Scholarlink

Purpose <p>Antenatal hydronephrosis is the most common prenatal urinary tract abnormality, and some cases, mainly due to ureteropelvic junction obstruction, require surgery. Noninvasive biomarkers are needed to improve diagnosis and guide management. To evaluate urinary exosomal long non-coding RNAs, specifically MALAT1 and HOTAIR, as biomarkers for predicting surgical necessity in children with unilateral antenatal hydronephrosis, and to develop an AI-based predictive model.</p> Methods <p>This retrospective case-control study included 88 children (38 requiring pyeloplasty, 26 non-obstructive dilatation, 24 controls). Urinary exosomes were isolated, and lncRNA expression was quantified by RT-qPCR. Clinical variables included anteroposterior pelvic diameter and split renal function. Diagnostic performance was assessed by ROC analysis and logistic regression. A SVM model was developed integrating biomarker and clinical data, with 5-fold cross-validation and an interactive Shiny web application for clinical translation.</p> Results <p>Preoperative MALAT1 and HOTAIR levels were significantly elevated compared to non-obstructive dilatation and controls and decreased after surgery (<i>p</i> &lt; 0.001). MALAT1 remained an independent predictor of surgical intervention (OR = 1.272, <i>p</i> &lt; 0.001) in multivariate analysis, alongside APD and SRF. MALAT1 showed an AUC of 0.745 (cut-off: 9.06), and HOTAIR an AUC of 0.685 (cut-off: 11.76). The SVM model achieved 91.9% accuracy, 89.5% sensitivity, 95.8% specificity, and an AUC of 0.951.</p> Conclusions <p>Urinary exosomal MALAT1 and HOTAIR are promising noninvasive biomarker for predicting surgical need in pediatric ureteropelvic junction obstruction. Integration with clinical parameters in an SVM-based model enhances diagnostic precision, and the development of a freely accessible web application supports real-time individualized risk prediction.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Can urinary exosomal lncRNAs HOTAIR and MALAT1 predict surgical need and outcomes in unilateral antenatal hydronephrosis?

  • 期刊论文

  • 发表日期：

    2026年7月3日

查看全文
  • Rifat Burak Ergül,
  • Canan Küçükgergin,
  • Ismail Selvi,
  • A. Fatih Aydın,
  • Şule Seçkin,
  • M. İrfan Dönmez,
  • Orhan Ziylan,
  • Tayfun Oktar

摘要

Purpose

Antenatal hydronephrosis is the most common prenatal urinary tract abnormality, and some cases, mainly due to ureteropelvic junction obstruction, require surgery. Noninvasive biomarkers are needed to improve diagnosis and guide management. To evaluate urinary exosomal long non-coding RNAs, specifically MALAT1 and HOTAIR, as biomarkers for predicting surgical necessity in children with unilateral antenatal hydronephrosis, and to develop an AI-based predictive model.

Methods

This retrospective case-control study included 88 children (38 requiring pyeloplasty, 26 non-obstructive dilatation, 24 controls). Urinary exosomes were isolated, and lncRNA expression was quantified by RT-qPCR. Clinical variables included anteroposterior pelvic diameter and split renal function. Diagnostic performance was assessed by ROC analysis and logistic regression. A SVM model was developed integrating biomarker and clinical data, with 5-fold cross-validation and an interactive Shiny web application for clinical translation.

Results

Preoperative MALAT1 and HOTAIR levels were significantly elevated compared to non-obstructive dilatation and controls and decreased after surgery (p < 0.001). MALAT1 remained an independent predictor of surgical intervention (OR = 1.272, p < 0.001) in multivariate analysis, alongside APD and SRF. MALAT1 showed an AUC of 0.745 (cut-off: 9.06), and HOTAIR an AUC of 0.685 (cut-off: 11.76). The SVM model achieved 91.9% accuracy, 89.5% sensitivity, 95.8% specificity, and an AUC of 0.951.

Conclusions

Urinary exosomal MALAT1 and HOTAIR are promising noninvasive biomarker for predicting surgical need in pediatric ureteropelvic junction obstruction. Integration with clinical parameters in an SVM-based model enhances diagnostic precision, and the development of a freely accessible web application supports real-time individualized risk prediction.