Advances and Challenges in Spondyloarthritis Imaging for Diagnosis and Assessment of Disease

The role of deep learning in diagnostic imaging of spondyloarthropathies: a systematic review

AIM

Diagnostic imaging is an integral part of identifying spondyloarthropathies (SpA), yet the interpretation of these images can be challenging. This review evaluated the use of deep learning models to enhance the diagnostic accuracy of SpA imaging.

METHODS

Following PRISMA guidelines, we systematically searched major databases up to February 2024, focusing on studies that applied deep learning to SpA imaging. Performance metrics, model types, and diagnostic tasks were extracted and analyzed. Study quality was assessed using QUADAS-2.

RESULTS

We analyzed 21 studies employing deep learning in SpA imaging diagnosis across MRI, CT, and X-ray modalities. These models, particularly advanced CNNs and U-Nets, demonstrated high accuracy in diagnosing SpA, differentiating arthritis forms, and assessing disease progression. Performance metrics frequently surpassed traditional methods, with some models achieving AUCs up to 0.98 and matching expert radiologist performance.

CONCLUSION

This systematic review underscores the effectiveness of deep learning in SpA imaging diagnostics across MRI, CT, and X-ray modalities. The studies reviewed demonstrated high diagnostic accuracy. However, the presence of small sample sizes in some studies highlights the need for more extensive datasets and further prospective and external validation to enhance the generalizability of these AI models.

KEY POINTS

Question How can deep learning models improve diagnostic accuracy in imaging for spondyloarthropathies (SpA), addressing challenges in early detection and differentiation from other forms of arthritis? Findings Deep learning models, especially CNNs and U-Nets, showed high accuracy in SpA imaging across MRI, CT, and X-ray, often matching or surpassing expert radiologists. Clinical relevance Deep learning models can enhance diagnostic precision in SpA imaging, potentially reducing diagnostic delays and improving treatment decisions, but further validation on larger datasets is required for clinical integration.

Axial Spondyloarthritis: Does Magnetic Resonance Imaging Classification Improve Report Interpretation

OBJECTIVE

The interpretation of magnetic resonance imaging (MRI) reports is crucial for the diagnosis of axial spondyloarthritis, but the subjective nature of narrative reports can lead to varying interpretations. This study presents a validation of a novel MRI reporting system for the sacroiliac joint in clinical practice.

METHODS

A historical review was conducted on 130 consecutive patients referred by 2 rheumatologists for initial MRI assessment of possible axial spondyloarthritis. The original MRI reports were interpreted by the rheumatologists and the radiologist who originally read the images and then categorized according to the novel system. Two musculoskeletal radiologists then reinterpreted the original MRI scans using the new system, and the resulting reports were interpreted and categorized by the same rheumatologists. The quality of the new framework was assessed by comparing the interpretations of both reports.

RESULTS

Ninety-two patients met the study criteria. The rheumatologists disagreed on the categorization of the original MRI reports in 12% of cases. The rheumatologists and original radiologists disagreed on the categorization of the initial report in 23.4% of cases. In contrast, there was 100% agreement between the rheumatologists and radiologists on the categorization of the new MRI report.

CONCLUSION

The new MRI categorization system significantly improved the agreement between the clinician and radiologist in report interpretation. The system provided a standard vocabulary for reporting, reduced variability in report interpretation, and may therefore improve clinical decision-making.

Characteristics of Late-Onset Spondyloarthritis: Data from the Moroccan Registry of Biological Therapies in Rheumatic Diseases

