Abstract
Objective:
To validate the ankle-specific Pediatric Arthritis Ultrasound Scoring System (PAUSS-ankle) in children with juvenile idiopathic arthritis (JIA).
Methods:
Patients with a diagnosis of JIA prospectively underwent a standard clinical assessment and musculoskeletal ultrasound (MSUS) of one or both ankles. B-mode and Power-Doppler mode MSUS images were acquired and scored according to the PAUSS-ankle protocol. A subset of patients received a contrast-enhanced MRI (ceMRI) of the affected ankle. ceMRI scoring for synovitis was performed according to the Rheumatoid Arthritis MRI System (RAMRIS). Test characteristics of the PAUSS-ankle scores were evaluated with ceMRI as reference. Associations between the findings on physical examination, PAUSS-ankle, and RAMRIS were investigated.
Results:
Thirty-two patients with JIA contributed 63 MSUS and 15 ceMRIs of the ankles. The PAUSS-ankle total B-mode score had a moderate correlation with physical examination findings (correlation (r)=0.43, p<0.001). The PAUSS-ankle B-mode score ≥1 exhibited a sensitivity of 79% and specificity of 100%, demonstrating excellent diagnostic accuracy with an area under the curve (AUC)= 0.89 (confidence intervals, CI, 0.78–1.00) while clinical assessment had a sensitivity of 57% and AUC= 0.71 (CI: 0.58–0.85). The PAUSS-ankle B-mode score had significant strong correlations (r=0.68–0.90, p<0.005) with the RAMRIS for the assessment of disease severity for each joint area and the ankle joint as a whole.
Conclusion:
Our findings demonstrate excellent diagnostic accuracy of the PAUSS-ankle in detecting the presence and severity of ankle synovitis when compared to ceMRI. The PAUSS-ankle holds significant promise as an accurate measurement that may complement current clinical standards.
Keywords: Juvenile idiopathic arthritis, ankle, diagnostic imaging, musculoskeletal ultrasound
INTRODUCTION
Over the past two decades, there has been a remarkable expansion in the array of treatment options available for juvenile idiopathic arthritis (JIA), the most prevalent autoimmune inflammatory arthropathy in children [1]. Consistent and standardized assessment of joint inflammation is essential to support timely recognition and appropriate management of arthritis in children affected by JIA. The role of an expert examiner is to detect the presence of arthritis and to assess the burden of joint inflammation in children with JIA to guide treatment options (i.e., escalation or de-escalation of therapy) while considering the side effects associated with disease-modifying antirheumatic drugs [2]. Clinical diagnosis of arthritis in children requires either the presence of joint swelling or in its absence, the presence of joint limited range of motion (LROM) and pain [3]. However, clinician assessment of arthritis is influenced, among others, by the capability of a child to report joint pain (which is an age-dependent report), the presence of hypermobile joints in children (which is patient-specific and challenges clinician’s determination of LROM), and the ability of a clinician to recognize joint swelling. To illustrate, studies have reported low agreement between expert clinicians, particularly in anatomically complex joints like ankles [4].
Contrast-enhanced magnetic resonance imaging (ceMRI) serves as the current reference standard for the determination of presence and severity of joint inflammation. However, ceMRI challenges related to its relatively high cost, limited availability, the need for IV placement and the use of contrast, and for some children the need for sedation for its acquisition, limit its routine use in all patients with JIA [5]. In recent years, musculoskeletal ultrasound (MSUS) has emerged as a surrogate bedside instrument to support accurate evaluation of the presence and extent of arthritis in real-time [6]. In pediatric rheumatology, pediatric joint-specific MSUS synovitis scoring systems have been shown to be reliable and sensitive to change over time [6]. Notably, pilot validation study for the knee joint indicated a high accuracy of the knee-specific MSUS scoring system in evaluating arthritis in JIA when compared to ceMRI [7].
The ankle-specific Pediatric Arthritis Ultrasound Scoring System (PAUSS-ankle) has been developed as a comprehensive and reliable scanning protocol and scoring system for the assessment of synovitis and tenosynovitis in children [8]. This cross-sectional pilot study aimed to further validate PAUSS-ankle as an assessment tool for the evaluation of the ankle joint in patients with JIA against the current gold standard of ceMRI.
