Akshay Chaudhari
Assistant Professor (Research) of Radiology (Integrative Biomedical Imaging Informatics at Stanford) and of Biomedical Data Science
Bio
Dr. Chaudhari is an Assistant Professor of Radiology and (by courtesy) Biomedical Data Science. Dr. Chaudhari leads the Machine Intelligence in Medical Imaging research group at Stanford focusing on improving both the acquisition and analysis of medical images and related healthcare data. His group develops new self-supervised and representation learning techniques for multi-modal deep learning for healthcare using vision, language, and medical records data. Dr. Chaudhari’s research is funded by the NIH, ARPA-H, and several industry partners. He has won the W.S. Moore Young Investigator Award and the Junior Fellow Award from the International Society for Magnetic Resonance in Medicine, and is inducted into the Academy of Radiology’s Council of Early Career Investigators in Imaging program. He also serves as the Co-Director of the Stanford Radiology AI Development and Evaluation Laboratory as well as the Associate Director of Research and Education at the Stanford AIMI Center.
Academic Appointments
-
Assistant Professor (Research), Radiology
-
Assistant Professor (Research), Department of Biomedical Data Science
-
Member, Bio-X
-
Member, Cardiovascular Institute
-
Faculty Affiliate, Institute for Human-Centered Artificial Intelligence (HAI)
-
Member, Wu Tsai Neurosciences Institute
Honors & Awards
-
Junior Fellow, International Society for Magnetic Resonance in Medicine (2020)
-
W.S. Moore Young Investigator Award, International Society for Magnetic Resonance in Medicine (2019)
-
Best Young Investigator Award, 12th Intl. Workshop on Osteoarthritis (2019)
-
Best Emerging Investigator, Imaging Elevated Symposium (2019)
-
2nd - 'Best Science' Presentation, ISMRM and RSNA Workshop on Value in MRI (2018)
-
2nd – ‘Best Value’ Presentation, ISMRM and RSNA Workshop on Value in MRI (2018)
-
2x Magna Cum Laude Merit Award, International Society for Magnetic Resonance in Medicine Annual Meeting (2018)
-
Best Healthcare Poster, NVIDIA GPU Technology Conference (2018)
-
Best Junior Investigator Abstract, 11th Intl. Workshop on Osteoarthritis (2018)
-
Best Overall Poster, NVIDIA GPU Technology Conference (2018)
-
Editor’s Monthly Pick, Magnetic Resonance in Medicine (2018)
-
Outstanding Teacher Award, International Society for Magnetic Resonance in Medicine Annual Meeting (2018)
-
Best Young Investigator Award, 10th Intl. Workshop on Osteoarthritis (2017)
-
Biodesign NEXT Fellow, Stanford Biodesign (2017)
-
Magna Cum Laude Merit Award, International Society for Magnetic Resonance in Medicine (2017)
-
Best Young Investigator Award, 9h Intl. Workshop on Osteoarthritis (2016)
-
Mobile Biodesign Innovation Award, Stanford Biodesign (2016)
-
Siebel Scholar for Engineering Leadership, Siebel Foundation (2016)
-
Award of Merit for Highly Rated Trainee Abstract, 8th Intl. Workshop on Osteoarthritis (2015)
-
Whitaker International Program Summer Fellow, Whitaker Foundation (2015)
-
Best Poster, Center for Biomedical Imaging at Stanford Symposium (2014)
-
Graduate Research Fellow, National Science Foundation (2012)
-
Best Undergraduate Research Poster, University of California San Diego Bioengineering Day (2011)
-
Chuao Chocolate Alumni Scholar, University of California San Diego (2010)
-
Most Informative Poster, Genentech Summer Intern Poster Expo (2010)
-
Outstanding UCSD Junior, Genentech Process Research and Development (2010)
-
Best Oral Presentation, Biomedical Engineering Society Lab Expo (2009)
-
Gordon Scholar, Jacobs School of Engineering (UCSD) (2009)
Current Research and Scholarly Interests
Dr. Chaudhari is interested in the application of artificial intelligence techniques to all aspects of medical imaging, including automated schedule and reading prioritization, image reconstruction, quantitative analysis, clinical decision support, and prediction of patient outcomes. His interests focus on the development and evaluation new self-supervised and representation learning techniques for multi-modal deep learning in healthcare using vision, language, and medical records data
2024-25 Courses
- Foundation Models for Healthcare
BIODS 271, RAD 271 (Spr) -
Independent Studies (8)
- Directed Investigation
BIOE 392 (Aut, Win, Spr, Sum) - Directed Reading and Research
BIOMEDIN 299 (Aut, Win, Spr, Sum) - Directed Reading in Radiology
RAD 299 (Aut, Win, Spr, Sum) - Directed Study
BIOE 391 (Aut, Win, Spr, Sum) - Graduate Research
BMP 399 (Aut, Win, Spr, Sum) - Graduate Research
RAD 399 (Aut, Win, Spr, Sum) - Ph.D. Research
CME 400 (Aut, Win, Spr, Sum) - Undergraduate Research
RAD 199 (Aut, Win, Spr, Sum)
- Directed Investigation
-
Prior Year Courses
2023-24 Courses
- Biomedical Data Science Student Seminar
BIODS 201, BIOMEDIN 201 (Sum) - Essentials of Deep Learning in Medicine
BIOS 407 (Spr) - Foundation Models for Healthcare
BIODS 271, CS 277, RAD 271 (Win)
2022-23 Courses
- Biomedical Informatics Student Seminar
BIODS 201, BIOMEDIN 201 (Sum)
- Biomedical Data Science Student Seminar
Stanford Advisees
-
Postdoctoral Faculty Sponsor
Anthony Gatti, Magdalini Paschali, Tian Tan, Yan-Ran (Joyce) Wang, McKenzie White -
Doctoral Dissertation Advisor (AC)
Louis Blankemeier, Stefania Moroianu, Anoosha Pai S -
Doctoral Dissertation Co-Advisor (AC)
Ashwin Kumar -
Postdoctoral Research Mentor
Robert Holland, Jiaming Liu, Sophie Ostmeier
All Publications
-
A vision-language foundation model for the generation of realistic chest X-ray images.
Nature biomedical engineering
2024
Abstract
The paucity of high-quality medical imaging datasets could be mitigated by machine learning models that generate compositionally diverse images that faithfully represent medical concepts and pathologies. However, large vision-language models are trained on natural images, and the diversity distribution of the generated images substantially differs from that of medical images. Moreover, medical language involves specific and semantically rich vocabulary. Here we describe a domain-adaptation strategy for large vision-language models that overcomes distributional shifts. Specifically, by leveraging publicly available datasets of chest X-ray images and the corresponding radiology reports, we adapted a latent diffusion model pre-trained on pairs of natural images and text descriptors to generate diverse and visually plausible synthetic chest X-ray images (as confirmed by board-certified radiologists) whose appearance can be controlled with free-form medical text prompts. The domain-adaptation strategy for the text-conditioned synthesis of medical images can be used to augment training datasets and is a viable alternative to the sharing of real medical images for model training and fine-tuning.
View details for DOI 10.1038/s41551-024-01246-y
View details for PubMedID 39187663
View details for PubMedCentralID 10131505
-
Merlin: A Vision Language Foundation Model for 3D Computed Tomography.
Research square
2024
Abstract
Over 85 million computed tomography (CT) scans are performed annually in the US, of which approximately one quarter focus on the abdomen. Given the current shortage of both general and specialized radiologists, there is a large impetus to use artificial intelligence to alleviate the burden of interpreting these complex imaging studies while simultaneously using the images to extract novel physiological insights. Prior state-of-the-art approaches for automated medical image interpretation leverage vision language models (VLMs) that utilize both the image and the corresponding textual radiology reports. However, current medical VLMs are generally limited to 2D images and short reports. To overcome these shortcomings for abdominal CT interpretation, we introduce Merlin - a 3D VLM that leverages both structured electronic health records (EHR) and unstructured radiology reports for pretraining without requiring additional manual annotations. We train Merlin using a high-quality clinical dataset of paired CT scans (6+ million images from 15,331 CTs), EHR diagnosis codes (1.8+ million codes), and radiology reports (6+ million tokens) for training. We comprehensively evaluate Merlin on 6 task types and 752 individual tasks. The non-adapted (off-the-shelf) tasks include zero-shot findings classification (31 findings), phenotype classification (692 phenotypes), and zero-shot cross-modal retrieval (image to findings and image to impressions), while model adapted tasks include 5-year chronic disease prediction (6 diseases), radiology report generation, and 3D semantic segmentation (20 organs). We perform internal validation on a test set of 5,137 CTs, and external validation on 7,000 clinical CTs and on two public CT datasets (VerSe, TotalSegmentator). Beyond these clinically-relevant evaluations, we assess the efficacy of various network architectures and training strategies to depict that Merlin has favorable performance to existing task-specific baselines. We derive data scaling laws to empirically assess training data needs for requisite downstream task performance. Furthermore, unlike conventional VLMs that require hundreds of GPUs for training, we perform all training on a single GPU. This computationally efficient design can help democratize foundation model training, especially for health systems with compute constraints. We plan to release our trained models, code, and dataset, pending manual removal of all protected health information.
View details for DOI 10.21203/rs.3.rs-4546309/v1
View details for PubMedID 38978576
View details for PubMedCentralID PMC11230513
-
ShapeMed-Knee: A Dataset and Neural Shape Model Benchmark for Modeling 3D Femurs.
medRxiv : the preprint server for health sciences
2024
Abstract
Analyzing anatomic shapes of tissues and organs is pivotal for accurate disease diagnostics and clinical decision-making. One prominent disease that depends on anatomic shape analysis is osteoarthritis, which affects 30 million Americans. To advance osteoarthritis diagnostics and prognostics, we introduce ShapeMed-Knee, a 3D shape dataset with 9,376 high-resolution, medical-imaging-based 3D shapes of both femur bone and cartilage. Besides data, ShapeMed-Knee includes two benchmarks for assessing reconstruction accuracy and five clinical prediction tasks that assess the utility of learned shape representations. Leveraging ShapeMed-Knee, we develop and evaluate a novel hybrid explicit-implicit neural shape model which achieves up to 40% better reconstruction accuracy than a statistical shape model and implicit neural shape model. Our hybrid models achieve state-of-the-art performance for preserving cartilage biomarkers; they're also the first models to successfully predict localized structural features of osteoarthritis, outperforming shape models and convolutional neural networks applied to raw magnetic resonance images and segmentations. The ShapeMed-Knee dataset provides medical evaluations to reconstruct multiple anatomic surfaces and embed meaningful disease-specific information. ShapeMed-Knee reduces barriers to applying 3D modeling in medicine, and our benchmarks highlight that advancements in 3D modeling can enhance the diagnosis and risk stratification for complex diseases. The dataset, code, and benchmarks will be made freely accessible.
View details for DOI 10.1101/2024.05.06.24306965
View details for PubMedID 38766040
-
Abdominal CT metrics in 17,646 patients reveal associations between myopenia, myosteatosis, and medical phenotypes: a phenome-wide association study
EBIOMEDICINE
2024; 103
View details for DOI 10.1051161016/j.ebiom.2024.105116
View details for Web of Science ID 001231570700001
-
Automated abdominal CT contrast phase detection using an interpretable and open-source artificial intelligence algorithm.
European radiology
2024
Abstract
To develop and validate an open-source artificial intelligence (AI) algorithm to accurately detect contrast phases in abdominal CT scans.Retrospective study aimed to develop an AI algorithm trained on 739 abdominal CT exams from 2016 to 2021, from 200 unique patients, covering 1545 axial series. We performed segmentation of five key anatomic structures-aorta, portal vein, inferior vena cava, renal parenchyma, and renal pelvis-using TotalSegmentator, a deep learning-based tool for multi-organ segmentation, and a rule-based approach to extract the renal pelvis. Radiomics features were extracted from the anatomical structures for use in a gradient-boosting classifier to identify four contrast phases: non-contrast, arterial, venous, and delayed. Internal and external validation was performed using the F1 score and other classification metrics, on the external dataset "VinDr-Multiphase CT".The training dataset consisted of 172 patients (mean age, 70 years ± 8, 22% women), and the internal test set included 28 patients (mean age, 68 years ± 8, 14% women). In internal validation, the classifier achieved an accuracy of 92.3%, with an average F1 score of 90.7%. During external validation, the algorithm maintained an accuracy of 90.1%, with an average F1 score of 82.6%. Shapley feature attribution analysis indicated that renal and vascular radiodensity values were the most important for phase classification.An open-source and interpretable AI algorithm accurately detects contrast phases in abdominal CT scans, with high accuracy and F1 scores in internal and external validation, confirming its generalization capability.Contrast phase detection in abdominal CT scans is a critical step for downstream AI applications, deploying algorithms in the clinical setting, and for quantifying imaging biomarkers, ultimately allowing for better diagnostics and increased access to diagnostic imaging.Digital Imaging and Communications in Medicine labels are inaccurate for determining the abdominal CT scan phase. AI provides great help in accurately discriminating the contrast phase. Accurate contrast phase determination aids downstream AI applications and biomarker quantification.
View details for DOI 10.1007/s00330-024-10769-6
View details for PubMedID 38683384
View details for PubMedCentralID 9700820
-
Abdominal CT metrics in 17,646 patients reveal associations between myopenia, myosteatosis, and medical phenotypes: aphenome-wide association study.
EBioMedicine
2024; 103: 105116
Abstract
BACKGROUND: Deep learning facilitates large-scale automated imaging evaluation of body composition. However, associations of body composition biomarkers with medical phenotypes have been underexplored. Phenome-wide association study (PheWAS) techniques search for medical phenotypes associated with biomarkers. A PheWAS integrating large-scale analysis of imaging biomarkers and electronic health record (EHR) data could discover previously unreported associations and validate expected associations. Here we use PheWAS methodology to determine the association of abdominal CT-based skeletal muscle metrics with medical phenotypes in a large North American cohort.METHODS: An automated deep learning pipeline was used to measure skeletal muscle index (SMI; biomarker of myopenia) and skeletal muscle density (SMD; biomarker of myosteatosis) from abdominal CT scans of adults between 2012 and 2018. A PheWAS was performed with logistic regression using patient sex and age as covariates to assess for associations between CT-derived muscle metrics and 611 common EHR-derived medical phenotypes. PheWAS P values were considered significant at a Bonferroni corrected threshold (alpha=0.05/1222).FINDINGS: 17,646 adults (mean age, 56 years±19 [SD]; 57.5% women) were included. CT-derived SMI was significantly associated with 268 medical phenotypes; SMD with 340 medical phenotypes. Previously unreported associations with the highest magnitude of significance included higher SMI with decreased cardiac dysrhythmias (OR [95% CI], 0.59 [0.55-0.64]; P<0.0001), decreased epilepsy (OR, 0.59 [0.50-0.70]; P<0.0001), and increased elevated prostate-specific antigen (OR, 1.84 [1.47-2.31]; P<0.0001), and higher SMD with decreased decubitus ulcers (OR, 0.36 [0.31-0.42]; P<0.0001), sleep disorders (OR, 0.39 [0.32-0.47]; P<0.0001), and osteomyelitis (OR, 0.43 [0.36-0.52]; P<0.0001).INTERPRETATION: PheWAS methodology reveals previously unreported associations between CT-derived biomarkers of myopenia and myosteatosis and EHR medical phenotypes. The high-throughput PheWAS technique applied on a population scale can generate research hypotheses related to myopenia and myosteatosis and can be adapted to research possible associations of other imaging biomarkers with hundreds of EHR medical phenotypes.FUNDING: National Institutes of Health, Stanford AIMI-HAI pilot grant, Stanford Precision Health and Integrated Diagnostics, Stanford Cardiovascular Institute, Stanford Center for Digital Health, and Stanford Knight-Hennessy Scholars.
