Jihye Baek

All Publications

• A 2D autocorrelation-based frequency estimator reflecting spatial tissue distribution to improve Ultrasound H-scan tissue characterization. ArXiv Baek, J., Brevett, T., Hyun, D., El Kaffas, A., Parker, K. J., Dahl, J. J. 2026

  Abstract

  H-scan is a promising quantitative ultrasound technique that estimates the frequency content of backscattered signals and maps the estimated frequencies onto a red/blue color scale to reflect underlying tissue properties. Although it relies on matched filters tuned to different frequencies, the broad spectral bandwidth of ultrasound produces noisy, granular displays. Here, we introduce an adaptive frequency estimator designed to suppress the noise within homogeneous regions while preserving sharpness across tissue boundaries. The method combines 2D autocorrelation with a matched filter. In the first stage, a matched-filter-based estimation yields an a priori map of the spatial distribution of frequencies. The local heterogeneity of these estimates then defines a 2D weighting function that guides a second estimation stage. Drawing on the concept of Loupas's blood velocity estimator, we apply autocorrelation over a 2D spatial kernel to recover the axial frequency components, employing a weighted summation that accounts for the spatial frequency distribution within the kernel. We benchmarked the proposed estimator against conventional approaches, including the short-time Fourier transform, the H-scan matched filter, and standard autocorrelation, using both Field II simulations and in vivo data from human subjects with hepatic steatosis. In simulation, our adaptive estimator reduced the noisy texture in homogeneous regions while retaining clear boundary delineation, whereas the other estimators could achieve only one of these objectives. Applied to the in vivo human liver, the estimator improved H-scan image quality by lowering noise and enhancing the discrimination of steatotic liver from adjacent gallbladder and skin layers.

  View details for PubMedID 42819009

  View details for PubMedCentralID PMC13622993

• Improved Nondestructive Ultrasound Molecular Imaging With Lightweight Convolutional Neural Network. IEEE transactions on medical imaging Baek, J., Hyun, D., Natarajan, A., Tabesh, F., Paulmurugan, R., Dahl, J. J. 2026; 45 (7): 3895-3907

  Abstract

  Ultrasound molecular imaging (USMI) is an imaging approach that utilizes targeted microbubbles (MBs) to highlight biomarkers of disease. While differential targeted enhancement (DTE) is the current state-of-the-art for USMI, its reliance on destructive pulses hinders real-time clinical application. We have developed a neural network-based nondestructive USMI, validated in vivo using a transgenic mouse model of spontaneous breast cancer. To enhance training efficacy despite a limited animal number (N=14), we utilized several augmentation strategies including the use of several targeted MB types for each animal to generate independent image and texture patterns, alternative DTE approaches (sham and injection DTE), and random patch selection, overall resulting in a total of 15,350 patches to train the network. The resulting nondestructive USMI produces an image of the pixelwise MB classification score of the presence of targeted MBs. Our nondestructive USMI achieved a correlation coefficient of 0.954 with DTE, a continuous dice coefficient of 0.863 for a molecular signal coverage of the lesion over 20%, and a higher AUC than DTE (0.954 vs. 0.845) compared to the reference image developed from the contrast enhanced ultrasound (CEUS) image and manual lesion contour. Nondestructive imaging during continuous motion of the transducer under elevation sweeps yielded fewer artifacts and higher AUC than DTE (0.953 vs. 0.892), compared to the reference image. This demonstrates the potential of free-hand and real-time nondestructive imaging. Overall, nondestructive imaging showed comparable performance to DTE under stationary conditions and superior performance to DTE under transducer motion, indicating its clinical imaging potential.

  View details for DOI 10.1109/TMI.2026.3687142

  View details for PubMedID 42043989

• Renal cancer early diagnosis using a novel B7-H3 targeted ultrasound contrast imaging Natarajan, A., Baek, J., Paulmurugan, R., Dahl, J. AMER ASSOC CANCER RESEARCH. 2026

  View details for DOI 10.1158/1538-7445.AM2026-7901

  View details for Web of Science ID 001734420600027

• Enhancing Ultrasound Molecular Imaging: Toward Real-Time RPCA-Based Filtering to Differentiate Bound and Free Microbubbles IEEE TRANSACTIONS ON ULTRASONICS Hashemi, H. S., Hyun, D., Nguyen, N., Baek, J., Natarajan, A., Tabesh, F., Andrzejek, A., Paulmurugan, R., Dahl, J. J. 2026; 73 (2): 183-193

