DA-UCT: Self-Supervised Domain-Adaptive Ultrasound Computed Tomography for Rapid Musculoskeletal Sound Speed Reconstruction

2026-05-24

AI Summary

This work addresses the challenges of ultrasound computed tomography (UCT) in musculoskeletal imaging—namely high computational cost, difficulty in converging nonlinear optimization, and scarcity of in vivo annotated data—by proposing the DA-UCT framework. DA-UCT introduces, for the first time, self-supervised domain adaptation into in vivo UCT, combining full-waveform inversion-based simulation pretraining with physics-informed self-supervised learning to enable efficient transfer from synthetic to real data. By integrating an attention-enhanced network (AttUCT) with low-rank adaptation (LoRA), the method achieves performance comparable to full fine-tuning while updating only 3% of parameters, enabling real-time 3D visualization. Experiments demonstrate a PSNR of 29.23 dB and SSIM of 0.928 on simulated data; in vivo reconstructions clearly resolve skin, fat, muscle, tendon, and bone structures, showing strong agreement with MRI, with single-frame reconstruction in just 5 milliseconds.

Abstract

Ultrasound computed tomography (UCT) via full waveform inversion (FWI) enables high-resolution quantitative imaging for tissue characterization and disease diagnosis. However, UCT suffers from large computational burden and severe convergence issues due to highly nonlinear optimization. Deep learning can accelerate UCT reconstruction, but supervised training requires large-scale labeled datasets difficult to obtain in vivo. To address these limitations, we propose SDA-UCT, a two-stage self-supervised domain-adaptive framework for rapid and accurate UCT imaging of musculoskeletal tissues. SDA-UCT employs an attention-enhanced network (AttUCT) pre-trained on simulation datasets and transfers to in-vivo data via physics-informed self-supervised learning, effectively bridging the simulation-to-real domain gap. A Low-Rank Adaptation (LoRA) mechanism is integrated to enable efficient adaptation across diverse clinical scenarios. Results showed that AttUCT achieved high-quality SOS reconstruction for simulated human forearm with a PSNR of 29.23 dB and SSIM of 0.928, outperforming conventional FWI and existing deep learning methods. Validated on in-vivo data, SDA-UCT successfully reconstructed SOS images revealing complex anatomical structures (skin, fat, muscle, tendon, bone and bone marrow) for human forearm, in high concordance with MRI references. The LoRA mechanism adjusting only 3% of parameters achieved comparable performance to full fine-tuning. The rapid reconstruction (5 ms per frame) enables real-time 3D visualization, achieving five-orders-of-magnitude improvement over traditional FWI. This work represents the first self-supervised domain-adaptive deep learning for rapid, high-resolution in-vivo UCT imaging, showing potential for musculoskeletal disease diagnosis.

Problem

• Ultrasound Computed Tomography
• Domain Adaptation
• Self-Supervised Learning
• Musculoskeletal Imaging
• Sound Speed Reconstruction

Innovation

• self-supervised learning
• domain adaptation
• ultrasound computed tomography
• Low-Rank Adaptation (LoRA)
• full waveform inversion