The International Workshop on Osteoarthritis Imaging Knee MRI Segmentation Challenge: A Multi-Institute Evaluation and Analysis Framework on a Standardized Dataset

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 (ρ < 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.

Effects of the Competitive Season and Off-Season on Knee Articular Cartilage in Collegiate Basketball Players Using Quantitative MRI: A Multicenter Study

BACKGROUND

Injuries to the articular cartilage in the knee are common in jumping athletes, particularly high-level basketball players. Unfortunately, these are often diagnosed at a late stage of the disease process, after tissue loss has already occurred.

PURPOSE/HYPOTHESIS

To evaluate longitudinal changes in knee articular cartilage and knee function in National Collegiate Athletic Association (NCAA) basketball players and their evolution over the competitive season and off-season.

STUDY TYPE

Longitudinal, multisite cohort study.

POPULATION

Thirty-two NCAA Division 1 athletes: 22 basketball players and 10 swimmers.

FIELD STRENGTH/SEQUENCE

Bilateral magnetic resonance imaging (MRI) using a combined T_1ρ and T₂ magnetization-prepared angle-modulated portioned k-space spoiled gradient-echo snapshots (MAPSS) sequence at 3T.

ASSESSMENT

We calculated T₂ and T_1ρ relaxation times to compare compositional cartilage changes between three timepoints: preseason 1, postseason 1, and preseason 2. Knee Osteoarthritis Outcome Scores (KOOS) were used to assess knee health.

STATISTICAL TESTS

One-way variance model hypothesis test, general linear model, and chi-squared test.

RESULTS

In the femoral articular cartilage of all athletes, we saw a global decrease in T₂ and T_1ρ relaxation times during the competitive season (all P < 0.05) and an increase in T₂ and T_1ρ relaxation times during the off-season (all P < 0.05). In the basketball players' femoral cartilage, the anterior and central compartments respectively had the highest T₂ and T_1ρ relaxation times following the competitive season and off-season. The basketball players had significantly lower KOOS measures in every domain compared with the swimmers: Pain (P < 0.05), Symptoms (P < 0.05), Function in Daily Living (P < 0.05), Function in Sport/Recreation (P < 0.05), and Quality of Life (P < 0.05).

CONCLUSION

Our results indicate that T₂ and T_1ρ MRI can detect significant seasonal changes in the articular cartilage of basketball players and that there are regional differences in the articular cartilage that are indicative of basketball-specific stress on the femoral cartilage. This study demonstrates the potential of quantitative MRI to monitor global and regional cartilage health in athletes at risk of developing cartilage problems.

LEVEL OF EVIDENCE

2 Technical Efficacy Stage: 2.

A neural network to predict the knee adduction moment in patients with osteoarthritis using anatomical landmarks obtainable from 2D video analysis

OBJECTIVE

The knee adduction moment (KAM) can inform treatment of medial knee osteoarthritis; however, measuring the KAM requires an expensive gait analysis laboratory. We evaluated the feasibility of predicting the peak KAM during natural and modified walking patterns using the positions of anatomical landmarks that could be identified from video analysis.

METHOD

Using inverse dynamics, we calculated the KAM for 86 individuals (64 with knee osteoarthritis, 22 without) walking naturally and with foot progression angle modifications. We trained a neural network to predict the peak KAM using the 3-dimensional positions of 13 anatomical landmarks measured with motion capture (3D neural network). We also trained models to predict the peak KAM using 2-dimensional subsets of the dataset to simulate 2-dimensional video analysis (frontal and sagittal plane neural networks). Model performance was evaluated on a held-out, 8-person test set that included steps from all trials.

RESULTS

The 3D neural network predicted the peak KAM for all test steps with r2( Murray et al., 2012) 2 = 0.78. This model predicted individuals' average peak KAM during natural walking with r2( Murray et al., 2012) 2 = 0.86 and classified which 15° foot progression angle modifications reduced the peak KAM with accuracy = 0.85. The frontal plane neural network predicted peak KAM with similar accuracy (r2( Murray et al., 2012) 2 = 0.85) to the 3D neural network, but the sagittal plane neural network did not (r2( Murray et al., 2012) 2 = 0.14).

