Musculoskeletal Imaging Research Articles

The promise and limitations of artificial intelligence in musculoskeletal imaging.

With the recent developments in deep learning and the rapid growth of convolutional neural networks, artificial intelligence has shown promise as a tool that can transform several aspects of the musculoskeletal imaging cycle. Its applications can involve both interpretive and non-interpretive tasks such as the ordering of imaging, scheduling, protocoling, image acquisition, report generation and communication of findings. However, artificial intelligence tools still face a number of challenges that can hinder effective implementation into clinical practice. The purpose of this review is to explore both the successes and limitations of artificial intelligence applications throughout the muscuskeletal imaging cycle and to highlight how these applications can help enhance the service radiologists deliver to their patients, resulting in increased efficiency as well as improved patient and provider satisfaction.

Musculoskeletal CT Imaging: State-of-the-Art Advancements and Future Directions.

CT is one of the most widely used modalities for musculoskeletal imaging. Recent advancements in the field include the introduction of four-dimensional CT, which captures a CT image during motion; cone-beam CT, which uses flat-panel detectors to capture the lower extremities in weight-bearing mode; and dual-energy CT, which operates at two different x-ray potentials to improve the contrast resolution to facilitate the assessment of tissue material compositions such as tophaceous gout deposits and bone marrow edema. Most recently, photon-counting CT (PCCT) has been introduced. PCCT is a technique that uses photon-counting detectors to produce an image with higher spatial and contrast resolution than conventional multidetector CT systems. In addition, postprocessing techniques such as three-dimensional printing and cinematic rendering have used CT data to improve the generation of both physical and digital anatomic models. Last, advancements in the application of artificial intelligence to CT imaging have enabled the automatic evaluation of musculoskeletal pathologies. In this review, the authors discuss the current state of the above CT technologies, their respective advantages and disadvantages, and their projected future directions for various musculoskeletal applications.

Highlights of the Annual Scientific Meeting of the Australasian Musculoskeletal Imaging Group (AMSIG) 2023, Perth, Western Australia.

Highlights of the Annual Scientific Meeting of the Australasian Musculoskeletal Imaging Group (AMSIG) 2023, Perth, Western Australia.

Musculoskeletal Magnetic Resonance Imaging Revisited – Does Tesla of Magnetic Resonance Imaging Machines Matter?

The field of modern medical science has been revolutionized by magnetic resonance imaging (MRI) which is the preferred modality for the investigation of a whole spectrum of musculoskeletal (MSK) conditions. MRI is a careful interplay between the temporal, spatial, and contrast resolution which forms the foundation for its improved diagnostic performance and value. There are a lot of aspects that improve the image quality and diagnostic performance, however, a higher magnet strength of 3-Tesla has the biggest impact within the current diagnostic range. However various advancements in the hardware and software parameters such as multichannel multi-phased array coils, advanced gradient systems and better post processing techniques have significantly improved image quality at 1.5T scanners as well. All the leading manufacturers offer MRI systems with a higher field strength of 3T which are increasingly being used in recent clinical settings. Scanning at 3T has the advantage of a better signal-to-noise ratio which translates into better spatial and temporal resolution with the added advantage of faster acquisition. Challenges of 3T scanning include higher magnetic susceptibility, chemical shift, and higher radiofrequency energy deposition. This is particularly important in the presence of orthopedic implants because of the two-fold increase in susceptibility artifacts resulting in significant periprosthetic signal loss, signal displacements with voids and pileups, and failed spectral fat suppression. Various modifications are needed to minimize the artifacts at 3T scanners to better utilize the improved spatial and contrast resolution achieved as a result of scanning at a higher field strength. This review discusses the technical features of scanning at 1.5 and 3T scanners along with their clinical implications and diagnostic usefulness in MSK imaging.

Frequency-domain full-waveform inversion-based musculoskeletal ultrasound computed tomography.

