Quantifying synovitis in rheumatoid arthritis using computer-assisted manual segmentation with 3 Tesla MRI scanning

Computer‐Assisted Image Analysis in Assessment of Peripheral Joint MRI in Inflammatory Arthritis: A Systematic Review and Meta‐analysis

Objective

To summarize the feasibility of computer‐assisted quantification of joint pathologies on magnetic resonance imaging (MRI) in patients with inflammatory arthritis by evaluating the published data on reliability, validity, and feasibility.

Methods

A systematic literature search was performed for original articles published from January 1, 1985, to January 1, 2021. We selected studies in which patients with inflammatory arthritis were enrolled, and arthritis‐related structural damage/synovitis in peripheral joints was assessed on non‐contrast‐enhanced, contrast‐enhanced (CE), or dynamic CE (DCE)‐MRI using (semi)automated methods. Data were pooled using random‐effects model.

Results

Twenty‐eight studies consisting of 1342 MRIs were included (mean age, 54.8 years; 66.7% female; duration of arthritis, 3.6 years). Among clinical/laboratory factors, synovial membrane volume (SV) was moderately correlated with erthrocyte sedimentation rate (ESR) level (P < 0.01). Pooled analysis showed an overall excellent intra‐ and inter‐reader reliability for computer‐aided quantification of bone erosion volume (BEV; r = 0.97 [95% CI: 0.92‐0.99], 0.93 [0.87‐0.97]), SV (r = 0.98 [95% CI: 0.90‐0.99], 0.86 [0.78‐0.91]), and DCE‐MRI perfusion parameters (r = 0.96‐0.99). Meta‐regression showed that computer‐aided and manual methods provide comparable reliability (P > 0.05). Computer‐aided measurement of BEV (r = 0.92), SV (r = 0.82), and DCE‐MRI biomarkers (r = 0.72 N‐total; r = 0.74 N‐plateau; r = 0.64 N‐washout) were significantly correlated with the Rheumatoid Arthritis Magnetic Resonance Imaging Score (RAMRIS; P < 0.01), allowing for earlier assessment of drug efficacy. On average, (semi)automated analysis of BEV/SV took 17 minutes (vs. 9 minutes for the RAMRIS) and DCE‐MRI took 4 minutes (vs. 33 minutes for manual assessment).

Conclusion

Computer‐aided image quantification technologies demonstrate excellent reliability and validity when used to quantify MRI pathologies of peripheral joints in patients with inflammatory arthritis. Computer‐aided evaluation of inflammatory arthritis is an emerging field and should be considered as a viable complement to conventional observer‐based scoring methods for clinical trials application.

MRI of Hands with Early Rheumatoid Arthritis: Usefulness of Three-Point Dixon Sequences to Quantitatively Assess Disease Activity

The use of efficient treatment with a treat-to-target strategy combined with early detection of the disease completely changed the imaging presentation and outcome of newly diagnosed rheumatoid arthritis (RA) patients. Magnetic Resonance Imaging (MRI) has become the reference technique in clinical research to detect and quantify inflammatory involvement of the soft tissues (synovitis and tenosynovitis) and bone marrow (osteitis) along with structural damages of the bone (erosions) in hands of patients with RA. Three-point Dixon MRI may be a valuable alternative to the currently recommended sequences as it yields effective fat signal suppression, high imaging quality and reproducible assessment of disease activity.

Novel quantitative digital image analysis methodology for assessment of inflammatory changes in MRI data in a post-hoc analysis of data acquired from a phase IIb study of baricitinib in patients with active rheumatoid arthritis

PURPOSE

To evaluate a novel quantitative methodology to assess inflammatory changes in magnetic resonance imaging (MRI) data from patients with rheumatoid arthritis (RA) and the impact of image quality on imaging outcomes compared to the RA Magnetic Resonance Imaging Score (RAMRIS).