PATIENTS AND METHODS
Patients
Children aged 2 – 18 years old meeting a diagnosis of JIA according with the International League of Associations for Rheumatology classification criteria and presenting to the Cincinnati Children’s Hospital Medical Center Rheumatology clinic with: 1. ankle arthritis determined by an expert rheumatologist, or 2. patient-reported ankle pain or stiffness were enrolled in this study [9]. Children of all races and ethnicities, male and female were included as equally as possible between June 2020 and April 2023. Subjects were excluded if they were unable to tolerate MSUS or if they had received an intraarticular glucocorticoid injection of the ankle within four weeks before the study visit. Participants received a standard clinical assessment for ankle arthritis, subtalar joint (STJ) arthritis, and tendon involvement by experienced pediatric rheumatologists before imaging studies were collected. Subjects could contribute one or both ankles, depending on the presence of ankle arthritis, pain or stiffness at the time of the study. Participants were offered a non-sedated ceMRI of the affected ankle, provided there was no contraindication to undergo this procedure. This study was approved by the CCHMC Institutional Review Board and written informed consent with patient assent was obtained from all subjects in the study (Study number: 2018–7939).
Ultrasound assessment
All subjects had a MSUS of the affected ankle(s) according with the ankle-specific scanning protocol previously proposed by the Childhood Arthritis and Rheumatology Research Alliance (CARRA) US group [8]. In brief, B-mode and Power-Doppler (PD) mode still images were obtained at the area with maximum findings at the tibiotalar joint (TTJ), talonavicular joint (TNJ), STJ (anterior (ASTJ) and posterior (PSTJ) recesses were examined), anterior, medial and lateral tendon groups, as well as at the Achilles insertion [8]. Given the lack of communication between the ASTJ and PSTJ and because both aspects can be affected in JIA, the ASTJ and PSTJ recesses were included in this study [10]. All MSUS imaging was done by one of two pediatric rheumatologists (TT, PVF) with more than 10 years of experience in MSUS who were blinded to clinical data of the subject. MSUS studies were collected using a General Electric US Logiq S8 XDclear or Fortis machines equipped with multifrequency linear probes (4–20 MHz). De-identified MSUS images were stored in a secured research platform managed by the Imaging Research Center at CCHMC.
De-identified MSUS images were independently evaluated by two pediatric rheumatologists (TT, PVF) who remained blinded to clinical information. MSUS images were scored according to the semiquantitative (grade range 0–3) PAUSS-ankle for synovitis proposed by the CARRA US group [8]. For the current study, images with a B-mode MSUS grade of 0 and 1 were classified as physiologic findings and grade 2 and 3 were classified as findings of synovitis. For the PAUSS-ankle B-mode score, MSUS grade 0 and 1 were given a score of 0, while a B-mode MSUS grade 2 finding (moderate synovitis) was given a score of 1 and a B-mode MSUS grade 3 finding (severe synovitis) was given a score of 2. Abnormal PD grades 1, 2, or 3 were assigned a score of 1, 2, or 3, respectively. Therefore, the TTJ B-mode score ranged from 0–2 and the PD-mode score ranged from 0–3. The TNJ B-mode score and PD-mode score range from 0–2 and 0–3, respectively. The STJ B-mode score ranged from 0–4 and PD-mode score ranged from 0–6 because both the ASTJ and PSTJ scores were used for the assessment of STJ synovitis.
To calculate the PAUSS-ankle total B-mode score (range 0–12) and the PAUSS-ankle total PD-mode score (range 0–18), the TTJ and the TNJ scores were multiplied by 2 to avoid giving extra weight to the STJ scores, since the STJ scores resulted from the addition of the ASTJ and PSTJ. Subsequently, the TTJ, TNJ and STJ scores were added. To illustrate, the PAUSS-ankle total B-mode= (TTJ B-mode score x 2) + (TNJ B-mode score x 2) + (ASTJ B-mode score + PSTJ B-mode score). Tenosynovitis was defined as per the proposed Outcome Measures in Rheumatology (OMERACT) US definition of tenosynovitis in JIA, i.e. abnormal anechoic or hypoechoic tendon sheath widening, which can be related to both the presence of tenosynovial abnormal effusion or hypertrophy [11]. Tenosynovitis was graded on a binary scale (1- presence, 0- absence).
