Blind Adaptive Local Denoising for CEST Imaging

2025-11-25

AI Summary

CEST MRI quantitative imaging (e.g., APT) suffers from spatially varying noise and heteroscedastic interference, leading to distorted contrast maps; conventional denoising methods often compromise biologically relevant signals and rely heavily on handcrafted priors. To address this, we propose an unsupervised, prior-free adaptive local denoising framework: for the first time, we exploit the self-similarity inherent in CEST data to construct an adaptive variance-stabilizing transform, coupled with a two-stage denoising pipeline and local singular value decomposition (SVD), enabling robust noise equalization and precise separation of molecular signals. Our method avoids spatial and spectral artifacts, consistently outperforms state-of-the-art approaches on multiple phantoms and in vivo datasets, and significantly improves quantitative accuracy in molecular concentration estimation and diagnostic performance in cancer detection.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Computer Vision: Medical and Biological Imaging
• Intelligent Robots: State Estimation

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods
• Responsible Web: Data and user privacy-enhancing technologies for the Web

Abstract

Chemical Exchange Saturation Transfer (CEST) MRI enables molecular-level visualization of low-concentration metabolites by leveraging proton exchange dynamics. However, its clinical translation is hindered by inherent challenges: spatially varying noise arising from hardware limitations, and complex imaging protocols introduce heteroscedasticity in CEST data, perturbing the accuracy of quantitative contrast mapping such as amide proton transfer (APT) imaging. Traditional denoising methods are not designed for this complex noise and often alter the underlying information that is critical for biomedical analysis. To overcome these limitations, we propose a new Blind Adaptive Local Denoising (BALD) method. BALD exploits the self-similar nature of CEST data to derive an adaptive variance-stabilizing transform that equalizes the noise distributions across CEST pixels without prior knowledge of noise characteristics. Then, BALD performs two-stage denoising on a linear transformation of data to disentangle molecular signals from noise. A local SVD decomposition is used as a linear transform to prevent spatial and spectral denoising artifacts. We conducted extensive validation experiments on multiple phantoms and extit{in vivo} CEST scans. In these experiments, BALD consistently outperformed state-of-the-art CEST denoisers in both denoising metrics and downstream tasks such as molecular concentration maps estimation and cancer detection.

Problem

• Addresses spatially varying noise in CEST MRI data
• Overcomes limitations of traditional denoising methods for biomedical analysis
• Improves accuracy of quantitative contrast mapping in molecular imaging

Innovation

• Blind adaptive local denoising for CEST imaging
• Variance-stabilizing transform equalizes noise distributions
• Local SVD decomposition prevents denoising artifacts