Recently, full-waveform inversion (FWI) has become a promising tool for ultrasound computed tomography (USCT). However, as a computationally intensive technique, FWI suffers from computational burden, especially in conventional time-domain full-waveform inversion (TDFWI). On the contrary, frequency-domain full-waveform inversion (FDFWI) provides a relatively high computational efficiency as the propagation of discrete frequencies is much cheaper than full time-domain modeling. FDFWI has already been applied in soft tissue imaging, such as breast, but for the musculoskeletal model with high impedance contrast between hard and soft tissues, there is still a lack of an effective source estimation method. In this paper, a water-referenced data calibration method is proposed to address the source estimation challenge in the presence of bones, which achieves consistency between the measured and simulated data before the FDFWI procedure. To avoid the cycle-skipping local minimum effect and facilitate the algorithm convergence, a starting frequency criterion for musculoskeletal FDFWI is further proposed. The feasibility of the proposed method is demonstrated by numerical studies on retrieving the anatomies of the leg models and different musculoskeletal lesions. The study extends the advanced FDFWI method to the musculoskeletal system and provides an alternative solution for musculoskeletal USCT imaging with high computational efficiency.

Poster 328: Pre-Operative Low-Intensity Blood Flow Restriction Training Does Not Improve Quadriceps Strength or Hypertrophy After Anterior Cruciate Ligament Reconstruction

Objectives: The purpose of this study is to evaluate the effect of low-intensity blood flow restriction (LI- BFR) training utilized in a pre-rehabilitative setting for patients undergoing anterior cruciate ligament reconstruction (ACLR) on postoperative outcomes by minimizing postoperative atrophy and impairment of the quadriceps. We hypothesize that postoperative strength and hypertrophy of the quadriceps will be greater in a population that undergoes preoperative LI-BFR training than that of a population that undergoes low-intensity sham control (LI-SHAM) training. Methods: Twenty-eight subjects undergoing primary ACLR underwent 2 weeks of either a BFR intervention or a SHAM intervention at a rate of 2 sessions per week before surgery. Quadriceps strength and cross-sectional area were measured as well as other exploratory outcomes were measured before and after the intervention before surgery and at 5 different time points after surgery, through 6 months postoperative. Quadriceps strength was assessed through handheld dynamometry and a leg press test and quadriceps cross-sectional area was assessed using musculoskeletal ultrasound imaging. Results: Adjusting for sex, time, and strength/hypertrophy at baseline, there were no significant differences found between treatment groups for the primary endpoints. Compared to 24-48 hours pre- op, this suggests that the post-op average hypertrophy remains significantly lower through 24 weeks post-op, while leg strength is similar at 24 weeks post-op as measured by handheld dynamometer and even significantly larger at 24 weeks as measured by the leg press test. Conclusions: A 2-week preoperative program of LI-BFR training may not be effective to improve postoperative quadriceps strength and hypertrophy after ACLR.

Settings and Indications of Ultrasound in Imaging of Shoulder, Foot, and Ankle

Ultrasonography is a well-established musculoskeletal imaging technique with a multitude of advantages when compared to other modalities. It provides great spatial resolution in the evaluation of superficial articular and peri-articular structures including tendons, ligaments, bursae, and nerves. Given that it is the only modality which allows dynamic assessment, it also plays a crucial role in diagnosing impingement, subluxation/dislocation, and instability. The purpose of this article is to review the settings and indications of US in the imaging of shoulder, foot, and ankle in particular. Relevant literature, predominantly in the form of peer-reviewed journal articles was obtained from the electronic databases such as PubMed and MEDLINE and reviewed in a structured manner. This was combined with background knowledge and expertise in this field.

Radiologists’ perceptions of knowledge and training required by radiographers in the interpretation of radiographic images: An explorative study in KwaZulu-Natal province, South Africa