MRI assessment
The ceMRI was obtained immediately upon completion of the MSUS imaging. All ceMRIs utilized a 3T Philips Ingenia scanner (Philips Healthcare,Best, The Netherlands) and an 8-channel dStream foot/ankle coil (Philips Healthcare,Best, The Netherlands) with the following protocol: axial T2 fat-suppressed (FS), coronal T2 FS, sagittal T2 FS, sagittal 3D mFFE, sagittal T1 FS postcontrast, and axial T1 FS post-contrast sequences.
De-identified ceMRI images were scored by a pediatric radiologist (AM) with more than 10 years of experience in musculoskeletal MRI according to the OMERACT Rheumatoid Arthritis Magnetic Resonance Imaging scoring (RAMRIS) system [12]. Synovitis was defined as an area in the synovial compartment that showed above-normal post-gadolinium enhancement of a thickness greater than the width of the normal synovium. The semiquantitative OMERACT RAMRIS scoring system ranges from 0 (normal) to 3 (severe) with each score representing one-third of the maximum volume of the enhancing tissue in the synovial compartment [13]. RAMRIS scores for the ankle joint (RAMRIS-ankle) of ≥ 1 were classified as consistent with synovitis. Tenosynovitis, defined as peritendinous effusion and/or post-contrast enhancement of the tendon sheath, was graded on a binary scale (presence or absence) [13]. Ankle structures analyzed by ceMRI included the TTJ, TNJ, STJ, and midfoot recesses; the anterior, medial, and lateral tendon compartments; and the Achilles tendon insertion. The disparities between US and MRI findings were examined after all imaging data was analyzed.
Statistical analysis
Descriptive characteristics were reported using frequencies (percentages) for categorical data and means (standard deviations (SD)) and medians (interquartile ranges (IQR)) for continuous data. Fisher’s exact tests (for dichotomous variables) and Wilcoxon ranked sum tests (for continuous variables) were used to determine statistically significant differences between groups. Each ankle is treated as an independent and was the unit of analysis. Agreement between the two pediatric rheumatologists in scoring synovitis of the MSUS images was evaluated by Cohen’s kappa coefficient and almost perfect agreement was defined as a kappa ≥0.80 [14]. Strength of associations among physical exam, MSUS, and MRI findings were evaluated using Odds Ratios for binary variables, or Point-Biserial and Spearman’s correlations for continuous variables. Probability values associated with each comparison are also reported. The strength of the Point-Biserial and Spearman’s correlation coefficients can be interpreted as follows: very weak (0.0–0.19), weak (0.2–0.39), moderate (0.4–0.59), strong (0.6–0.79), or very strong (0.8–1.0) [15]. To investigate the association between PAUSS-ankle and RAMRIS-ankle scores, the RAMRIS scores of the TTJ and TNJ were doubled to not overcount the STJ findings (RAMRIS-ankle score = (2xTTJ) + (2xTNJ) + ASTJ + PSTJ). Receiver operating characteristic (ROC) curves are presented to show the area under the curve (AUC). An open-sourced ROCPLOT SAS macro was used to determine the MSUS cut-off score for best sensitivity and specificity in predicting MRI-detected synovitis. An alpha level of p < 0.05 showed statistical significance, and 95% confidence interval (CI) was provided. The SAS version 9.4 was used for statistical analysis.
RESULTS
Patients Characteristics.
A total of 32 patients (mean age 11.5 years) with JIA contributed a total of 63 ankles for this study. Demographic and clinical features of enrolled patients are shown in Table 1. Interrater reliability for the PAUSS-ankle was perfect (kappa=1) for the TTJ and STJ , and 0.75 (0.4–1.0) for the TNJ. Given the very low frequency of abnormal PD findings of arthritis (n=3) only B-mode data is presented in the current report.
Table 1.
