Logo: to the web site of Uppsala University

Most ultrasound imaging techniques necessitate the fundamental step of converting temporal signals received from transducer elements into a spatial echogenecity map. This beamforming (BF) step requires the knowledge of speed-of-sound (SoS) value in the imaged medium. An incorrect assumption of BF SoS leads to aberration artifacts, not only deteriorating the quality and resolution of conventional brightness mode (B-mode) images, hence limiting their clinical usability, but also impairing other ultrasound modalities such as elastography and spatial SoS reconstructions, which rely on faithfully beamformed images as their input. In this work, we propose an analytical method for estimating BF SoS. We show that pixel-wise relative shifts between frames beamformed with an assumed SoS is a function of geometric disparities of the transmission paths and the error in such SoS assumption. Using this relation, we devise an analytical model, the closed form solution of which yields the difference between the assumed and the true SoS in the medium. Based on this, we correct the BF SoS, which can also be applied iteratively. Both in simulations and experiments, lateral B-mode resolution is shown to be improved by ≈ 25% compared to that with an initial SoS assumption error of 3.3% (50 m/s), while localization artifacts from beamforming are also corrected. After 5 iterations, our method achieves BF SoS errors of under 0.6 m/s in simulations. Residual time-delay errors in beamforming 32 numerical phantoms are shown to reduce down to 0.07 µs, with average improvements of up to 21 folds compared to initial inaccurate assumptions. We additionally show the utility of the proposed method in imaging local SoS maps, where using our correction method reduces reconstruction root-mean-square errors substantially, down to their lower-bound with actual BF SoS.

Objectives: The aim is to assess the feasibility and accuracy of a novel quantitative ultrasound (US) method based on global speed-of-sound (g-SoS) measurement using conventional US machines, for breast density assessment in comparison to mammographic ACR (m-ACR) categories.

Materials and methods: In a prospective study, g-SoS was assessed in the upper-outer breast quadrant of 100 women, with 92 of them also having m-ACR assessed by two radiologists across the entire breast. For g-SoS, ultrasonic waves were transmitted from varying transducer locations and the image misalignments between these were then related analytically to breast SoS. To test reproducibility, two consecutive g-SoS acquisitions each were taken at two similar breast locations by the same operator.

Results: Measurements were found highly repeatable, with a mean absolute difference +/- standard deviation of 3.16 +/- 3.79 m/s. Multiple measurements were combined yielding a single g-SoS estimate per each patient, which strongly correlated to m-ACR categories (Spearman's = 0.773). The g-SoS values for categories A-D were 1459.6 +/- 0.74, 1475.6 +/- 15.92, 1515.6 +/- 27.10, and 1545.7 +/- 20.62, with all groups (except A-B) being significantly different from each other. Dense breasts (m-ACR C&D) were classified with 100% specificity at 78% sensitivity, with an area under the curve (AUC) of 0.931. Extremely dense breasts (m-ACR D) were classified with 100% sensitivity at 77.5% specificity (AUC = 0.906).

Conclusion: Quantitative g-SoS measurement of the breast was shown feasible and repeatable using conventional US machines, with values correlating strongly with m-ACR assessments.

Key Points: Breast density is a strong predictor of risk for breast cancer, which frequently develops in dense tissue regions. Therefore, density assessment calls for refined non-ionizing methods.

Findings: Quantitative global speed-of-sound (g-SoS) measurement of the breast is shown to be feasible using conventional US machines, repeatable, and able to classify breast density with high accuracy.

Clinical relevance: Being effective in classifying dense breasts, where mammography has reduced sensitivity, g-SoS can help stratify patients for alternative modalities. Ideal day for mammography or MRI can be determined by monitoring g-SoS. Furthermore, g-SoS can be integrated into personalized risk assessment.

Speed of sound (SoS) is an emerging quantitative biomechanical marker for tissue characterization. It is also used for beamforming in various ultrasound imaging solutions, the success of which hence depend critically on the accuracy of SoS assumption. Several robust global (single-value) SoS estimation techniques have been proposed. However, in heterogeneous in-vivo tissue, localized SoS estimation is desired. In this work, we propose a meta-method that extends beamforming-based global SoS estimation approaches to enable windowed SoS estimation within a desired region of interest (ROI) that is SoS-wise homogeneous. By dividing the ROI into sub-windows, we introduce a differential formulation that eliminates the influence of superficial layers for deducing the SoS within the ROI. We demonstrate this based on a recent beamforming-based SoS estimator using simulations and phantom experiments, showing that our method can accurately estimate ROI-window SoS under varying superficial layer conditions. Our window-based correction method reduces the SoS errors by over 80% in numerical simulations and over 60% in phantom experiments compared to applying global SoS estimation naively within the ROI. Our window-correction meta-method can extend to other beamforming-based global SoS estimators that are robust enough within sub-windows of a ROI.

Speed-of-sound (SoS) is an emerging ultrasound contrast modality, where pulse-echo techniques using conventional transducers offer multiple benefits. For estimating tissue SoS distributions, spatial domain reconstruction from relative speckle shifts between different beamforming sequences is a promising approach. This operates based on a forward model that relates the sought local values of SoS to observed speckle shifts, for which the associated image reconstruction inverse problem is solved. The reconstruction accuracy thus highly depends on the hand-crafted forward imaging model. In this work, we propose to learn the SoS imaging model based on data. We introduce a convolutional formulation of the pulse-echo SoS imaging problem such that the entire field-of-view requires a single unified kernel, the learning of which is then tractable and robust. We present least-squares estimation of such convolutional kernel, which can further be constrained and regularized for numerical stability. In experiments, we show that a forward model learned from k-Wave simulations reduces the contrast error of SoS reconstruc- tions by 38%, compared to a conventional hand-crafted line-based wave-path model. This simulation-learned model generalizes successfully to acquired phantom data, reducing the contrast error compared to the conventional hand-crafted alternative. We successfully demonstrate the feasibility of learning machine-specific kernels as well as one-shot learning from a single image. On in-vivo data of a cancerous breast tumor, the phantom-learned model exhibits an SoS contrast of 34.6 m/s, as an impressive improvement over the conventional model contrast of merely 3.4 m/s.

Medical imaging aims to recover underlying tissue properties, using inexact (simplified/linearized) imaging models and often from inaccurate and incomplete measurements. Analytical reconstruction methods rely on hand-crafted regularization, sensitive to noise assumptions and parameter tuning. Among deep learning alternatives, plug-and-play (PnP) approaches learn regularization while incorporating imaging physics during inference, outperforming purely data-driven methods. The performance of all these approaches, however, still strongly depends on measurement quality and imaging model accuracy. In this work, we propose DenOiS, a framework that denoises both input observations and resulting solution in their respective domains. It consists of an observation refinement strategy that corrects degraded measurements while compensating for imaging model simplifications, and a diffusion-based PnP reconstruction approach that remains robust under missing measurements. DenOiS enables generalization to real data from training only in simulations, resulting in high-fidelity image reconstruction with noisy observations and inexact imaging models. We demonstrate this for speed-of-sound imaging as a challenging setting of quantitative ultrasound image reconstruction.