IntroductionThe health sector of South Africa is burdened by the shortage of radiologists leading to the under-reporting of radiographic images and ultimately mismanagement of patients. Previous studies have recommended training of radiographers in radiographic image interpretation in order to improve the reporting. There is paucity of information regarding the knowledge and training required by radiographers to interpret radiographic images. The purpose of this study was therefore to explore the knowledge and training required by diagnostic radiographers, according to radiologists, for the interpretation of radiographs. MethodA qualitative descriptive study employing criterion sampling to select qualified radiologists practicing in the eThekwini district of the KwaZulu Natal province, was conducted. One-on-one and in-depth, semi-structured interviews were used to collect data from three participants. The interviews were not carried out face to face as a result of the Covid 19 pandemic and the regulation of social distancing. This did not permit engagement with research communities. The data from the interviews were analysed using Tesch's eight steps for analysing qualitative data. ResultsFindings revealed that radiologists supported the interpretation of radiographic images by radiographers in rural settings, and for the radiographer's scope of practice to be restructured to include the reporting of chest and the musculoskeletal system images. The themes that emerged from the analysis included knowledge, training, clinical competencies and medico-legal responsibilities required by radiographers in the interpretation of radiographic images. ConclusionAlthough the radiologists support the training of radiographers in the interpretation of radiographic images, radiologists think that the scope of practice should be limited to the interpretation of the chest and musculoskeletal systems in rural areas only.

Ultrasound versus videos: A comparative study on the effectiveness of musculoskeletal anatomy education and student cognition.

Ultrasound imaging is a dynamic imaging technique that uses high-frequency sound waves to capture live images of the structures beneath the skin. In addition to its growing use in diagnosis and interventions, ultrasound imaging has the potential to reinforce concepts in the undergraduate medical curriculum. However, research assessing the impact of ultrasound on anatomy learning and student cognition is scarce. The purpose of this study was to compare the impact of ultrasound-based instruction versus narrated videos on students' understanding of anatomical relationships, as well as the role of intrinsic motivation, self-efficacy beliefs, and students' attitudes in this process. A booster course on anterior leg and wrist anatomy was offered to second-year medical students. A randomized controlled trial with a cross-over design allocated students to either an ultrasound-based teaching condition (cohort A) or a narrated anatomy video condition (cohort B). Next, participants were crossed to the alternative intervention. At the start of the study (T0), baseline anatomy knowledge, intrinsic motivation, self-efficacy beliefs, and spatial ability were measured. After the first intervention (T1) and at the end of the study (T2), both cohorts were administered an anatomy test, an intrinsic motivation scale, and a self-efficacy scale. In addition, each student was asked to fill out a perception survey after the ultrasound intervention. Finally, building on the cross-over design, the most optimal sequence of interventions was examined. A total of 181 students participated (cohort A: n = 82, cohort B: n = 99). Both cohorts performed comparably on the baseline anatomy knowledge test, spatial ability test, intrinsic motivation, and self-efficacy scale. At T1, cohort B outperformed cohort A on the anatomy test (p = 0.019), although only a small effect size could be detected (Cohen's d = 0.34). Intrinsic motivation and self-efficacy of both cohorts were similar at T1. At T2, the anatomy test, intrinsic motivation, and self-efficacy scale did not reflect an effect after studying either sequence of the interventions. Students reported favorably about the ultrasound experience, but also mentioned a steep learning curve. Medical students found the hands-on ultrasound sessions to be valuable, increasing their interest in musculoskeletal anatomy and ultrasound imaging. However, the addition of ultrasound did not result in superior spatial anatomy understanding compared to watching anatomy videos. In addition, ultrasound teaching did not have a major effect on student cognition. Ultrasound-based teaching of musculoskeletal anatomy is regarded as difficult to learn, and therefore it is hypothesized that too high levels of cognitive load might explain the presented results.

Dual-Energy CT in Cardiothoracic Imaging: Current Developments.

This article describes the technical principles and clinical applications of dual-energy computed tomography (DECT) in the context of cardiothoracic imaging with a focus on current developments and techniques. Since the introduction of DECT, different vendors developed distinct hard and software approaches for generating multi-energy datasets and multiple DECT applications that were developed and clinically investigated for different fields of interest. Benefits for various clinical settings, such as oncology, trauma and emergency radiology, as well as musculoskeletal and cardiovascular imaging, were recently reported in the literature. State-of-the-art applications, such as virtual monoenergetic imaging (VMI), material decomposition, perfused blood volume imaging, virtual non-contrast imaging (VNC), plaque removal, and virtual non-calcium (VNCa) imaging, can significantly improve cardiothoracic CT image workflows and have a high potential for improvement of diagnostic accuracy and patient safety.

