Congrats

Congratulations to former PULSE Lab postdoc Lingyi Zhao! Her first-author journal paper entitled Detection of COVID-19 features in lung ultrasound images using deep neural networks was published in Communications Medicine, which is a Nature Publishing group journal). This paper is the first to demonstrate that simulations can be used to train deep neural networks to detect COVID-19 features in lung ultrasound images of patients. Therefore, DNNs trained with simulated and in vivo data are promising alternatives to training with only real or only simulated data when segmenting in vivo COVID-19 lung ultrasound features.

We offer public access (https://gitlab.com/pulselab/covid19) to the datasets and code described in the paper “Detection of COVID-19 features in lung ultrasound images using deep neural networks.” Communications Medicine, 2024. https://www.nature.com/articles/s43856-024-00463-5.

Access includes simulated B-mode images containing A-line, B-line, and consolidation features with paired ground truth segmentations, as well as our segmentation annotations of publicly available point of care ultrasound (POCUS) datasets (originating from https://github.com/jannisborn/covid19_ultrasound).

If you find our datasets and/or code useful, please cite the following references:

1. L. Zhao, T.C. Fong, M.A.L. Bell, “Detection of COVID-19 features in lung ultrasound images using deep neural networks”, Communications Medicine, 2024. https://www.nature.com/articles/s43856-024-00463-5
2. L. Zhao, M.A.L. Bell (2023). Code for the paper “Detection of COVID-19 features in lung ultrasound images using deep neural networks”. Zenodo. https://doi.org/10.5281/zenodo.10324042

JHU Hub Press Release

MSN

Tech Times

Published less than two years ago in June 2020, Application of the generalized contrast-to-noise ratio to assess photoacoustic image quality received the honor of being a top-cited paper in the journal Biomedical Optics Express in January 2022. This paper investigates a newly developed, probability-based, generalized contrast-to-noise (gCNR) when applied to photoacoustic images. We recommend gCNR as a new standard for assessment of novel photoacoustic beamforming and image formation techniques.

Congratulations to Kelley, Michelle, Mardava, Theron, and Prof. Bell for contributing this foundational development for the photoacoustics and biomedical optics community!

Within its first month of publication, Photoacoustic-Guided Surgery from Head to Toe, an invited review co-authored by Alycen Wiacek and Prof. Bell, received the honor of being a top 10 download from the journal Biomedical Optics Express. This invited review covers multiple aspects of the use of photoacoustic imaging to guide both surgical and related non-surgical interventions and includes a discussion of complete systems and tools needed to maximize success.

Congratulations to Alycen and Prof. Bell for capturing the attention of the biomedical optics community!

Congratulations to Eduardo González! His first-author journal paper entitled Combined ultrasound and photoacoustic image guidance of spinal pedicle cannulation demonstrated with intact ex vivo specimens was accepted to the journal IEEE Transactions on Biomedical Engineering.

This paper presents the first known combined ultrasound and photoacoustic image guidance system with software capabilities that are optimized for pedicle cannulation in posterior spinal fusion surgery. We demonstrate that both amplitude- and coherence-based beamforming methods are mutually beneficial for this task. Specifically, coherence-based beamforming of ultrasound images improved the visualization of bone for ultrasound-to-CT registration, while coherence-based beamforming of photoacoustic images improved target localization, which is important for tracking tool tips during pedicle hole creation. As shown in the figure, amplitude-based photoacoustic beamforming differentiated signals associated with an optical fiber place inside a pedicle hole (which is ideal for screw placement) from signals associated with an optical fiber touching cortical bone (which is characteristic of an impending bone breach that needs to be avoided). This proposed combination of imaging modalities and beamforming methods is promising to assist surgeons with identifying and avoiding accidental bone breaches during spinal fusion surgeries.

These new findings nicely complement our previous findings demonstrating that photoacoustic-based differentiation is possible prior to the creation of any holes, which is advantageous for correctly determining an appropriate starting point for hole creation. Together, these findings represent a complete system that can be used prior to and during pedicle hole creation for spinal fusion surgeries.