Subject characteristics (n=32)
| Total number of ankles in the study, n | 63 | ||
| Age in years, mean (SD) | 11.5 (4.1) | ||
| Age in years at diagnosis, median (IQR) | 7.1 (4.1 – 13.2) | ||
| Disease duration in years, median (IQR) | 2.1 (0.9 – 6) | ||
| Gender, n (%) | Female | 24 (75) | |
| Male | 8 (25) | ||
| JIA subtype, n (%) | Oligoarticular JIA, persistent | 6 (18.7) | |
| Oligoarticular JIA, extended | 3 (9.4) | ||
| Polyarticular JIA, RF negative | 12 (37.5) | ||
| Polyarticular JIA, RF positive | 2 (6.2) | ||
| Psoriatic arthritis | 3 (9.4) | ||
| Enthesitis related arthritis | 3 (9.4) | ||
| Systemic JIA | 2 (6.2) | ||
| Undifferentiated JIA | 1 (3.1) | ||
| Physician global assessment of disease activity a , median (IQR) | 2 (0.5–3) | ||
| Treatment, n (%) | No | 1 (3.1) | |
|
Yesb
31 (96.9) |
NSAIDs | 23 (71.9) | |
| Conventional DMARDs | 20 (62.3) | ||
| Biologics DMARDs | 18 (56.3) | ||
| Oral glucocorticoid | 9 (28.1) | ||
SD: standard deviation, IQR: interquartile range, JIA: juvenile idiopathic arthritis, RF: rheumatoid factor, NSAID: non-steroidal anti-inflammatory drug, DMARD: disease-modifying antirheumatic drug.
Physician global assessment of disease activity ranges from 0 (no active disease) to 10 (very active disease);
list is not mutually exclusive.
PAUSS-ankle and clinical assessment of arthritis.
Approximately half of the ankles (n=32) were noted to have clinical presence of arthritis (Table 2). The median number of joints with an abnormal MSUS findings per ankle was 0 (IQR= 0–2; mean= 0.90, SD= 1.19) (Supplementary Figures 1–3). About 20% (7/31) of ankles that were considered to be free of inflammation by clinical examination were found to have moderate-to-severe synovitis per MSUS. Conversely, 65% (21/32) of the ankles that were thought to have active arthritis based on clinical examination failed to exhibit sonographic findings of arthritis. Of 44 ankles assessed for clinical presence of STJ arthritis, 5 (11%) were noted to have STJ arthritis clinically but only 3 of these 5 ankles were also positive for STJ synovitis on MSUS. The association between clinical exam and MSUS imaging of arthritis was statistically significant in that it was 6.5 times more likely to have a positive MSUS consistent with synovitis if physical examination was positive for arthritis (Table 2). The association between clinical and sonographic assessment of STJ arthritis was not statistically significant. The PAUSS-ankle total B-mode score showed a moderate correlation with physical examination findings (r=0.43; p<0.001).
Table 2.
Association between MSUS findings and clinical assessment of ankle arthritis
| Negative for clinical evidence of ankle arthritis (n=31) | Positive for clinical evidence of ankle arthritis (n=32) | Odds Ratios (95% CI) p-value | |
|---|---|---|---|
| Negative for MSUS findings of TTJ, TNJ, and STJ synovitis, n (%) | 24 (77.4) | 11 (34.4) | OR=6.5 (2.14–19.94) p<0.001 |
| Positive for MSUS findings of TTJ, TNJ, and STJ synovitis, n (%) | 7 (22.6) | 21 (65.6) | |
| Negative for STJ arthritis per PE (n=39) | Positive for STJ arthritis per PE (n=5) | Odds Ratios (95% CI) p-value | |
| Negative for MSUS findings of STJ synovitis, n (%) | 29 (74.4) | 2 (40.0) | OR=4.3 (0.63–29.91) p=0.123 |
| Positive for MSUS findings of STJ synovitis, n (%) | 10 (25.6) | 3 (60.0) | |
| Negative for clinical evidence of ankle arthritis, median (IQR) | Positive for clinical evidence of ankle arthritis, median (IQR) | Correlation (95% CI) p-valuea | |
| PAUSS-ankle total B-mode score | 0 (0–0) | 1 (0–4.5) | r=0.43 (0.21–0.62) p<0.001 |
MSUS: musculoskeletal ultrasound, CI: confidence interval, TTJ: tibiotalar joint, TNJ: talonavicular joint, STJ: subtalar joint, IQR: interquartile range, PAUSS-ankle: pediatric arthritis ultrasound scoring system of the ankle.