CONCLUSION

Using the positions of anatomical landmarks from motion capture, a neural network accurately predicted the peak KAM during natural and modified walking. This study demonstrates the feasibility of measuring the peak KAM using positions obtainable from 2D video analysis.

MR method for measuring microscopic histologic soft tissue textures

PURPOSE

Provide a direct, non-invasive diagnostic measure of microscopic tissue texture in the size scale between tens of microns and the much larger scale measurable by clinical imaging. This paper presents a method and data demonstrating the ability to measure these microscopic pathologic tissue textures (histology) in the presence of subject motion in an MR scanner. This size range is vital to diagnosing a wide range of diseases.

THEORY/METHODS

MR micro-Texture (MRµT) resolves these textures by a combination of measuring a targeted set of k-values to characterize texture-as in diffraction analysis of materials, performing a selective internal excitation to isolate a volume of interest (VOI), applying a high k-value phase encode to the excited spins in the VOI, and acquiring each individual k-value data point in a single excitation-providing motion immunity and extended acquisition time for maximizing signal-to-noise ratio. Additional k-value measurements from the same tissue can be made to characterize the tissue texture in the VOI-there is no need for these additional measurements to be spatially coherent as there is no image to be reconstructed. This method was applied to phantoms and tissue specimens including human prostate tissue.

RESULTS

Data demonstrating resolution <50 µm, motion immunity, and clearly differentiating between normal and cancerous tissue textures are presented.

CONCLUSION

The data reveal textural differences not resolvable by standard MR imaging. As MRµT is a pulse sequence, it is directly translatable to MRI scanners currently in clinical practice to meet the need for further improvement in cancer imaging.

Diffusion Tensor Imaging of Skeletal Muscle Contraction Using Oscillating Gradient Spin Echo

Diffusion tensor imaging (DTI) measures water diffusion in skeletal muscle tissue and allows for muscle assessment in a broad range of neuromuscular diseases. However, current DTI measurements, typically performed using pulsed gradient spin echo (PGSE) diffusion encoding, are limited to the assessment of non-contracted musculature, therefore providing limited insight into muscle contraction mechanisms and contraction abnormalities. In this study, we propose the use of an oscillating gradient spin echo (OGSE) diffusion encoding strategy for DTI measurements to mitigate the effect of signal voids in contracted muscle and to obtain reliable diffusivity values. Two OGSE sequences with encoding frequencies of 25 and 50 Hz were tested in the lower leg of five healthy volunteers with relaxed musculature and during active dorsiflexion and plantarflexion, and compared with a conventional PGSE approach. A significant reduction of areas of signal voids using OGSE compared with PGSE was observed in the tibialis anterior for the scans obtained in active dorsiflexion and in the soleus during active plantarflexion. The use of PGSE sequences led to unrealistically elevated axial diffusivity values in the tibialis anterior during dorsiflexion and in the soleus during plantarflexion, while the corresponding values obtained using the OGSE sequences were significantly reduced. Similar findings were seen for radial diffusivity, with significantly higher diffusivity measured in plantarflexion in the soleus muscle using the PGSE sequence. Our preliminary results indicate that DTI with OGSE diffusion encoding is feasible in human musculature and allows to quantitatively assess diffusion properties in actively contracting skeletal muscle. OGSE holds great potential to assess microstructural changes occurring in the skeletal muscle during contraction, and for non-invasive assessment of contraction abnormalities in patients with muscle diseases.

Detection of knee synovitis using non-contrast-enhanced qDESS compared with contrast-enhanced MRI

Background

To assess diagnostic accuracy of quantitative double-echo in steady-state (qDESS) MRI for detecting synovitis in knee osteoarthritis (OA).

Methods

Patients with different degrees of radiographic knee OA were included prospectively. All underwent MRI with both qDESS and contrast-enhanced T1-weighted magnetic resonance imaging (CE-MRI). A linear combination of the two qDESS images can be used to create an image that displays contrast between synovium and the synovial fluid. Synovitis on both qDESS and CE-MRI was assessed semi-quantitatively, using a whole-knee synovitis sum score, indicating no/equivocal, mild, moderate, and severe synovitis. The correlation between sum scores of qDESS and CE-MRI (reference standard) was determined using Spearman’s rank correlation coefficient and intraclass correlation coefficient for absolute agreement. Receiver operating characteristic analysis was performed to assess the diagnostic performance of qDESS for detecting different degrees of synovitis, with CE-MRI as reference standard.