Citation: González E, Jain A, Bell MAL, Combined ultrasound and photoacoustic image guidance of spinal pedicle cannulation demonstrated with intact ex vivo specimens, IEEE Transactions on Biomedical Engineering (accepted December 17, 2020) [pdf]

Prof. Bell shares her thoughts on Photoacoustic imaging for surgical guidance: Principles, applications, and outlook in an invited Perspective that was recently published in the Journal of Applied Physics. According to the journal’s website: Perspective articles are written to present an expert viewpoint on topics currently generating a lot of interest in the research community. While perspectives generally provide a brief overview of the topic, their main purpose is to provide a forward looking view on where progress in a particular research area is heading. 

Citation: Bell MAL, Photoacoustic imaging for surgical guidance: Principles, applications, and outlook, Journal of Applied Physics, 128(6):060904, 2020 [pdf]

Congratulations to Kelley Kempski! Her first-author journal paper entitled Application of the generalized contrast-to-noise ratio to assess photoacoustic image quality was published in Biomedical Optics Express.

This paper investigates a newly developed, probability-based, generalized contrast-to-noise (gCNR) when applied to photoacoustic images. More traditional metrics experience large variations when a target is fully detectable with additional increases bearing no impact on photoacoustic target detectability. In addition, gCNR is robust to changes in traditional metrics introduced by applying a minimum threshold to image amplitudes. Therefore, gCNR has promising potential to provide additional insight, particularly when designing new beamformers and image formation techniques and when reporting quantitative performance without an opportunity to qualitatively assess corresponding images (e.g., in text-only abstracts). We recommend gCNR as a new standard for assessment of novel photoacoustic beamforming and image formation techniques.

Citation: Kempski KM, Graham MT, Gubbi MR, Palmer T, Bell MAL, Application of the generalized contrast-to-noise ratio to assess photoacoustic image quality, Biomedical Optics Express, 11(7), 3684-3698, 2020 [pdf]

Congratulations to Eduardo González! His first-author journal paper entitled GPU implementation of photoacoustic short-lag spatial coherence imaging for improved image-guided interventions was accepted to Journal of Biomedical Optics.

This paper introduces the first known real-time implementation of short-lag spatial coherence (SLSC) beamforming for photoacoustic imaging and applies this real-time algorithm to improve signal segmentation during photoacoustic-based visual servoing with low-energy lasers. Results are promising for the use of low-energy, miniaturized lasers to perform GPU-SLSC photoacoustic-based visual servoing in the operating room with laser pulse repetition frequencies as high as 41.2 Hz.

Citation: González E, Bell MAL, GPU implementation of photoacoustic short-lag spatial coherence imaging for improved image-guided interventions, Journal of BiomedicalOptics, 25(7):077002, 2020 [pdf]

Congratulations to Michelle Graham! Her first-author journal paper entitled Simulations and human cadaver head studies to identify optimal acoustic receiver locations for minimally invasive photoacoustic-guided neurosurgery was accepted to the journal Photoacoustics.

This paper presents simulation and experimental studies performed with both an intact human skull (which was cleaned from tissue attachments) and a complete human cadaver head (with contents and surrounding tissue intact) to investigate optimal locations for ultrasound probe placement during photoacoustic-guided surgeries of the skull base. The combined simulation and experimental results newly introduce the eye as a suitable acoustic window for surgical guidance. We also demonstrate a light delivery design that is suitable for patient use. Results are generally promising toward identifying, quantifying, and overcoming major system design barriers for progression to future patient testing.

Citation: Graham MT, Huang J, Creighton FX, Bell MAL, Simulations and human cadaver head studies to identify optimal acoustic receiver locations for minimally invasive photoacoustic-guided neurosurgery, Photoacoustics, 19:100183, 2020 [pdf]

Congratulations to Michelle Graham! Her first-author journal paper entitled Photoacoustic Spatial Coherence Theory and Applications to Coherence-Based Image Contrast and Resolution was accepted to the journal IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control.

This paper presents the first known theoretical derivation for photoacoustic spatial coherence functions, demonstrating excellent agreement with experimental results, particularly in the short spatial lag region, which is represented as a percentage of the receive aperture (click on the .gif to see for yourself). The coherence functions theory can be described by relating the van Cittert Zernike theorem to photoacoustics. We achieved the associated incoherent source requirement by modeling a photoacoustic target as a collection of spatially incoherent absorbers. This theory was then used to hypothesize and test previously unexplored principles for optimizing photoacoustic short-lag spatial coherence (SLSC) images, including the influence of the incident light profile on photoacoustic spatial coherence functions and associated SLSC image contrast and resolution.