View details for DOI 10.1016/j.ebiom.2024.105116
View details for PubMedID 38636199
-
Adapted large language models can outperform medical experts in clinical text summarization.
Nature medicine
2024
Abstract
Analyzing vast textual data and summarizing key information from electronic health records imposes a substantial burden on how clinicians allocate their time. Although large language models (LLMs) have shown promise in natural language processing (NLP) tasks, their effectiveness on a diverse range of clinical summarization tasks remains unproven. Here we applied adaptation methods to eight LLMs, spanning four distinct clinical summarization tasks: radiology reports, patient questions, progress notes and doctor-patient dialogue. Quantitative assessments with syntactic, semantic and conceptual NLP metrics reveal trade-offs between models and adaptation methods. A clinical reader study with 10 physicians evaluated summary completeness, correctness and conciseness; in most cases, summaries from our best-adapted LLMs were deemed either equivalent (45%) or superior (36%) compared with summaries from medical experts. The ensuing safety analysis highlights challenges faced by both LLMs and medical experts, as we connect errors to potential medical harm and categorize types of fabricated information. Our research provides evidence of LLMs outperforming medical experts in clinical text summarization across multiple tasks. This suggests that integrating LLMs into clinical workflows could alleviate documentation burden, allowing clinicians to focus more on patient care.
View details for DOI 10.1038/s41591-024-02855-5
View details for PubMedID 38413730
View details for PubMedCentralID 5593724
-
Self-Supervised Learning Improves Accuracy and Data Efficiency for IMU-Based Ground Reaction Force Estimation.
IEEE transactions on bio-medical engineering
2024; PP
Abstract
OBJECTIVE: Recent deep learning techniques hold promise to enable IMU-driven kinetic assessment; however, they require large extents of ground reaction force (GRF) data to serve as labels for supervised model training. We thus propose using existing self-supervised learning (SSL) techniques to leverage large IMU datasets to pre-train deep learning models, which can improve the accuracy and data efficiency of IMU-based GRF estimation.METHODS: We performed SSL by masking a random portion of the input IMU data and training a transformer model to reconstruct the masked portion. We systematically compared a series of masking ratios across three pre-training datasets that included real IMU data, synthetic IMU data, or a combination of the two. Finally, we built models that used pre-training and labeled data to estimate GRF during three prediction tasks: overground walking, treadmill walking, and drop landing.RESULTS: When using the same amount of labeled data, SSL pre-training significantly improved the accuracy of 3-axis GRF estimation during walking compared to baseline models trained by conventional supervised learning. Fine-tuning SSL model with 1-10% of walking data yielded comparable accuracy to training baseline model with 100% of walking data. The optimal masking ratio for SSL is 6.25-12.5%.CONCLUSION: SSL leveraged large real and synthetic IMU datasets to increase the accuracy and data efficiency of deep-learning-based GRF estimation, reducing the need for labeled data.SIGNIFICANCE: This work, with its open-source code and models, may unlock broader use cases of IMU-driven kinetic assessment by mitigating the scarcity of GRF measurements in practical applications.
View details for DOI 10.1109/TBME.2024.3361888
View details for PubMedID 38315597
-
Self-Supervised Learning Improves Accuracy and Data Efficiency for IMU-Based Ground Reaction Force Estimation.
bioRxiv : the preprint server for biology
2024
Abstract
Recent deep learning techniques hold promise to enable IMU-driven kinetic assessment; however, they require large extents of ground reaction force (GRF) data to serve as labels for supervised model training. We thus propose using existing self-supervised learning (SSL) techniques to leverage large IMU datasets to pre-train deep learning models, which can improve the accuracy and data efficiency of IMU-based GRF estimation.We performed SSL by masking a random portion of the input IMU data and training a transformer model to reconstruct the masked portion. We systematically compared a series of masking ratios across three pre-training datasets that included real IMU data, synthetic IMU data, or a combination of the two. Finally, we built models that used pre-training and labeled data to estimate GRF during three prediction tasks: overground walking, treadmill walking, and drop landing.When using the same amount of labeled data, SSL pre-training significantly improved the accuracy of 3-axis GRF estimation during walking compared to baseline models trained by conventional supervised learning. Fine-tuning SSL model with 1-10% of walking data yielded comparable accuracy to training baseline model with 100% of walking data. The optimal masking ratio for SSL is 6.25-12.5%.SSL leveraged large real and synthetic IMU datasets to increase the accuracy and data efficiency of deep-learning-based GRF estimation, reducing the need for labeled data.This work, with its open-source code and models, may unlock broader use cases of IMU-driven kinetic assessment by mitigating the scarcity of GRF measurements in practical applications.
View details for DOI 10.1101/2023.10.25.564057
View details for PubMedID 38328126
View details for PubMedCentralID PMC10849467
-
Opportunistic assessment of ischemic heart disease risk using abdominopelvic computed tomography and medical record data: a multimodal explainable artificial intelligence approach.
Scientific reports
2023; 13 (1): 21034
Abstract
Current risk scores using clinical risk factors for predicting ischemic heart disease (IHD) events-the leading cause of global mortality-have known limitations and may be improved by imaging biomarkers. While body composition (BC) imaging biomarkers derived from abdominopelvic computed tomography (CT) correlate with IHD risk, they are impractical to measure manually. Here, in a retrospective cohort of 8139 contrast-enhanced abdominopelvic CT examinations undergoing up to 5 years of follow-up, we developed multimodal opportunistic risk assessment models for IHD by automatically extracting BC features from abdominal CT images and integrating these with features from each patient's electronic medical record (EMR). Our predictive methods match and, in some cases, outperform clinical risk scores currently used in IHD risk assessment. We provide clinical interpretability of our model using a new method of determining tissue-level contributions from CT along with weightings of EMR features contributing to IHD risk. We conclude that such a multimodal approach, which automatically integrates BC biomarkers and EMR data, can enhance IHD risk assessment and aid primary prevention efforts for IHD. To further promote research, we release the Opportunistic L3 Ischemic heart disease (OL3I) dataset, the first public multimodal dataset for opportunistic CT prediction of IHD.
View details for DOI 10.1038/s41598-023-47895-y
View details for PubMedID 38030716
View details for PubMedCentralID 7734661
-
Skeletal Muscle Area on CT: Determination of an Optimal Height Scaling Power and Testing for Mortality Risk Prediction.
AJR. American journal of roentgenology
2023
Abstract
BACKGROUND: Sarcopenia is commonly assessed on CT using the skeletal muscle index (SMI), calculated as skeletal muscle area (SMA) at L3 divided by patient height squared (i.e., height scaling power of 2). OBJECTIVE: To determine the optimal height scaling power for SMA measurements on CT, and to test the influence of the derived optimal scaling power on the utility of SMI in predicting all-cause mortality. METHODS: This retrospective study included 16,575 patients (mean age, 56.4 years; 6985 men, 9590 women) who underwent abdominal CT from December 2012 through October 2018. SMA at L3 was determined using automated software. The sample was stratified into 5459 patients without major medical conditions (using ICD-9 and ICD-10 codes) for determining an optimal height scaling power, and 11,116 patients with major medical conditions for testing this power. The optimal scaling power was determined by allometric analysis (whereby regression coefficients were fitted to log-linear sex-specific models relating height to SMA) and by analysis of statistical independence of SMI from height across scaling powers. Cox proportional hazards models were used to test the derived optimal scaling power's influence on utility of SMI in predicting all-cause mortality. RESULTS: In allometric analysis, the regression coefficient of log(height) in patients ≤40 years was 1.02 in men and 1.08 in women, and in patients >40 years was 1.07 in men and 1.10 in women (all p<.05 vs regression coefficient of 2). In analyses for statistical independence of SMI from height, the optimal height scaling power (i.e., those yielding correlations closest to 0) was, in patients ≤40 years, 0.97 in men and 1.08 in women, and in patients >40 years, 1.03 in men and 1.09 in women. In the Cox model used for testing, SMI predicted all-cause mortality with greater concordance index using a height scaling power of 1 than 2 in men (0.675 vs 0.663, p<.001) and women (0.664 vs 0.653, p<.001). CONCLUSION: The findings support a height scaling power of 1, rather than conventional power of 2, for SMI computation. CLINICAL IMPACT: A revised height scaling power for SMI could impact the utility of CT-based sarcopenia diagnoses in risk assessment.
View details for DOI 10.2214/AJR.23.29889
View details for PubMedID 37877596
-
Noise2Recon: Enabling SNR-robust MRI reconstruction with semi-supervised and self-supervised learning.
Magnetic resonance in medicine
2023
Abstract
PURPOSE: To develop a method for building MRI reconstruction neural networks robust to changes in signal-to-noise ratio (SNR) and trainable with a limited number of fully sampled scans.METHODS: We propose Noise2Recon, a consistency training method for SNR-robust accelerated MRI reconstruction that can use both fully sampled (labeled) and undersampled (unlabeled) scans. Noise2Recon uses unlabeled data by enforcing consistency between model reconstructions of undersampled scans and their noise-augmented counterparts. Noise2Recon was compared to compressed sensing and both supervised and self-supervised deep learning baselines. Experiments were conducted using retrospectively accelerated data from the mridata three-dimensional fast-spin-echo knee and two-dimensional fastMRI brain datasets. All methods were evaluated in label-limited settings and among out-of-distribution (OOD) shifts, including changes in SNR, acceleration factors, and datasets. An extensive ablation study was conducted to characterize the sensitivity of Noise2Recon to hyperparameter choices.RESULTS: In label-limited settings, Noise2Recon achieved better structural similarity, peak signal-to-noise ratio, and normalized-RMS error than all baselines and matched performance of supervised models, which were trained with 14 * $$ 14\times $$ more fully sampled scans. Noise2Recon outperformed all baselines, including state-of-the-art fine-tuning and augmentation techniques, among low-SNR scans and when generalizing to OOD acceleration factors. Augmentation extent and loss weighting hyperparameters had negligible impact on Noise2Recon compared to supervised methods, which may indicate increased training stability.CONCLUSION: Noise2Recon is a label-efficient reconstruction method that is robust to distribution shifts, such as changes in SNR, acceleration factors, and others, with limited or no fully sampled training data.
View details for DOI 10.1002/mrm.29759
View details for PubMedID 37427449
-
Self-supervised learning for medical image classification: a systematic review and implementation guidelines.
NPJ digital medicine
2023; 6 (1): 74
Abstract
Advancements in deep learning and computer vision provide promising solutions for medical image analysis, potentially improving healthcare and patient outcomes. However, the prevailing paradigm of training deep learning models requires large quantities of labeled training data, which is both time-consuming and cost-prohibitive to curate for medical images. Self-supervised learning has the potential to make significant contributions to the development of robust medical imaging models through its ability to learn useful insights from copious medical datasets without labels. In this review, we provide consistent descriptions of different self-supervised learning strategies and compose a systematic review of papers published between 2012 and 2022 on PubMed, Scopus, and ArXiv that applied self-supervised learning to medical imaging classification. We screened a total of 412 relevant studies and included 79 papers for data extraction and analysis. With this comprehensive effort, we synthesize the collective knowledge of prior work and provide implementation guidelines for future researchers interested in applying self-supervised learning to their development of medical imaging classification models.
View details for DOI 10.1038/s41746-023-00811-0
View details for PubMedID 37100953
-
A scoping review of portable sensing for out-of-lab anterior cruciate ligament injury prevention and rehabilitation.
NPJ digital medicine
2023; 6 (1): 46
Abstract
Anterior cruciate ligament (ACL) injury and ACL reconstruction (ACLR) surgery are common. Laboratory-based biomechanical assessment can evaluate ACL injury risk and rehabilitation progress after ACLR; however, lab-based measurements are expensive and inaccessible to most people. Portable sensors such as wearables and cameras can be deployed during sporting activities, in clinics, and in patient homes. Although many portable sensing approaches have demonstrated promising results during various assessments related to ACL injury, they have not yet been widely adopted as tools for out-of-lab assessment. The purpose of this review is to summarize research on out-of-lab portable sensing applied to ACL and ACLR and offer our perspectives on new opportunities for future research and development. We identified 49 original research articles on out-of-lab ACL-related assessment; the most common sensing modalities were inertial measurement units, depth cameras, and RGB cameras. The studies combined portable sensors with direct feature extraction, physics-based modeling, or machine learning to estimate a range of biomechanical parameters (e.g., knee kinematics and kinetics) during jump-landing tasks, cutting, squats, and gait. Many of the reviewed studies depict proof-of-concept methods for potential future clinical applications including ACL injury risk screening, injury prevention training, and rehabilitation assessment. By synthesizing these results, we describe important opportunities that exist for clinical validation of existing approaches, using sophisticated modeling techniques, standardization of data collection, and creation of large benchmark datasets. If successful, these advances will enable widespread use of portable-sensing approaches to identify ACL injury risk factors, mitigate high-risk movements prior to injury, and optimize rehabilitation paradigms.
View details for DOI 10.1038/s41746-023-00782-2
View details for PubMedID 36934194
-
Improving Data-Efficiency and Robustness of Medical Imaging Segmentation Using Inpainting-Based Self-Supervised Learning.
Bioengineering (Basel, Switzerland)
2023; 10 (2)
Abstract
We systematically evaluate the training methodology and efficacy of two inpainting-based pretext tasks of context prediction and context restoration for medical image segmentation using self-supervised learning (SSL). Multiple versions of self-supervised U-Net models were trained to segment MRI and CT datasets, each using a different combination of design choices and pretext tasks to determine the effect of these design choices on segmentation performance. The optimal design choices were used to train SSL models that were then compared with baseline supervised models for computing clinically-relevant metrics in label-limited scenarios. We observed that SSL pretraining with context restoration using 32 × 32 patches and Poission-disc sampling, transferring only the pretrained encoder weights, and fine-tuning immediately with an initial learning rate of 1 × 10-3 provided the most benefit over supervised learning for MRI and CT tissue segmentation accuracy (p < 0.001). For both datasets and most label-limited scenarios, scaling the size of unlabeled pretraining data resulted in improved segmentation performance. SSL models pretrained with this amount of data outperformed baseline supervised models in the computation of clinically-relevant metrics, especially when the performance of supervised learning was low. Our results demonstrate that SSL pretraining using inpainting-based pretext tasks can help increase the robustness of models in label-limited scenarios and reduce worst-case errors that occur with supervised learning.
View details for DOI 10.3390/bioengineering10020207
View details for PubMedID 36829701
-
RaLEs: a Benchmark for Radiology Language Evaluations
NEURAL INFORMATION PROCESSING SYSTEMS (NIPS). 2023
View details for Web of Science ID 001228825101034
-
Generalizability of Deep Learning Segmentation Algorithms for Automated Assessment of Cartilage Morphology and MRI Relaxometry.