  Abstract

  Ultrasound molecular imaging (UMI) is an advanced imaging modality that shows promise in detecting cancer at early stages. It uses microbubbles as contrast agents, which are functionalized to bind to cancer biomarkers overexpressed on endothelial cells. A major challenge in UMI is isolating bound microbubble signal, which represents the molecular imaging signal, from that of free-floating microbubbles, which is considered background noise. In this work, we propose a fast GPU-based method using robust principal component analysis (RPCA) to distinguish bound microbubbles from free-floating ones. We explore the method using simulations and measure the accuracy using the Dice coefficient and RMS error as functions of the number of frames used in RPCA reconstruction. Experiments using stationary and flowing microbubbles in tissue-mimicking phantoms were used to validate the method. Additionally, the method was applied to data from ten transgenic mouse models of breast cancer development, injected with B7-H3 targeted microbubbles, and two mice injected with non-targeted microbubbles. The results showed that RPCA using 20 frames achieved a Dice score of 0.95 and a computation time of 0.2 seconds, indicating that 20 frames is potentially suitable for real-time implementation. On in vivo data, RPCA using 20 frames achieved a Dice score of 0.82 with DTE, indicating good agreement between the two, given the limitations of each method.

  View details for DOI 10.1109/TUSON.2025.3647590

  View details for Web of Science ID 001815829500011

  View details for PubMedID 42078652

  View details for PubMedCentralID PMC13132560

• Real-Time Convolutional Neural Network for Nondestructive Detection of Targeted Microbubbles in Ultrasound Molecular Imaging Hashemi, H. S., Nguyen, N., Baek, J., Abdullah, R., Natarajan, A., Tabesh, F., Hyun, D., Paulmurugan, R., Dahl, J. J. edited by Boehm, C., Mehrmohammadi, M., Xiang, S. L. SPIE-INT SOC OPTICAL ENGINEERING. 2026

  View details for DOI 10.1117/12.3086168

  View details for Web of Science ID 001767133000033

• Enhancing Ultrasound Molecular Imaging: Toward Real-Time RPCA-Based Filtering to Differentiate Bound and Free Microbubbles. IEEE transactions on ultrasonics Hashemi, H. S., Hyun, D., Nguyen, N., Baek, J., Natarajan, A., Tabesh, F., Andrzejek, A., Paulmurugan, R., Dahl, J. J. 2025

  Abstract

  Ultrasound molecular imaging (UMI) is an advanced imaging modality that shows promise in detecting cancer at early stages. It uses microbubbles as contrast agents, which are functionalized to bind to cancer biomarkers overexpressed on endothelial cells. A major challenge in UMI is isolating bound microbubble signal, which represents the molecular imaging signal, from that of free-floating microbubbles, which is considered background noise. In this work, we propose a fast GPU-based method using robust principal component analysis (RPCA) to distinguish bound microbubbles from free-floating ones. We explore the method using simulations and measure the accuracy using the Dice coefficient and RMS error as functions of the number of frames used in RPCA reconstruction. Experiments using stationary and flowing microbubbles in tissue-mimicking phantoms were used to validate the method. Additionally, the method was applied to data from ten transgenic mouse models of breast cancer development, injected with B7-H3 targeted microbubbles, and two mice injected with non-targeted microbubbles. The results showed that RPCA using 20 frames achieved a Dice score of 0.95 and a computation time of 0.2 seconds, indicating that 20 frames is potentially suitable for real-time implementation. On in vivo data, RPCA using 20 frames achieved a Dice score of 0.82 with DTE, indicating good agreement between the two, given the limitations of each method.

  View details for DOI 10.1109/tuson.2025.3647590

  View details for PubMedID 42078652

  View details for PubMedCentralID PMC13132560

• A first-in-human study of quantitative ultrasound to assess transplant kidney fibrosis NATURE MEDICINE Hysi, E., Baek, J., Koven, A., He, X., Severino, L., Wu, Y., Kek, K., Huang, S., Krizova, A., Farcas, M., Ordon, M., Fok, K., Stewart, R., Pace, K. T., Kolios, M. C., Parker, K. J., Yuen, D. A. 2025; 31 (3): 970-978