Results

In the 31 patients included, very strong correlation was found between synovitis sum scores on qDESS and CE-MRI (ρ = 0.96, p < 0.001), with high absolute agreement (0.84 (95%CI 0.14–0.95)). Mean sum score (SD) values on qDESS 5.16 (3.75) were lower than on CE-MRI 7.13 (4.66), indicating systematically underestimated synovitis severity on qDESS. For detecting mild synovitis or higher, high sensitivity and specificity were found for qDESS (1.00 (95%CI 0.80–1.00) and 0.909 (0.571–1.00), respectively). For detecting moderate synovitis or higher, sensitivity and specificity were good (0.727 (95%CI 0.393–0.927) and 1.00 (0.800–1.00), respectively).

Conclusion

qDESS MRI is able to, however with an underestimation, detect synovitis in patients with knee OA.

Utilizing shared information between gradient-spoiled and RF-spoiled steady-state MRI signals

This work presents an analytical relationship between gradient-spoiled and RF-spoiled steady-state signals. The two echoes acquired in double-echo in steady-state scans are shown to lie on a line in the signal plane, where the two axes represent the amplitudes of each echo. The location along the line depends on the amount of spoiling and the diffusivity. The line terminates in a point corresponding to an RF-spoiled signal. In addition to the main contribution of demonstrating this signal relationship, we also include the secondary contribution of preliminary results from an example application of the relationship, in the form of a heuristic denoising method when both types of scans are performed. This is investigated in simulations, phantom scans, and in vivo scans. For the signal model, the main topic of this study, simulations confirmed its accuracy and explored its dependency on signal parameters and image noise. For the secondary topic of its preliminary application to reduce noise, simulations demonstrated the denoising method giving a reduction in noise-induced standard deviation of about 30%. The relative effect of the method on the signals is shown to depend on the slope of the described line, which is demonstrated to be zero at the Ernst angle. The phantom scans show a similar effect as the simulations. In vivo scans showed a slightly lower average improvement of about 28%.

Rapid Knee MRI Acquisition and Analysis Techniques for Imaging Osteoarthritis

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.

Identifying Musculoskeletal Pain Generators Using Clinical PET

Identifying the source of a person's pain is a significant clinical challenge because the physical sensation of pain is believed to be subjective and difficult to quantify. The experience of pain is not only modulated by the individual's threshold to painful stimuli but also a product of the person's affective contributions, such as fear, anxiety, and previous experiences. Perhaps then to quantify pain is to examine the degree of nociception and pro-nociceptive inflammation, that is, the extent of cellular, chemical, and molecular changes that occur in pain-generating processes. Measuring changes in the local density of receptors, ion channels, mediators, and inflammatory/immune cells that are involved in the painful phenotype using targeted, highly sensitive, and specific positron emission tomography (PET) radiotracers is therefore a promising approach toward objectively identifying peripheral pain generators. Although several preclinical radiotracer candidates are being developed, a growing number of ongoing clinical PET imaging approaches can measure the degree of target concentration and thus serve as a readout for sites of pain generation. Further, when PET is combined with the spatial and contrast resolution afforded by magnetic resonance imaging, nuclear medicine physicians and radiologists can potentially identify pain drivers with greater accuracy and confidence. Clinical PET imaging approaches with fluorine-18 fluorodeoxyglucose, fluorine-18 sodium fluoride, and sigma-1 receptor PET radioligand and translocator protein radioligands to isolate the source of pain are described here.

Non-ionic CT contrast solutions rapidly alter bovine cartilage and meniscus mechanics

OBJECTIVE

To evaluate effects of a common CT contrast agent (iohexol) on the mechanical behaviors of cartilage and meniscus.