METHODS

Three-dimensional, T1-weighted, fat-suppressed MRI sequences of the hand/wrist before and after intravenous Gadolinium contrast from patients with RA in a placebo-controlled clinical trial (NCT01185353) were re-evaluated post hoc. The methodology was integrated into proprietary software (DYNAMIKA®) and assessed inflammation through pixelated measurements of the contrast-enhancing (inflammatory) volume. A semi-automatic approach outlined contrast-enhancing synovial tissue in the wrist and second to fifth metacarpophalangeal joints with a rough region of interest (ROI); quantitative imaging biomarkers were generated by means of quantitative total volume of inflammation and quantitative degree of inflammation relative to the signal in a 1 cm in diameter ROI in the center of the thenar or lumbrical muscle for internal reference. The time from Gadolinium injection to finalization of the post-contrast images was calculated from the images' Digital Imaging and Communications in Medicine header. An experienced reader graded image quality as poor, acceptable, or good.

RESULTS

Results from this quantitative methodology, especially when excluding images with poor quality scores (14-32%), provided a more pronounced and monotonically increasing dose-response than the original RAMRIS results on synovitis and osteitis.

CONCLUSIONS

This computer-aided quantitative scoring method provided continuous measures of inflammatory changes relative to muscle and may be more sensitive and interpretable concerning dose/response separation between RA treatment groups.

Computer-aided diagnosis based on hand thermal, RGB images, and grip force using artificial intelligence as screening tool for rheumatoid arthritis in women

Rheumatoid arthritis (RA) is an autoimmune disorder that typically affects people between 23 and 60 years old causing chronic synovial inflammation, symmetrical polyarthritis, destruction of large and small joints, and chronic disability. Clinical diagnosis of RA is stablished by current ACR-EULAR criteria, and it is crucial for starting conventional therapy in order to minimize damage progression. The 2010 ACR-EULAR criteria include the presence of swollen joints, elevated levels of rheumatoid factor or anti-citrullinated protein antibodies (ACPA), elevated acute phase reactant, and duration of symptoms. In this paper, a computer-aided system for helping in the RA diagnosis, based on quantitative and easy-to-acquire variables, is presented. The participants in this study were all female, grouped into two classes: class I, patients diagnosed with RA (n = 100), and class II corresponding to controls without RA (n = 100). The novel approach is constituted by the acquisition of thermal and RGB images, recording their hand grip strength or gripping force. The weight, height, and age were also obtained from all participants. The color layout descriptors (CLD) were obtained from each image for having a compact representation. After, a wrapper forward selection method in a range of classification algorithms included in WEKA was performed. In the feature selection process, variables such as hand images, grip force, and age were found relevant, whereas weight and height did not provide important information to the classification. Our system obtains an AUC ROC curve greater than 0.94 for both thermal and RGB images using the RandomForest classifier. Thirty-eight subjects were considered for an external test in order to evaluate and validate the model implementation. In this test, an accuracy of 94.7% was obtained using RGB images; the confusion matrix revealed our system provides a correct diagnosis for all participants and failed in only two of them (5.3%). Graphical abstract.

MRI and ultrasound of the hands and wrists in rheumatoid arthritis. I. Imaging findings

The management of patients with rheumatoid arthritis (RA) has rapidly evolved with the development of newer disease-modifying drugs and the recognition that long-term damage can be mitigated by an earlier and more-informed use of these medications. Historically, radiographs were the mainstay of imaging in RA patients, but radiographic joint narrowing and erosions are late and insensitive findings in the disease. MRI (with intravenous contrast agent) and ultrasound (with power Doppler interrogation) of the hands and wrists are able to demonstrate erosions earlier and with greater sensitivity than radiographs. More importantly, these imaging studies also depict synovitis and active soft-tissue inflammation, which represents a precursor to structural damage. Additionally, MRI can show inflammation within the bones (osteitis), which is proving to be the most important prognosticator of an aggressive disease course. Part I of this review discusses the imaging techniques, pitfalls, definitions, and comparative studies of MRI and ultrasound for identifying and quantifying erosions, synovitis, and osteitis. Part II will demonstrate how these imaging findings influence the clinical management of RA patients throughout their disease course, from presentation through clinical remission.