  Abstract

  Kidney transplantation is the optimal treatment for renal failure. In the United States, a biopsy at the time of organ procurement is often used to assess kidney quality to decide whether it should be used for transplant. This assessment is focused on renal fibrotic burden, because fibrosis is an important measure of irreversible kidney injury. Unfortunately, biopsy at the time of transplant is plagued by problems, including bleeding risk, inaccuracies introduced by sampling bias and rapid sample preparation, and the need for round-the-clock pathology expertise. We developed a quantitative algorithm, called renal H-scan, that can be added to standard ultrasound workflows to quickly and noninvasively measure renal fibrotic burden in preclinical animal models and human transplant kidneys. Furthermore, we provide evidence that biopsy-based fibrosis estimates, because of their highly localized nature, are inaccurate measures of whole-kidney fibrotic burden and do not associate with kidney function post-transplant. In contrast, we show that whole-kidney H-scan fibrosis estimates associate closely with post-transplant renal function. Taken together, our data suggest that the addition of H-scan to standard ultrasound workflows could provide a safe, rapid and easy-to-perform method for accurate quantification of transplant kidney fibrotic burden, and thus better prediction of post-transplant renal outcomes.

  View details for DOI 10.1038/s41591-024-03417-5

  View details for Web of Science ID 001435587000001

  View details for PubMedID 40033112

  View details for PubMedCentralID PMC11922760

• MBFormer: A Transformer model for 3D Time-Series Data Processing to Improve Bound Bubble Detection in Nondestructive Ultrasound Molecular Imaging Baek, J., Lee, J., Hashemi, H. S., Natarajan, A., Tabesh, F., Paulmurugan, R., Dahl, J. J., IEEE IEEE. 2025

  View details for DOI 10.1109/IUS62464.2025.11201536

  View details for Web of Science ID 001719884800271

• Sound Speed Contribution to Quantitative Ultrasound for Improving Liver Fat Assessment Baek, J., Frey, B., Guardado, L., Dahl, J. J., IEEE IEEE. 2025

  View details for DOI 10.1109/IUS62464.2025.11201389

  View details for Web of Science ID 001719884800125

• Ultrasound Molecular Imaging: A Quantitative Metric for Vascular-Independent Targeted Microbubble Binding Hashemi, H. S., Baek, J., Nguyen, N., Natarajan, A., Tabesh, F., Paulmurugan, R., Dahl, J. J., IEEE IEEE. 2025

  View details for DOI 10.1109/IUS62464.2025.11201656

  View details for Web of Science ID 001719884800388

• Multiparametric quantification and visualization of liver fat using ultrasound. WFUMB ultrasound open Baek, J., Kaffas, A. E., Kamaya, A., Hoyt, K., Parker, K. J. 2024; 2 (1)

  Abstract

  Several ultrasound measures have shown promise for assessment of steatosis compared to traditional B-scan, however clinicians may be required to integrate information across the parameters. Here, we propose an integrated multiparametric approach, enabling simple clinical assessment of key information from combined ultrasound parameters.We have measured 13 parameters related to ultrasound and shear wave elastography. These were measured in 30 human subjects under a study of liver fat. The 13 individual measures are assessed for their predictive value using independent magnetic resonance imaging-derived proton density fat fraction (MRI-PDFF) measurements as a reference standard. In addition, a comprehensive and fine-grain analysis is made of all possible combinations of sub-sets of these parameters to determine if any subset can be efficiently combined to predict fat fraction.We found that as few as four key parameters related to ultrasound propagation are sufficient to generate a linear multiparametric parameter with a correlation against MRI-PDFF values of greater than 0.93. This optimal combination was found to have a classification area under the curve (AUC) approaching 1.0 when applying a threshold for separating steatosis grade zero from higher classes. Furthermore, a strategy is developed for applying local estimates of fat content as a color overlay to produce a visual impression of the extent and distribution of fat within the liver.In principle, this approach can be applied to most clinical ultrasound systems to provide the clinician and patient with a rapid and inexpensive estimate of liver fat content.

  View details for DOI 10.1016/j.wfumbo.2024.100045

  View details for PubMedID 41676455

  View details for PubMedCentralID PMC12889900

• Enhancing Ultrasound Molecular Imaging: RPCA-Based Filtering to Differentiate Tumor-Bound and Free Microbubbles Hashemi, H. S., Hyun, D., Baek, J., Natarajan, A., Tabesh, F., Paulmurugan, R., Dahl, J. J., IEEE IEEE. 2024

  View details for DOI 10.1109/UFFC-JS60046.2024.10793633

  View details for Web of Science ID 001428150100142

• A multi-parametric model for progression of metabolic dysfunction-associated steatohepatitis (MASH) in humans Baek, J., Sanabria, S., Oyarzabal, I., Echevarria-Uraga, J. J., Quesada, C., Dahl, J., Parker, K. J., IEEE IEEE. 2024