METHODS

Indentation responses of juvenile bovine cartilage and meniscus were monitored following exposure to undiluted contrast agent (100% CA), 50% CA/water, 50% CA/Phosphate Buffered Saline (PBS) or PBS alone, and during re-equilibration in PBS. The normalized peak force (Fpk¯), effective osmotic strain (εosm), and normalized effective contact modulus (Ec¯) were calculated for every cycle, with time constants determined for both exposure and recovery via mono- or biexponential fits to Fpk¯.

RESULTS

All cartilage CA groups exhibited long-term increases in Fpk¯ following exposure, although the hyperosmolal 100% CA and 50% CA/PBS groups showed an initial transient decrease. Meniscus presented opposing trends, with decreasing Fpk¯ for all CA groups. Re-equilibration in PBS for 1hr after exposure to 100% CA produced recovery to baseline Fpk¯ in cartilage but not in meniscus, and extended tests indicated that meniscus required ∼2.5 h to recover halfway. Ec¯ increased with CA exposure time for cartilage but decreased for meniscus, suggesting an increased effective stiffness for cartilage and decreased stiffness for meniscus. Long-term changes to εosm in both tissues were consistent with changes in Ec¯.

CONCLUSION

Exposure to iohexol solutions affected joint tissues differentially, with increased cartilage stiffness, likely relating to competing hyperosmotic and hypotonic interactions with tissue fixed charges, and decreased meniscus stiffness, likely dominated by hyperosmolarity. These altered tissue mechanics could allow non-physiological deformation during ambulatory weight-bearing, resulting in an increased risk of tissue or cell damage.

Rapid volumetric gagCEST imaging of knee articular cartilage at 3 T: evaluation of improved dynamic range and an osteoarthritic population

Chemical exchange saturation transfer of glycosaminoglycans, gagCEST, is a quantitative MR technique that has potential for assessing cartilage proteoglycan content at field strengths of 7 T and higher. However, its utility at 3 T remains unclear. The objective of this work was to implement a rapid volumetric gagCEST sequence with higher gagCEST asymmetry at 3 T to evaluate its sensitivity to osteoarthritic changes in knee articular cartilage and in comparison with T₂ and T_1ρ measures. We hypothesize that gagCEST asymmetry at 3 T decreases with increasing severity of osteoarthritis (OA). Forty-two human volunteers, including 10 healthy subjects and 32 subjects with medial OA, were included in the study. Knee Injury and Osteoarthritis Outcome Scores (KOOS) were assessed for all subjects, and Kellgren-Lawrence grading was performed for OA volunteers. Healthy subjects were scanned consecutively at 3 T to assess the repeatability of the volumetric gagCEST sequence at 3 T. For healthy and OA subjects, gagCEST asymmetry and T₂ and T_1ρ relaxation times were calculated for the femoral articular cartilage to assess sensitivity to OA severity. Volumetric gagCEST imaging had higher gagCEST asymmetry than single-slice acquisitions (p = 0.015). The average scan-rescan coefficient of variation was 6.8%. There were no significant differences in average gagCEST asymmetry between younger and older healthy controls (p = 0.655) or between healthy controls and OA subjects (p = 0.310). T₂ and T_1ρ relaxation times were elevated in OA subjects (p < 0.001 for both) compared with healthy controls and both were moderately correlated with total KOOS scores (rho = -0.181 and rho = -0.332 respectively). The gagCEST technique developed here, with volumetric scan times under 10 min and high gagCEST asymmetry at 3 T, did not vary significantly between healthy subjects and those with mild-moderate OA. This further supports a limited utility for gagCEST imaging at 3 T for assessment of early changes in cartilage composition in OA.