Fast volumetric ultrasound facilitates high-resolution 3D mapping of tissue compartments.

Volumetric ultrasound imaging has the potential for operator-independent acquisition and enhanced field of view. Panoramic acquisition has many applications across ultrasound; spanning musculoskeletal, liver, breast, and pediatric imaging; and image-guided therapy. Challenges in high-resolution human imaging, such as subtle motion and the presence of bone or gas, have limited such acquisition. These issues can be addressed with a large transducer aperture and fast acquisition and processing. Programmable, ultrafast ultrasound scanners with a high channel count provide an unprecedented opportunity to optimize volumetric acquisition. In this work, we implement nonlinear processing and develop distributed beamformation to achieve fast acquisition over a 47-centimeter aperture. As a result, we achieve a 50-micrometer -6-decibel point spread function at 5 megahertz and resolve in-plane targets. A large volume scan of a human limb is completed in a few seconds, and in a 2-millimeter dorsal vein, the image intensity difference between the vessel center and surrounding tissue was ~50 decibels, facilitating three-dimensional reconstruction of the vasculature.

Ultrasound Fusion: Applications in Musculoskeletal Imaging.

Ultrasound fusion is an established technique that pairs real time B-scan ultrasound (US) with other forms of cross-sectional imaging, including computed tomography (CT), magnetic resonance imaging (MRI) and positron emission tomography (PET). Each of these imaging modalities has distinct advantages. CT provides superior anatomic resolution, with improved imaging of bone and calcified structures; MRI has superior contrast resolution; and PET provides physiologic information, identifying processes that are metabolically active (i.e., tumor, inflammatory conditions). However, these modalities are static. A key highlight of ultrasound is its capability of dynamic, real-time scanning. The ability to pair CT, MRI or PET with ultrasound can have significant advantages, both in diagnostic evaluation and when performing difficult or challenging image-guided interventions. Percutaneous interventions using ultrasound fusion have been described in the abdominal imaging literature; however, there have been very few musculoskeletal applications detailed in the literature. The purpose of this article is to review the basic concepts of real-time ultrasound fusion, and to detail, through the use of multiple case examples, its potential use as a safe and effective method for performing image-guided musculoskeletal interventions.

Generation of Musculoskeletal Ultrasound Images with Diffusion Models

The recent advances in deep learning have revolutionised computer-aided diagnosis in medical imaging. However, deep learning approaches to unveil their full potential require significant amounts of data, which can be a challenging task in some scientific fields, such as musculoskeletal ultrasound imaging, in which data privacy and security reasons can lead to important limitations in the acquisition and the distribution process of patients’ data. For this reason, different generative methods have been introduced to significantly reduce the required amount of real data by generating synthetic images, almost indistinguishable from the real ones. In this study, the power of the diffusion models is incorporated for the generation of realistic data from a small set of musculoskeletal ultrasound images in four different muscles. Afterwards, the similarity of the generated and real images is assessed with different types of qualitative and quantitative metrics that correspond well with human judgement. In particular, the histograms of pixel intensities of the two sets of images have demonstrated that the two distributions are statistically similar. Additionally, the well-established LPIPS, SSIM, FID, and PSNR metrics have been used to quantify the similarity of these sets of images. The two sets of images have achieved extremely high similarity scores in all these metrics. Subsequently, high-level features are extracted from the two types of images and visualized in a two-dimensional space for inspection of their structure and to identify patterns. From this representation, the two sets of images are hard to distinguish. Finally, we perform a series of experiments to assess the impact of the generated data for training a highly efficient Attention-UNet for the important clinical application of muscle thickness measurement. Our results depict that the synthetic data play a significant role in the model’s final performance and can lead to the improvement of the deep learning systems in musculoskeletal ultrasound.

