多发性骨髓瘤合并侵袭性真菌感染危险因素及其列线图预测模型构建

Risk factors of multiple myeloma complicated with IFI and construction of nomogram prediction model

目的探讨多发性骨髓瘤(MM)患者合并侵袭性真菌感染(IFI)的潜在危险因素,并据此构建一个预测模型。方法回顾性分析了2019年2月至2023年6月期间157例MM患者的临床资料。通过多因素logistic回归分析,识别出MM合并IFI的独立危险因素,并据此构建了一个列线图预测模型。随后,对模型的效能进行分析和验证。结果在157例患者中,有61例合并IFI,发生率为38.85%。G试验阳性、单核细胞绝对值降低以及高敏C反应蛋白(hs-CRP)水平≥10 mg/L是MM合并IFI的独立危险因素(P<0.05)。所构建的列线图预测模型的曲线下面积(AUC)为0.755(95%置信区间:0.699~0.851),显示出比单一因素更高的预测价值。通过Bootstrap法进行的内部验证和决策分析表明,该预测模型具有良好的效能和较高的临床应用价值。结论MM合并IFI的发生率相对较高。基于G试验、单核细胞绝对值和hs-CRP水平构建的预测模型具有良好的判别效度,可作为早期识别MM患者发生IFI的重要理论依据。

Objective To investigate the risk factors of multiple myeloma(MM)combined with invasive fungal infection(IFI),and to establish a prediction model based on this.Methods A retrospective analysis was carried out on the clinical data of 157 MM patients from February 2019 to June 2023.Independent risk factors for MM combined with IFI were identified through multiple logistic regression analysis,and a column chart prediction model was constructed based on this.Subsequently,the effectiveness of the model was analyzed and validated.Results Among 157 patients,61 cases were complicated with IFI,and incidence rate was 38.85%.Positive G test,decreased absolute value of monocytes and hs-CRP≥10 mg/L were independent risk factors for MM and IFI(P<0.05).The area under the curve(AUC)of the constructed column chart prediction model was 0.755(95%confidence interval:0.699–0.851),indicating a higher predictive value than a single factor.Internal validation and decision analysis through Bootstrap method indicated that the prediction model had good performance and high clinical application value.Conclusion The incidence of MM complicated with IFI is high.Predictive model constructed based on G-test,the absolute value of monocytes,and hs-CRP level has good discriminant validity and can serve as an important theoretical basis for early identification of IFI in MM patients.