The prevalence of femoroacetabular impingement anatomy in Division 1 aquatic athletes who tread water

Femoroacetabular impingement (FAI) is a disorder that causes hip pain and disability in young patients, particularly athletes. Increased stress on the hip during development has been associated with increased risk of cam morphology. The specific forces involved are unclear, but may be due to continued rotational motion, like the eggbeater kick. The goal of this prospective cohort study was to use magnetic resonance imaging (MRI) to identify the prevalence of FAI anatomy in athletes who tread water and compare it to the literature on other sports. With university IRB approval, 20 Division 1 water polo players and synchronized swimmers (15 female, 5 male), ages 18-23 years (mean age 20.7 ± 1.4), completed the 33-item International Hip Outcome Tool and underwent non-contrast MRI scans of both hips using a 3 Tesla scanner. Recruitment was based on sport, with both symptomatic and asymptomatic individuals included. Cam and pincer morphology were identified. The Wilcoxon Signed-Rank/Rank Sum tests were used to assess outcomes. Seventy per cent (14/20) of subjects reported pain in their hips yet only 15% (3/20) sought clinical evaluation. Cam morphology was present in 67.5% (27/40) of hips, while 22.5% (9/40) demonstrated pincer morphology. The prevalence of cam morphology in water polo players and synchronized swimmers is greater than that reported for the general population and at a similar level as some other sports. From a clinical perspective, acknowledgment of the high prevalence of cam morphology in water polo players and synchronized swimmers should be considered when these athletes present with hip pain.

Automated Classification of Radiographic Knee Osteoarthritis Severity Using Deep Neural Networks

Purpose

To develop an automated model for staging knee osteoarthritis severity from radiographs and to compare its performance to that of musculoskeletal radiologists.

Materials and Methods

Radiographs from the Osteoarthritis Initiative staged by a radiologist committee using the Kellgren-Lawrence (KL) system were used. Before using the images as input to a convolutional neural network model, they were standardized and augmented automatically. The model was trained with 32 116 images, tuned with 4074 images, evaluated with a 4090-image test set, and compared to two individual radiologists using a 50-image test subset. Saliency maps were generated to reveal features used by the model to determine KL grades.

Results

With committee scores used as ground truth, the model had an average F1 score of 0.70 and an accuracy of 0.71 for the full test set. For the 50-image subset, the best individual radiologist had an average F1 score of 0.60 and an accuracy of 0.60; the model had an average F1 score of 0.64 and an accuracy of 0.66. Cohen weighted κ between the committee and model was 0.86, comparable to intraexpert repeatability. Saliency maps identified sites of osteophyte formation as influential to predictions.

Conclusion

An end-to-end interpretable model that takes full radiographs as input and predicts KL scores with state-of-the-art accuracy, performs as well as musculoskeletal radiologists, and does not require manual image preprocessing was developed. Saliency maps suggest the model's predictions were based on clinically relevant information. Supplemental material is available for this article. © RSNA, 2020.

Utility of deep learning super-resolution in the context of osteoarthritis MRI biomarkers

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 naïve interpolation and may improve image quality without biasing quantitative biomarkers.

LEVEL OF EVIDENCE

2 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2020;51:768-779.

pyKNEEr: An image analysis workflow for open and reproducible research on femoral knee cartilage

Transparent research in musculoskeletal imaging is fundamental to reliably investigate diseases such as knee osteoarthritis (OA), a chronic disease impairing femoral knee cartilage. To study cartilage degeneration, researchers have developed algorithms to segment femoral knee cartilage from magnetic resonance (MR) images and to measure cartilage morphology and relaxometry. The majority of these algorithms are not publicly available or require advanced programming skills to be compiled and run. However, to accelerate discoveries and findings, it is crucial to have open and reproducible workflows. We present pyKNEEr, a framework for open and reproducible research on femoral knee cartilage from MR images. pyKNEEr is written in python, uses Jupyter notebook as a user interface, and is available on GitHub with a GNU GPLv3 license. It is composed of three modules: 1) image preprocessing to standardize spatial and intensity characteristics; 2) femoral knee cartilage segmentation for intersubject, multimodal, and longitudinal acquisitions; and 3) analysis of cartilage morphology and relaxometry. Each module contains one or more Jupyter notebooks with narrative, code, visualizations, and dependencies to reproduce computational environments. pyKNEEr facilitates transparent image-based research of femoral knee cartilage because of its ease of installation and use, and its versatility for publication and sharing among researchers. Finally, due to its modular structure, pyKNEEr favors code extension and algorithm comparison. We tested our reproducible workflows with experiments that also constitute an example of transparent research with pyKNEEr, and we compared pyKNEEr performances to existing algorithms in literature review visualizations. We provide links to executed notebooks and executable environments for immediate reproducibility of our findings.