  View details for DOI 10.1109/UFFC-JS60046.2024.10793524

  View details for Web of Science ID 001428150100042

• Differentiable Beamforming for Distributed Attenuation Estimation and Spatial Gain Compensation (SGC) Frey, B. N., Hyun, D., Simson, W., Brevett, T., Zhuang, L., Baek, J., Sanabria, S. J., Dahl, J. J., IEEE IEEE. 2024

  View details for DOI 10.1109/UFFC-JS60046.2024.10794091

  View details for Web of Science ID 001428150100552

• Frequency estimator to improve H-scan tissue characterization Baek, J., Brevett, T., Hyun, D., El Kaffas, A., Parker, K. J., Dahl, J., IEEE IEEE. 2024

  View details for DOI 10.1109/UFFC-JS60046.2024.10793751

  View details for Web of Science ID 001428150100216

• Nondestructive ultrasound molecular imaging based on a neural network approach utilizing post-processed ultrasound images Baek, J., Hyun, D., Nataraj, A., Tabesh, F., Paulmurugan, R., Dahl, J. J., IEEE IEEE. 2024

  View details for DOI 10.1109/UFFC-JS60046.2024.10793498

  View details for Web of Science ID 001428150100019

• H-Scan Discrimination for Tumor Microenvironmental Heterogeneity in Melanoma ULTRASOUND IN MEDICINE AND BIOLOGY Baek, J., Qin, S. S., Prieto, P. A., Parker, K. J. 2024; 50 (2): 268-276

  Abstract

  Melanoma is a form of malignant skin cancer that exhibits significant inter-tumoral differences in the tumor microenvironment (TME) secondary to genetic mutations. The heterogeneity may be subtle but can complicate the treatment of metastatic melanoma, contributing to a high mortality rate. Therefore, developing an accurate and non-invasive procedure to discriminate microenvironmental heterogeneity to facilitate therapy selection is an important goal.In vivo murine melanoma models that recapitulate human disease using synchronous implanted YUMM 1.7 (Yale University Mouse Melanoma) and YUMMER 1.7 (Yale University Mouse Melanoma Exposed to Radiation) murine melanoma lines were investigated. Mice were treated with antibodies to modulate the immune response and longitudinally scanned with ultrasound (US). US radiofrequency data were processed using the H-scan analysis, attenuation estimation and B-mode processing to extract five US features. The measures were used to compare different TMEs (YUMMER vs. YUMM) and responses to immunomodulatory therapies with CD8 depletion or programmed cell death protein 1 (PD-1) inhibition.Multiparametric analysis produced a combined H-scan parameter, resolving significant differences (i) between untreated YUMMER and YUMM and (ii) between untreated, PD-1-treated and CD8-treated YUMMER. However, more importantly, the B-mode and attenuation measures failed to differentiate YUMMER and YUMM and to monitor treatment responses, indicating that H-scan is required to differentiate subtle differences within the TME.We anticipate that the H-scan analysis could discriminate heterogeneous melanoma metastases and guide diagnosis and treatment selection, potentially reducing the need for invasive biopsies or immunologic procedures.

  View details for DOI 10.1016/j.ultrasmedbio.2023.10.012

  View details for Web of Science ID 001140031400001

  View details for PubMedID 37993356

  View details for PubMedCentralID PMC10794040

• Multiparametric ultrasound imaging for early-stage steatosis: Comparison with magnetic resonance imaging-based proton density fat fraction MEDICAL PHYSICS Baek, J., Basavarajappa, L., Margolis, R., Arthur, L., Li, J., Hoyt, K., Parker, K. J. 2024; 51 (2): 1313-1325