Introduction The Assessment of SpondyloArthritis International Society (ASAS) criteria for axial and peripheral spondyloarthritis (SpA) allow for the classification of patients with an age of onset of disease of less than 45 years. However, SpA can start after this age. This study aimed to assess the characteristics of late-onset SpA (SpA>45 years) in the Moroccan registry of biological therapies in rheumatic diseases (RBSMR). Methods A cross-sectional study was conducted using the baseline data of the RBSMR. The protocol for the original RBSMR study was reviewed and approved by the Ethics Committee for Biomedical Research Mohammed V University - Rabat, Faculty of Medicine and Pharmacy of Rabat (approval number for the study was 958/09/19, and the date of approval was September 11, 2019), and all patients had given their written consent. Patients who met the 2009 ASAS criteria for SpA were included. They were divided into two groups: early-onset SpA (≤ 45 years) and late-onset SpA (>45 years). Clinical, biological, radiological, and therapy data of the two groups were compared. Statistical analysis was performed using SPSS v25 software (IBM Corp. Armonk, NY). Parameters with a p-value ≤0.05 were considered significant. Results Our population consisted of 194 patients. Thirty-one patients (16%) had late-onset SpA. Comparison between patients with early-onset (≤45 years) and late-onset SpA (>45 years) revealed that late-onset SpA had a higher tender joint count (p=0.01), a higher swollen joint count (p=0.02), depression (p=0.00), fibromyalgia (p=0.001), hypercholesterolemia (p=0.01), and a lower frequency of coxitis (p=0.008). Logistic regression analysis confirmed that late-onset SpA was associated with a higher tender joint count (OR=0.93, CI 95%: 0.88-0.98), a higher swollen joint count (OR=0.92, CI 95%: 0.85-0.99), depression ( OR=0.19, CI 95%:0.04-0.38), fibromyalgia (OR=1.75, CI 95%: 1.74-17.85), and a lower frequency of coxitis ( OR=0.29, CI 95%: 0.11-0.75). Conclusion As life expectancy increases, late-onset SpA will become increasingly common. It is therefore imperative to determine its characteristics. In the RBSMR study, late-onset SpA was associated with a high number of tender and swollen joints, depression, fibromyalgia, and a lower frequency of coxitis.

Diagnostic utility of whole spine and thoracic spine MRI corner inflammatory lesions in axial spondyloarthritis

Background:

The presence of ⩾3 corner inflammatory lesions has been proposed as the definition of a positive spinal magnetic resonance imaging (MRI) for axial spondyloarthritis (axSpA), but subsequent studies showed inconclusive findings. Our objective was to evaluate whether locations of corner inflammatory lesions (CILs) would affect the diagnostic utility of MRI in axSpA.

Method:

Two groups were consecutively recruited from eight rheumatology centers in Hong Kong. The ‘axSpA’ group included 369 participants with a known diagnosis of axSpA. The ‘non-specific back pain’ (NSBP) control group consisted of 117 participants. Clinical, biochemical, and radiological parameters were collected and all patients underwent MRI of the spine and sacroiliac joints. CILs were assessed based on their locations (cervical, thoracic or lumbar) to determine the optimal cutoff for diagnosis.

Results:

The cutoff of ⩾5 whole spine CILs (W-CILs) and ⩾3 thoracic spine CILs (T-CILs) had comparable specificity to MRI sacroiliitis. Of 85/369 axSpA patients without sacroiliitis on conventional radiograph or MRI, 7 had ⩾5 W-CILs and 11 had ⩾3 T-CILs. Incorporating the proposed cutoffs into Assessment of SpondyloArthritis international Society axSpA criteria, ⩾5 W-CILs and ⩾3 T-CILs had similar performance when added to the imaging criteria for sacroiliitis (sensitivity 0.79 versus 0.80, specificity 0.92 versus 0.91).

Conclusion:

Spinal MRI provided little incremental diagnostic value in unselected axSpA patients. However, in patients without sacroiliitis on MRI or radiographs, 8–13% might be diagnosed by spinal MRI. Thoracic and whole spine MRI had similar diagnostic performance using the proposed cutoff of ⩾5 W-CILs and ⩾3 T-CILs.

Ankylosing spondylitis and axial spondyloarthritis: recent insights and impact of new classification criteria

Development of the Assessment in Spondyloarthritis International Society (ASAS) classification criteria for axial spondyloarthritis (SpA) was one of the major breakthroughs in the field over the past decade. Despite some concerns related to the specificity of the criteria, they stimulated research into the early stage of the disease. This resulted in major advances in the understanding of the course of the disease, revealing predictors of progression, improvement in early diagnosis and treatment in axial SpA. In this review, we summarize the recent developments resulting from the introduction of the ASAS classification criteria for axial SpA and the implications for research and clinical practice.