Synovitis detected by MSUS vs. ceMRI.
Fifteen children completed a ceMRI of the ankle. All of the ankles that were found to have an abnormal joint effusion were also found to meet the RAMRIS criteria for synovitis. Analysis of MSUS diagnostic test accuracy vs. ceMRI, with ceMRI as the reference, is presented in Table 3. Although the TTJ MSUS failed to identify four cases that had synovitis per ceMRI, MSUS did not overcall TTJ pathology. When examining MSUS STJ synovitis as the sum of the scores of the ASTJ and PSTJ, MSUS detected all subjects with STJ synovitis per MRI and only misclassified one subject as positive for STJ synovitis. The reasons for discrepancy of synovitis between MSUS and ceMRI were examined and are presented in Supplementary Table 1. Compared to ceMRI, PAUSS-ankle B-mode synovitis, defined as a PAUSS-ankle total B-mode score of ≥ 1, had excellent diagnostic accuracy (AUC (95% CI)=0.89 (0.78–1.00) (Figure 1). The correlations between the MSUS joint-specific scoring system for the ankle and the RAMRIS-ankle score per joint area are presented in Table 4. Statistically significant, strong correlations ranging from 0.68 – 0.81 between MSUS and MRI severity scores were observed for the TTJ, TNJ, and STJ. The PAUSS-ankle total B-mode score had a very strong correlation with the RAMRIS-ankle score (Supplementary Figure 4).
Table 3.
Sensitivity and specificity of MSUS for the detection of ankle synovitis when compared to ceMRI (n=15)
| MSUS | ceMRI | Sensitivity, specificity (95% CI) | |||
|---|---|---|---|---|---|
| Negative for synovitis, n (%) | Positive for synovitis, n (%) | ||||
| TTJ | Negative for synovitis, n (%) | 6 (100) | 4 (44.4) | 0.56 (0.23–0.88), 1.00 (1.00–1.00) |
|
| Positive for synovitis, n (%) | 0 (0.0) | 5 (55.6) | |||
| TNJ | Negative for synovitis, n (%) | 8 (88.9) | 2 (33.3) | 0.67 (0.29–1.00), 0.89 (0.68–1.00) |
|
| Positive for synovitis, n (%) | 1 (11.1) | 4 (66.7) | |||
| STJ | Negative for synovitis, n (%) | 5 (83.3) | 0 (0.0) | 1.00 (1.00–1.00), 0.83 (0.53–1.00) |
|
| Positive for synovitis, n (%) | 1 (16.7) | 9 (100) | |||
| STJ | Anterior recess of STJ | Negative for synovitis, n (%) | 9 (81.8) | 1 (25.0) | 0.75 (0.32–1.00), 0.82 (0.59–1.00) |
| Positive for synovitis, n (%) | 2 (18.2) | 3 (75.0) | |||
| Posterior recess of STJ | Negative for synovitis, n (%) | 6 (85.7) | 0 (0.0) | 1.00 (1.00–1.00), 0.86 (0.42–1.00) |
|
| Positive for synovitis, n (%) | 1 (14.3) | 8 (100) | |||
| Tenosynovitis | Negative, n (%) | 5 (83.3) | 5 (55.6) | 0.44 (0.14–0.79), 0.83 (0.36–1.00) |
|
| Positive, n (%) | 1 (16.7) | 4 (44.4) | |||
| PAUSS-ankle total B-mode score a | Negative for synovitis (score = 0, n (%) |
1 (100) | 3 (21.4) | 0.79 (0.57–1.00), 1.00 (1.00–1.00) |
|
| Positive for synovitis (score ≥ 1), n (%) |
0 (0.0) | 11 (78.6) | |||
MSUS: musculoskeletal ultrasound, ceMRI: contrast-enhanced magnetic resonance imaging, CI: confidence interval, TTJ: tibiotalar joint, TNJ: talonavicular joint, STJ: subtalar joint, PAUSS-ankle: pediatric arthritis ultrasound scoring system of the ankle.