Journal of magnetic resonance imaging : JMRI
2022
Abstract
BACKGROUND: Deep learning (DL)-based automatic segmentation models can expedite manual segmentation yet require resource-intensive fine-tuning before deployment on new datasets. The generalizability of DL methods to new datasets without fine-tuning is not well characterized.PURPOSE: Evaluate the generalizability of DL-based models by deploying pretrained models on independent datasets varying by MR scanner, acquisition parameters, and subject population.STUDY TYPE: Retrospective based on prospectively acquired data.POPULATION: Overall test dataset: 59 subjects (26 females); Study 1: 5 healthy subjects (zero females), Study 2: 8 healthy subjects (eight females), Study 3: 10 subjects with osteoarthritis (eight females), Study 4: 36 subjects with various knee pathology (10 females).FIELD STRENGTH/SEQUENCE: A 3-T, quantitative double-echo steady state (qDESS).ASSESSMENT: Four annotators manually segmented knee cartilage. Each reader segmented one of four qDESS datasets in the test dataset. Two DL models, one trained on qDESS data and another on Osteoarthritis Initiative (OAI)-DESS data, were assessed. Manual and automatic segmentations were compared by quantifying variations in segmentation accuracy, volume, and T2 relaxation times for superficial and deep cartilage.STATISTICAL TESTS: Dice similarity coefficient (DSC) for segmentation accuracy. Lin's concordance correlation coefficient (CCC), Wilcoxon rank-sum tests, root-mean-squared error-coefficient-of-variation to quantify manual vs. automatic T2 and volume variations. Bland-Altman plots for manual vs. automatic T2 agreement. A P value<0.05 was considered statistically significant.RESULTS: DSCs for the qDESS-trained model, 0.79-0.93, were higher than those for the OAI-DESS-trained model, 0.59-0.79. T2 and volume CCCs for the qDESS-trained model, 0.75-0.98 and 0.47-0.95, were higher than respective CCCs for the OAI-DESS-trained model, 0.35-0.90 and 0.13-0.84. Bland-Altman 95% limits of agreement for superficial and deep cartilage T2 were lower for the qDESS-trained model, ±2.4msec and ±4.0msec, than the OAI-DESS-trained model, ±4.4msec and ±5.2msec.DATA CONCLUSION: The qDESS-trained model may generalize well to independent qDESS datasets regardless of MR scanner, acquisition parameters, and subject population.EVIDENCE LEVEL: 1 TECHNICAL EFFICACY: Stage 1.
View details for DOI 10.1002/jmri.28365
View details for PubMedID 35852498
-
Opportunistic Incidence Prediction of Multiple Chronic Diseases from Abdominal CT Imaging Using Multi-task Learning
SPRINGER INTERNATIONAL PUBLISHING AG. 2022: 309-318
View details for DOI 10.1007/978-3-031-16449-1_30
View details for Web of Science ID 000867568000030
-
Low-count whole-body PET with deep learning in a multicenter and externally validated study.
NPJ digital medicine
2021; 4 (1): 127
Abstract
More widespread use of positron emission tomography (PET) imaging is limited by its high cost and radiation dose. Reductions in PET scan time or radiotracer dosage typically degrade diagnostic image quality (DIQ). Deep-learning-based reconstruction may improve DIQ, but such methods have not been clinically evaluated in a realistic multicenter, multivendor environment. In this study, we evaluated the performance and generalizability of a deep-learning-based image-quality enhancement algorithm applied to fourfold reduced-count whole-body PET in a realistic clinical oncologic imaging environment with multiple blinded readers, institutions, and scanner types. We demonstrate that the low-count-enhanced scans were noninferior to the standard scans in DIQ (p<0.05) and overall diagnostic confidence (p<0.001) independent of the underlying PET scanner used. Lesion detection for the low-count-enhanced scans had a high patient-level sensitivity of 0.94 (0.83-0.99) and specificity of 0.98 (0.95-0.99). Interscan kappa agreement of 0.85 was comparable to intrareader (0.88) and pairwise inter-reader agreements (maximum of 0.72). SUV quantification was comparable in the reference regions and lesions (lowest p-value=0.59) and had high correlation (lowest CCC=0.94). Thus, we demonstrated that deep learning can be used to restore diagnostic image quality and maintain SUV accuracy for fourfold reduced-count PET scans, with interscan variations in lesion depiction, lower than intra- and interreader variations. This method generalized to an external validation set of clinical patients from multiple institutions and scanner types. Overall, this method may enable either dose or exam-duration reduction, increasing safety and lowering the cost of PET imaging.
View details for DOI 10.1038/s41746-021-00497-2
View details for PubMedID 34426629
-
The International Workshop on Osteoarthritis Imaging Knee MRI Segmentation Challenge: A Multi-Institute Evaluation and Analysis Framework on a Standardized Dataset.
Radiology. Artificial intelligence
2021; 3 (3): e200078
Abstract
Purpose: To organize a multi-institute knee MRI segmentation challenge for characterizing the semantic and clinical efficacy of automatic segmentation methods relevant for monitoring osteoarthritis progression.Materials and Methods: A dataset partition consisting of three-dimensional knee MRI from 88 retrospective patients at two time points (baseline and 1-year follow-up) with ground truth articular (femoral, tibial, and patellar) cartilage and meniscus segmentations was standardized. Challenge submissions and a majority-vote ensemble were evaluated against ground truth segmentations using Dice score, average symmetric surface distance, volumetric overlap error, and coefficient of variation on a holdout test set. Similarities in automated segmentations were measured using pairwise Dice coefficient correlations. Articular cartilage thickness was computed longitudinally and with scans. Correlation between thickness error and segmentation metrics was measured using the Pearson correlation coefficient. Two empirical upper bounds for ensemble performance were computed using combinations of model outputs that consolidated true positives and true negatives.Results: Six teams (T 1-T 6) submitted entries for the challenge. No differences were observed across any segmentation metrics for any tissues (P = .99) among the four top-performing networks (T 2, T 3, T 4, T 6). Dice coefficient correlations between network pairs were high (> 0.85). Per-scan thickness errors were negligible among networks T 1-T 4 (P = .99), and longitudinal changes showed minimal bias (< 0.03 mm). Low correlations (rho < 0.41) were observed between segmentation metrics and thickness error. The majority-vote ensemble was comparable to top-performing networks (P = .99). Empirical upper-bound performances were similar for both combinations (P = .99).Conclusion: Diverse networks learned to segment the knee similarly, where high segmentation accuracy did not correlate with cartilage thickness accuracy and voting ensembles did not exceed individual network performance.See also the commentary by Elhalawani and Mak in this issue.Keywords: Cartilage, Knee, MR-Imaging, Segmentation © RSNA, 2020Supplemental material is available for this article.
View details for DOI 10.1148/ryai.2021200078
View details for PubMedID 34235438
-
Accuracy and longitudinal reproducibility of quantitative femorotibial cartilage measures derived from automated U-Net-based segmentation of two different MRI contrasts: data from the osteoarthritis initiative healthy reference cohort.
Magma (New York, N.Y.)
2020
Abstract
OBJECTIVE: To evaluate the agreement, accuracy, and longitudinal reproducibility of quantitative cartilage morphometry from 2D U-Net-based automated segmentations for 3T coronal fast low angle shot (corFLASH) and sagittal double echo at steady-state (sagDESS) MRI.METHODS: 2D U-Nets were trained using manual, quality-controlled femorotibial cartilage segmentations available for 92 Osteoarthritis Initiative healthy reference cohort participants from both corFLASH and sagDESS (n=50/21/21 training/validation/test-set). Cartilage morphometry was computed from automated and manual segmentations for knees from the test-set. Agreement and accuracy were evaluated from baseline visits (dice similarity coefficient: DSC, correlation analysis, systematic offset). The longitudinal reproducibility was assessed from year-1 and -2 follow-up visits (root-mean-squared coefficient of variation, RMSCV%).RESULTS: Automated segmentations showed high agreement (DSC 0.89-0.92) and high correlations (r≥0.92) with manual ground truth for both corFLASH and sagDESS and only small systematic offsets (≤10.1%). The automated measurements showed a similar test-retest reproducibility over 1year (RMSCV% 1.0-4.5%) as manual measurements (RMSCV% 0.5-2.5%).DISCUSSION: The 2D U-Net-based automated segmentation method yielded high agreement compared with manual segmentation and also demonstrated high accuracy and longitudinal test-retest reproducibility for morphometric analysis of articular cartilage derived from it, using both corFLASH and sagDESS.
View details for DOI 10.1007/s10334-020-00889-7
View details for PubMedID 33025284
-
Prospective Deployment of Deep Learning in MRI: A Framework for Important Considerations, Challenges, and Recommendations for Best Practices.
Journal of magnetic resonance imaging : JMRI
2020
Abstract
Artificial intelligence algorithms based on principles of deep learning (DL) have made a large impact on the acquisition, reconstruction, and interpretation of MRI data. Despite the large number of retrospective studies using DL, there are fewer applications of DL in the clinic on a routine basis. To address this large translational gap, we review the recent publications to determine three major use cases that DL can have in MRI, namely, that of model-free image synthesis, model-based image reconstruction, and image or pixel-level classification. For each of these three areas, we provide a framework for important considerations that consist of appropriate model training paradigms, evaluation of model robustness, downstream clinical utility, opportunities for future advances, as well recommendations for best current practices. We draw inspiration for this framework from advances in computer vision in natural imaging as well as additional healthcare fields. We further emphasize the need for reproducibility of research studies through the sharing of datasets and software. LEVEL OF EVIDENCE: 5 TECHNICAL EFFICACY STAGE: 2.
View details for DOI 10.1002/jmri.27331
View details for PubMedID 32830874
-
Diagnostic Accuracy of Quantitative Multi-Contrast 5-Minute Knee MRI Using Prospective Artificial Intelligence Image Quality Enhancement.
AJR. American journal of roentgenology
2020
Abstract
Potential approaches for abbreviated knee MRI, including prospective acceleration with deep learning, have achieved limited clinical implementation to date.The objective of this study was to evaluate the inter-reader agreement of conventional knee MRI and a 5-minute 3D quantitative double-echo steady-state (qDESS) sequence with automatic T2 mapping and deep-learning super-resolution (DLSR) augmentation, as well as to compare the diagnostic performance of the two methods with respect to findings from arthroscopic surgery.A total of 51 patients with knee pain underwent knee MRI that included an additional 3D qDESS sequence with automatic T2 mapping. Fourier interpolation was followed by prospective DLSR to enhance qDESS slice-resolution twofold. A musculoskeletal radiologist and a radiology resident performed retrospective independent evaluations of the articular cartilage, menisci, ligaments, bones, extensor mechanism, and synovium using conventional MRI. Following a two-month washout period, the readers reviewed qDESS images alone, followed by qDESS with the automatic T2 maps. Inter-reader agreement between conventional MRI and qDESS was computed using percent agreement and Cohen's Kappa. The sensitivity and specificity of conventional MRI, qDESS alone, and qDESS+T2 were compared with arthroscopic findings using exact McNemar's tests.Conventional MRI and qDESS demonstrated 92% agreement in evaluation of articular cartilage, menisci, ligaments, bones, extensor mechanism, and synovium combined. Kappa was 0.79 (0.76-0.81) across all imaging findings. In the 43/51 patients who underwent arthroscopy, sensitivity and specificity were not significantly different (p=0.23-1.00) between conventional MRI (sensitivity: 58%-93%; specificity: 27%-87%) and qDESS alone (sensitivity: 54%-90%; specificity: 23%-91%) for cartilage, menisci, ligaments, and synovium. Sensitivity and specificity for grade 1 cartilage lesions were 33%/56% for conventional MRI, 23%/53% for qDESS (p=0.81), and 46%/39% for qDESS+T2 (p=0.80); for grade 2A lesions, 27%/53% for conventional MRI, 26%/52% for qDESS (p=0.02), and 58%/40% for qDESS+T2 (p<0.001).qDESS prospectively enhanced with deep learning had strong inter-reader agreement with conventional knee MRI and near-equivalent diagnostic performance with respect to arthroscopy. The ability of qDESS to automatically generate T2 maps increases sensitivity for cartilage abnormalities. Clinical Impact: qDESS using prospective artificial intelligence image quality enhancement may facilitate an abbreviated knee MRI protocol while generating quantitative T2 maps.
View details for DOI 10.2214/AJR.20.24172
View details for PubMedID 32755384
-
Rapid Knee MRI Acquisition and Analysis Techniques for Imaging Osteoarthritis.
Journal of magnetic resonance imaging : JMRI
2019
Abstract
Osteoarthritis (OA) of the knee is a major source of disability that has no known treatment or cure. Morphological and compositional MRI is commonly used for assessing the bone and soft tissues in the knee to enhance the understanding of OA pathophysiology. However, it is challenging to extend these imaging methods and their subsequent analysis techniques to study large population cohorts due to slow and inefficient imaging acquisition and postprocessing tools. This can create a bottleneck in assessing early OA changes and evaluating the responses of novel therapeutics. The purpose of this review article is to highlight recent developments in tools for enhancing the efficiency of knee MRI methods useful to study OA. Advances in efficient MRI data acquisition and reconstruction tools for morphological and compositional imaging, efficient automated image analysis tools, and hardware improvements to further drive efficient imaging are discussed in this review. For each topic, we discuss the current challenges as well as potential future opportunities to alleviate these challenges. LEVEL OF EVIDENCE: 5 TECHNICAL EFFICACY STAGE: 3.
View details for DOI 10.1002/jmri.26991
View details for PubMedID 31755191
-
Utility of deep learning super-resolution in the context of osteoarthritis MRI biomarkers.
Journal of magnetic resonance imaging : JMRI
2019
Abstract
BACKGROUND: Super-resolution is an emerging method for enhancing MRI resolution; however, its impact on image quality is still unknown.PURPOSE: To evaluate MRI super-resolution using quantitative and qualitative metrics of cartilage morphometry, osteophyte detection, and global image blurring.STUDY TYPE: Retrospective.POPULATION: In all, 176 MRI studies of subjects at varying stages of osteoarthritis.FIELD STRENGTH/SEQUENCE: Original-resolution 3D double-echo steady-state (DESS) and DESS with 3* thicker slices retrospectively enhanced using super-resolution and tricubic interpolation (TCI) at 3T.ASSESSMENT: A quantitative comparison of femoral cartilage morphometry was performed for the original-resolution DESS, the super-resolution, and the TCI scans in 17 subjects. A reader study by three musculoskeletal radiologists assessed cartilage image quality, overall image sharpness, and osteophytes incidence in all three sets of scans. A referenceless blurring metric evaluated blurring in all three image dimensions for the three sets of scans.STATISTICAL TESTS: Mann-Whitney U-tests compared Dice coefficients (DC) of segmentation accuracy for the DESS, super-resolution, and TCI images, along with the image quality readings and blurring metrics. Sensitivity, specificity, and diagnostic odds ratio (DOR) with 95% confidence intervals compared osteophyte detection for the super-resolution and TCI images, with the original-resolution as a reference.RESULTS: DC for the original-resolution (90.2±1.7%) and super-resolution (89.6±2.0%) were significantly higher (P<0.001) than TCI (86.3±5.6%). Segmentation overlap of super-resolution with the original-resolution (DC = 97.6±0.7%) was significantly higher (P<0.0001) than TCI overlap (DC = 95.0±1.1%). Cartilage image quality for sharpness and contrast levels, and the through-plane quantitative blur factor for super-resolution images, was significantly (P<0.001) better than TCI. Super-resolution osteophyte detection sensitivity of 80% (76-82%), specificity of 93% (92-94%), and DOR of 32 (22-46) was significantly higher (P<0.001) than TCI sensitivity of 73% (69-76%), specificity of 90% (89-91%), and DOR of 17 (13-22).DATA CONCLUSION: Super-resolution appears to consistently outperform naive interpolation and may improve image quality without biasing quantitative biomarkers.LEVEL OF EVIDENCE: 2 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2019.