Time-saving opportunities in knee osteoarthritis: T2 mapping and structural imaging of the knee using a single 5-min MRI scan

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 ± 4.3 ms 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 ± 4.7 ms 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 T₂values 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.

Electronic supplementary material

The online version of this article (10.1007/s00330-019-06542-9) contains supplementary material, which is available to authorized users.

Does Injection of Hyaluronic Acid Protect Against Early Cartilage Injury Seen After Marathon Running? A Randomized Controlled Trial Utilizing High-Field Magnetic Resonance Imaging

BACKGROUND

Previous studies have shown that runners demonstrate elevated T2 and T1ρ values on magnetic resonance imaging (MRI) after running a marathon, with the greatest changes in the patellofemoral and medial compartment, which can persist after 3 months of reduced activity. Additionally, marathon running has been shown to increase serum inflammatory markers. Hyaluronic acid (HA) purportedly improves viscoelasticity of synovial fluid, serving as a lubricant while also having chondroprotective and anti-inflammatory effects.

PURPOSE/HYPOTHESIS

The purpose was to investigate whether intra-articular HA injection can protect articular cartilage from injury attributed to marathon running. The hypothesis was that the addition of intra-articular HA 1 week before running a marathon would reduce the magnitude of early cartilage breakdown measured by MRI.

STUDY DESIGN

Randomized controlled trial; Level of evidence, 2.

METHODS

After institutional review board approval, 20 runners were randomized into receiving an intra-articular injection of HA or normal saline (NS) 1 week before running a marathon. Exclusionary criteria included any prior knee injury or surgery and having run >3 prior marathons. Baseline 3-T knee MRI was obtained within 48 hours before the marathon (approximately 5 days after injection). Follow-up 3-T MRI scans of the same knee were obtained 48 to 72 hours and 3 months after the marathon. The T2 and T1ρ relaxation times of articular cartilage were measured in 8 locations-the medial and lateral compartments (including 2 areas of each femoral condyle) and the patellofemoral joint. The statistical analysis compared changes in T2 and T1ρ relaxation times (ms) from baseline to immediate and 3-month postmarathon scans between the HA and NS groups with repeated measures analysis of variance.

RESULTS

Fifteen runners completed the study: 6 women and 2 men in the HA group (mean age, 31 years; range, 23-50 years) and 6 women and 1 man in the NS group (mean age, 27 years; range, 20-49 years). There were no gross morphologic MRI changes after running the marathon. Postmarathon studies revealed no statistically significant changes between the HA and NS groups in all articular cartilage areas of the knee on both T2 and T1ρ relaxation times.

CONCLUSION

Increased T2 and T1ρ relaxation times have been observed in marathon runners, suggesting early cartilage injury. The addition of intra-articular HA did not significantly affect relaxation times in all areas of the knee when compared with an NS control.

Applications of PET-MRI in musculoskeletal disease

New integrated PET-MRI systems potentially provide a complete imaging modality for diagnosis and evaluation of musculoskeletal disease. MRI is able to provide excellent high-resolution morphologic information with multiple contrast mechanisms that has made it the imaging modality of choice in evaluation of many musculoskeletal disorders. PET offers incomparable abilities to provide quantitative information about molecular and physiologic changes that often precede structural and biochemical changes. In combination, hybrid PET-MRI can enhance imaging of musculoskeletal disorders through early detection of disease as well as improved diagnostic sensitivity and specificity. The purpose of this article is to review emerging applications of PET-MRI in musculoskeletal disease. Both clinical applications of malignant musculoskeletal disease as well as new opportunities to incorporate the molecular capabilities of nuclear imaging into studies of nononcologic musculoskeletal disease are discussed. Lastly, we discuss some of the technical considerations and challenges of PET-MRI as they specifically relate to musculoskeletal disease.

LEVEL OF EVIDENCE

5 TECHNICAL EFFICACY: Stage 3 J. Magn. Reson. Imaging 2018;48:27-47.

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

BACKGROUND

Clinical knee MRI protocols require upwards of 15 minutes of scan time.

