Scholarlink

<p>Early diagnosis of Goodpasture Syndrome (GPS), an autoimmune disease caused by anti-glomerular basement membrane (anti-GBM) antibodies, is still clinically problematic because of its swift development and paucity of diagnostic indicators. In this paper, a novel framework is proposed for patient-focused detection and treatment planning. The framework functions on three synergistic levels: (1) Computational Simulation Layer which simulates protein–antibody interactions through molecular dynamics and graph-based protein interaction networks to produce virtual biomarker signatures; (2) a Quantum-Inspired Data Fusion and Pattern Recognition Layer that combines genomics, and imaging data through quantum annealing-inspired clustering and probabilistic recognition for features extraction; and (3) a Predictive Digital Triplet Layer that integrates disease progression simulations through multi-agent simulation. Comparative assessment with LSTM (long-short term memory), Transformer, and Spectral FNO (Fourier neural operator) baselines shows our framework with higher true positives (TP = 94 vs. 77, 80, and 83 respectively), lower false positives (FP = 6 vs. 21, 18, and 14), and 2.8× improved energy efficiency and 3.2× reduced latency.</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

Quantum-enhanced data fusion framework for early detection and intervention of goodpasture syndrome (GPS)

  • 期刊论文

  • 发表日期：

    2026年1月9日

查看全文
  • B. Senthilnayaki,
  • S. Karthik,
  • M. S. S. Sasikumar,
  • Pravallika Koppula,
  • S. Dhanasekar,
  • Kishore Balasubramanian

摘要

Early diagnosis of Goodpasture Syndrome (GPS), an autoimmune disease caused by anti-glomerular basement membrane (anti-GBM) antibodies, is still clinically problematic because of its swift development and paucity of diagnostic indicators. In this paper, a novel framework is proposed for patient-focused detection and treatment planning. The framework functions on three synergistic levels: (1) Computational Simulation Layer which simulates protein–antibody interactions through molecular dynamics and graph-based protein interaction networks to produce virtual biomarker signatures; (2) a Quantum-Inspired Data Fusion and Pattern Recognition Layer that combines genomics, and imaging data through quantum annealing-inspired clustering and probabilistic recognition for features extraction; and (3) a Predictive Digital Triplet Layer that integrates disease progression simulations through multi-agent simulation. Comparative assessment with LSTM (long-short term memory), Transformer, and Spectral FNO (Fourier neural operator) baselines shows our framework with higher true positives (TP = 94 vs. 77, 80, and 83 respectively), lower false positives (FP = 6 vs. 21, 18, and 14), and 2.8× improved energy efficiency and 3.2× reduced latency.