratory volume in 1 second(FEV1%),krebs von den Lungen 6(KL-6),positive rate of anti-melanoma differentiation-associated gene 5(MDA5)antibody,application rates of immunosuppressant and immunoglobulin were lower than those in the validation group,while C-reactive protein(CRP),creatine kinase(CK),creatine kinase isoenzyme(CK-MB)and procalcitonin(PCT)were higher than those in the validation group.There was no significant difference in reminding indexes between the derivation group and validation group.②In the derivative group,MCK model could not distinguish the prognostic risk stratification of low risk,medium risk and high risk well(logrank test,P>0.05).③Elder,lower ejection fraction(EF),higher alanine transaminase(ALT),lower albumin(ALB),higher alkaline phosphatase(ALP),higher neuron-specific enolase(NSE)and negative anti-aminoacyl tRNA synthetase(anti-ARS antibody)were independent risk factors for prognosis of PM/DM-ILD patients.④Nomogram model based on them distinguish low,moderate or high risk of mortality well in derivative group(Logrank test,P<0.01;C-index=0.898)and verification group(Logrank test,P<0.01;C-index=0.766).Conclusions1.Elder,lower EF,higher ALT,lower ALB,higher ALP,higher NSE and negative anti-ARS antibodies were independent prognostic risk factors for PM/DM-ILD patients.2.5AEN nomogram model based on age,EF,ALT,ALB,ALP,NSE and anti-ARS antibodies may help to predict the prognosis of PM/DM-ILD patients.