PURPOSE/HYPOTHESIS

To compare the imaging appearance of knee abnormalities depicted with a 5-minute 3D double-echo in steady-state (DESS) sequence with separate echo images, with that of a routine clinical knee MRI protocol. A secondary goal was to compare the imaging appearance of knee abnormalities depicted with 5-minute DESS paired with a 2-minute coronal proton-density fat-saturated (PDFS) sequence.

STUDY TYPE

Prospective.

SUBJECTS

Thirty-six consecutive patients (19 male) referred for a routine knee MRI.

FIELD STRENGTH/SEQUENCES

DESS and PDFS at 3T.

ASSESSMENT

Five musculoskeletal radiologists evaluated all images for the presence of internal knee derangement using DESS, DESS+PDFS, and the conventional imaging protocol, and their associated diagnostic confidence of the reading.

STATISTICAL TESTS

Differences in positive and negative percent agreement (PPA and NPA, respectively) and 95% confidence intervals (CIs) for DESS and DESS+PDFS compared with the conventional protocol were calculated and tested using exact McNemar tests. The percentage of observations where DESS or DESS+PDFS had equivalent confidence ratings to DESS+Conv were tested with exact symmetry tests. Interreader agreement was calculated using Krippendorff's alpha.

RESULTS

DESS had a PPA of 90% (88-92% CI) and NPA of 99% (99-99% CI). DESS+PDFS had increased PPA of 99% (95-99% CI) and NPA of 100% (99-100% CI) compared with DESS (both P < 0.001). DESS had equivalent diagnostic confidence to DESS+Conv in 94% of findings, whereas DESS+PDFS had equivalent diagnostic confidence in 99% of findings (both P < 0.001). All readers had moderate concordance for all three protocols (Krippendorff's alpha 47-48%).

DATA CONCLUSION

Both 1) 5-minute 3D-DESS with separated echoes and 2) 5-minute 3D-DESS paired with a 2-minute coronal PDFS sequence depicted knee abnormalities similarly to a routine clinical knee MRI protocol, which may be a promising technique for abbreviated knee MRI.

LEVEL OF EVIDENCE

2 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2018.

Applications of PET-Computed Tomography-Magnetic Resonance in the Management of Benign Musculoskeletal Disorders

Although computed tomography (CT) and MR imaging alone have been used extensively to evaluate various musculoskeletal disorders, hybrid imaging modalities of PET-CT and PET-MR imaging were recently developed, combining the advantages of each method: molecular information from PET and anatomical information from CT or MR imaging. Furthermore, different radiotracers can be used in PET to uncover different disease mechanisms. In this article, potential applications of PET-CT and PET-MR imaging for benign musculoskeletal disorders are organized by benign cell proliferation/dysplasia, diabetic foot complications, joint prostheses, degeneration, inflammation, and trauma, metabolic bone disorders, and pain (acute and chronic) and peripheral nerve imaging.

Patellofemoral cartilage stresses are most sensitive to variations in vastus medialis muscle forces

The purpose of this study was to evaluate the effects of variations in quadriceps muscle forces on patellofemoral stress. We created subject-specific finite element models for 21 individuals with chronic patellofemoral pain and 16 pain-free control subjects. We extracted three-dimensional geometries from high resolution magnetic resonance images and registered the geometries to magnetic resonance images from an upright weight bearing squat with the knees flexed at 60°. We estimated quadriceps muscle forces corresponding to 60° knee flexion during a stair climb task from motion analysis and electromyography-driven musculoskeletal modelling. We applied the quadriceps muscle forces to our finite element models and evaluated patellofemoral cartilage stress. We quantified cartilage stress using an energy-based effective stress, a scalar quantity representing the local stress intensity in the tissue. We used probabilistic methods to evaluate the effects of variations in quadriceps muscle forces from five trials of the stair climb task for each subject. Patellofemoral effective stress was most sensitive to variations in forces in the two branches of the vastus medialis muscle. Femur cartilage effective stress was most sensitive to variations in vastus medialis forces in 29/37 (78%) subjects, and patella cartilage effective stress was most sensitive to variations in vastus medialis forces in 21/37 (57%) subjects. Femur cartilage effective stress was more sensitive to variations in vastus medialis longus forces in subjects classified as maltrackers compared to normal tracking subjects (p = 0.006). This study provides new evidence of the importance of the vastus medialis muscle in the treatment of patellofemoral pain.