View details for DOI 10.1002/jmri.26872
View details for PubMedID 31313397
-
Combined 5-minute double-echo in steady-state with separated echoes and 2-minute proton-density-weighted 2D FSE sequence for comprehensive whole-joint knee MRI assessment
JOURNAL OF MAGNETIC RESONANCE IMAGING
2019; 49 (7): E183–E194
View details for DOI 10.1002/jmri.26582
View details for Web of Science ID 000474612300018
-
Super-resolution musculoskeletal MRI using deep learning.
Magnetic resonance in medicine
2018
Abstract
PURPOSE: To develop a super-resolution technique using convolutional neural networks for generating thin-slice knee MR images from thicker input slices, and compare this method with alternative through-plane interpolation methods.METHODS: We implemented a 3D convolutional neural network entitled DeepResolve to learn residual-based transformations between high-resolution thin-slice images and lower-resolution thick-slice images at the same center locations. DeepResolve was trained using 124 double echo in steady-state (DESS) data sets with 0.7-mm slice thickness and tested on 17 patients. Ground-truth images were compared with DeepResolve, clinically used tricubic interpolation, and Fourier interpolation methods, along with state-of-the-art single-image sparse-coding super-resolution. Comparisons were performed using structural similarity, peak SNR, and RMS error image quality metrics for a multitude of thin-slice downsampling factors. Two musculoskeletal radiologists ranked the 3 data sets and reviewed the diagnostic quality of the DeepResolve, tricubic interpolation, and ground-truth images for sharpness, contrast, artifacts, SNR, and overall diagnostic quality. Mann-Whitney U tests evaluated differences among the quantitative image metrics, reader scores, and rankings. Cohen's Kappa (kappa) evaluated interreader reliability.RESULTS: DeepResolve had significantly better structural similarity, peak SNR, and RMS error than tricubic interpolation, Fourier interpolation, and sparse-coding super-resolution for all downsampling factors (p<.05, except 4*and 8*sparse-coding super-resolution downsampling factors). In the reader study, DeepResolve significantly outperformed (p<.01) tricubic interpolation in all image quality categories and overall image ranking. Both readers had substantial scoring agreement (kappa=0.73).CONCLUSION: DeepResolve was capable of resolving high-resolution thin-slice knee MRI from lower-resolution thicker slices, achieving superior quantitative and qualitative diagnostic performance to both conventionally used and state-of-the-art methods.
View details for PubMedID 29582464
-
Five-minute knee MRI for simultaneous morphometry and T2 relaxometry of cartilage and meniscus and for semiquantitative radiological assessment using double-echo in steady-state at 3T.
Journal of magnetic resonance imaging : JMRI
2018; 47 (5): 1328–41
Abstract
Biomarkers for assessing osteoarthritis activity necessitate multiple MRI sequences with long acquisition times.To perform 5-minute simultaneous morphometry (thickness/volume measurements) and T2 relaxometry of both cartilage and meniscus, and semiquantitative MRI Osteoarthritis Knee Scoring (MOAKS).Prospective.Fifteen healthy volunteers for morphometry and T2 measurements, and 15 patients (five each Kellgren-Lawrence grades 0/2/3) for MOAKS assessment.A 5-minute double-echo steady-state (DESS) sequence was evaluated for generating quantitative and semiquantitative osteoarthritis biomarkers at 3T.Flip angle simulations evaluated tissue signals and sensitivity of T2 measurements. Morphometry and T2 reproducibility was compared against morphometry-optimized and relaxometry-optimized sequences. Repeatability was assessed by scanning five volunteers twice. MOAKS reproducibility was compared to MOAKS derived from a clinical knee MRI protocol by two readers.Coefficients of variation (CVs), concordance confidence intervals (CCI), and Wilcoxon signed-rank tests compared morphometry and relaxometry measurements with their reference standards. DESS MOAKS positive percent agreement (PPA), negative percentage agreement (NPA), and interreader agreement was calculated using the clinical protocol as a reference. Biomarker variations between Kellgren-Lawrence groups were evaluated using Wilcoxon rank-sum tests.Cartilage thickness (P = 0.65), cartilage T2 (P = 0.69), and meniscus T2 (P = 0.06) did not significantly differ from their reference standard (with a 20° DESS flip angle). DESS slightly overestimated meniscus volume (P < 0.001). Accuracy and repeatability CVs were <3.3%, except the meniscus T2 accuracy (7.6%). DESS MOAKS had substantial interreader agreement and high PPA/NPA values of 87%/90%. Bone marrow lesions and menisci had slightly lower PPAs. Cartilage and meniscus T2 , and MOAKS (cartilage surface area, osteophytes, cysts, and total score) was higher in Kellgren-Lawrence groups 2 and 3 than group 0 (P < 0.05).The 5-minute DESS sequence permits MOAKS assessment for a majority of tissues, along with repeatable and reproducible simultaneous cartilage and meniscus T2 relaxometry and morphometry measurements.2 Technical Efficacy: Stage 1 J. Magn. Reson. Imaging 2018;47:1328-1341.
View details for PubMedID 29090500
View details for PubMedCentralID PMC5899635
-
Deep Learning Super-Resolution Enables Rapid Simultaneous Morphological and Quantitative Magnetic Resonance Imaging
SPRINGER INTERNATIONAL PUBLISHING AG. 2018: 3–11
View details for DOI 10.1007/978-3-030-00129-2_1
View details for Web of Science ID 000477767500001
-
connective tissues in the knee using ultrashort echo-time double-echo steady-state (UTEDESS).
Magnetic resonance in medicine
2017
Abstract
To develop a radial, double-echo steady-state (DESS) sequence with ultra-short echo-time (UTE) capabilities for T2 measurement of short-T2 tissues along with simultaneous rapid, signal-to-noise ratio (SNR)-efficient, and high-isotropic-resolution morphological knee imaging.THe 3D radial UTE readouts were incorporated into DESS, termed UTEDESS. Multiple-echo-time UTEDESS was used for performing T2 relaxometry for short-T2 tendons, ligaments, and menisci; and for Dixon water-fat imaging. In vivo T2 estimate repeatability and SNR efficiency for UTEDESS and Cartesian DESS were compared. The impact of coil combination methods on short-T2 measurements was evaluated by means of simulations. UTEDESS T2 measurements were compared with T2 measurements from Cartesian DESS, multi-echo spin-echo (MESE), and fast spin-echo (FSE).UTEDESS produced isotropic resolution images with high SNR efficiency in all short-T2 tissues. Simulations and experiments demonstrated that sum-of-squares coil combinations overestimated short-T2 measurements. UTEDESS measurements of meniscal T2 were comparable to DESS, MESE, and FSE measurements while the tendon and ligament measurements were less biased than those from Cartesian DESS. Average UTEDESS T2 repeatability variation was under 10% in all tissues.The T2 measurements of short-T2 tissues and high-resolution morphological imaging provided by UTEDESS makes it promising for studying the whole knee, both in routine clinical examinations and longitudinal studies. Magn Reson Med, 2017. © 2017 International Society for Magnetic Resonance in Medicine.
View details for DOI 10.1002/mrm.26577
View details for PubMedID 28074498
-
Spectral Graph Sample Weighting for Interpretable Sub-cohort Analysis in Predictive Models for Neuroimaging.
PRedictive Intelligence in MEdicine. PRIME (Workshop)
2025; 15155: 24-34
Abstract
Recent advancements in medicine have confirmed that brain disorders often comprise multiple subtypes of mechanisms, developmental trajectories, or severity levels. Such heterogeneity is often associated with demographic aspects (e.g., sex) or disease-related contributors (e.g., genetics). Thus, the predictive power of machine learning models used for symptom prediction varies across subjects based on such factors. To model this heterogeneity, one can assign each training sample a factor-dependent weight, which modulates the subject's contribution to the overall objective loss function. To this end, we propose to model the subject weights as a linear combination of the eigenbases of a spectral population graph that captures the similarity of factors across subjects. In doing so, the learned weights smoothly vary across the graph, highlighting sub-cohorts with high and low predictability. Our proposed sample weighting scheme is evaluated on two tasks. First, we predict initiation of heavy alcohol drinking in young adulthood from imaging and neuropsychological measures from the National Consortium on Alcohol and NeuroDevelopment in Adolescence (NCANDA). Next, we detect Dementia vs. Mild Cognitive Impairment (MCI) using imaging and demographic measurements in subjects from the Alzheimer's Disease Neuroimaging Initiative (ADNI). Compared to existing sample weighting schemes, our sample weights improve interpretability and highlight sub-cohorts with distinct characteristics and varying model accuracy.
View details for DOI 10.1007/978-3-031-74561-4_3
View details for PubMedID 39525051
View details for PubMedCentralID PMC11549025
-
Artificial intelligence tools trained on human-labeled data reflect human biases: a case study in a large clinical consecutive knee osteoarthritis cohort.
Scientific reports
2024; 14 (1): 26782
Abstract
Humans have been shown to have biases when reading medical images, raising questions about whether humans are uniform in their disease gradings. Artificial intelligence (AI) tools trained on human-labeled data may have inherent human non-uniformity. In this study, we used a radiographic knee osteoarthritis external validation dataset of 50 patients and a six-year retrospective consecutive clinical cohort of 8,273 patients. An FDA-approved and CE-marked AI tool was tested for potential non-uniformity in Kellgren-Lawrence grades between the right and left sides of the images. We flipped the images horizontally so that a left knee looked like a right knee and vice versa. According to human review, the AI tool showed non-uniformity with 20-22% disagreements on the external validation dataset and 13.6% on the cohort. However, we found no evidence of a significant difference in the accuracy compared to senior radiologists on the external validation dataset, or age bias or sex bias on the cohort. AI non-uniformity can boost the evaluated performance against humans, but image areas with inferior performance should be investigated.
View details for DOI 10.1038/s41598-024-75752-z
View details for PubMedID 39500908
-
Multiscale hamstring muscle adaptations following 9 weeks of eccentric training.
Journal of sport and health science
2024: 100996
Abstract
Eccentric training, such as Nordic hamstring exercise (NHE) training, is commonly used as a preventive measure for hamstring strains. Eccentric training is believed to induce lengthening of muscle fascicles and to be associated with the addition of sarcomeres in series within muscle fibers. However, the difficulty in measuring sarcomere adaptation in human muscles has severely limited information about the precise mechanisms of adaptation. This study addressed this limitation by measuring the multiscale hamstring muscle adaptations in response to 9 weeks of NHE training and 3 weeks of detraining.Twelve participants completed 9 weeks of supervised NHE training, followed by a 3-week detraining period. We assessed biceps femoris long-head (BFlh) muscle fascicle length, sarcomere length, and serial sarcomere number in the central and distal regions of the muscle. Additionally, we measured muscle volume and eccentric strength at baseline, post-training, and post-detraining.NHE training over 9 weeks induced significant architectural and strength adaptations in the BFlh muscle. Fascicle length increased by 19% in the central muscle region and 33% in the distal muscle region. NHE also induced increases in serial sarcomere number (25% in the central region and 49% in the distal region). BFlh muscle volume increased by 8%, and knee flexion strength increased by 40% with training. Following 3 weeks of detraining, fascicle length decreased by 12% in the central region and 16% in the distal region along with reductions in serial sarcomere number.Nine weeks of NHE training produced substantial, region-specific increases in BFlh muscle fascicle length, muscle volume, and force generation. The direct measurement of sarcomere lengths revealed that the increased fascicle length was accompanied by the addition of sarcomeres in series within the muscle fascicles.
View details for DOI 10.1016/j.jshs.2024.100996
View details for PubMedID 39461588
-
Quantification Of 3D Knee Morphology In Patients With Patellar Instability
LIPPINCOTT WILLIAMS & WILKINS. 2024: 61-62
View details for Web of Science ID 001315123200134
-
Evaluation of an automated laminar cartilage T2 relaxation time analysis method in an early osteoarthritis model.
Skeletal radiology
2024
Abstract
A fully automated laminar cartilage composition (MRI-based T2) analysis method was technically and clinically validated by comparing radiographically normal knees with (CL-JSN) and without contra-lateral joint space narrowing or other signs of radiographic osteoarthritis (OA, CL-noROA).2D U-Nets were trained from manually segmented femorotibial cartilages (n = 72) from all 7 echoes (AllE), or from the 1st echo only (1stE) of multi-echo-spin-echo (MESE) MRIs acquired by the Osteoarthritis Initiative (OAI). Because of its greater accuracy, only the AllE U-Net was then applied to knees from the OAI healthy reference cohort (n = 10), CL-JSN (n = 39), and (1:1) matched CL-noROA knees (n = 39) that all had manual expert segmentation, and to 982 non-matched CL-noROA knees without expert segmentation.The agreement (Dice similarity coefficient) between automated vs. manual expert cartilage segmentation was between 0.82 ± 0.05/0.79 ± 0.06 (AllE/1stE) and 0.88 ± 0.03/0.88 ± 0.03 (AllE/1stE) across femorotibial cartilage plates. The deviation between automated vs. manually derived laminar T2 reached up to - 2.2 ± 2.6 ms/ + 4.1 ± 10.2 ms (AllE/1stE). The AllE U-Net showed a similar sensitivity to cross-sectional laminar T2 differences between CL-JSN and CL-noROA knees in the matched (Cohen's D ≤ 0.54) and the non-matched (D ≤ 0.54) comparison as the matched manual analyses (D ≤ 0.48). Longitudinally, the AllE U-Net also showed a similar sensitivity to CL-JSN vs. CS-noROA differences in the matched (D ≤ 0.51) and the non-matched (D ≤ 0.43) comparison as matched manual analyses (D ≤ 0.41).The fully automated T2 analysis showed a high agreement, acceptable accuracy, and similar sensitivity to cross-sectional and longitudinal laminar T2 differences in an early OA model, compared with manual expert analysis.Clinicaltrials.gov identification: NCT00080171.
View details for DOI 10.1007/s00256-024-04786-1
View details for PubMedID 39230576
View details for PubMedCentralID 3485166
-
Generating Synthetic Data for Medical Imaging.
Radiology
2024; 312 (3): e232471
Abstract
Artificial intelligence (AI) models for medical imaging tasks, such as classification or segmentation, require large and diverse datasets of images. However, due to privacy and ethical issues, as well as data sharing infrastructure barriers, these datasets are scarce and difficult to assemble. Synthetic medical imaging data generated by AI from existing data could address this challenge by augmenting and anonymizing real imaging data. In addition, synthetic data enable new applications, including modality translation, contrast synthesis, and professional training for radiologists. However, the use of synthetic data also poses technical and ethical challenges. These challenges include ensuring the realism and diversity of the synthesized images while keeping data unidentifiable, evaluating the performance and generalizability of models trained on synthetic data, and high computational costs. Since existing regulations are not sufficient to guarantee the safe and ethical use of synthetic images, it becomes evident that updated laws and more rigorous oversight are needed. Regulatory bodies, physicians, and AI developers should collaborate to develop, maintain, and continually refine best practices for synthetic data. This review aims to provide an overview of the current knowledge of synthetic data in medical imaging and highlights current key challenges in the field to guide future research and development.
View details for DOI 10.1148/radiol.232471
View details for PubMedID 39254456
-
The design of a sample rapid magnetic resonance imaging (MRI) acquisition protocol supporting assessment of multiple articular tissues and pathologies in knee osteoarthritis.
