文献速递慢性肾脏病遗传性与小管间质肾病

利用深度学习与先进影像生物标志物刻画ADPKD-IFT140表型特征

Characterizing the ADPKD-IFT140 Phenotypic Signature With Deep Learning and Advanced Imaging Biomarkers.
PubMed 文献 · Kidney Int Rep · 2025-05-07

这项回顾性队列研究纳入携带IFT140、非截短型PKD1或PKD2致病变异的患者,按性别、年龄和身高校正总肾脏体积匹配,开发深度学习模型和实用模型区分ADPKD-IFT140,并在内部和外部队列验证。ADPKD-IFT140患者囊肿较少但较大,88.9%无肝囊肿,eGFR下降较ADPKD-PKD1NT慢。

要点

  • ADPKD-IFT140患者囊肿数量中位数42个、平均囊肿体积12.1 ml,88.9%无肝囊肿。
  • ADPKD-IFT140的eGFR下降速度为-0.69 ml/min per 1.73 m2/yr,较ADPKD-PKD1NT的-1.62慢(P=0.006)。
  • 深度学习模型在开发队列区分ADPKD-IFT140的AUC为0.949,内部队列特异度88.9%。

核心结论

  • 研究详细刻画了ADPKD-IFT140表型,可用实用或深度学习分割模型在多样临床环境中区分。

研究设计与样本

  • 回顾性队列研究,纳入携带IFT140、非截短型PKD1(PKD1NT)或PKD2致病变异的患者,按性别(48.1%男性)、年龄(均值[SD]:57.7±13.3岁)和身高校正总肾脏体积(中位数[Q1-Q3]:572.9[314.1-1137.9] ml/m)匹配。
  • 在开发队列(n=81)开发两种预测模型:纳入囊肿-实质表面积(CPSA)和囊性指数的深度学习模型,以及使用2个最大囊肿占TKV椭球体百分比的实用模型,囊肿体积由囊肿直径按公式V=π/6(d1^3+d2^3)估算。
  • 模型在内部特异度队列(n=569)和外部敏感度队列(n=36)验证。

主要发现

  • ADPKD-IFT140患者囊肿数量中位数42个、平均囊肿体积12.1 ml,88.9%无肝囊肿,与ADPKD-PKD1NT和ADPKD-PKD2相比有差异。
  • ADPKD-IFT140的eGFR下降为-0.69 ml/min per 1.73 m2/yr,较ADPKD-PKD1NT的-1.62慢(P=0.006),与ADPKD-PKD2的-0.90相比P=0.737。
  • 深度学习模型在开发队列区分ADPKD-IFT140的AUC为0.949,内部队列特异度88.9%。
  • 体积与TKV比值≥18.6%识别ADPKD-IFT140的AUC为0.814,外部队列敏感度72.2%。

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: ADPKD-IFT140 is the third most common disease-causing variant in autosomal dominant polycystic kidney disease (ADPKD) after ADPKD-PKD1 and ADPKD-PKD2. This study aimed to characterize the clinical presentation, progression, and distinctive imaging phenotype of ADPKD-IFT140. METHODS: This retrospective cohort study included patients with disease-causing variants in IFT140, nontruncating PKD1 (PKD1NT), or PKD2. Patients were matched by sex (48.1% male), age (mean [SD]: 57.7 ± 13.3 years), and height-adjusted total kidney volume (TKV; htTKV) (median [Q1-Q3]: 572.9 [314.1-1137.9] ml/m). Two predictive models were developed in the development cohort (n = 81): a deep-learning model incorporating cyst-parenchymal surface area (CPSA) and cystic index, and a practical model using percentage of TKVellipsoid occupied by the 2 largest cysts, with cyst volumes estimated from cyst diameters using the formula V = π 6 ( d 1 3 + d 2 3 ) . Models were validated in an internal specificity cohort (n = 569) and an external sensitivity cohort (n = 36). RESULTS: Patients with ADPKD-IFT140 exhibited fewer (median cyst number: 42) but larger cysts (average cyst volume: 12.1 ml), with 88.9% having no liver cysts, compared with ADPKD-PKD1NT and ADPKD-PKD2. The estimated glomerular filtration rate (eGFR) of decline was slower in ADPKD-IFT140 (-0.69 ml/min per 1.73 m2/yr) than in ADPKD-PKD1NT (-1.62, P = 0.006) and in ADPKD-PKD2 (-0.90, P = 0.737). The deep-learning model demonstrated an area-under-the-curve (AUC) of 0.949 for distinguishing ADPKD-IFT140 patients in the development cohort, and 88.9% specificity in the internal cohort. A volume-to-TKV ratio ≥ 18.6% identified ADPKD-IFT140 with an AUC of 0.814 and demonstrated 72.2% sensitivity in the external cohort. CONCLUSION: We provide a detailed characterization of the ADPKD-IFT140 phenotype that can be distinguished using a practical or deep-learning segmentation model applicable in diverse clinical settings.

来源:Kidney Int Rep,查看原文。版权归原作者所有。

引用

  1. Ghanem A, Munairdjy Debeh FG, Borghol AH, et al. Characterizing the ADPKD-IFT140 Phenotypic Signature With Deep Learning and Advanced Imaging Biomarkers. Kidney Int Rep. 2025. PMID: 40814641. doi:10.1016/j.ekir.2025.04.062

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。