Does evaluation of the ligamentous compartment enhance diagnostic utility of sacroiliac joint MRI in axial spondyloarthritis?

Introduction

Inflammation of the sacroiliac joints (SIJ) is a fundamental clinical feature of axial spondyloarthritis (SpA). The anatomy of the irregularly shaped SIJ is complex with an antero-inferior cartilaginous compartment containing central hyaline and peripheral fibrocartilage, and a dorso-superior ligamentous compartment. Several scoring modules to systematically assess SIJ magnetic resonance imaging (MRI) in SpA have been developed. Nearly all of them are based on the cartilaginous joint compartment alone. However, there are only limited data about the frequency of inflammatory lesions in the ligamentous compartment and their potential diagnostic utility in axial SpA. We therefore aimed to evaluate the ligamentous compartment on sacroiliac joint MRI for lesion distribution and potential incremental value towards diagnosis of SpA over and above the traditional assessment of the cartilaginous compartment alone.

Methods

Two independent cohorts of 69 and 88 consecutive back pain patients ≤50 years were referred for suspected SpA (cohort A) or acute anterior uveitis plus back pain (cohort B). Patients were classified according to rheumatologist expert opinion based on clinical, radiographic and laboratory examination as having nonradiographic axial SpA (nr-axSpA; n = 51), ankylosing spondylitis (n = 34), or nonspecific back pain (NSBP; n = 72). Five blinded readers assessed SIJ MRI globally for presence/absence of SpA. Bone marrow edema (BME) and fat metaplasia were recorded in the cartilaginous and ligamentous compartment. The incremental value of evaluating the ligamentous additionally to the cartilaginous compartment alone for diagnosis of SpA was graded qualitatively. We determined the lesion distribution between the two compartments, and the impact of the ligamentous compartment evaluation on diagnostic utility.

Results

MRI bone marrow lesions solely in the ligamentous compartment in the absence of lesions in the cartilaginous compartment were reported in just 0–2.0/0–4.0 % (BME/fat metaplasia) of all subjects. Additional assessment of the ligamentous compartment was regarded as essential for diagnosis in 0 and 0.6 %, and as contributory in 28.0 and 7.7 % of nr-axSpA patients in cohorts A and B, respectively. Concomitant BME in both compartments was evident in 11.6–42.0 % of nr-axSpA and 2.1–2.4 % of NSBP patients.

Conclusion

Assessing the ligamentous compartment on SIJ MRI provided no incremental value for diagnosis of axial SpA. However, concomitant BME in both compartments may help discriminate nr-axSpA from NSBP.

Does spinal MRI add incremental diagnostic value to MRI of the sacroiliac joints alone in patients with non-radiographic axial spondyloarthritis?

OBJECTIVE

To assess the incremental diagnostic value of spine MRI evaluated separately from and combined with sacroiliac joint (SIJ) MRI in non-radiographic axial spondyloarthritis (nr-axSpA) compared with SIJ MRI alone.

METHODS

The study sample comprised two independent cohorts A/B of 130 consecutive patients aged ≤50 years with back pain, newly referred to two university clinics, and 20 healthy controls. Patients were classified according to clinical examination and pelvic radiographs as having nr-axSpA (n=50), ankylosing spondylitis (n=33), or non-specific back pain (n=47). Four readers assessed SIJ and spine MRI separately 6 months apart, and 1-12 months later both scans simultaneously using standardised modules. Readers recorded presence/absence of SpA and their level of confidence in this conclusion on a 0-10 scale (0=definitely not; 10=definite). We analysed differences between SIJ MRI versus spine MRI alone, and SIJ MRI alone versus combined MRI, descriptively by the number/percentage of subjects according to the mean of four readers.

RESULTS

In cohorts A/B, 15.8%/24.2% of patients with nr-axSpA having a negative SIJ MRI were reclassified as being positive for SpA by global evaluation of combined scans. However, 26.8%/11.4% of non-specific back pain controls and 17.5% of healthy volunteers with a negative SIJ MRI were falsely reclassified as having SpA by combined MRI. Low confidence in a diagnosis of SpA by SIJ MRI increased to high confidence by combined MRI in 6.6%/7.3% of patients with nr-axSpA.