Quantitative and qualitative image quality assessment in shoulder examinations with a first-generation photon-counting detector CT

Photon-counting detector (PCD) CT allows for ultra-high-resolution (UHR) examinations of the shoulder without requiring an additional post-patient comb filter to narrow the detector aperture. This study was designed to compare the PCD performance with a high-end energy-integrating detector (EID) CT. Sixteen cadaveric shoulders were examined with both scanners using dose-matched 120 kVp acquisition protocols (low-dose/full-dose: CTDIvol = 5.0/10.0 mGy). Specimens were scanned in UHR mode with the PCD-CT, whereas EID-CT examinations were conducted in accordance with the clinical standard as “non-UHR”. Reconstruction of EID data employed the sharpest kernel available for standard-resolution scans (ρ50 = 12.3 lp/cm), while PCD data were reconstructed with both a comparable kernel (11.8 lp/cm) and a sharper dedicated bone kernel (16.5 lp/cm). Six radiologists with 2–9 years of experience in musculoskeletal imaging rated image quality subjectively. Interrater agreement was analyzed by calculation of the intraclass correlation coefficient in a two-way random effects model. Quantitative analyses comprised noise recording and calculating signal-to-noise ratios based on attenuation measurements in bone and soft tissue. Subjective image quality was higher in UHR-PCD-CT than in EID-CT and non-UHR-PCD-CT datasets (all p < 0.001). While low-dose UHR-PCD-CT was considered superior to full-dose non-UHR studies on either scanner (all p < 0.001), ratings of low-dose non-UHR-PCD-CT and full-dose EID-CT examinations did not differ (p > 0.99). Interrater reliability was moderate, indicated by a single measures intraclass correlation coefficient of 0.66 (95% confidence interval: 0.58–0.73; p < 0.001). Image noise was lowest and signal-to-noise ratios were highest in non-UHR-PCD-CT reconstructions at either dose level (p < 0.001). This investigation demonstrates that superior depiction of trabecular microstructure and considerable denoising can be realized without additional radiation dose by employing a PCD for shoulder CT imaging. Allowing for UHR scans without dose penalty, PCD-CT appears as a promising alternative to EID-CT for shoulder trauma assessment in clinical routine.

Reliability of transversus abdominis and internal oblique muscle thickness measurement using B-mode ultrasound imaging in standing and single leg standing positions

Background: The thickness of the transversus abdominis (TrA) and internal oblique (IO) muscles are measured with high reliability by musculoskeletal ultrasound imaging (MSUSI) in the supine hook-lying and walking position. However, the reliability of the procedure to measure the TrA and IO muscle thickness measurement in the functional positions such as standing and single leg standing positions have not been well established. Objectives: To examine the reliability of MSUSI for the TrA and IO muscles thickness measurement during standing and single leg standing positions in healthy participants. Materials and methods: Ten healthy participants (aged 24.10±2.47 years) were recruited to perform MSUSI measurement of TrA and IO muscle thickness during the resting and abdominal drawing in maneuver (ADIM) in the standing and single leg standing positions. The intra-tester reliability of MSUSI measurement was performed by the same investigator in 2 different times (24 to 72 hours apart), while the inter-tester reliability was performed by 2 investigators within the same day. The intraclass correlation coefficient (ICC) whereby ICC3,3 was used to determine the intra-tester and ICC 2,3 was used to determine the inter-tester reliability of TrA and IO muscles thickness measurement using MSUSI. The standard error of measurement (SEM) and coefficient of variation (CV) were calculated. Results: The MSUSI measurements for TrA and IO muscles thickness showed high reliability in both standing and single leg standing positions (ICC3,3 and ICC2,3 &gt;0.9). Moreover, the values of SEM ranged between 0.055 to 0.662 mm, MDC ranged between 0.152 to 1.834 mm and CV ranged between 0.543 to 14.001%. Conclusion: The MSUSI measurement of TrA and IO muscles thickness could be performed with high intra and inter-tester reliability in standing and single leg standing positions among healthy individuals. The study findings on the MSUSI procedure to measure muscle thickness of TrA and IO would be useful for practice guidance in various research and clinical settings.