Fully automated segmentation of wrist bones on T2-weighted fat-suppressed MR images in early rheumatoid arthritis

Background

Magnetic resonance imaging (MRI) allows accurate determination of soft tissue and bone inflammation in rheumatoid arthritis. Inflammation can be measured semi-quantitatively using the well-established RA-MRI scoring system (RAMRIS), but its application is time consuming in routine clinical practice. To fully realize an automated quantitation of inflammation scoring for clinical use, automatic segmentation of the wrist bones on MR imaging is needed. Most previous studies extracted the wrist bones on T1-weighted (T1W) MR images, and then used registration to segment T2W fat-suppressed images for bone marrow oedema quantification, introducing spatial errors into the process. Relatively little work has tried segmentation directly from T2W fat-suppressed images and no prior study have used convolution neural network (CNN) to segment the wrist bones. The purpose of this study is to develop a CNN-based algorithm for automated segmentation of the wrist bones in early rheumatoid arthritis (ERA) patients on T2W fat-saturated MR images.

Methods

As preliminary tests indicated that out-of-the-box segmentation CNN U-net performance was compromised by close apposition of wrist tendons and bone, we separated the volumes prior to segmentation by using classification CNN Inception V3 to group images with similar features. The classified images were then segmented by individually trained U-net. We trained the networks on 40 cases and tested them on 11 cases derived from an MR imaging dataset of 51 patients with varying severity of ERA.

Results

We obtained a wrist bone segmentation with an average dice similarity coefficient (DICE) of 0.888±0.014, when compared to a manually drawn label. These results are comparable to existing atlas-based methods.

Conclusions

We have developed a fully automatic method to segment the wrist bones in ERA patients of varying severity directly from T2W fat-suppressed MR images. This compares well with manually drawn labels.

ERAMRS: a new MR scoring system for early rheumatoid arthritis of the wrist

PURPOSE

To (i) devise a new semi-quantitative scoring system known as Early Rheumatoid Arthritis Magnetic Resonance Score (ERAMRS) to assess inflammation of the wrist on magnetic resonance imaging in early rheumatoid arthritis and to (ii) test ERAMRS and other MR scoring systems against everyday used clinical scorings.

MATERIALS AND METHODS

One hundred six treatment-naïve patients (81 females, 25 males, mean age 53 ± 12 years) with early rheumatoid arthritis (ERA) underwent clinical/serological testing as well as 3-T MRI examination of the most symptomatic wrist. Clinical assessment included Disease Activity Score-28 and Health Assessment Questionnaire; erythrocyte sedimentation rate and C-reactive protein were measured. MR imaging data was scored in all patients using three devised MR semi-quantitative scoring systems, namely, the (a) ERAMRS system, (b) Rheumatoid Arthritis Magnetic Resonance Imaging Score (RAMRIS) system, and the (c) McQueen Score system.

RESULTS

Synovitis was present in 106 (100%), tenosynovitis in 98 (92%), and bone marrow edema in 84 (79%) of 106 ERA wrists. ERAMRS had the highest correlation with clinical disease activity scores (r = 0.476, p < 0.001) and serological parameters (r = 0.562, p < 0.001). RAMRIS system had the lowest correlation (r = 0.369, p < 0.001 for clinical disease activity; r = 0.436, p < 0.001 for serological parameters). RAMRIS synovitis subscore had a lower correlation than ERAMRS for clinical disease activity (r = 0.410, p < 0.001) and for serological parameters (r = 0.456, p < 0.001).

CONCLUSION

The ERAMRS system, designed to grade inflammation on wrist MRI in ERA, provided the best correlation with all clinical scoring systems and serological parameters, indicating its improved clinical relevance over other MR scoring systems.

KEY POINTS

• We devised a clinically relevant, easy-to-use semi-quantitative scoring system for scoring inflammation on MRI of the wrist in patients with early rheumatoid arthritis. • ERAMRS system showed better correlation with all clinical and serological assessment of inflammation in patients with early rheumatoid arthritis indicating its improved clinical relevance over other MR scoring systems.