PET-MRI for the Study of Metabolic Bone Disease

PURPOSE OF REVIEW

This review article attempts to summarize the current state and applications of the hybrid imaging modality of PET-MRI to metabolic bone diseases. The advances of PET and MRI are also discussed for metabolic bone diseases as potentially applied via PET-MRI.

RECENT FINDINGS

Etiologies and mechanisms of metabolic bone disease can be complex where molecular changes precede structural changes. Although PET-MRI has yet to be applied directly to metabolic bone disease, possible applications exist since PET, specifically ¹⁸F-NaF PET, can quantitatively track changes in bone metabolism and is useful for assessing treatment, while MRI can give detailed information on bone water concentration, porosity, and architecture through novel techniques such as UTE and ZTE MRI. Earlier detection and further understanding of metabolic bone disease via PET and MRI could lead to better treatment and prevention. More research using this modality is needed to further understand how it can be implemented in this realm.

Super-resolution musculoskeletal MRI using deep learning

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 (κ) 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 (κ = 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.

SNR-weighted regularization of ADC estimates from double-echo in steady-state (DESS)

PURPOSE

To improve the homogeneity and consistency of apparent diffusion coefficient (ADC) estimates in cartilage from the double-echo in steady-state (DESS) sequence by applying SNR-weighted regularization during post-processing.

METHODS

An estimation method that linearizes ADC estimates from DESS is used in conjunction with a smoothness constraint to suppress noise-induced variation in ADC estimates. Simulations, phantom scans, and in vivo scans are used to demonstrate how the method reduces ADC variability. Conventional diffusion-weighted echo-planar imaging (DW EPI) maps are acquired for comparison of mean and standard deviation (SD) of the ADC estimate.

RESULTS

Simulations and phantom scans demonstrated that the SNR-weighted regularization can produce homogenous ADC maps at varying levels of SNR, whereas non-regularized maps only estimate ADC accurately at high SNR levels. The in vivo maps showed that the SNR-weighted regularization produced ADC maps with similar heterogeneity to maps produced with standard DW EPI, but without the distortion of such reference scans.

CONCLUSION

A linear approximation of a simplified model of the relationship between DESS signals allows for fast SNR-weighted regularization of ADC maps that reduces estimation error in relatively short T₂ tissue such as cartilage.

Physical activity is associated with changes in knee cartilage microstructure

OBJECTIVE

The purpose of this study was to determine if there is an association between objectively measured physical activity and longitudinal changes in knee cartilage microstructure.

METHODS

We used accelerometry and T₂-weighted magnetic resonance imaging (MRI) data from the Osteoarthritis Initiative, restricting the analysis to men aged 45-60 years, with a body mass index (BMI) of 25-27 kg/m² and no radiographic evidence of knee osteoarthritis. After computing 4-year changes in mean T₂ relaxation time for six femoral cartilage regions and mean daily times spent in the sedentary, light, moderate, and vigorous activity ranges, we performed canonical correlation analysis (CCA) to find a linear combination of times spent in different activity intensity ranges (Activity Index) that was maximally correlated with a linear combination of regional changes in cartilage microstructure (Cartilage Microstructure Index). We used leave-one-out pre-validation to test the robustness of the model on new data.

RESULTS

Nineteen subjects satisfied the inclusion criteria. CCA identified an Activity Index and a Cartilage Microstructure Index that were significantly correlated (r = .82, P < .0001 on test data). Higher levels of sedentary time and vigorous activity were associated with greater medial-lateral differences in longitudinal T₂ changes, whereas light activity was associated with smaller differences.

CONCLUSIONS

Physical activity is better associated with an index that contrasts microstructural changes in different cartilage regions than it is with univariate or cumulative changes, likely because this index separates the effect of activity, which is greater in the medial loadbearing region, from that of patient-specific natural aging.