Citation: Graham MT, Bell MAL, Photoacoustic Spatial Coherence Theory and Applications to Coherence-Based Image Contrast and Resolution, IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control (accepted May 26, 2020) [pdf]

Congratulations to Arun Nair! His first-author journal paper entitled Deep learning to obtain simultaneous ultrasound image and segmentation outputs from a single input of raw channel data was accepted to IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control. The paper will appear in the journal’s special issue on Deep Learning in Medical Ultrasound – from image formation to image analysis.

This paper explores the use of deep neural network (DNNs) as alternatives to delay-and-sum beamforming. The DNNs learned information directly from raw channel data to simultaneously generate both a segmentation map for automated ultrasound tasks and a corresponding ultrasound B-mode image for interpretable supervision of the automation. Although the focus was visualization and segmentation of anechoic targets surrounded by tissue, the concept can be adapted to any specific task of interest. Overall, the DNNs successfully translated feature representations learned from simulated data to phantom and in vivo data, which is promising for this novel approach to simultaneous ultrasound image formation and segmentation.

Citation: Nair AA, Washington K, Tran T, Reiter A, Bell MAL, Deep learning to obtain simultaneous ultrasound image and segmentation outputs from a single input of raw channel data, IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control (accepted May 6, 2020) [pdf]

Congratulations to Alycen Wiacek! Her first-author journal paper entitled CohereNet: A Deep Learning Architecture for Ultrasound Spatial Correlation Estimation and Coherence-Based Beamforming was accepted to IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control. The paper will appear in the journal’s special issue on Deep Learning in Medical Ultrasound – from image formation to image analysis.

This paper presents details of a novel deep neural network (DNN) architecture, named CohereNet, that was trained to estimate spatial correlation functions. The DNN-estimated correlation functions were then used to create short-lag spatial coherence ultrasound images at a faster rate than a CPU approach and with more accuracy than a GPU approach. Results were generalizable across multiple phantoms, in vivo datasets, ultrasound transducers, and ultrasound system manufacturers not included during training. CohereNet has additional potential benefits in low-power DNN-based FPGA implementations of coherence-based beamforming for miniaturized ultrasound imaging systems. In addition, CohereNet has potential utility in other areas of ultrasound imaging that require fundamental cross-correlation calculations, including elastography, speckle tracking, sound speed correction, and other advanced beamforming algorithms, such as minimum variance beamforming.

Citation: A. Wiacek, E. González and M. A. L. Bell, “CohereNet: A Deep Learning Architecture for Ultrasound Spatial Correlation Estimation and Coherence-Based Beamforming,” IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, accepted March 20, 2020 [pdf]

Congratulations to Alycen Wiacek! Her first-author journal paper entitled Coherence-based beamforming increases the diagnostic certainty of distinguishing fluid from solid masses in breast ultrasound exams was accepted to Ultrasound in Medicine & Biology. A major highlight of this paper is the inclusion of a task-based user study to determine the ability of coherence-based beamforming, specifically robust short-lag spatial coherence (R-SLSC) imaging, to assess breast mass content and ultimately impact clinical care.

The information from R-SLSC images reduced the uncertainty of fluid mass content from 47.5% to 15.8%, and the number of fluid-filled masses recommended for biopsy was reduced from 43.3% to 13.3%. This work is the first to investigate coherence-based beamforming in breast ultrasound to inform clinical decision making, highlighting the potential of this novel technique to improve diagnostic certainty and to reduce the number of unnecessary biopsies of of fluid-filled breast masses.

Citation: A Wiacek, E Oluyemi, K Myers, L Mulen, MAL Bell, Coherence-based beamforming increases the diagnostic certainty of distinguishing fluid from solid masses in breast ultrasound exams, Ultrasound in Medicine and Biology, accepted January 20, 2020 [pdf]

Congratulations to Alycen Wiacek! Her first-author journal paper entitled “Robust Short-Lag Spatial Coherence Imaging of Breast Ultrasound Data: Initial Clinical Results” was accepted to the IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control. This paper will appear in the Special Issue on Pilot Clinical Translation of New Medical Ultrasound Methodologies.