  Abstract

  The prevalence of liver diseases, especially steatosis, requires a more convenient and noninvasive tool for liver diagnosis, which can be a surrogate for the gold standard biopsy. Magnetic resonance (MR) measurement offers potential, however ultrasound (US) has better accessibility than MR.This study aims to suggest a multiparametric US approach which demonstrates better quantification and imaging performance than MR imaging-based proton density fat fraction (MRI-PDFF) for hepatic steatosis assessment.We investigated early-stage steatosis to evaluate our approach. An in vivo (within the living) animal study was performed. Fat inclusions were accumulated in the animal livers by feeding a methionine and choline deficient (MCD) diet for 2 weeks. The animals (n = 19) underwent US and MR imaging, and then their livers were excised for histological staining. From the US, MR, and histology images, fat accumulation levels were measured and compared: multiple US parameters; MRI-PDFF; histology fat percentages. Seven individual US parameters were extracted using B-mode measurement, Burr distribution estimation, attenuation estimation, H-scan analysis, and shear wave elastography. Feature selection was performed, and the selected US features were combined, providing quantification of fat accumulation. The combined parameter was used for visualizing the localized probability of fat accumulation level in the liver; This procedure is known as disease-specific imaging (DSI).The combined US parameter can sensitively assess fat accumulation levels, which is highly correlated with histology fat percentage (R = 0.93, p-value < 0.05) and outperforms the correlation between MRI-PDFF and histology (R = 0.89, p-value < 0.05). Although the seven individual US parameters showed lower correlation with histology compared to MRI-PDFF, the multiparametric analysis enabled US to outperform MR. Furthermore, this approach allowed DSI to detect and display gradual increases in fat accumulation. From the imaging output, we measured the color-highlighted area representing fatty tissues, and the fat fraction obtained from DSI and histology showed strong agreement (R = 0.93, p-value < 0.05).We demonstrated that fat quantification utilizing a combination of multiple US parameters achieved higher performance than MRI-PDFF; therefore, our multiparametric analysis successfully combined selected features for hepatic steatosis characterization. We anticipate clinical use of our proposed multiparametric US analysis, which could be beneficial in assessing steatosis in humans.

  View details for DOI 10.1002/mp.16648

  View details for Web of Science ID 001037702600001

  View details for PubMedID 37503961

  View details for PubMedCentralID PMC11238269

• Disease specific imaging utilizing support vector machine: assessment of steatosis Baek, J., Basavarajappa, L., Hoyt, K., Parker, K. J., IEEE IEEE. 2023

  View details for DOI 10.1109/IUS51837.2023.10307150

  View details for Web of Science ID 001794732200229

• Multiparametric ultrasound analysis for diagnosis of hepatic steatosis in human subjects Baek, J., Basavarajappa, L., El Kaffas, A., Kamaya, A., Hoyt, K., Parker, K. J., IEEE IEEE. 2023

  View details for DOI 10.1109/IUS51837.2023.10306523

  View details for Web of Science ID 001794732200061

• Improving breast cancer diagnosis by incorporating raw ultrasound parameters into machine learning MACHINE LEARNING-SCIENCE AND TECHNOLOGY Baek, J., O'Connell, A. M., Parker, K. J. 2022; 3 (4): 045013

  Abstract

  The improved diagnostic accuracy of ultrasound breast examinations remains an important goal. In this study, we propose a biophysical feature-based machine learning method for breast cancer detection to improve the performance beyond a benchmark deep learning algorithm and to furthermore provide a color overlay visual map of the probability of malignancy within a lesion. This overall framework is termed disease-specific imaging. Previously, 150 breast lesions were segmented and classified utilizing a modified fully convolutional network and a modified GoogLeNet, respectively. In this study multiparametric analysis was performed within the contoured lesions. Features were extracted from ultrasound radiofrequency, envelope, and log-compressed data based on biophysical and morphological models. The support vector machine with a Gaussian kernel constructed a nonlinear hyperplane, and we calculated the distance between the hyperplane and each feature's data point in multiparametric space. The distance can quantitatively assess a lesion and suggest the probability of malignancy that is color-coded and overlaid onto B-mode images. Training and evaluation were performed on in vivo patient data. The overall accuracy for the most common types and sizes of breast lesions in our study exceeded 98.0% for classification and 0.98 for an area under the receiver operating characteristic curve, which is more precise than the performance of radiologists and a deep learning system. Further, the correlation between the probability and Breast Imaging Reporting and Data System enables a quantitative guideline to predict breast cancer. Therefore, we anticipate that the proposed framework can help radiologists achieve more accurate and convenient breast cancer classification and detection.

