Evidential Perfusion Physics-Informed Neural Networks with Residual Uncertainty Quantification

2026-03-10

AI Summary

This work addresses the unreliability and lack of uncertainty quantification in deconvolution-based CT perfusion imaging, which often arises from violations of physical constraints. The authors propose a novel approach integrating evidential deep learning with physics-informed modeling. Perfusion parameters are estimated using coordinate neural networks, and a normal-inverse-gamma distribution is introduced over the physics residuals, enabling voxel-wise aleatoric and epistemic uncertainty quantification within a physics-informed neural network framework—without requiring Bayesian sampling or ensemble inference. Physiological constraints are explicitly embedded to enhance optimization robustness. Experiments demonstrate that the method outperforms conventional deconvolution and existing physics-informed neural network approaches across digital phantoms, the ISLES 2018 dataset, and clinical data, particularly under sparse sampling and low signal-to-noise conditions, while achieving the highest sensitivity for ischemic core detection.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Probabilistic Inference
• Computer Vision: Low Level & Physics-based Vision

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

Physics-informed neural networks (PINNs) have shown promise in addressing the ill-posed deconvolution problem in computed tomography perfusion (CTP) imaging for acute ischemic stroke assessment. However, existing PINN-based approaches remain deterministic and do not quantify uncertainty associated with violations of physics constraints, limiting reliability assessment. We propose Evidential Perfusion Physics-Informed Neural Networks (EPPINN), a framework that integrates evidential deep learning with physics-informed modeling to enable uncertainty-aware perfusion parameter estimation. EPPINN models arterial input, tissue concentration, and perfusion parameters using coordinate-based networks, and places a Normal--Inverse--Gamma distribution over the physics residual to characterize voxel-wise aleatoric and epistemic uncertainty in physics consistency without requiring Bayesian sampling or ensemble inference. The framework further incorporates physiologically constrained parameterization and stabilization strategies to promote robust per-case optimization. We evaluate EPPINN on digital phantom data, the ISLES 2018 benchmark, and a clinical cohort. On the evaluated datasets, EPPINN achieves lower normalized mean absolute error than classical deconvolution and PINN baselines, particularly under sparse temporal sampling and low signal-to-noise conditions, while providing conservative uncertainty estimates with high empirical coverage. On clinical data, EPPINN attains the highest voxel-level and case-level infarct-core detection sensitivity. These results suggest that evidential physics-informed learning can improve both accuracy and reliability of CTP analysis for time-critical stroke assessment.

Problem

• CT perfusion
• uncertainty quantification
• physics-informed neural networks
• deconvolution
• acute ischemic stroke

Innovation

• evidential deep learning
• physics-informed neural networks
• uncertainty quantification
• CT perfusion
• stroke imaging