文献速递肾小球疾病慢性肾脏病

梅奥MGRS预测工具计算单克隆丙种球蛋白病患者肾活检发现具有肾脏意义的单克隆丙种球蛋白病的风险

The Mayo MGRS Prediction Tool calculates the risk of finding monoclonal gammopathy of renal significance in a kidney biopsy in patients with monoclonal gammopathy.
PubMed 文献 · Kidney Int · 2025-05-20

本研究基于2013年至2023年10月在梅奥诊所接受肾活检的280例慢性肾脏病合并单克隆丙种球蛋白病患者,构建梅奥MGRS预测工具,最终模型含8个预测因子,计算及乐观校正AUC分别为0.896和0.836,外部验证队列109例表现良好。

要点

  • 梅奥MGRS预测工具含8个预测因子,计算AUC为0.896,乐观校正AUC为0.836。
  • 阈值概率0.10时敏感性98.9%、特异性50.5%;阈值概率0.25时敏感性88.0%、特异性70.2%。
  • 该工具在明尼苏达大学109例外部队列中验证表现良好。

核心结论

  • 梅奥MGRS预测工具有助于临床医生预测肾活检发现MGRS病变的概率。

研究设计与样本

  • 纳入2013年至2023年10月在梅奥诊所接受肾活检的患者,排除因肿瘤负荷等需要特定治疗的血液系统疾病如多发性骨髓瘤者,最终纳入280例。
  • 采用多因素logistic回归向前变量选择拟合模型,并使用明尼苏达大学109例外部队列验证。

主要发现

  • 280例中92例(32.9%)存在MGRS病变,最常见的病变为淀粉样轻链型或原发性淀粉样变性,共38例(41.3%)。
  • 最终模型包含8个预测因子:糖尿病、受累/未受累游离轻链比值、尿蛋白(克/24小时)、尿蛋白电泳或免疫固定电泳阳性、血清肌酐、C3水平(毫克/分升)、血尿和收缩压。
  • 计算及乐观校正AUC分别为0.896和0.836。
  • 决策曲线分析显示模型在所有阈值下均提供净获益。
  • 阈值概率0.10时敏感性98.9%、特异性50.5%;阈值概率0.25时敏感性88.0%、特异性70.2%。

关键数值与推荐

  • MGRS病变检出率32.9%(92/280),其中淀粉样轻链型或原发性淀粉样变性占41.3%(38/92)。
  • 模型计算AUC 0.896,乐观校正AUC 0.836。
  • 阈值概率0.10:敏感性98.9%,特异性50.5%;阈值概率0.25:敏感性88.0%,特异性70.2%。

临床怎么用

  • 梅奥MGRS预测工具可用于辅助临床医生预测肾活检发现MGRS病变的概率。

局限性

  • 原文未报告局限性。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

INTRODUCTION: Monoclonal gammopathy of renal significance (MGRS) is increasingly recognized as an important cause of kidney failure. A kidney biopsy remains an important diagnostic measure. However, a kidney biopsy is not without risks. Here, we devised the Mayo MGRS Prediction Tool to assess the probability of finding MGRS in patients with chronic kidney disease with monoclonal gammopathy (MG). METHODS: We included patients from 2013 to October 2023 who underwent a kidney biopsy at the Mayo Clinic excluding those whose hematological condition required specific treatment due to tumor burden such as multiple myeloma. RESULTS: Of 280 patients, 92 (32.9%) had MGRS lesions with amyloid light chain or primary amyloidosis being the most common lesion in 38 patients (41.3%). We performed multivariable logistic regression with forward variable selection to fit the model. The final model included eight predictors: diabetes, affected/unaffected free light chain ratio, urinary protein (g/24 hours), positive urine protein electrophoresis or immunofixation electrophoresis, serum creatinine, C3 level (mg/dL), hematuria, and systolic blood pressure. The calculated and optimism-corrected area under the curve was 0.896 and 0.836, respectively. The decision curve analysis showed that the model provided net benefit across all thresholds. A threshold probability of 0.10 was 98.9% sensitive and 50.5% specific, and a threshold probability of 0.25 was 88.0% sensitive and 70.2% specific. The model was validated using an external cohort of 109 patients from the University of Minnesota with good performance. CONCLUSIONS: Our Mayo MGRS Prediction Tool is useful in assisting clinicians in predicting the probability of finding an MGRS lesion in a kidney biopsy.

来源:Kidney Int,查看原文。版权归原作者所有。

引用

  1. Klomjit N, Evans MD, Vargas M, et al. The Mayo MGRS Prediction Tool calculates the risk of finding monoclonal gammopathy of renal significance in a kidney biopsy in patients with monoclonal gammopathy. Kidney Int. 2025. PMID: 40403931. doi:10.1016/j.kint.2025.04.018

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。