Osteoarthritis and cartilage open
2024; 6 (3): 100505
Abstract
This expert opinion paper proposes a design for a state-of-the-art magnetic resonance image (MRI) acquisition protocol for knee osteoarthritis clinical trials in early and advanced disease. Semi-quantitative and quantitative imaging endpoints are supported, partly amendable to automated analysis. Several (peri-) articular tissues and pathologies are covered, including synovitis.A PubMed literature search was conducted, with focus on the past 5 years. Further, osteoarthritis imaging experts provided input. Specific MRI sequences, orientations, spatial resolutions and parameter settings were identified to align with study goals. We strived for implementation on standard clinical scanner hardware, with a net acquisition time ≤30 min.Short- and long-term longitudinal MRIs should be obtained at ≥1.5T, if possible without hardware changes during the study. We suggest a series of gradient- and spin-echo-sequences, supporting MOAKS, quantitative analysis of cartilage morphology and T2, and non-contrast-enhanced depiction of synovitis. These sequences should be properly aligned and positioned using localizer images. One of the sequences may be repeated in each participant (re-test), optimally at baseline and follow-up, to estimate within-study precision. All images should be checked for quality and protocol-adherence as soon as possible after acquisition. Alternative approaches are suggested that expand on the structural endpoints presented.We aim to bridge the gap between technical MRI acquisition guides and the wealth of imaging literature, proposing a balance between image acquisition efficiency (time), safety, and technical/methodological diversity. This approach may entertain scientific innovation on tissue structure and composition assessment in clinical trials on disease modification of knee osteoarthritis.
View details for DOI 10.1016/j.ocarto.2024.100505
View details for PubMedID 39183946
View details for PubMedCentralID PMC11342198
-
Deep learning for accelerated and robust MRI reconstruction.
Magma (New York, N.Y.)
2024
Abstract
Deep learning (DL) has recently emerged as a pivotal technology for enhancing magnetic resonance imaging (MRI), a critical tool in diagnostic radiology. This review paper provides a comprehensive overview of recent advances in DL for MRI reconstruction, and focuses on various DL approaches and architectures designed to improve image quality, accelerate scans, and address data-related challenges. It explores end-to-end neural networks, pre-trained and generative models, and self-supervised methods, and highlights their contributions to overcoming traditional MRI limitations. It also discusses the role of DL in optimizing acquisition protocols, enhancing robustness against distribution shifts, and tackling biases. Drawing on the extensive literature and practical insights, it outlines current successes, limitations, and future directions for leveraging DL in MRI reconstruction, while emphasizing the potential of DL to significantly impact clinical imaging practices.Affiliations [3 and 6] has been split into two different affiliations. Please check if action taken is appropriate and amend if necessary.looks good.
View details for DOI 10.1007/s10334-024-01173-8
View details for PubMedID 39042206
View details for PubMedCentralID 4459721
-
Marker Data Enhancement For Markerless Motion Capture.
bioRxiv : the preprint server for biology
2024
Abstract
Human pose estimation models can measure movement from videos at a large scale and low cost; however, open-source pose estimation models typically detect only sparse keypoints, which leads to inaccurate joint kinematics. OpenCap, a freely available service for researchers to measure movement from videos, addresses this issue using a deep learning model-the marker enhancer-that transforms sparse keypoints into dense anatomical markers. However, OpenCap performs poorly on movements not included in the training data. Here, we create a much larger and more diverse training dataset and develop a more accurate and generalizable marker enhancer.We compiled marker-based motion capture data from 1176 subjects and synthesized 1433 hours of keypoints and anatomical markers to train the marker enhancer. We evaluated its accuracy in computing kinematics using both benchmark movement videos and synthetic data representing unseen, diverse movements.The marker enhancer improved kinematic accuracy on benchmark movements (mean error: 4.1°, max: 8.7°) compared to using video keypoints (mean: 9.6°, max: 43.1°) and OpenCap's original enhancer (mean: 5.3°, max: 11.5°). It also better generalized to unseen, diverse movements (mean: 4.1°, max: 6.7°) than OpenCap's original enhancer (mean: 40.4°, max: 252.0°).Our marker enhancer demonstrates both accuracy and generalizability across diverse movements.We integrated the marker enhancer into OpenCap, thereby offering its thousands of users more accurate measurements across a broader range of movements.
View details for DOI 10.1101/2024.07.13.603382
View details for PubMedID 39071421
View details for PubMedCentralID PMC11275905
-
Sociodemographic Differences Among Patients Receiving Coronary Artery Calcium Imaging vs Nongated Chest Computed Tomography Imaging.
JACC. Advances
2024; 3 (7): 100963
View details for DOI 10.1016/j.jacadv.2024.100963
View details for PubMedID 39129975
View details for PubMedCentralID PMC11312303
-
Applications of Artificial Intelligence for Pediatric Cancer Imaging.
AJR. American journal of roentgenology
2024
Abstract
Artificial intelligence (AI) is transforming medical imaging of adult patients. However, its utilization in pediatric oncology imaging remains constrained, in part due to the inherent data scarcity associated with childhood cancers. Pediatric cancers are rare, and imaging technologies are evolving rapidly, leading to insufficient data of a particular type to effectively train these algorithms. The small market size of pediatrics compared to adults could also contribute to this challenge, as market size is a driver of commercialization. This article provides an overview of the current state of AI applications for pediatric cancer imaging, including applications for medical image acquisition, processing, reconstruction, segmentation, diagnosis, staging, and treatment response monitoring. While current developments are promising, impediments due to diverse anatomies of growing children and nonstandardized imaging protocols have led to limited clinical translation thus far. Opportunities include leveraging reconstruction algorithms to achieve accelerated low-dose imaging and automating the generation of metric-based staging and treatment monitoring scores. Transfer-learning of adult-based AI models to pediatric cancers, multi-institutional data sharing, and ethical data privacy practices for pediatric patients with rare cancers will be keys to unlocking AI's full potential for clinical translation and improved outcomes for these young patients.
View details for DOI 10.2214/AJR.24.31076
View details for PubMedID 38809123
-
Missing Wedge Completion via Unsupervised Learning with Coordinate Networks.
International journal of molecular sciences
2024; 25 (10)
Abstract
Cryogenic electron tomography (cryoET) is a powerful tool in structural biology, enabling detailed 3D imaging of biological specimens at a resolution of nanometers. Despite its potential, cryoET faces challenges such as the missing wedge problem, which limits reconstruction quality due to incomplete data collection angles. Recently, supervised deep learning methods leveraging convolutional neural networks (CNNs) have considerably addressed this issue; however, their pretraining requirements render them susceptible to inaccuracies and artifacts, particularly when representative training data is scarce. To overcome these limitations, we introduce a proof-of-concept unsupervised learning approach using coordinate networks (CNs) that optimizes network weights directly against input projections. This eliminates the need for pretraining, reducing reconstruction runtime by 3-20× compared to supervised methods. Our in silico results show improved shape completion and reduction of missing wedge artifacts, assessed through several voxel-based image quality metrics in real space and a novel directional Fourier Shell Correlation (FSC) metric. Our study illuminates benefits and considerations of both supervised and unsupervised approaches, guiding the development of improved reconstruction strategies.
View details for DOI 10.3390/ijms25105473
View details for PubMedID 38791508
-
Reproducibility of Quantitative Double-Echo Steady-State T2 Mapping of Knee Cartilage.
Journal of magnetic resonance imaging : JMRI
2024
Abstract
Cartilage T2 can detect joints at risk of developing osteoarthritis. The quantitative double-echo steady state (qDESS) sequence is attractive for knee cartilage T2 mapping because of its acquisition time of under 5 minutes. Understanding the reproducibility errors associated with qDESS T2 is essential to profiling the technical performance of this biomarker.To examine the combined acquisition and segmentation reproducibility of knee cartilage qDESS T2 using two different regional analysis schemes: 1) manual segmentation of subregions loaded during common activities and 2) automatic subregional segmentation.Prospective.11 uninjured participants (age: 28 ± 3 years; 8 (73%) female).3-T, qDESS.Test-retest T2 maps were acquired twice on the same day and with a 1-week interval between scans. For each acquisition, average cartilage T2 was calculated in four manually segmented regions encompassing tibiofemoral contact areas during common activities and 12 automatically segmented regions from the deep-learning open-source framework for musculoskeletal MRI analysis (DOSMA) encompassing medial and lateral anterior, central, and posterior tibiofemoral regions. Test-retest T2 values from matching regions were used to evaluate reproducibility.Coefficients of variation (%CV), root-mean-square-average-CV (%RMSA-CV), and intraclass correlation coefficients (ICCs) assessed test-retest T2 reproducibility. The median of test-retest standard deviations was used for T2 precision. Bland-Altman (BA) analyses examined test-retest biases. The smallest detectable difference (SDD) was defined as the BA limit of agreement of largest magnitude. Significance was accepted for P < 0.05.All cartilage regions across both segmentation schemes demonstrated intraday and interday qDESS T2 CVs and RMSA-CVs of ≤5%. T2 ICC values >0.75 were observed in the majority of regions but were more variable in interday tibial comparisons. Test-retest T2 precision was <1.3 msec. The T2 SDD was 3.8 msec.Excellent CV and RMSA-CV reproducibility may suggest that qDESS T2 increases or decreases >5% (3.8 msec) could represent changes to cartilage composition.Stage 2.
View details for DOI 10.1002/jmri.29431
View details for PubMedID 38703134
-
Missing Wedge Completion via Unsupervised Learning with Coordinate Networks.
bioRxiv : the preprint server for biology
2024
Abstract
Cryogenic electron tomography (cryoET) is a powerful tool in structural biology, enabling detailed 3D imaging of biological specimens at a resolution of nanometers. Despite its potential, cryoET faces challenges such as the missing wedge problem, which limits reconstruction quality due to incomplete data collection angles. Recently, supervised deep learning methods leveraging convolutional neural networks (CNNs) have considerably addressed this issue; however, their pretraining requirements render them susceptible to inaccuracies and artifacts, particularly when representative training data is scarce. To overcome these limitations, we introduce a proof-of-concept unsupervised learning approach using coordinate networks (CNs) that optimizes network weights directly against input projections. This eliminates the need for pretraining, reducing reconstruction runtime by 3 - 20× compared to supervised methods. Our in silico results show improved shape completion and reduction of missing wedge artifacts, assessed through several voxel-based image quality metrics in real space and a novel directional Fourier Shell Correlation (FSC) metric. Our study illuminates benefits and considerations of both supervised and unsupervised approaches, guiding the development of improved reconstruction strategies.
View details for DOI 10.1101/2024.04.12.589090
View details for PubMedID 38712113
View details for PubMedCentralID PMC11071277
-
Biomarkers of Body Composition.
Seminars in musculoskeletal radiology
2024; 28 (1): 78-91
Abstract
The importance and impact of imaging biomarkers has been increasing over the past few decades. We review the relevant clinical and imaging terminology needed to understand the clinical and research applications of body composition. Imaging biomarkers of bone, muscle, and fat tissues obtained with dual-energy X-ray absorptiometry, computed tomography, magnetic resonance imaging, and ultrasonography are described.
View details for DOI 10.1055/s-0043-1776430
View details for PubMedID 38330972
-
Differences in Anatomic Adaptation and Injury Patterns Related to Valgus Extension Overload in Overhead Throwing Athletes.
Diagnostics (Basel, Switzerland)
2024; 14 (2)
Abstract
The purpose of our study was to determine differences in adaptative and injury patterns in the elbow related to valgus extension overload (VEO) in overhead throwing athletes by age. A total of 86 overhead throwing athletes and 23 controls underwent MRI or MR arthrography (MRA) of the elbow. Throwing athletes were divided by age into three groups: ≤16 years (26 subjects), 17-19 years (25 subjects), and ≥20 years (35 subjects). Consensus interpretation of each MRI was performed, with measurements of ulnar collateral ligament (UCL) thickness and subchondral sclerosis at the radial head, humeral trochlea, and olecranon process. A higher frequency of apophyseal and stress injuries was seen in adolescent athletes and increased incidence of soft tissue injuries was observed in older athletes. Early adaptive and degenerative changes were observed with high frequency independent of age. Significant differences were observed between athletes and controls for UCL thickness (p < 0.001) and subchondral sclerosis at the radial head (p < 0.001), humeral trochlea (p < 0.001), and olecranon process (p < 0.001). Significant differences based on athlete age were observed for UCL thickness (p < 0.001) and subchondral sclerosis at the olecranon process (p = 0.002). Our study highlights differences in anatomic adaptations related to VEO at the elbow between overhead throwing athletes and control subjects, as well as across age in throwing athletes.
View details for DOI 10.3390/diagnostics14020217
View details for PubMedID 38275464
-
MEDALIGN: A Clinician-Generated Dataset for Instruction Following with Electronic Medical Records
ASSOC ADVANCEMENT ARTIFICIAL INTELLIGENCE. 2024: 22021-22030
View details for Web of Science ID 001239985800017
-
Accelerated Musculoskeletal Magnetic Resonance Imaging.
Journal of magnetic resonance imaging : JMRI
2023
Abstract
With a substantial growth in the use of musculoskeletal MRI, there has been a growing need to improve MRI workflow, and faster imaging has been suggested as one of the solutions for a more efficient examination process. Consequently, there have been considerable advances in accelerated MRI scanning methods. This article aims to review the basic principles and applications of accelerated musculoskeletal MRI techniques including widely used conventional acceleration methods, more advanced deep learning-based techniques, and new approaches to reduce scan time. Specifically, conventional accelerated MRI techniques, including parallel imaging, compressed sensing, and simultaneous multislice imaging, and deep learning-based accelerated MRI techniques, including undersampled MR image reconstruction, super-resolution imaging, artifact correction, and generation of unacquired contrast images, are discussed. Finally, new approaches to reduce scan time, including synthetic MRI, novel sequences, and new coil setups and designs, are also reviewed. We believe that a deep understanding of these fast MRI techniques and proper use of combined acceleration methods will synergistically improve scan time and MRI workflow in daily practice. EVIDENCE LEVEL: 3 TECHNICAL EFFICACY: Stage 1.
View details for DOI 10.1002/jmri.29205
View details for PubMedID 38156716
-
AI in osteoarthritis: illuminating the meandering path forward.
Osteoarthritis and cartilage
2023
View details for DOI 10.1016/j.joca.2023.11.009
View details for PubMedID 38013138
-
Reconsidering Conclusions of Bias Assessment in Medical Imaging Foundation Models.
Radiology. Artificial intelligence
2023; 5 (6): e230432
View details for DOI 10.1148/ryai.230432
View details for PubMedID 38074780
View details for PubMedCentralID PMC10698581
-
Reconsidering Conclusions of Bias Assessment in Medical Imaging Foundation Models
RADIOLOGY-ARTIFICIAL INTELLIGENCE
2023; 5 (6)
View details for Web of Science ID 001172477100012
-
Clinical Text Summarization: Adapting Large Language Models Can Outperform Human Experts.
Research square
2023
Abstract
Sifting through vast textual data and summarizing key information from electronic health records (EHR) imposes a substantial burden on how clinicians allocate their time. Although large language models (LLMs) have shown immense promise in natural language processing (NLP) tasks, their efficacy on a diverse range of clinical summarization tasks has not yet been rigorously demonstrated. In this work, we apply domain adaptation methods to eight LLMs, spanning six datasets and four distinct clinical summarization tasks: radiology reports, patient questions, progress notes, and doctor-patient dialogue. Our thorough quantitative assessment reveals trade-offs between models and adaptation methods in addition to instances where recent advances in LLMs may not improve results. Further, in a clinical reader study with ten physicians, we show that summaries from our best-adapted LLMs are preferable to human summaries in terms of completeness and correctness. Our ensuing qualitative analysis highlights challenges faced by both LLMs and human experts. Lastly, we correlate traditional quantitative NLP metrics with reader study scores to enhance our understanding of how these metrics align with physician preferences. Our research marks the first evidence of LLMs outperforming human experts in clinical text summarization across multiple tasks. This implies that integrating LLMs into clinical workflows could alleviate documentation burden, empowering clinicians to focus more on personalized patient care and the inherently human aspects of medicine.