Automated assessment of synovitis in 0.2T magnetic resonance images of the wrist

According to the current recommendations in diagnosis of rheumatoid arthritis (RA), Magnetic Resonance (MR) images of wrist joints are used to evaluate three main types of lesions: synovitis, bone edema and bone erosions. In the clinical practice, the RA-related lesions seen in MR images are assessed manually with the semi-quantitative RAMRIS scoring system. In this paper we present an automated method for inflamed synovial membrane volume determination, based on the analysis of pre- and post-contrast MR images and segmentation of wrist bones seen in MR images. We found that the correlation between the automatically quantified volume of synovitis and RAMRIS scores was in the range from 0.76 to 0.87 for the total RAMRIS synovitis score. This can be compared with the correlation between the manually quantified volume of synovitis and RAMRIS scores which was in the range from 0.75 to 0.81 for the total synovitis score. The results of the study demonstrate that computer assisted methods for assessment of synovitis have great potential for clinical applications.

Computer-aided and manual quantifications of MRI synovitis, bone marrow edema-like lesions, erosion and cartilage loss in rheumatoid arthritis of the wrist

PURPOSE

To investigate the reliability and validity of computer-aided automated and manual quantification as well as semiquantitative analysis for MRI synovitis, bone marrow edema-like lesions, erosion and cartilage loss of the wrist in rheumatoid arthritis (RA) compared to the OMERACT-RAMRIS.

METHODS AND MATERIALS

Wrist MRI was performed at 3 T in 16 patients with RA. Synovial volume and perfusion, bone marrow edema-like lesion (BMEL) volume, signal intensity and perfusion, and erosion dimensions were measured manually and using an in-house-developed automated software algorithm; findings were correlated with the OMERAC-RAMRIS gradings. In addition, a semiquantitative MRI cartilage loss score system was developed. Intraclass correlation coefficients (ICCs) were used to test the reproducibility of these quantitative and semiquantitative techniques. Spearman correlation coefficients were calculated between lesion quantifications and RAMRIS and between the MRI cartilage score and radiographic Sharp van der Heijde joint space narrowing scores.

RESULTS

The intra- and interobserver ICCs were excellent for synovial, BMEL and erosion quantifications and cartilage loss grading (all >0.89). The synovial volume, BMEL volume and signal intensity, and erosion dimensions were significantly correlated with the corresponding RAMRIS (r = 0.727 to 0.900, p < 0.05). Synovial perfusion parameter maximum enhancement (Emax) was significantly correlated with synovitis RAMRIS (r = 0.798). BMEL perfusion parameters were not correlated with the RAMRIS BME score. Cartilage loss gradings from MRI were significantly correlated with the Sharp joint space narrowing scores (r = 0.635, p = 0.008).

CONCLUSION

The computer-aided, manual and semiquantitative methods presented in this study can be used to evaluate MRI pathologies in RA with excellent reproducibility. Significant correlations with standard RAMRIS were found in the measurements using these methods.

Segmentation of bones in magnetic resonance images of the wrist

PURPOSE

Rheumatoid arthritis (RA) is a disease characterized by progressive and irreversible destruction of bones and joints. According to current recommendations, magnetic resonance imaging (MRI) is used to asses three main signs of RA based on manual evaluation of MR images: synovitis, bone edema and bone erosions. The key feature of a future computer-assisted diagnostic system for evaluation RA lesions is accurate segmentation of 15 wrist bones. In the present paper, we focus on developing a wrist bones segmentation framework.

METHOD

The segmentation procedure consisted of three stages: segmentation of the distal parts of ulna and radius, segmentation of the proximal parts of metacarpal bones and segmentation of carpal bones. At every stage, markers of bones were determined first, using an atlas-based approach. Then, given markers of bones and a marker of background, a watershed from markers algorithm was applied to find the final segmentation.

RESULTS

The MR data for 37 cases were analyzed. The automated segmentation results were compared with gold-standard manual segmentations using a few well-established metrics: area under ROC curve AUC, mean similarity MS and mean absolute distance MAD. The mean (standard deviation) values of AUC, MS and MAD were 0.97 (0.04), 0.93 (0.09) and 1.23 (0.28), respectively.