PAUSS-ankle total B-mode score results from: (2 x TTJ score) + (2 x TNJ score) + (anterior recess of STJ) + (posterior recess of STJ).
Figure 1. Receiver operating curve to predict positive findings of arthritis on ceMRI when using PAUSS-ankle total B-mode score.
ceMRI: contrast-enhanced magnetic resonance imaging, PAUSS-ankle: pediatric arthritis ultrasound scoring system of the ankle, ROC: receiver operating curve.
PAUSS-ankle total B-mode score ≥ 1 has a sensitivity of 78% and a specificity of 100% in detecting positive findings of arthritis on MRI (AUC (95% CI): 0.89 (0.78–1.00)).
Table 4.
Correlation of MSUS-ankle scoring system and RAMRIS-ankle scores per joint area (n=15)
| RAMRIS-ankle score per joint area (95% CI) (p-valuea) | |
|---|---|
| TTJ B-mode MSUS score | 0.81 (0.51–0.93) (<0.001) |
| TNJ B-mode MSUS score | 0.68 (0.25–0.88) (0.004) |
| STJ B-mode MSUS score | 0.77 (0.42–0.92) (0.001) |
| RAMRIS-ankle score (95% CI) (p-value a ) | |
| PAUSS-ankle total B-mode score b | 0.90 (0.71–0.97) (<0.001) |
MSUS: musculoskeletal ultrasound, RAMRIS-ankle: rheumatoid arthritis magnetic resonance imaging scoring system of the ankle, CI: confidence interval, TTJ: tibiotalar joint, TNJ: talonavicular joint, STJ: subtalar joint, PAUSS-ankle: pediatric arthritis ultrasound scoring system of the ankle.
The strength of the correlation calculated using Spearman rank order correlations as follows: very weak: 0.0–0.19, weak: 0.2–0.39, moderate: 0.4–0.59, strong: 0.6–0.79, very strong: 0.8–1.0.
PAUSS-ankle total B-mode score results from: (2 x TTJ score) + (2 x TNJ score) + (anterior recess of STJ) + (posterior recess of STJ).
The sensitivity and specificity of MSUS in identifying tenosynovitis in any tendon compartment was 44% and 83%, respectively (Table 3). Tenosynovitis by ceMRI was mainly observed over the medial tendon compartment. MSUS did not have any false positive results per ceMRI findings for the medial tendon compartment and only misclassified one subject for presence of tenosynovitis in the lateral tendon compartment (Supplementary Table 2).
PAUSS-ankle, ceMRI assessment and clinical examination.
Table 5 compares the sensitivity and specificity of physical examination and MSUS against ceMRI as the reference. Briefly, clinical examination misclassified 7 of the 15 cases (sensitivity= 57%) while MSUS failed to identify pathology in only 3 of the 15 cases (sensitivity= 79%). When MSUS findings were used to complement the physical examination findings, the sensitivity for the detection of synovitis as per ceMRI increased to 86% but AUC decreased to 0.43 (95% CI=0.33–0.52).
Table 5.
Test characteristics of clinical assessment and MSUS assessment for findings of ankle arthritis as determined by ceMRIa
| Negative for ceMRI synovitis (n=1) | Positive for ceMRI synovitis (n=14) | Sensitivity, specificity (95% CI) | AUC (95% CI) | |
|---|---|---|---|---|
| Negative for clinical evidence of ankle arthritis, n (%) | 0 (0.0) | 6 (42.9) | 0.57 (0.28–0.82), 0.00 (0.00–0.00) |
0.71 (0.58–0.85) |
| Positive for clinical evidence of ankle arthritis, n (%) | 1 (100) | 8 (57.1) | ||
| Negative for MSUS synovitis, n (%) | 1 (100) | 3 (21.4) | 0.79 (0.57–1.00) 1.00 (1.00–1.00) |
0.89 (0.78–1.00) |
| Positive for MSUS synovitis, n (%) | 0 (0.0) | 11 (78.6) | ||
| Negative for clinical evidence of ankle arthritis and/or MSUS synovitis, n (%) | 0 (0.0) | 2 (14.3) | 0.86 (0.67–1.00) 0.00 (0.00–0.00) |
0.43 (0.33–0.52) |
| Positive for clinical evidence of ankle arthritis and/or MSUS synovitis, n (%) | 1 (100) | 12 (85.7) |
MSUS: musculoskeletal ultrasound, ceMRI: contrast-enhanced magnetic resonance imaging, AUC: under area curve, CI: confidence interval.