  View details for DOI 10.1088/2632-2153/ac9bcc

  View details for Web of Science ID 000880819300001

  View details for PubMedID 36698865

  View details for PubMedCentralID PMC9855672

• Disease-Specific Imaging Utilizing Support Vector Machine Classification of H-Scan Parameters: Assessment of Steatosis in a Rat Model IEEE TRANSACTIONS ON ULTRASONICS FERROELECTRICS AND FREQUENCY CONTROL Baek, J., Basavarajappa, L., Hoyt, K., Parker, K. J. 2022; 69 (2): 720-731

  Abstract

  In medical imaging, quantitative measurements have shown promise in identifying diseases by classifying normal versus pathological parameters from tissues. The support vector machine (SVM) has shown promise as a supervised classification algorithm and has been widely used. However, the classification results typically identify a category of abnormal tissues but do not necessarily differentiate progressive stages of a disease. Moreover, the classification result is typically provided independently as a supplement to medical images, which contributes to an overload of information sources in the clinic. Hence, we propose a new imaging method utilizing the SVM to integrate classification results into medical images. This framework is called disease-specific imaging (DSI) that produces a color overlaid highlight on B-mode ultrasound images indicating the type, location, and severity of pathology from different conditions. In this article, the SVM training was performed to construct hyperplanes that can differentiate normal, fibrosis, steatosis, and pancreatic ductal adenocarcinoma (PDAC) metastases in livers based on ultrasound echoes. Also, cluster centroids for specific diseases define unique disease axes, and the inner product between measured features and any disease axis selected by the SVM quantifies the disease progression. The features were measured from 2794 ultrasound frames using the H-scan analysis, attenuation estimation, and B-mode image analysis. The performance of our proposed DSI method was evaluated for a preclinical model of steatosis ( n = 400 frames). The contribution of each feature was assessed, and the results were compared with ground truth from histology. Moreover, the images generated by our DSI were compared with earlier imaging methods of B-mode, H-scan, and histology. The comparisons demonstrate that DSI images yield higher sensitivity to monitor progressive steatosis than B-mode and H-scan and provide a comparable performance with the histology. For the parameter comparison, DSI and H-scan resulted in similar correlation with histology ( rs = 0.83 ) but higher than attenuation ( rs = 0.73 ) and B-mode ( rs = 0.47 ). Therefore, we conclude that DSI utilizing the SVM applied to steatosis can visually represent the classification results with color highlighting, which can simplify the interpretation of classification compared to the traditional SVM result. We expect that the proposed DSI can be used for any medical imaging modality that can estimate multiple quantitative parameters at high resolution.

  View details for DOI 10.1109/TUFFC.2021.3137644

  View details for Web of Science ID 000748372800029

  View details for PubMedID 34936555

  View details for PubMedCentralID PMC8908945

• Breast lesion detection and visualization utilizing artificial intelligence and the H-scan Baek, J., O'Connell, A. M., Parker, K. J., IEEE IEEE. 2022

  View details for DOI 10.1109/IUS54386.2022.9957993

  View details for Web of Science ID 000896080400280

• Detecting kidney fibrosis using H-scan Baek, J., Hysi, E., He, X., Yuen, D. A., Kolios, M. C., Parker, K. J., IEEE IEEE. 2022

  View details for DOI 10.1109/IUS54386.2022.9957217

  View details for Web of Science ID 000896080400027

• CLUSTERS OF ULTRASOUND SCATTERING PARAMETERS FOR THE CLASSIFICATION OF STEATOTIC AND NORMAL LIVERS ULTRASOUND IN MEDICINE AND BIOLOGY Baek, J., Poul, S. S., Basavarajappa, L., Reddy, S., Tai, H., Hoyt, K., Parker, K. J. 2021; 47 (10): 3014-3027

  View details for DOI 10.1016/j.ultrasmedbio.2021.06.010

  View details for Web of Science ID 000693230700020

• H-scan trajectories indicate the progression of specific diseases MEDICAL PHYSICS Baek, J., Parker, K. J. 2021; 48 (9): 5047-5058

  Abstract

  The ability of ultrasound to assess pathology is increasing with the development of quantitative parameters. Among these are a set of parameters derived from the recent H-scan analysis of subresolvable scattering. The emergence of these quantitative measures of tissue/ultrasound interactions now enables a study of the unique trajectories of multiparametric features in multidimensional space, representing the progression of specific diseases over time. We develop the mathematical and visual tools that are effective for classifying, quantifying, and visualizing the steady progression of several diseases from independent studies, all within a uniform framework.After applying the H-scan analysis of ultrasound echoes, we trained a support vector machine (SVM) to classify the unique trajectories of progressive liver disease from fibrosis, steatosis, and pancreatic ductal adenocarcinoma (PDAC) metastasis. Our approaches include the development of trajectory maps and disease-specific color imaging stains.The multidimensional SVM image classification reached 100% accuracy across the three different studies.H-scan trajectories can be useful to track the progression of multiple classes of diseases, improving diagnosis, staging, and assessing the response to therapy.