This work is the first to investigate the application of short-lag spatial coherence (SLSC) imaging — and two new variations of this method developed in our lab (i.e., M-Weighted SLSC imaging and Robust SLSC imaging) — to breast ultrasound data. The two newer approaches revisit the lag summation step of SLSC imaging to achieve additional robustness to coherence outliers through weighted summation of individual coherence images (i.e., M-weighting) and through the application of robust principal component analysis (i.e., Robust SLSC, or R-SLSC).

An interesting finding from this initial investigation is that solid breast masses, which appear hypoechoic in traditional B-mode images, have similarly high coherence to that of surrounding tissue in these coherence-based images. This finding represents the first known instance of this phenomenon in hypoechoic ultrasound data from any type of tissue (including simulated, phantom, and in vivo liver data). This work shows great promise for implementing SLSC, M-Weighted SLSC, and/or R-SLSC imaging to distinguish between fluid-filled and solid hypoechoic breast masses, which has implications for improved breast cancer detection and screening, more streamlined diagnostic work ups, and reduced patient anxiety over suspicious breast mass findings.

Citation: A Wiacek, OMH Rindal, E Falomo, K Myers, K Fabrega-Foster, S Harvey, MAL Bell, Robust Short-Lag Spatial Coherence Imaging of Breast Ultrasound Data: Initial Clinical Results, IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, accepted November 20, 2018 [pdf]

Also Available on Journal Website: https://ieeexplore.ieee.org/document/8548567

Our journal paper entitled Photoacoustic-based visual servoing of a needle tip was accepted for publication in Scientific Reports. Check out this video for a demonstration of our robotic photoacoustic assistant!

Citation: Bell MAL, Shubert J, Photoacoustic-based visual servoing of a needle tip, Scientific Reports, 8:15519, 2018 [pdf]

Our journal paper entitled Additive noise models for photoacoustic spatial coherence theory was accepted for publication in Biomedical Optics Express. This paper introduces two noise models for photoacoustic spatial coherence theory that compare favorably with experimental measurements. It will appear in the journal’s feature issue Topics in Biomedical Optics from the OSA Biophotonics Congress 2018.

Citation: B Stephanian, MT Graham, H Hou, MAL Bell, Additive noise models for photoacoustic spatial coherence theory, Biomedical Optics Express, 9(11):5566-5582, 2018 [pdf]

Our journal paper entitled Photoacoustic imaging of a human vertebra: implications for guiding spinal fusion surgeries was accepted for publication in Physics in Medicine and Biology. This paper will appear in the journal’s special issue Focus on Interventional Photoacoustic Imaging.  The accepted version of the manuscript is currently accessible online (ahead of print).

Citation: J Shubert and MAL Bell, Photoacoustic imaging of a human vertebra: Implications for guiding spinal fusion surgeries, Physics in Medicine and Biology (accepted) [pdf]

Congratulations to Derek Allman! His paper entitled “Photoacoustic Source Detection and Reflection Artifact Removal Enabled by Deep Learning” was accepted to the IEEE Transactions on Medical Imaging. This paper is expected to appear in the Special Issue on Machine Learning for Image Reconstruction.

This work is the first to use deep convolutional neural networks (CNNs) as an alternative to the photoacoustic beamforming and image reconstruction process. We used simulations to train CNNs to identify sources and reflection artifacts in raw photoacoustic channel data, reformatted the network outputs to usable images that we call CNN-Based images, and transferred these trained networks to operate on experimental data. Multiple parameters were varied during training (e.g., channel noise, number of sources, number of artifacts, sound speed, signal amplitude, transducer model, lateral and axial locations of sources and artifacts, and spacing between sources and artifacts). The classification accuracy of simulation and experimental data  ranged from 96-100% when the channel signal-to-noise ratio was -9 dB or greater and when sources were located in trained locations. Over 99% of the results had submillimeter location accuracy. Our CNN-Based images have high contrast,  no artifacts, and resolution that rivals the traditional photoacoustic image resolution of low-frequency ultrasound probes.

Citation: D Allman, A Reiter, MAL Bell, Photoacoustic Source Detection and Reflection Artifact Removal Enabled by Deep Learning, IEEE Transactions on Medical Imaging, 37(6):1464-1477, 2018 [pdf | datasets | code]