Evidence of ankle arthritis by clinical examination, MSUS, and ceMRI includes findings of arthritis over the tibiotalar, talonavicular, and/or subtalar joints.
DISCUSSION
Our findings demonstrated an excellent diagnostic accuracy (AUC=0.89) of the PAUSS-ankle total for detection of ankle synovitis when compared to ceMRI as the gold standard. In addition, a strong-to-very strong correlation of the PAUSS-ankle total B-mode score with the RAMRIS-ankle was determined. These results underscore the effectiveness of the PAUSS-ankle in assessing ankle disease activity in children with JIA.
Identification of the specific structure involved in JIA is fundamental as it may guide JIA disease classification as well as treatment approach. The limited ability of physical examination to determine the presence of ankle, TTJ, TNJ, and STJ arthritis against MSUS has been previously reported [16–19]. Lanni et al., even after applying cutoff values to minimize bias in the interpretation of US findings, found poor concordance between clinical evaluation of MSUS for all joint compartments of the ankle [16]. Similarly, Rooney et al. found that clinical examination failed to recognize the presence of tenosynovitis at the ankle in JIA [20]. In the current study, 20 to 25% of the ankles classified as normal by physical examination were found to have MSUS findings consistent with moderate to severe synovitis. To further validate our findings, we investigated the sensitivity of physical examination and MSUS to detect ankle arthritis when compared to ceMRI as reference standard. In line with our initial results, MSUS performed better than physical examination in detecting the presence of ankle arthritis and identifying the specific structures affected when comparing to ceMRI. The better accuracy of the PAUSS-ankle for detection of arthritis compared to physical examination highlights the value of MSUS as a complementary tool for a comprehensive evaluation of ankle arthritis in clinical practice.
In children with JIA, ceMRI has been recommended as the reference standard to evaluate joint and tendon inflammation [21]. The OMERACT-RAMRIS criteria measure changes related to active disease in the joint area (such as synovial thickening and bone marrow edema) as well as changes related to chronic inflammation (cartilage loss, bone erosions). Studies in children with JIA have successfully applied the RAMRIS criteria for evaluation of JIA disease activity [5,22]. When comparing clinical examination and ultrasound (US) findings of arthritis against MRI (as the reference standard), Wakefield et al. [23]. found in patients with rheumatoid arthritis (RA) that although physical examination was sensitive for the detection of synovitis, US was more specific. In their study, the sensitivity/specificity of US for detection of synovitis over the TTJ, TNJ, medial aspect of STJ, and lateral aspect of STJ were 63/72%, 89/69%, 50/80%, and 83/60%, respectively. The role of US as a complementary instrument to clinical examination was not explored [23]. Our study showed that grade 2 (moderate) and grade 3 (severe) B-mode findings, as per the PAUSS-ankle protocol, are highly accurate (sensitivity of 79%; specificity of 100%) to capture synovitis as defined by ceMRI, while the sensitivity and specificity of physical examination findings of arthritis was only 57% and 0% as compared to ceMRI. The PAUSS-ankle B-mode score aims to support the assessment of disease severity with a higher score indicative of severe disease activity. In our study, the PAUSS-ankle B-mode score and RAMRIS-ankle score per joint area and ankle as a whole showed strong-to-very strong correlations. Hence, the PAUSS-ankle emerges as a promising tool for capturing both the presence and severity of ankle arthritis, given that routine clinical surveillance of arthritis with ceMRI is cost prohibitive and not feasible. The potential inclusion of MSUS for the monitoring of ankle arthritis holds promise as an effective outcome measurement instrument.