CONCLUSION

The results of the present study demonstrate that automated segmentation of wrist bones is feasible. The proposed algorithm can be the first stage for the detection of early lesions like bone edema or synovitis.

Insights into rheumatoid arthritis from use of MRI

Magnetic resonance imaging (MRI) is ideal for imaging the joints of rheumatoid arthritis (RA) patients. It produces anatomically detailed images of bone, cartilage, tendons and synovial membrane. It can reveal structural damage, in the form of bone erosion, cartilage thinning and/or tendon rupture, and regions of inflammation, using sequences that reveal water content and vascularity. MRI synovitis, tenosynovitis and bone oedema/osteitis all have prognostic significance, and MRI studies of RA have helped elucidate the mechanisms whereby bone and synovial inflammation lead to joint damage. Bone oedema/osteitis has become an important imaging biomarker, and can be used to help predict progression from undifferentiated arthritis to definite RA. Recent MRI studies have confirmed that subclinical inflammation is often present in patients in clinical remission, and these data may affect disease management. Finally, recent clinical trials are reviewed, in which MRI outcome measures are being established as sensitive response markers.

MRI in rheumatoid arthritis: a useful tool for the clinician?

Over the last two decades, MRI has emerged as an important clinical tool to assist in the diagnosis and management of rheumatic disease. In rheumatoid arthritis (RA), MRI has improved our understanding of the pathological basis of disease and has provided new information about imaging features that reflect joint inflammation and damage. Using MRI, we can now directly observe inflammation involving the synovial membrane and tenosynovium, plus joint damage including bone erosion and cartilage thinning. Inflammation of bone beneath the joint (osteitis) appears as bone oedema which is a feature unique to MRI and yields important diagnostic and prognostic information in patients with inflammatory arthritis. With the introduction of biologics to rheumatology clinical practice, sensitive tools are required to monitor disease activity and progression, so that the disease suppressing effect of these new agents can be measured. MRI fits the bill for this role as it can inform the clinician about the development of bone erosions well before plain radiography, and its ability to reveal cartilage damage is emerging. The use of MRI as a marker of outcome in clinical trials is being paralleled by its increasing role in the clinic. Both extremity and high field MRI have clinical applications in RA and need to be considered along with other advanced imaging techniques as useful tools to add to the clinician's armamentarium. This review will summarise recent advances in this field and will apply current knowledge to specific clinical scenarios relevant to modern rheumatology practice.

Spinal level of myelomeningocele lesion as a contributing factor in posterior fossa volume, intracranial cerebellar volume, and cerebellar ectopia

OBJECT

McLone and Knepper's unified theory of Chiari malformation Type II (CM-II) describes how the loss of CSF via the open posterior neuropore fails to create adequate distending pressure for the developing rhomboencephalic vesicle. The authors of the present article describe the relationship between the posterior fossa volume and intracranial cerebellar volume as being related to the distance from the obex of the fourth ventricle to the myelomeningocele lesion using a common mathematical model, the Hagen-Poiseuille law.

METHODS

All newborns who required closure of a myelomeningocele at the authors' institution between 2008 and 2011 and who were between 4 weeks premature and 2 months, corrected gestational age, at the time of MRI were included in this study. Volumes and measurements were obtained from axial and sagittal T2-weighted MR images of the brain and spine.

RESULTS

A total of 56 newborn infants met the inclusion criteria. There was a direct linear relationship between both posterior fossa volume and cerebellar volume and the spinal level of the myelomeningocele lesion (p = 0.0012 and p = 0.0041, respectively). There was a negative linear relationship between the cerebellar descent, the spinal level of the lesion, and posterior fossa volume and cerebellar volume. These relationships strengthen in patients with no syringomyelia and are not significant in those groups with syringomyelia. The results of a 1-way ANOVA for the 3 groups did not reach significance.

