Q-PART: Quasi-Periodic Adaptive Regression with Test-time Training for Pediatric Left Ventricular Ejection Fraction Regression

2025-03-06

AI Summary

Existing methods for continuous left ventricular ejection fraction (LVEF) regression in pediatric echocardiography suffer from poor test-time adaptation and inadequate modeling of the quasi-periodic nature of cardiac motion. Method: We propose the first quasi-periodic adaptive regression framework specifically for LVEF assessment. It integrates a quasi-periodic neural network—built upon parameterized spiral trajectories and neural controlled differential equations (NCDEs)—to explicitly model heartbeat rhythm; a component-aware fine-tuning strategy distinguishing periodic and non-periodic features; and a theoretically grounded variance-constrained regression loss. Our approach synergistically combines quasi-periodic signal decomposition, test-time training (TTT), robust variance minimization, and image degradation augmentation. Results: Evaluated on three independent pediatric cohorts, our method significantly outperforms state-of-the-art approaches, achieving a clinical screening mAUROC of 0.9747. Critically, it demonstrates gender-agnostic performance across all metrics, validating its clinical robustness and practical utility.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Computer Vision: Learning & Optimization for CV
• Intelligent Robots: State Estimation

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

In this work, we address the challenge of adaptive pediatric Left Ventricular Ejection Fraction (LVEF) assessment. While Test-time Training (TTT) approaches show promise for this task, they suffer from two significant limitations. Existing TTT works are primarily designed for classification tasks rather than continuous value regression, and they lack mechanisms to handle the quasi-periodic nature of cardiac signals. To tackle these issues, we propose a novel extbf{Q}uasi- extbf{P}eriodic extbf{A}daptive extbf{R}egression with extbf{T}est-time Training (Q-PART) framework. In the training stage, the proposed Quasi-Period Network decomposes the echocardiogram into periodic and aperiodic components within latent space by combining parameterized helix trajectories with Neural Controlled Differential Equations. During inference, our framework further employs a variance minimization strategy across image augmentations that simulate common quality issues in echocardiogram acquisition, along with differential adaptation rates for periodic and aperiodic components. Theoretical analysis is provided to demonstrate that our variance minimization objective effectively bounds the regression error under mild conditions. Furthermore, extensive experiments across three pediatric age groups demonstrate that Q-PART not only significantly outperforms existing approaches in pediatric LVEF prediction, but also exhibits strong clinical screening capability with high mAUROC scores (up to 0.9747) and maintains gender-fair performance across all metrics, validating its robustness and practical utility in pediatric echocardiography analysis.

Problem

• Adaptive pediatric LVEF assessment challenges
• Handling quasi-periodic nature of cardiac signals
• Improving regression accuracy in echocardiogram analysis

Innovation

• Quasi-Period Network decomposes echocardiogram signals
• Variance minimization strategy for image augmentations
• Differential adaptation rates for signal components