View details for DOI 10.21203/rs.3.rs-3483777/v1
View details for PubMedID 37961377
View details for PubMedCentralID PMC10635391
-
OpenCap: Human movement dynamics from smartphone videos.
PLoS computational biology
2023; 19 (10): e1011462
Abstract
Measures of human movement dynamics can predict outcomes like injury risk or musculoskeletal disease progression. However, these measures are rarely quantified in large-scale research studies or clinical practice due to the prohibitive cost, time, and expertise required. Here we present and validate OpenCap, an open-source platform for computing both the kinematics (i.e., motion) and dynamics (i.e., forces) of human movement using videos captured from two or more smartphones. OpenCap leverages pose estimation algorithms to identify body landmarks from videos; deep learning and biomechanical models to estimate three-dimensional kinematics; and physics-based simulations to estimate muscle activations and musculoskeletal dynamics. OpenCap's web application enables users to collect synchronous videos and visualize movement data that is automatically processed in the cloud, thereby eliminating the need for specialized hardware, software, and expertise. We show that OpenCap accurately predicts dynamic measures, like muscle activations, joint loads, and joint moments, which can be used to screen for disease risk, evaluate intervention efficacy, assess between-group movement differences, and inform rehabilitation decisions. Additionally, we demonstrate OpenCap's practical utility through a 100-subject field study, where a clinician using OpenCap estimated musculoskeletal dynamics 25 times faster than a laboratory-based approach at less than 1% of the cost. By democratizing access to human movement analysis, OpenCap can accelerate the incorporation of biomechanical metrics into large-scale research studies, clinical trials, and clinical practice.
View details for DOI 10.1371/journal.pcbi.1011462
View details for PubMedID 37856442
View details for PubMedCentralID PMC10586693
-
LIVER VOLUME PREDICTS MORTALITY IN ALCOHOL ASSOCIATED LIVER DISEASE
LIPPINCOTT WILLIAMS & WILKINS. 2023: S1622-S1623
View details for Web of Science ID 001094865404050
-
Association of Coronary Artery Calcium Detected by Routine Ungated CT Imaging With Cardiovascular Outcomes.
Journal of the American College of Cardiology
2023; 82 (12): 1192-1202
Abstract
Coronary artery calcium (CAC) is a strong predictor of cardiovascular events across all racial and ethnic groups. CAC can be quantified on nonelectrocardiography (ECG)-gated computed tomography (CT) performed for other reasons, allowing for opportunistic screening for subclinical atherosclerosis.The authors investigated whether incidental CAC quantified on routine non-ECG-gated CTs using a deep-learning (DL) algorithm provided cardiovascular risk stratification beyond traditional risk prediction methods.Incidental CAC was quantified using a DL algorithm (DL-CAC) on non-ECG-gated chest CTs performed for routine care in all settings at a large academic medical center from 2014 to 2019. We measured the association between DL-CAC (0, 1-99, or ≥100) with all-cause death (primary outcome), and the secondary composite outcomes of death/myocardial infarction (MI)/stroke and death/MI/stroke/revascularization using Cox regression. We adjusted for age, sex, race, ethnicity, comorbidities, systolic blood pressure, lipid levels, smoking status, and antihypertensive use. Ten-year atherosclerotic cardiovascular disease risk was calculated using the pooled cohort equations.Of 5,678 adults without ASCVD (51% women, 18% Asian, 13% Hispanic/Latinx), 52% had DL-CAC >0. Those with DL-CAC ≥100 had an average 10-year ASCVD risk of 24%; yet, only 26% were on statins. After adjustment, patients with DL-CAC ≥100 had increased risk of death (HR: 1.51; 95% CI: 1.28-1.79), death/MI/stroke (HR: 1.57; 95% CI: 1.33-1.84), and death/MI/stroke/revascularization (HR: 1.69; 95% CI: 1.45-1.98) compared with DL-CAC = 0.Incidental CAC ≥100 was associated with an increased risk of all-cause death and adverse cardiovascular outcomes, beyond traditional risk factors. DL-CAC from routine non-ECG-gated CTs identifies patients at increased cardiovascular risk and holds promise as a tool for opportunistic screening to facilitate earlier intervention.
View details for DOI 10.1016/j.jacc.2023.06.040
View details for PubMedID 37704309
-
Federated benchmarking of medical artificial intelligence with MedPerf
NATURE MACHINE INTELLIGENCE
2023
View details for DOI 10.1038/s42256-023-00652-2
View details for Web of Science ID 001033508400003
-
Federated benchmarking of medical artificial intelligence with MedPerf.
Nature machine intelligence
2023; 5 (7): 799-810
Abstract
Medical artificial intelligence (AI) has tremendous potential to advance healthcare by supporting and contributing to the evidence-based practice of medicine, personalizing patient treatment, reducing costs, and improving both healthcare provider and patient experience. Unlocking this potential requires systematic, quantitative evaluation of the performance of medical AI models on large-scale, heterogeneous data capturing diverse patient populations. Here, to meet this need, we introduce MedPerf, an open platform for benchmarking AI models in the medical domain. MedPerf focuses on enabling federated evaluation of AI models, by securely distributing them to different facilities, such as healthcare organizations. This process of bringing the model to the data empowers each facility to assess and verify the performance of AI models in an efficient and human-supervised process, while prioritizing privacy. We describe the current challenges healthcare and AI communities face, the need for an open platform, the design philosophy of MedPerf, its current implementation status and real-world deployment, our roadmap and, importantly, the use of MedPerf with multiple international institutions within cloud-based technology and on-premises scenarios. Finally, we welcome new contributions by researchers and organizations to further strengthen MedPerf as an open benchmarking platform.
View details for DOI 10.1038/s42256-023-00652-2
View details for PubMedID 38706981
View details for PubMedCentralID PMC11068064
-
Patellar Tracking: An Old Problem with New Insights.
Radiographics : a review publication of the Radiological Society of North America, Inc
2023; 43 (6): e220177
Abstract
Patellofemoral pain and instability are common indications for imaging that are encountered in everyday practice. The authors comprehensively review key aspects of patellofemoral instability pertinent to radiologists that can be seen before the onset of osteoarthritis, highlighting the anatomy, clinical evaluation, diagnostic imaging, and treatment. Regarding the anatomy, the medial patellofemoral ligament (MPFL) is the primary static soft-tissue restraint to lateral patellar displacement and is commonly reconstructed surgically in patients with MPFL dysfunction and patellar instability. Osteoarticular abnormalities that predispose individuals to patellar instability include patellar malalignment, trochlear dysplasia, and tibial tubercle lateralization. Clinically, patients with patellar instability may be divided into two broad groups with imaging findings that sometimes overlap: patients with a history of overt patellar instability after a traumatic event (eg, dislocation, subluxation) and patients without such a history. In terms of imaging, radiography is generally the initial examination of choice, and MRI is the most common cross-sectional examination performed preoperatively. For all imaging techniques, there has been a proliferation of published radiologic measurement methods. The authors summarize the most common validated measurements for patellar malalignment, trochlear dysplasia, and tibial tubercle lateralization. Given that static imaging is inherently limited in the evaluation of patellar motion, dynamic imaging with US, CT, or MRI may be requested by some surgeons. The primary treatment strategy for patellofemoral pain is conservative. Surgical treatment options include MPFL reconstruction with or without osseous corrections such as trochleoplasty and tibial tubercle osteotomy. Postoperative complications evaluated at imaging include patellar fracture, graft failure, graft malposition, and medial patellar subluxation. ©RSNA, 2023 Quiz questions for this article are available in the supplemental material.
View details for DOI 10.1148/rg.220177
View details for PubMedID 37261964
-
[Formula: see text] Field inhomogeneity correction for qDESS [Formula: see text] mapping: application to rapid bilateral knee imaging.
Magma (New York, N.Y.)
2023
Abstract
[Formula: see text] mapping is a powerful tool for studying osteoarthritis (OA) changes and bilateral imaging may be useful in investigating the role of between-knee asymmetry in OA onset and progression. The quantitative double-echo in steady-state (qDESS) can provide fast simultaneous bilateral knee [Formula: see text] and high-resolution morphometry for cartilage and meniscus. The qDESS uses an analytical signal model to compute [Formula: see text] relaxometry maps, which require knowledge of the flip angle (FA). In the presence of [Formula: see text] inhomogeneities, inconsistencies between the nominal and actual FA can affect the accuracy of [Formula: see text] measurements. We propose a pixel-wise [Formula: see text] correction method for qDESS [Formula: see text] mapping exploiting an auxiliary [Formula: see text] map to compute the actual FA used in the model.The technique was validated in a phantom and in vivo with simultaneous bilateral knee imaging. [Formula: see text] measurements of femoral cartilage (FC) of both knees of six healthy participants were repeated longitudinally to investigate the association between [Formula: see text] variation and [Formula: see text].The results showed that applying the [Formula: see text] correction mitigated [Formula: see text] variations that were driven by [Formula: see text] inhomogeneities. Specifically, [Formula: see text] left-right symmetry increased following the [Formula: see text] correction ([Formula: see text] = 0.74 > [Formula: see text] = 0.69). Without the [Formula: see text] correction, [Formula: see text] values showed a linear dependence with [Formula: see text]. The linear coefficient decreased using the [Formula: see text] correction (from 24.3 ± 1.6 ms to 4.1 ± 1.8) and the correlation was not statistically significant after the application of the Bonferroni correction (p value > 0.01).The study showed that [Formula: see text] correction could mitigate variations driven by the sensitivity of the qDESS [Formula: see text] mapping method to [Formula: see text], therefore, increasing the sensitivity to detect real biological changes. The proposed method may improve the robustness of bilateral qDESS [Formula: see text] mapping, allowing for an accurate and more efficient evaluation of OA pathways and pathophysiology through longitudinal and cross-sectional studies.
View details for DOI 10.1007/s10334-023-01094-y
View details for PubMedID 37142852
View details for PubMedCentralID 2268124
-
Towards Automatic Cartilage Quantification in Clinical Trials - Continuing from the 2019 IWOAI Knee Segmentation Challenge.
Osteoarthritis imaging
2023; 3 (1)
Abstract
To evaluate whether the deep learning (DL) segmentation methods from the six teams that participated in the IWOAI 2019 Knee Cartilage Segmentation Challenge are appropriate for quantifying cartilage loss in longitudinal clinical trials.We included 556 subjects from the Osteoarthritis Initiative study with manually read cartilage volume scores for the baseline and 1-year visits. The teams used their methods originally trained for the IWOAI 2019 challenge to segment the 1130 knee MRIs. These scans were anonymized and the teams were blinded to any subject or visit identifiers. Two teams also submitted updated methods. The resulting 9,040 segmentations are available online.The segmentations included tibial, femoral, and patellar compartments. In post-processing, we extracted medial and lateral tibial compartments and geometrically defined central medial and lateral femoral sub-compartments. The primary study outcome was the sensitivity to measure cartilage loss as defined by the standardized response mean (SRM).For the tibial compartments, several of the DL segmentation methods had SRMs similar to the gold standard manual method. The highest DL SRM was for the lateral tibial compartment at 0.38 (the gold standard had 0.34). For the femoral compartments, the gold standard had higher SRMs than the automatic methods at 0.31/0.30 for medial/lateral compartments.The lower SRMs for the DL methods in the femoral compartments at 0.2 were possibly due to the simple sub-compartment extraction done during post-processing. The study demonstrated that state-of-the-art DL segmentation methods may be used in standardized longitudinal single-scanner clinical trials for well-defined cartilage compartments.
View details for DOI 10.1016/j.ostima.2023.100087
View details for PubMedID 39036792
View details for PubMedCentralID PMC11258861
-
Efficient Diagnosis Assignment Using Unstructured Clinical Notes
ASSOC COMPUTATIONAL LINGUISTICS-ACL. 2023: 485-494
View details for Web of Science ID 001181088800042
-
Exploring Image Augmentations for Siamese Representation Learning with Chest X-Rays
JMLR-JOURNAL MACHINE LEARNING RESEARCH. 2023: 444-467
View details for Web of Science ID 001221108600027
-
ViLLA: Fine-Grained Vision-Language Representation Learning from Real-World Data
IEEE COMPUTER SOC. 2023: 22168-22178
View details for DOI 10.1109/ICCV51070.2023.02031
View details for Web of Science ID 001169500506073
-
Radiology Decision Support System for Selecting Appropriate CT Imaging Titles Using Machine Learning Techniques Based on Electronic Medical Records
IEEE ACCESS
2023; 11: 99222-99236
View details for DOI 10.1109/ACCESS.2023.3314380
View details for Web of Science ID 001070603300001
-
Developing medical imaging AI for emerging infectious diseases.
Nature communications
2022; 13 (1): 7060
View details for DOI 10.1038/s41467-022-34234-4
View details for PubMedID 36400764
-
The KNee OsteoArthritis Prediction (KNOAP2020) Challenge: An image analysis challenge to predict incident symptomatic radiographic knee osteoarthritis from MRI and X-ray images.
Osteoarthritis and cartilage
2022
Abstract
The KNee OsteoArthritis Prediction (KNOAP2020) challenge was organized to objectively compare methods for the prediction of incident symptomatic radiographic knee osteoarthritis within 78 months on a test set with blinded ground truth.The challenge participants were free to use any available data sources to train their models. A test set of 423 knees from the Prevention of Knee Osteoarthritis in Overweight Females (PROOF) study consisting of magnetic resonance imaging (MRI) and X-ray image data along with clinical risk factors at baseline was made available to all challenge participants. The ground truth outcomes, i.e., which knees developed incident symptomatic radiographic knee osteoarthritis (according to the combined ACR criteria) within 78 months, were not provided to the participants. To assess the performance of the submitted models, we used the area under the receiver operating characteristic curve (ROCAUC) and balanced accuracy (BACC).Seven teams submitted 23 entries in total. A majority of the algorithms were trained on data from the Osteoarthritis Initiative. The model with the highest ROCAUC (0.64 (95% confidence interval (CI): 0.57-0.70)) used deep learning to extract information from X-ray images combined with clinical variables. The model with the highest BACC (0.59 (95% CI: 0.52-0.65)) ensembled three different models that used automatically extracted X-ray and MRI features along with clinical variables.The KNOAP2020 challenge established a benchmark for predicting incident symptomatic radiographic knee osteoarthritis. Accurate prediction of incident symptomatic radiographic knee osteoarthritis is a complex and still unsolved problem requiring additional investigation.
View details for DOI 10.1016/j.joca.2022.10.001
View details for PubMedID 36243308
-
Preliminary Clinical Validation Results of a Deep Learning Approach for Ankle Brachial Index Prediction in Noncompressible Tibial Vessels
MOSBY-ELSEVIER. 2022: E85
View details for Web of Science ID 000888852300001
-
A method for measuring B0 field inhomogeneity using quantitative double-echo in steady-state.