CONCLUSIONS

Combined spine and SIJ MRI added little incremental value compared with SIJ MRI alone for diagnosing patients with nr-axSpA and enhancing confidence in this diagnosis.

Assessing structural changes in axial spondyloarthritis using a low-dose biplanar imaging system

OBJECTIVES

Patients with axial SpA experience repeated spine imaging. EOS is a new low-dose imaging system with significantly lower irradiation than conventional radiography (CR). The objective was to explore the EOS performances compared with CR for the classification and follow-up of SpA.

METHODS

We performed an observational, cross-sectional, single-centre study including SpA patients (definite diagnosis by expert opinion) and control patients [definite chronic mechanical low back pain (cLBP)]. All patients underwent pelvic and frontal and lateral CR of the entire spine and two-dimensional (2D) EOS imaging on the same day. Images were blindly assessed for sacroiliitis [modified New York criteria (mNY)] and for ankylosis of the spine [modified Stoke AS Spine Score (mSASSS)]. Global ease of interpretation was rated on a scale of 0-10. The primary outcome was intermodality agreement, with an a priori defined non-inferiority limit of 0.7. Interobserver, intra-observer and intermodality agreement were measured by kappa, weighted kappa, intraclass correlation coefficient and Bland-Altman plots.

RESULTS

Forty-eight SpA patients [mean age 47.6 years (s.d. 14.9), symptom duration 21.4 years (s.d. 13.3), 35 (70%) men] and 48 cLBP controls [mean age 49.1 years (s.d. 10.7), 9 (22.5%) men] were included. Intermodality agreement between EOS and CR was 0.50 (95% CI 0.26, 0.75) and 0.97 (95% CI 0.95, 0.98) for sacroiliitis and mSASSS, respectively. Ease of interpretation was greater for CR [8.2 (s.d. 0.9)] compared with EOS [7.2 (s.d. 0.8), P < 0.0001).

CONCLUSION

Our results suggest that EOS could replace CR for the follow-up of structural damage of the spine, but its place in the classification of sacroiliitis needs to be further explored.

Fat infiltration on magnetic resonance imaging of the sacroiliac joints has limited diagnostic utility in nonradiographic axial spondyloarthritis

OBJECTIVE

To explore whether morphological features of fat infiltration (FI) on sacroiliac joint (SIJ) magnetic resonance imaging (MRI) contribute to diagnostic utility in 2 inception cohorts of patients with nonradiographic axial spondyloarthritis (nr-axSpA).

METHODS

Four blinded readers assessed SIJ MRI in 2 cohorts (A/B) of 157 consecutive patients with back pain who were ≤ 50 years old, and in 20 healthy controls. Patients were classified according to clinical examination and pelvic radiography as having nr-axSpA (n = 51), ankylosing spondylitis (n = 34), or nonspecific back pain (n = 72). Readers recorded FI, bone marrow edema (BME), and erosion, predefined morphological features of FI (distinct border, homogeneity, subchondral location), and anatomical distribution of SIJ FI. The proportion of SIJ quadrants affected by FI and frequencies of various SIJ FI features were analyzed descriptively. We calculated positive/negative likelihood ratios (LR) to estimate the diagnostic utility of various features of FI, with and without associated BME, and erosion.

RESULTS

Of the patients with nr-axSpA in cohorts A/B, 45.0%/48.4% had FI in ≥ 2 SIJ quadrants. Of those, 25.0%/22.6% and 20.0%/25.8% showed FI with distinct border or homogeneous pattern, respectively, and 50% to 100% of those patients displayed concomitant BME or erosion. FI per se in ≥ 2 SIJ quadrants had no diagnostic utility (LR+ 1.62/1.91). FI with distinct border (LR+ 8.29/2.13) or homogeneity (LR+ 6.24/3.78) demonstrated small to moderate diagnostic utility.

CONCLUSION

SIJ FI per se was not of clinical utility in recognition of nr-axSpA. Distinct border or homogeneity of FI on SIJ MRI showed small to moderate diagnostic utility in nr-axSpA, but were strongly associated with concomitant BME or erosion, highlighting the contextual interpretation of SIJ MRI.