Diagnostic Radiology Fellowship Applicant Selection Criteria: A Survey of Program Directors

Introduction In this study, we aimed to investigate the importance of various characteristics used by diagnostic radiology subspecialty fellowship programs when selecting candidates for their fellowship programs. Methods An online survey was sent to program directors from multiple radiology fellowship programs in the United States. Multiple questions were asked, including applicant gender, personal statement, interview performance, research experience, United States Medical Licensing Exam (USMLE) performance, and residency program attended. Survey recipients were asked to rank these characteristics from least to most important. Results A total of 119 responses were collected (response rate of 26.2%), with a relatively balanced representation of the different subspecialties. An aggregate analysis of all the subspecialties showed interview performance, letters of recommendation, and residency program attended as the three most important factors considered by the program directors when selecting a candidate for their fellowship program. In contrast, the three least important factors were gender, prior subspecialty training, and minority status. The applicant's visa status, personal statement, and USMLE Step 3 performance were the only statistically significant (p<0.05) differences between the subspecialties. The women's and body imaging subspecialties rated the personal statement more important than musculoskeletal imaging and neuroradiology. Cardiothoracic and musculoskeletal imaging rated the applicant's visa status more important than neuroradiology. Women's and cardiothoracic imaging rated USMLE Step 3 performance as more important than musculoskeletal imaging. Conclusion Selecting the ideal candidate for a fellowship in radiology can be difficult. Our findings outline which applicant characteristics program directors believe are the most and least important when selecting an ideal candidate. For instance, personal characteristics such as interview performance and letters of recommendation are more valued than extracurricular activities such as research experience. Identifying these characteristics provides a better understanding of the fellowship application process and may guide future applicants.

Whole-Body MRI in Musculoskeletal Oncology: A Comprehensive Review with Recommendations.

Whole-body (WB) MRI has emerged as an attractive method for oncologic evaluation, potentially replacing conventional imaging modalities and providing a one-step wide-coverage assessment of both the skeleton and soft tissues. In addition to providing anatomic information, WB MRI may also yield a functional analysis with the inclusion of diffusion-weighted imaging (DWI). DWI translates microstructural changes, resulting in an excellent alternative to fluorodeoxyglucose PET/CT. WB MRI (with DWI) offers comparable accuracy to PET/CT and has the advantage of avoiding ionizing radiation. Technological advances and the development of faster protocols have prompted greater accessibility of WB MRI, with growing applications in routine practice for the diagnosis, staging, and follow-up of cancer. This review discusses the technical considerations, clinical applications, and accuracy of WB MRI in musculoskeletal oncology. Keywords: Pediatrics, MR Imaging, Skeletal-Axial, Skeletal-Appendicular, Soft Tissues/Skin, Bone Marrow, Extremities, Oncology, Musculoskeletal Imaging © RSNA, 2023.

Seeing More with Less: Clinical Benefits of Photon-counting Detector CT

Photon-counting detector (PCD) CT is an emerging technology that has led to continued innovation and progress in diagnostic imaging after it was approved by the U.S. Food and Drug Administration for clinical use in September 2021. Conventional energy-integrating detector (EID) CT measures the total energy of x-rays by converting photons to visible light and subsequently using photodiodes to convert visible light to digital signals. In comparison, PCD CT directly records x-ray photons as electric signals, without intermediate conversion to visible light. The benefits of PCD CT systems include improved spatial resolution due to smaller detector pixels, higher iodine image contrast, increased geometric dose efficiency to allow high-resolution imaging, reduced radiation dose for all body parts, multienergy imaging capabilities, and reduced artifacts. To recognize these benefits, diagnostic applications of PCD CT in musculoskeletal, thoracic, neuroradiologic, cardiovascular, and abdominal imaging must be optimized and adapted for specific diagnostic tasks. The diagnostic benefits and clinical applications resulting from PCD CT in early studies have allowed improved visualization of key anatomic structures and radiologist confidence for some diagnostic tasks, which will continue as PCD CT evolves and clinical use and applications grow. ©RSNA, 2023 Quiz questions for this article are available in the supplemental material. See the invited commentary by Ananthakrishnan in this issue.