Magnetic resonance in medicine
2022
Abstract
To develop and validate a method for B 0 $$ {B}_0 $$ mapping for knee imaging using the quantitative Double-Echo in Steady-State (qDESS) exploiting the phase difference ( Δ θ $$ \Delta \theta $$ ) between the two echoes acquired. Contrary to a two-gradient-echo (2-GRE) method, Δ θ $$ \Delta \theta $$ depends only on the first echo time.Bloch simulations were applied to investigate robustness to noise of the proposed methodology and all imaging studies were validated with phantoms and in vivo simultaneous bilateral knee acquisitions. Two phantoms and five healthy subjects were scanned using qDESS, water saturation shift referencing (WASSR), and multi-GRE sequences. Δ B 0 $$ \Delta {B}_0 $$ maps were calculated with the qDESS and the 2-GRE methods and compared against those obtained with WASSR. The comparison was quantitatively assessed exploiting pixel-wise difference maps, Bland-Altman (BA) analysis, and Lin's concordance coefficient ( ρ c $$ {\rho}_c $$ ). For in vivo subjects, the comparison was assessed in cartilage using average values in six subregions.The proposed method for measuring Δ B 0 $$ \Delta {B}_0 $$ inhomogeneities from a qDESS acquisition provided Δ B 0 $$ \Delta {B}_0 $$ maps that were in good agreement with those obtained using WASSR. Δ B 0 $$ \Delta {B}_0 $$ ρ c $$ {\rho}_c $$ values were ≥ $$ \ge $$ 0.98 and 0.90 in phantoms and in vivo, respectively. The agreement between qDESS and WASSR was comparable to that of a 2-GRE method.The proposed method may allow B0 correction for qDESS T 2 $$ {T}_2 $$ mapping using an inherently co-registered Δ B 0 $$ \Delta {B}_0 $$ map without requiring an additional B0 measurement sequence. More generally, the method may help shorten knee imaging protocols that require an auxiliary Δ B 0 $$ \Delta {B}_0 $$ map by exploiting a qDESS acquisition that also provides T 2 $$ {T}_2 $$ measurements and high-quality morphological imaging.
View details for DOI 10.1002/mrm.29465
View details for PubMedID 36161727
-
Imaging of Sarcopenia.
Radiologic clinics of North America
2022; 60 (4): 575-582
Abstract
Sarcopenia is currently underdiagnosed and undertreated, but this is expected to change because sarcopenia is now recognized with a specific diagnosis code that can be used for billing in some countries, as well as an expanding body of research on prevention, diagnosis, and management. This article focuses on practical issues of increasing interest by highlighting 3 hot topics fundamental to understanding sarcopenia in older adults: definitions and terminology, current diagnostic imaging techniques, and the emerging role of opportunistic computed tomography.
View details for DOI 10.1016/j.rcl.2022.03.001
View details for PubMedID 35672090
-
Scale-Equivariant Unrolled Neural Networks for Data-Efficient Accelerated MRI Reconstruction
SPRINGER INTERNATIONAL PUBLISHING AG. 2022: 737-747
View details for DOI 10.1007/978-3-031-16446-0_70
View details for Web of Science ID 000867434800070
-
TorchXRayVision: A library of chest X-ray datasets and models
JMLR-JOURNAL MACHINE LEARNING RESEARCH. 2022: 231-249
View details for Web of Science ID 001227587200015
-
ViLMedic: a framework for research at the intersection of vision and language in medical AI
ASSOC COMPUTATIONAL LINGUISTICS-ACL. 2022: 23-34
View details for Web of Science ID 000828759800003
-
MRSaiFE: An AI-Based Approach Toward the Real-Time Prediction of Specific Absorption Rate (vol 9, pg 140824, 2021)
IEEE ACCESS
2022; 10: 19925
View details for DOI 10.1109/ACCESS.2022.3152277
View details for Web of Science ID 000764066800001
-
Validation of Deep Learning-based Augmentation for Reduced 18F-FDG Dose for PET/MRI in Children and Young Adults with Lymphoma.
Radiology. Artificial intelligence
2021; 3 (6): e200232
Abstract
Purpose: To investigate if a deep learning convolutional neural network (CNN) could enable low-dose fluorine 18 (18F) fluorodeoxyglucose (FDG) PET/MRI for correct treatment response assessment of children and young adults with lymphoma.Materials and Methods: In this secondary analysis of prospectively collected data (ClinicalTrials.gov identifier: NCT01542879), 20 patients with lymphoma (mean age, 16.4 years ± 6.4 [standard deviation]) underwent 18F-FDG PET/MRI between July 2015 and August 2019 at baseline and after induction chemotherapy. Full-dose 18F-FDG PET data (3 MBq/kg) were simulated to lower 18F-FDG doses based on the percentage of coincidence events (representing simulated 75%, 50%, 25%, 12.5%, and 6.25% 18F-FDG dose [hereafter referred to as 75%Sim, 50%Sim, 25%Sim, 12.5%Sim, and 6.25%Sim, respectively]). A U.S. Food and Drug Administration-approved CNN was used to augment input simulated low-dose scans to full-dose scans. For each follow-up scan after induction chemotherapy, the standardized uptake value (SUV) response score was calculated as the maximum SUV (SUVmax) of the tumor normalized to the mean liver SUV; tumor response was classified as adequate or inadequate. Sensitivity and specificity in the detection of correct response status were computed using full-dose PET as the reference standard.Results: With decreasing simulated radiotracer doses, tumor SUVmax increased. A dose below 75%Sim of the full dose led to erroneous upstaging of adequate responders to inadequate responders (43% [six of 14 patients] for 75%Sim; 93% [13 of 14 patients] for 50%Sim; and 100% [14 of 14 patients] below 50%Sim; P < .05 for all). CNN-enhanced low-dose PET/MRI scans at 75%Sim and 50%Sim enabled correct response assessments for all patients. Use of the CNN augmentation for assessing adequate and inadequate responses resulted in identical sensitivities (100%) and specificities (100%) between the assessment of 100% full-dose PET, augmented 75%Sim, and augmented 50%Sim images.Conclusion: CNN enhancement of PET/MRI scans may enable 50% 18F-FDG dose reduction with correct treatment response assessment of children and young adults with lymphoma.Keywords: Pediatrics, PET/MRI, Computer Applications Detection/Diagnosis, Lymphoma, Tumor Response, Whole-Body Imaging, Technology AssessmentClinical trial registration no: NCT01542879 Supplemental material is available for this article. ©RSNA, 2021.
View details for DOI 10.1148/ryai.2021200232
View details for PubMedID 34870211
-
Author Correction: Low-count whole-body PET with deep learning in a multicenter and externally validated study.
NPJ digital medicine
2021; 4 (1): 139
View details for DOI 10.1038/s41746-021-00512-6
View details for PubMedID 34521985
-
Open Source Software for Automatic Subregional Assessment of Knee Cartilage Degradation Using Quantitative T2 Relaxometry and Deep Learning.
Cartilage
2021: 19476035211042406
Abstract
OBJECTIVE: We evaluated a fully automated femoral cartilage segmentation model for measuring T2 relaxation values and longitudinal changes using multi-echo spin-echo (MESE) magnetic resonance imaging (MRI). We open sourced this model and developed a web app available at https://kl.stanford.edu into which users can drag and drop images to segment them automatically.DESIGN: We trained a neural network to segment femoral cartilage from MESE MRIs. Cartilage was divided into 12 subregions along medial-lateral, superficial-deep, and anterior-central-posterior boundaries. Subregional T2 values and four-year changes were calculated using a radiologist's segmentations (Reader 1) and the model's segmentations. These were compared using 28 held-out images. A subset of 14 images were also evaluated by a second expert (Reader 2) for comparison.RESULTS: Model segmentations agreed with Reader 1 segmentations with a Dice score of 0.85 ± 0.03. The model's estimated T2 values for individual subregions agreed with those of Reader 1 with an average Spearman correlation of 0.89 and average mean absolute error (MAE) of 1.34 ms. The model's estimated four-year change in T2 for individual subregions agreed with Reader 1 with an average correlation of 0.80 and average MAE of 1.72 ms. The model agreed with Reader 1 at least as closely as Reader 2 agreed with Reader 1 in terms of Dice score (0.85 vs. 0.75) and subregional T2 values.CONCLUSIONS: Assessments of cartilage health using our fully automated segmentation model agreed with those of an expert as closely as experts agreed with one another. This has the potential to accelerate osteoarthritis research.
View details for DOI 10.1177/19476035211042406
View details for PubMedID 34496667
-
Synthesizing Quantitative T2 Maps in Right Lateral Knee Femoral Condyles from Multicontrast Anatomic Data with a Conditional Generative Adversarial Network.
Radiology. Artificial intelligence
2021; 3 (5): e200122
Abstract
Purpose: To develop a proof-of-concept convolutional neural network (CNN) to synthesize T2 maps in right lateral femoral condyle articular cartilage from anatomic MR images by using a conditional generative adversarial network (cGAN).Materials and Methods: In this retrospective study, anatomic images (from turbo spin-echo and double-echo in steady-state scans) of the right knee of 4621 patients included in the 2004-2006 Osteoarthritis Initiative were used as input to a cGAN-based CNN, and a predicted CNN T2 was generated as output. These patients included men and women of all ethnicities, aged 45-79 years, with or at high risk for knee osteoarthritis incidence or progression who were recruited at four separate centers in the United States. These data were split into 3703 (80%) for training, 462 (10%) for validation, and 456 (10%) for testing. Linear regression analysis was performed between the multiecho spin-echo (MESE) and CNN T2 in the test dataset. A more detailed analysis was performed in 30 randomly selected patients by means of evaluation by two musculoskeletal radiologists and quantification of cartilage subregions. Radiologist assessments were compared by using two-sided t tests.Results: The readers were moderately accurate in distinguishing CNN T2 from MESE T2, with one reader having random-chance categorization. CNN T2 values were correlated to the MESE values in the subregions of 30 patients and in the bulk analysis of all patients, with best-fit line slopes between 0.55 and 0.83.Conclusion: With use of a neural network-based cGAN approach, it is feasible to synthesize T2 maps in femoral cartilage from anatomic MRI sequences, giving good agreement with MESE scans.See also commentary by Yi and Fritz in this issue.Keywords: Cartilage Imaging, Knee, Experimental Investigations, Quantification, Vision, Application Domain, Convolutional Neural Network (CNN), Deep Learning Algorithms, Machine Learning Algorithms© RSNA, 2021.
View details for DOI 10.1148/ryai.2021200122
View details for PubMedID 34617020
-
Challenges in ensuring the generalizability of image quantitation methods for MRI.
Medical physics
2021
Abstract
Image quantitation methods including quantitative MRI, multiparametric MRI, and radiomics, offer great promise for clinical use. However, many of these methods have limited clinical adoption, in part due to issues of generalizability, i.e., the ability to translate methods and models across institutions. Researchers can assess generalizability through measurement of repeatability and reproducibility, thus quantifying different aspects of measurement variance. In this article, we review the challenges to ensuring repeatability and reproducibility of image quantitation methods as well as present strategies to minimize their variance to enable wider clinical implementation. We present possible solutions for achieving clinically acceptable performance of image quantitation methods and briefly discuss the impact of minimizing variance and achieving generalizability towards clinical implementation and adoption. This article is protected by copyright. All rights reserved.
View details for DOI 10.1002/mp.15195
View details for PubMedID 34455593
-
Sarcopenia in rheumatic disorders: what the radiologist and rheumatologist should know.
Skeletal radiology
2021
Abstract
Sarcopenia is defined as the loss of muscle mass, strength, and function. Increasing evidence shows that sarcopenia is common in patients with rheumatic disorders. Although sarcopenia can be diagnosed using bioelectrical impedance analysis or DXA, increasingly it is diagnosed using CT, MRI, and ultrasound. In rheumatic patients, CT and MRI allow "opportunistic" measurement of body composition, including surrogate markers of sarcopenia, from studies obtained during routine patient care. Recognition of sarcopenia is important in rheumatic patients because sarcopenia can be associated with disease progression and poor outcomes. This article reviews how opportunistic evaluation of sarcopenia in rheumatic patients can be accomplished and potentially contribute to improved patient care.
View details for DOI 10.1007/s00256-021-03863-z
View details for PubMedID 34268590
-
Non-contrast MRI of synovitis in the knee using quantitative DESS.
European radiology
2021
Abstract
OBJECTIVES: To determine whether synovitis graded by radiologists using hybrid quantitative double-echo in steady-state (qDESS) images can be utilized as a non-contrast approach to assess synovitis in the knee, compared against the reference standard of contrast-enhanced MRI (CE-MRI).METHODS: Twenty-two knees (11 subjects) with moderate to severe osteoarthritis (OA) were scanned using CE-MRI, qDESS with a high diffusion weighting (qDESSHigh), and qDESS with a low diffusion weighting (qDESSLow). Four radiologists graded the overall impression of synovitis, their diagnostic confidence, and regional grading of synovitis severity at four sites (suprapatellar pouch, intercondylar notch, and medial and lateral peripatellar recesses) in the knee using a 4-point scale. Agreement between CE-MRI and qDESS, inter-rater agreement, and intra-rater agreement were assessed using a linearly weighted Gwet's AC2.RESULTS: Good agreement was seen between CE-MRI and both qDESSLow (AC2=0.74) and qDESSHigh (AC2=0.66) for the overall impression of synovitis, but both qDESS sequences tended to underestimate the severity of synovitis compared to CE-MRI. Good inter-rater agreement was seen for both qDESS sequences (AC2=0.74 for qDESSLow, AC2=0.64 for qDESSHigh), and good intra-rater agreement was seen for both sequences as well (qDESSLow AC2=0.78, qDESSHigh AC2=0.80). Diagnostic confidence was moderate to high for qDESSLow (mean=2.36) and slightly less than moderate for qDESSHigh (mean=1.86), compared to mostly high confidence for CE-MRI (mean=2.73).CONCLUSIONS: qDESS shows potential as an alternative MRI technique for assessing the severity of synovitis without the use of a gadolinium-based contrast agent.KEY POINTS: The use of the quantitative double-echo in steady-state (qDESS) sequence for synovitis assessment does not require the use of a gadolinium-based contrast agent. Preliminary results found that low diffusion-weighted qDESS (qDESSLow) shows good agreement to contrast-enhanced MRI for characterization of the severity of synovitis, with a relative bias towards underestimation of severity. Preliminary results also found that qDESSLow shows good inter- and intra-rater agreement for the depiction of synovitis, particularly for readers experienced with the sequence.
View details for DOI 10.1007/s00330-021-08025-2
View details for PubMedID 33993332
-
Measuring Robustness in Deep Learning Based Compressive Sensing
JMLR-JOURNAL MACHINE LEARNING RESEARCH. 2021
View details for Web of Science ID 000683104602041
-
Upstream Machine Learning in Radiology.
Radiologic clinics of North America
2021; 59 (6): 967-985
Abstract
Machine learning (ML) and Artificial intelligence (AI) has the potential to dramatically improve radiology practice at multiple stages of the imaging pipeline. Most of the attention has been garnered by applications focused on improving the end of the pipeline: image interpretation. However, this article reviews how AI/ML can be applied to improve upstream components of the imaging pipeline, including exam modality selection, hardware design, exam protocol selection, data acquisition, image reconstruction, and image processing. A breadth of applications and their potential for impact is shown across multiple imaging modalities, including ultrasound, computed tomography, and MRI.
View details for DOI 10.1016/j.rcl.2021.07.009
View details for PubMedID 34689881
-
MRSaiFE: An AI-Based Approach Towards the Real-Time Prediction of Specific Absorption Rate
IEEE ACCESS
2021; 9: 140824-140834
View details for DOI 10.1109/ACCESS.2021.3118290
View details for Web of Science ID 000709061400001
-
Improving in vivo human cerebral cortical surface reconstruction using data-driven super-resolution.
