Construction and validation of a nomogram prediction model for multiple myeloma complicated with nervous system damage

MI Qiuwen; LUO Lixiu; WU Yanling; WEN Xiaoying; ZHAO Yu; LIU Yang; CHEN Xiaomin; HUANG Chunlan

doi:10.3969/j.issn.1674-3806.2025.11.17

Abstract：Objective To construct and validate a nomogram prediction model for multiple myeloma(MM)complicated with nervous system damage.Methods A retrospective analysis was conducted on the clinical data of 202 patients with MM who were admitted to the Affiliated Hospital of Southwest Medical University from October 2019 to October 2023.The patients were divided into the occurrence group(86 patients)and the non-occurrence group(116 patients)according to the occurrence of nervous system damage during the follow-up period.Multivariate logistic regression was used to analyze the influencing factors of MM complicated with nervous system damage.A nomogram prediction model for MM complicated with nervous system damage was constructed.The predictive efficacy of the model for MM complicated with nervous system damage was evaluated using receiver operating characteristic(ROC)curve,calibration curve,Hosmer-Lemeshow goodness-of-fit test and decision curve analysis(DCA).Results The incidence of MM complicated with nervous system damage was 42.57％.The results of multivariate logistic regression analysis showed that older age,higher levels of lactate dehydrogenase(LDH)and immunoglobulin(Ig)G,and International Staging System(ISS)stage Ⅲ were independent risk factors for MM complicated with nervous system damage(P＜0.05),and longer activated partial thromboplastin time(APTT),longer prothrombin time(PT),and higher hemoglobin level were independent protective factors for MM complicated with nervous system damage(P＜0.05).Based on the above 7 influencing factors,a nomogram prediction model for MM complicated with nervous system damage was constructed.The results of ROC curve analysis showed that the area under the curve of the model for predicting MM complicated with nervous system damage was 0.917(95％CI:0.861-0.956),and the sensitivity and specificity were 95.35％and 87.85％,respectively.The Hosmer-Lemeshow goodness-of-fit test results(x2=2.944,P=0.872)proved that the model had an adequate goodness-of-fit.The calibration curve showed that the predicted probability of the model was close to the actual probability,indicating a good calibration.DCA showed that net clinical benefits could be achieved when the prediction threshold probability of the model was between 20％and 93％,or between 94％and 100％.Conclusion The nomogram prediction model which is constructed for MM complicated with nervous system damage based on 7 influencing factors including age,LDH,IgG,APTT,PT,hemoglobin and ISS stage has good predictive value.