CONCLUSIONS

Using a linear equation derived from the Hagen-Poiseuille law that describes pressure in the fourth ventricle as being directly related to the length of the central canal from the obex to the myelomeningocele lesion, the authors were able to explain the directly observed linear relation between posterior fossa volume, intracranial cerebellar volume, and cerebellar descent to the level of the spinal lesion. As this model assumes a uniform radius of the central canal they were able to validate this model when they observed a strengthening in relationships in the no syringomyelia group and statistically insignificant correlations in the groups with syringomyelia. They therefore propose that the spinal level of the lesion is one of the major determinants of posterior fossa volume, intracranial cerebellar volume, and cerebellar ectopia.

Bilateral MR imaging of the hand and wrist in early and very early inflammatory arthritis: tenosynovitis is associated with progression to rheumatoid arthritis

PURPOSE

To identify bilateral hand and wrist findings of synovial inflammation associated with progression to rheumatoid arthritis (RA) in very-early-RA cohort (VERA) (duration, <3 months) and early-RA cohort (ERA) (duration, <12 but >3 months), to test tenosynovitis as a magnetic resonance (MR) imaging additional parameter for improving diagnostic accuracy of the 2010 American College of Rheumatology/European League Against Rheumatism (ACR/EULAR) RA classification criteria, and to evaluate the symmetry of joint and tendon involvement.

MATERIALS AND METHODS

With institutional review board approval and informed consent, 32 women and three men (mean age, 45 years) with untreated recent-onset inflammatory arthritis participated in this prospective study and underwent MR imaging of both wrists and hands. After 12-month follow-up, 25 patients fulfilled the criteria for RA (10 VERA and 15 ERA patients). Ten patients did not fulfill the criteria for RA (non-RA [control] group). Possible associations between synovitis for each joint and tendon and RA diagnosis at 12 months were tested (univariate logistic regression analysis). Diagnostic performance of the ACR/EULAR RA classification criteria was evaluated (receiver operating characteristic curve analysis). Asymmetry prevalence (all joints and tendons in the analysis) was calculated.

RESULTS

Tenosynovitis of the extensor carpi ulnaris (odds ratio, 3.21) and flexor tendons of the second finger (odds ratio, 14.61) in VERA group and synovitis of the radioulnar joint (odds ratio, 8.79) and tenosynovitis of flexor tendons of the second finger (odds ratio, 9.60) in ERA group were significantly associated with progression to RA (P < .05). Consideration of tenosynovitis improved areas under the receiver operating characteristic curve of ACR/EULAR criteria performance for the diagnosis of RA from 0.942 (P < .0001; sensitivity, 52%; specificity, 100%) to 0.972 (P < .0001; sensitivity, 76%; specificity, 100%), with cutoff score of 6 or greater. Asymmetry was found in 80.0% (62 of 77) (VERA patients) and 69.3% (106 of 153) (ERA patients) of joint or tendon pairs (P < .05).

CONCLUSION

Tenosynovitis is an imaging finding in early RA, and its inclusion as a scoring criterion might contribute for a better diagnostic performance of the 2010 ACR/EULAR classification; early RA is an asymmetric disease.

Magnetic resonance imaging of inflammatory arthropathies of peripheral joints

As the treatment of inflammatory arthropathies has advanced with new therapies that can slow or even halt the development of disabling disease, early and accurate diagnosis has become imperative. Magnetic resonance (MR) imaging has proved to be very sensitive in the detection of erosions, but more importantly, it can demonstrate pre-erosive changes. Detection of synovitis and edema-like bone marrow lesions for initial diagnosis and as an indicator of disease progression can provide crucial information leading to therapeutic interventions before permanent joint damage occurs. Understanding the characteristic intra-articular and extra-articular MR imaging findings of the inflammatory arthritides allows the radiologist to provide appropriate consultations in the care of these patients. The MR appearances of both intra-articular and extra-articular findings of inflammatory arthritis are presented. Despite the advances in imaging, however, many of the MR findings remain nonspecific, and radiologists must avoid overdiagnosis by synthesizing all of the clinical information available into their interpretations.