Synthetic Inflammation Imaging with PatchGAN Deep Learning Networks.

Background: Gadolinium (Gd)-enhanced Magnetic Resonance Imaging (MRI) is crucial in several applications, including oncology, cardiac imaging, and musculoskeletal inflammatory imaging. One use case is rheumatoid arthritis (RA), a widespread autoimmune condition for which Gd MRI is crucial in imaging synovial joint inflammation, but Gd administration has well-documented safety concerns. As such, algorithms that could synthetically generate post-contrast peripheral joint MR images from non-contrast MR sequences would have immense clinical utility. Moreover, while such algorithms have been investigated for other anatomies, they are largely unexplored for musculoskeletal applications such as RA, and efforts to understand trained models and improve trust in their predictions have been limited in medical imaging. Methods: A dataset of 27 RA patients was used to train algorithms that synthetically generated post-Gd IDEAL wrist coronal T1-weighted scans from pre-contrast scans. UNets and PatchGANs were trained, leveraging an anomaly-weighted L1 loss and global generative adversarial network (GAN) loss for the PatchGAN. Occlusion and uncertainty maps were also generated to understand model performance. Results: UNet synthetic post-contrast images exhibited stronger normalized root mean square error (nRMSE) than PatchGAN in full volumes and the wrist, but PatchGAN outperformed UNet in synovial joints (UNet nRMSEs: volume = 6.29 ± 0.88, wrist = 4.36 ± 0.60, synovial = 26.18 ± 7.45; PatchGAN nRMSEs: volume = 6.72 ± 0.81, wrist = 6.07 ± 1.22, synovial = 23.14 ± 7.37; n = 7). Occlusion maps showed that synovial joints made substantial contributions to PatchGAN and UNet predictions, while uncertainty maps showed that PatchGAN predictions were more confident within those joints. Conclusions: Both pipelines showed promising performance in synthesizing post-contrast images, but PatchGAN performance was stronger and more confident within synovial joints, where an algorithm like this would have maximal clinical utility. Image synthesis approaches are therefore promising for RA and synthetic inflammatory imaging.

Consensus-Driven Definition for Unequivocal Sacroiliitis on Radiographs in Juvenile Spondyloarthritis.

Radiography is still used worldwide for the detection of sacroiliitis in juvenile spondyloarthritis (JSpA), despite its low sensitivity and reliability. We aimed to define unequivocal evidence of sacroiliitis on pelvic radiography in skeletally immature youth for use in classification criteria when magnetic resonance imaging (MRI) is unavailable. Subjects were a retrospective cohort of juvenile patients with spondyloarthritis with a radiograph and MRI as part of a diagnostic evaluation for axial disease. Six musculoskeletal imaging experts underwent an iterative consensus process to define unequivocal sacroiliitis on radiography in skeletally immature youth. Radiographs were graded using the modified New York (mNY) criteria and the unequivocal sacroiliitis criteria. Interrater agreement was assessed with the Fleiss [Formula: see text] statistic. Specificity, area under the receiver operator characteristic curve (AUROC), and sensitivity of the 2 measures were tested using 2 MRI reference standards. A total of 112 subjects, with a median age of 14.9 (range 6.7-20.1) years, were included. The Fleiss [Formula: see text] was fair for the mNY criteria (0.54, 95% CI 0.42-0.67) and the unequivocal sacroiliitis criteria (0.58, 95% CI 0.46-0.69). The unequivocal sacroiliitis criteria achieved > 90% specificity using both MRI reference standards. Sensitivity (59.26 and 57.14 vs 44.83 and 43.33) and AUROC (0.76 and 0.76 vs 0.71 and 0.71) were higher, for both reference standards, for the unequivocal sacroiliitis in youth definition than for the mNY criteria, respectively. In this study, we propose the first consensus-derived definition to our knowledge of unequivocal sacroiliitis by radiography in skeletally immature youth. This definition achieved excellent specificity and had higher AUROC and sensitivity values than the mNY criteria using both MRI reference standards. This definition has applicability to the JSpA axial disease classification imaging criterion when MRI is unavailable.