  View details for DOI 10.1002/mp.15108

  View details for Web of Science ID 000680580800001

  View details for PubMedID 34287952

  View details for PubMedCentralID PMC8455433

• Diagnostic Performance of An Artificial Intelligence System in Breast Ultrasound JOURNAL OF ULTRASOUND IN MEDICINE O'Connell, A. M., Bartolotta, T. V., Orlando, A., Jung, S., Baek, J., Parker, K. J. 2022; 41 (1): 97-105

  Abstract

  We study the performance of an artificial intelligence (AI) program designed to assist radiologists in the diagnosis of breast cancer, relative to measures obtained from conventional readings by radiologists.A total of 10 radiologists read a curated, anonymized group of 299 breast ultrasound images that contained at least one suspicious lesion and for which a final diagnosis was independently determined. Separately, the AI program was initialized by a lead radiologist and the computed results compared against those of the radiologists.The AI program's diagnoses of breast lesions had concordance with the 10 radiologists' readings across a number of BI-RADS descriptors. The sensitivity, specificity, and accuracy of the AI program's diagnosis of benign versus malignant was above 0.8, in agreement with the highest performing radiologists and commensurate with recent studies.The trained AI program can contribute to accuracy of breast cancer diagnoses with ultrasound.

  View details for DOI 10.1002/jum.15684

  View details for Web of Science ID 000625766400001

  View details for PubMedID 33665833

• Multiparametric ultrasound imaging for the assessment of normal versus steatotic livers SCIENTIFIC REPORTS Basavarajappa, L., Baek, J., Reddy, S., Song, J., Tai, H., Rijal, G., Parker, K. J., Hoyt, K. 2021; 11 (1): 2655

  Abstract

  Liver disease is increasing in prevalence across the globe. We present here a multiparametric ultrasound (mpUS) imaging approach for assessing nonalcoholic fatty liver disease (NALFD). This study was performed using rats (N = 21) that were fed either a control or methionine and choline deficient (MCD) diet. A mpUS imaging approach that includes H-scan ultrasound (US), shear wave elastography, and contrast-enhanced US measurements were then performed at 0 (baseline), 2, and 6 weeks. Thereafter, animals were euthanized and livers excised for histological processing. A support vector machine (SVM) was used to find a decision plane that classifies normal and fatty liver conditions. In vivo mpUS results from control and MCD diet fed animals reveal that all mpUS measures were different at week 6 (P < 0.05). Principal component analysis (PCA) showed that the H-scan US data contributed the highest percentage to the classification among the mpUS measurements. The SVM resulted in 100% accuracy for classification of normal and high fat livers and 92% accuracy for classification of normal, low fat, and high fat livers. Histology findings found considerable steatosis in the MCD diet fed animals. This study suggests that mpUS examinations have the potential to provide a comprehensive estimation of the main components of early stage NAFLD.

  View details for DOI 10.1038/s41598-021-82153-z

  View details for Web of Science ID 000616803100038

  View details for PubMedID 33514796

  View details for PubMedCentralID PMC7846566

• Disease-specific imaging with H-scan trajectories and support vector machine to visualize the progression of liver diseases Baek, J., Parker, K. J., IEEE IEEE. 2021

  View details for DOI 10.1109/IUS52206.2021.9593627

  View details for Web of Science ID 000832095000295

• H-scan imaging and quantitative measurement to distinguish melanoma metastasis Baek, J., Qin, S. S., Prieto, P. A., Parker, K. J., IEEE IEEE. 2021

  View details for DOI 10.1109/IUS52206.2021.9593760

  View details for Web of Science ID 000832095000417

• H-SCAN, SHEAR WAVE AND BIOLUMINESCENT ASSESSMENT OF THE PROGRESSION OF PANCREATIC CANCER METASTASES IN THE LIVER ULTRASOUND IN MEDICINE AND BIOLOGY Baek, J., Ahmed, R., Ye, J., Gerber, S. A., Parker, K. J., Doyley, M. M. 2020; 46 (12): 3369-3378

  Abstract

  The non-invasive quantification of tumor burden and the response to therapies remain an important objective for imaging modalities. To characterize the performance of two newly optimized ultrasound-based analyses, we applied shear wave and H-scan scattering analyses to repeated trans-abdominal ultrasound scans of a murine model of metastatic pancreatic cancer. In addition, bioluminescence measurements were obtained as an alternative reference. The tumor metastases grow aggressively and result in death at approximately 4 wk if untreated, but longer for those treated with chemotherapy. We found that our three imaging methods (shear wave speed, H-scan, bioluminescence) trended toward increasing output measures with time during tumor growth, and these measures were delayed for the group receiving chemotherapy. The relative sensitivity of H-scan tracked closely with bioluminescence measurements, particularly in the early to mid-stages of tumor growth. The correlation between H-scan and bioluminescence was found to be strong, with a Spearman's rank correlation coefficient greater than 0.7 across the entire series. These preliminary results suggest that non-invasive ultrasound imaging analyses are capable of tracking the response of tumor models to therapeutic agents.