Cerebral cortex (New York, N.Y. : 1991)
2020
Abstract
Accurate and automated reconstruction of the in vivo human cerebral cortical surface from anatomical magnetic resonance (MR) images facilitates the quantitative analysis of cortical structure. Anatomical MR images with sub-millimeter isotropic spatial resolution improve the accuracy of cortical surface and thickness estimation compared to the standard 1-millimeter isotropic resolution. Nonetheless, sub-millimeter resolution acquisitions require averaging multiple repetitions to achieve sufficient signal-to-noise ratio and are therefore long and potentially vulnerable to subject motion. We address this challenge by synthesizing sub-millimeter resolution images from standard 1-millimeter isotropic resolution images using a data-driven supervised machine learning-based super-resolution approach achieved via a deep convolutional neural network. We systematically characterize our approach using a large-scale simulated dataset and demonstrate its efficacy in empirical data. The super-resolution data provide improved cortical surfaces similar to those obtained from native sub-millimeter resolution data. The whole-brain mean absolute discrepancy in cortical surface positioning and thickness estimation is below 100mum at the single-subject level and below 50mum at the group level for the simulated data, and below 200mum at the single-subject level and below 100mum at the group level for the empirical data, making the accuracy of cortical surfaces derived from super-resolution sufficient for most applications.
View details for DOI 10.1093/cercor/bhaa237
View details for PubMedID 32887984
-
Layer-specific analysis of femorotibial cartilage t2 relaxation time based on registration of segmented double echo steady state (dess) to multi-echo-spin-echo (mese) images.
Magma (New York, N.Y.)
2020
Abstract
OBJECTIVE: To develop and validate a 3D registration approach by which double echo steady state (DESS) MR images with cartilage thickness segmentations are used to extract the cartilage transverse relaxation time (T2) from multi-echo-spin-echo (MESE) MR images, without direct segmentations for MESE.MATERIALS AND METHODS: Manual DESS segmentations of 89 healthy reference knees (healthy) and 60 knees with early radiographic osteoarthritis (early ROA) from the Osteoarthritis Initiative were registered to corresponding MESE images that had independent direct T2 segmentations. For validation purposes, (a) regression analysis of deep and superficial cartilage T2 was performed and (b) between-group differences between healthy vs. early ROA knees were compared for registered vs. direct MESE analysis.RESULTS: Moderate to high correlations were observed for the deep (r=0.80) and the superficial T2 (r=0.81), with statistically significant between-group differences (ROA vs. healthy) of+1.4ms (p=0.002) vs.+1.3ms (p<0.001) for registered vs. direct T2 segmentation in the deep, and+1.3ms (p=0.002) vs.+2.3ms (p<0.001) in the superficial layer.DISCUSSION: This registration approach enables extracting cartilage T2 from MESE scans using DESS (cartilage thickness) segmentations, avoiding the need for direct MESE T2 segmentations.
View details for DOI 10.1007/s10334-020-00852-6
View details for PubMedID 32458188
-
Preoperative MRI of Articular Cartilage in the Knee: A Practical Approach.
The journal of knee surgery
2020; 33 (11): 1088–99
Abstract
Articular cartilage of the knee can be evaluated with high accuracy by magnetic resonance imaging (MRI) in preoperative patients with knee pain, but image quality and reporting are variable. This article discusses the normal MRI appearance of articular cartilage as well as the common MRI abnormalities of knee cartilage that may be considered for operative treatment. This article focuses on a practical approach to preoperative MRI of knee articular cartilage using routine MRI techniques. Current and future directions of knee MRI related to articular cartilage are also discussed.
View details for DOI 10.1055/s-0040-1716719
View details for PubMedID 33124010
-
MRSaiFE: Tissue Heating Prediction for MRI: a Feasibility Study
IEEE. 2020
View details for DOI 10.1109/IMBIOC47321.2020.9385044
View details for Web of Science ID 000675835900045
-
A Deep Learning Automated Segmentation Algorithm Accurately Detects Differences in Longitudinal Cartilage Thickness Loss - Data from the FNIH Biomarkers Study of the Osteoarthritis Initiative.
Arthritis care & research
2020
Abstract
To study the longitudinal performance of fully automated cartilage segmentation in knees with radiographic osteoarthritis (ROA). We evaluate the sensitivity to change in progressor knees from the Foundation National Institutes of Health OA Biomarkers Consortium between the automated and previously reported manual expert segmentation, and whether differences in progression rates between predefined cohorts can be detected by the fully automated approach.The Osteoarthritis Initiative Biomarker Consortium was a nested case-control study. Progressor knees had both medial tibiofemoral radiographic joint space width loss (≥0.7 mm) and a persistent increase in WOMAC pain (≥9 on a 0-100 scale) after two years from baseline (n=194), whereas non-progressor knees did not have either of both (n=200). Deep learning automated algorithms trained on ROA or healthy reference (HRC) knees were used to automatically segment medial (MFTC) and lateral femorotibial cartilage on baseline and two-year follow-up MRIs. Findings were compared with previously published manual expert segmentation.The MFTC cartilage loss in the progressor cohort was -181±245µm by manual (SRM=-0.74), -144±200µm by ROA-based model (SRM=-0.72), and -69±231µm by HRC-based model segmentation (SRM=-0.30). The Cohen's D for rates of progression between progressor vs. non-progressor cohort was -0.84 (p<0.001) for manual, -0.68 (p<0.001) for automated ROA-model, and -0.14 (p=0.18) for automated HRC-model segmentation.A fully automated deep learning segmentation approach not only displayed similar sensitivity to change of longitudinal cartilage thickness loss in knee OA as manual expert segmentation, but also effectively differentiates longitudinal rates of cartilage thickness loss between cohorts with different progression profiles.
View details for DOI 10.1002/acr.24539
View details for PubMedID 33337584
-
Time-saving opportunities in knee osteoarthritis: T2 mapping and structural imaging of the knee using a single 5-min MRI scan.
European radiology
2019
Abstract
OBJECTIVES: To assess the discriminative power of a 5-min quantitative double-echo steady-state (qDESS) sequence for simultaneous T2 measurements of cartilage and meniscus, and structural knee osteoarthritis (OA) assessment, in a clinical OA population, using radiographic knee OA as reference standard.METHODS: Fifty-three subjects were included and divided over three groups based on radiographic and clinical knee OA: 20 subjects with no OA (Kellgren-Lawrence grade (KLG) 0), 18 with mild OA (KLG2), and 15 with moderate OA (KLG3). All patients underwent a 5-min qDESS scan. We measured T2 relaxation times in four cartilage and four meniscus regions of interest (ROIs) and performed structural OA evaluation with the MRI Osteoarthritis Knee Score (MOAKS) using qDESS with multiplanar reformatting. Between-group differences in T2 values and MOAKS were calculated using ANOVA. Correlations of the reference standard (i.e., radiographic knee OA) with T2 and MOAKS were assessed with correlation analyses for ordinal variables.RESULTS: In cartilage, mean T2 values were 36.1±SD 4.3, 40.6±5.9, and 47.1±4.3ms for no, mild, and moderate OA, respectively (p<0.001). In menisci, mean T2 values were 15±3.6, 17.5±3.8, and 20.6±4.7ms for no, mild, and moderate OA, respectively (p<0.001). Statistically significant correlations were found between radiographic OA and T2 and between radiographic OA and MOAKS in all ROIs (p<0.05).CONCLUSION: Quantitative T2 and structural assessment of cartilage and meniscus, using a single 5-min qDESS scan, can distinguish between different grades of radiographic OA, demonstrating the potential of qDESS as an efficient tool for OA imaging.KEY POINTS: Quantitative T 2values of cartilage and meniscus as well as structural assessment of the knee with a single 5-min quantitative double-echo steady-state (qDESS) scan can distinguish between different grades of knee osteoarthritis (OA). Quantitative and structural qDESS-based measurements correlate significantly with the reference standard, radiographic degree of OA, for all cartilage and meniscus regions. By providing quantitative measurements and diagnostic image quality in one rapid MRI scan, qDESS has great potential for application in large-scale clinical trials in knee OA.
View details for DOI 10.1007/s00330-019-06542-9
View details for PubMedID 31844957
-
Evaluation of a Flexible 12-Channel Screen-printed Pediatric MRI Coil
RADIOLOGY
2019; 291 (1): 179–84
View details for DOI 10.1148/radiol.2019181883
View details for Web of Science ID 000465222600036
-
Simultaneous NODDI and GFA parameter map generation from subsampled q-space imaging using deep learning
MAGNETIC RESONANCE IN MEDICINE
2019; 81 (4): 2399–2411
View details for DOI 10.1002/mrm.27568
View details for Web of Science ID 000462092100015
-
Clinical evaluation of fully automated thigh muscle and adipose tissue segmentation using a U-Net deep learning architecture in context of osteoarthritic knee pain.
Magma (New York, N.Y.)
2019
Abstract
Segmentation of thigh muscle and adipose tissue is important for the understanding of musculoskeletal diseases such as osteoarthritis. Therefore, the purpose of this work is (a) to evaluate whether a fully automated approach provides accurate segmentation of muscles and adipose tissue cross-sectional areas (CSA) compared with manual segmentation and (b) to evaluate the validity of this method based on a previous clinical study.The segmentation method is based on U-Net architecture trained on 250 manually segmented thighs from the Osteoarthritis Initiative (OAI). The clinical evaluation is performed on a hold-out test set bilateral thighs of 48 subjects with unilateral knee pain.The segmentation time of the method is < 1 s and demonstrated high agreement with the manual method (dice similarity coeffcient: 0.96 ± 0.01). In the clinical study, the automated method shows that similar to manual segmentation (- 5.7 ± 7.9%, p < 0.001, effect size: 0.69), painful knees display significantly lower quadriceps CSAs than contralateral painless knees (- 5.6 ± 7.6%, p < 0.001, effect size: 0.73).Automated segmentation of thigh muscle and adipose tissues has high agreement with manual segmentations and can replicate the effect size seen in a clinical study on osteoarthritic pain.
View details for DOI 10.1007/s10334-019-00816-5
View details for PubMedID 31872357
-
3D Ultrashort TE MRI for Evaluation of Cartilaginous Endplate of Cervical Disk In Vivo: Feasibility and Correlation With Disk Degeneration in T2-Weighted Spin-Echo Sequence
AMERICAN JOURNAL OF ROENTGENOLOGY
2018; 210 (5): 1131–40
Abstract
The purpose of this study was to evaluate the feasibility of 3D ultrashort TE (UTE) MRI in depicting the cartilaginous endplate (CEP) and its abnormalities and to investigate the association between CEP abnormalities and disk degeneration on T2-weighted spin-echo (SE) MR images in cervical disks in vivo.Eight healthy volunteers and 70 patients were examined using 3-T MRI with the 3D UTE cones trajectory technique (TR/TE, 16.1/0.032, 6.6). In the volunteer study, quantitative and qualitative assessments of CEP depiction were conducted for the 3D UTE and T2-weighted SE imaging. In the patient study, CEP abnormalities were analyzed. Intersequence agreement between the images obtained with the first-echo 3D UTE sequence and the images created by subtracting the second-echo from the first-echo 3D UTE sequence (subtracted 3D UTE) and the intraobserver and interobserver agreements for 3D UTE overall were also tested. The CEP abnormalities on the 3D UTE images correlated with the Miyazaki grading of the T2-weighted SE images.In the volunteer study, the CEP was well visualized on 3D UTE images but not on T2-weighted SE images (p < 0.001). In the patient study, for evaluation of CEP abnormalities, intersequence agreements were substantial to almost perfect, intraobserver agreements were substantial to almost perfect, and interobserver agreements were moderate to substantial (p < 0.001). All of the CEP abnormalities correlated with the Miyazaki grade with statistical significance (p < 0.001).Three-dimensional UTE MRI feasibly depicts the CEP and CEP abnormalities, which may be associated with the severity of disk degeneration on T2-weighted SE MRI.
View details for PubMedID 29629793
-
Simultaneous bilateral-knee MR imaging.
Magnetic resonance in medicine
2018; 80 (2): 529–37
Abstract
To demonstrate and evaluate the scan time and quantitative accuracy of simultaneous bilateral-knee imaging compared with single-knee acquisitions.Hardware modifications and safety testing was performed to enable MR imaging with two 16-channel flexible coil arrays. Noise covariance and sensitivity-encoding g-factor maps for the dual-coil-array configuration were computed to evaluate coil cross-talk and noise amplification. Ten healthy volunteers were imaged on a 3T MRI scanner with both dual-coil-array bilateral-knee and single-coil-array single-knee configurations. Two experienced musculoskeletal radiologists compared the relative image quality between blinded image pairs acquired with each configuration. Differences in T2 relaxation time measurements between dual-coil-array and single-coil-array acquisitions were compared with the standard repeatability of single-coil-array measurements using a Bland-Altman analysis.The mean g-factors for the dual-coil-array configuration were low for accelerations up to 6 in the right-left direction, and minimal cross-talk was observed between the two coil arrays. Image quality ratings of various joint tissues showed no difference in 89% (95% confidence interval: 85-93%) of rated image pairs, with only small differences ("slightly better" or "slightly worse") in image quality observed. The T2 relaxation time measurements between the dual-coil-array configuration and the single-coil configuration showed similar limits of agreement and concordance correlation coefficients (limits of agreement: -0.93 to 1.99 ms; CCC: 0.97 (95% confidence interval: 0.96-0.98)), to the repeatability of single-coil-array measurements (limits of agreement: -2.07 to 1.96 ms; CCC: 0.97 (95% confidence interval: 0.95-0.98)).A bilateral coil-array setup can image both knees simultaneously in similar scan times as conventional unilateral knee scans, with comparable image quality and quantitative accuracy. This has the potential to improve the value of MRI knee evaluations. Magn Reson Med 80:529-537, 2018. © 2017 International Society for Magnetic Resonance in Medicine.
View details for PubMedID 29250856
View details for PubMedCentralID PMC5910219
-
A simple analytic method for estimating T2 in the knee from DESS.
Magnetic resonance imaging
2016; 38: 63-70
Abstract
To introduce a simple analytical formula for estimating T2 from a single Double-Echo in Steady-State (DESS) scan.Extended Phase Graph (EPG) modeling was used to develop a straightforward linear approximation of the relationship between the two DESS signals, enabling accurate T2 estimation from one DESS scan. Simulations were performed to demonstrate cancellation of different echo pathways to validate this simple model. The resulting analytic formula was compared to previous methods for T2 estimation using DESS and fast spin-echo scans in agar phantoms and knee cartilage in three volunteers and three patients. The DESS approach allows 3D (256×256×44) T2-mapping with fat suppression in scan times of 3-4min.The simulations demonstrated that the model approximates the true signal very well. If the T1 is within 20% of the assumed T1, the T2 estimation error was shown to be less than 5% for typical scans. The inherent residual error in the model was demonstrated to be small both due to signal decay and opposing signal contributions. The estimated T2 from the linear relationship agrees well with reference scans, both for the phantoms and in vivo. The method resulted in less underestimation of T2 than previous single-scan approaches, with processing times 60 times faster than using a numerical fit.A simplified relationship between the two DESS signals allows for rapid 3D T2 quantification with DESS that is accurate, yet also simple. The simplicity of the method allows for immediate T2 estimation in cartilage during the MRI examination.
View details for DOI 10.1016/j.mri.2016.12.018
View details for PubMedID 28017730
View details for PubMedCentralID PMC5360502