Deep Few-view High-resolution Photon-counting Extremity CT at Halved Dose for a Clinical Trial

2024-03-19

arXiv.org

AI Summary

To address three key challenges in low-dose reconstruction for high-resolution photon-counting CT (PCCT)—GPU memory constraints, scarcity of real clinical training data, and simulation-to-reality domain shift—this work proposes a patch-based volumetric refinement network integrating synthetic-data pretraining with model-driven iterative refinement. The method achieves, for the first time, diagnostic-quality 3D reconstruction from few-view PCCT acquisitions (50% dose reduction and 2× faster scanning) of extremities. Evaluated on eight real clinical PCCT datasets, blind assessments by three radiologists show that reconstructed images are diagnostically equivalent to or superior to full-angle standard reconstructions. Phantom and simulation studies further demonstrate robustness across varying scan conditions and domains. This work establishes a generalizable technical paradigm enabling clinical deployment of low-dose PCCT.

Technology Category

• Computer Vision: Low Level & Physics-based Vision
• Machine Learning: Calibration & Uncertainty Quantification
• Search and Optimization: Mixed Discrete/Continuous Search

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

The latest X-ray photon-counting computed tomography (PCCT) for extremity allows multi-energy high-resolution (HR) imaging for tissue characterization and material decomposition. However, both radiation dose and imaging speed need improvement for contrast-enhanced and other studies. Despite the success of deep learning methods for 2D few-view reconstruction, applying them to HR volumetric reconstruction of extremity scans for clinical diagnosis has been limited due to GPU memory constraints, training data scarcity, and domain gap issues. In this paper, we propose a deep learning-based approach for PCCT image reconstruction at halved dose and doubled speed in a New Zealand clinical trial. Particularly, we present a patch-based volumetric refinement network to alleviate the GPU memory limitation, train network with synthetic data, and use model-based iterative refinement to bridge the gap between synthetic and real-world data. The simulation and phantom experiments demonstrate consistently improved results under different acquisition conditions on both in- and off-domain structures using a fixed network. The image quality of 8 patients from the clinical trial are evaluated by three radiologists in comparison with the standard image reconstruction with a full-view dataset. It is shown that our proposed approach is essentially identical to or better than the clinical benchmark in terms of diagnostic image quality scores. Our approach has a great potential to improve the safety and efficiency of PCCT without compromising image quality.

Problem

• Reducing radiation dose in photon-counting CT imaging
• Overcoming memory limitations in deep learning reconstruction
• Bridging domain gap between synthetic and clinical data

Innovation

• Patch-based network reduces GPU memory burden
• Synthetic data training overcomes data scarcity
• Model-based refinement bridges clinical domain gap