  View details for DOI 10.1016/j.ultrasmedbio.2020.08.006

  View details for Web of Science ID 000581178900019

  View details for PubMedID 32907773

  View details for PubMedCentralID PMC9066934

• SCATTERING SIGNATURES OF NORMAL VERSUS ABNORMAL LIVERS WITH SUPPORT VECTOR MACHINE CLASSIFICATION ULTRASOUND IN MEDICINE AND BIOLOGY Baek, J., Poul, S. S., Swanson, T. A., Tuthill, T., Parker, K. J. 2020; 46 (12): 3379-3392

  Abstract

  Fifty years of research on the nature of backscatter from tissues has resulted in a number of promising diagnostic parameters. We recently introduced two analyses tied directly to the biophysics of ultrasound scattering: the H-scan, based on a matched filter approach to distinguishing scattering transfer functions, and the Burr distribution for quantification of speckle patterns. Together, these analyses can produce at least five parameters that are directly linked to the mathematics of ultrasound in tissue. These have been measured in vivo in 35 rat livers under normal conditions and after exposure to compounds that induce inflammation, fibrosis, and steatosis in varying combinations. A classification technique, the support vector machine, is employed to determine clusters of the five parameters that are signatures of the different liver conditions. With the multiparametric measurement approach and determination of clusters, the different types of liver pathology can be discriminated with 94.6% accuracy.

  View details for DOI 10.1016/j.ultrasmedbio.2020.08.009

  View details for Web of Science ID 000581178900020

  View details for PubMedID 32917469

  View details for PubMedCentralID PMC9386788

• Fine-tuning the H-scan for discriminating changes in tissue scatterers BIOMEDICAL PHYSICS & ENGINEERING EXPRESS Parker, K. J., Baek, J. 2020; 6 (4): 045012

  Abstract

  The H-scan approach ('H' denoting hue, or Hermite) is a recent matched filter methodology that aims to add information to the traditional ultrasound B-scan. The theory is based on the differences in the echoes produced by different classes of reflectors or scatterers. Matched filters can be created for different types of scatterers, whereby the maximum output indicates a match, and color schemes can be used to indicate the class of scatterer responsible for echoes, providing a visual interpretation of the results. However, within the theory of weak scattering from a variety of shapes, small changes in the size of the inhomogeneous objects will create shifts in the scattering transfer function. In this paper, we argue for a general power law transfer function as the canonical model for transfer functions from most normal soft vascularized tissues, at least over some bandpass spectrum illuminated by the incident pulse. In cases where scatterer size and distributions change, this produces a corresponding shift in center frequency, along with time and frequency domain characteristics of echoes, and these are captured by matched filters to distinguish and visualize in color the major characteristics of scattering types. With this general approach, the H-scan matched filters can be set to elicit more fine grain shifts in scattering types, commensurate with more subtle changes in tissue morphology. Compensation for frequency-dependent attenuation is helpful for avoiding beam softening effects with increasing depths. Examples from phantoms and normal and pathological tissues are provided to demonstrate that the H-scan analysis and displays are sensitive to scatterer size and morphology, and can be adapted to conventional imaging systems.

  View details for DOI 10.1088/2057-1976/ab9206

  View details for Web of Science ID 000536804800001

  View details for PubMedID 33444273

  View details for PubMedCentralID PMC8042758

• Placental H-scan sonography demonstrates feasibility of an ultrasound-based method for generation of a placental biomarker Hollenbach, S. J., Baek, J., Thornburg, L. L., Feltovich, H., Miller, R. K., Parker, K., McAleavey, S. MOSBY-ELSEVIER. 2020: S629

  View details for DOI 10.1016/j.ajog.2019.11.1026

  View details for Web of Science ID 000504997301332

• Support vector machine (SVM) based liver classification: fibrosis, steatosis, and inflammation Baek, J., Swanson, T. A., Tuthill, T., Parker, K. J., IEEE IEEE. 2020

  View details for Web of Science ID 000635688900305