In our study, when synovitis extended throughout the joint recess, MSUS was successful in recognizing the presence of pathology as identified per ceMRI. The limitation of MSUS compared to ceMRI was evident when pathology was localized to a certain area of the joint recess. In this setting MSUS did not detect abnormal findings possibly related to the limited acoustic window. In addition, given the conservative approach encouraged by the PAUSS, findings were classified as non-pathologic when borderline MSUS findings (grade 1 vs. grade 2) were appreciated. On the other hand, the conservative approach of the PAUSS-ankle led to very few false positive findings by MSUS. Our analysis highlighted the importance of dynamic sonographic examination of the entire area of interest to facilitate recognition of pathology and the role of the posterior view of the ankle to inspect the TTJ and PSTJ. In addition, it is to be considered that the acquisition of video recording vs. still images may increase the accuracy of MSUS.
Tenosynovitis in the ankle joint is frequently seen in JIA and may even be the dominant or the only pathologic finding [20,24]. Diagnosis and differentiation of tenosynovitis from synovitis by physical examination is challenging [16]. Isolated tenosynovitis may cause an overdiagnosis of arthritis which may lead to an incorrect classification of JIA (i.e. joint counts for the differentiation of oligo and polyarticular JIA) [24]. The sensitivity and specificity of MSUS for the detection of tenosynovitis in our study was similar to that reported by Wakefield et al. in patients with RA [23]. The reasons for discordance for tenosynovitis positive on ceMRI but negative on MSUS were involvement of the tendon in a deep area of the plantar aspect where MSUS has limited penetration and limited distention of the tenosynovium given vicinity with medial malleoli which was not detected by MSUS.
The small number of patients included in this pilot study, primarily those that completed a ceMRI, is the most important limitation of our study related to the considerable discomfort of receiving a ceMRI, its duration and/or need for IV placement. Additionally, it is worth noting that although PD imaging was performed in all MSUS examinations, the low prevalence of PD positivity limited the analysis of the PAUSS-ankle PD-score test characteristics. Further, the enrollment of JIA patients who presented with ankle arthritis as determined by expert clinician or with ankle pain and stiffness may introduce selection bias but the administration of ceMRI to a non-symptomatic child was against obtained ethical approval. Another point to consider is the use of the RAMRIS criteria, a non-validated instrument in JIA, for the assessment of ceMRI findings of the ankle. However, ceMRI was scored by a pediatric radiologist expert in musculoskeletal MRI using the RAMRIS criteria as previously reported in JIA [5]. A strength of our study is that we utilized a well-defined study population including different categories of JIA and that all study procedures were performed on the same day eliminated fluctuations in disease activity. Post-mortem meetings including pediatric rheumatologists and a pediatric radiologist provided important insights about the discrepancies between MSUS and MRI. This cross-sectional pilot study was designed to provide initial validity of the PAUSS-ankle as a measure for the assessment of presence and severity of ankle synovitis in children with JIA using ceMRI as reference standard. Future larger longitudinal studies will help to address the above-mentioned limitations and delineate the therapeutic and prognostic implications of the PAUSS-ankle.
Conclusion
In conclusion, as ankle involvement in JIA is an important cause of disability and has a predictive role in disease progression [25,26], accurate and early detection of inflammation is desirable. MSUS enhances the assessment of ankle involvement in patients with JIA beyond what is achievable through clinical evaluation alone. The PAUSS-ankle was found to be accurate in detecting both the presence and severity of ankle synovitis when compared to ceMRI. Taken together, these findings suggest that PAUSS-ankle holds significant promise as an objective and accurate measurement that may complement current clinical standards and may have diagnostic and therapeutic implications looking to improve the outcome of children with JIA and ankle involvement. Future and larger prospective multicenter studies are needed.
Supplementary Material
Sources of funding
This work was supported by the Center for Clinical & Translational Science & Training at the University of Cincinnati funded by the National Institutes of Health Clinical and Translational Science Award program [2UL1TR001425-05A1 and 2KL2TR001426-05A to PVF]; the National Institutes of Arthritis and Musculoskeletal Skin Diseases [P30AR076316 and K23AR081424A to PVF]. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH; and the Diversity and Health Disparity Award by the Cincinnati Children’s Hospital Medical Center (to PVF).
Footnotes
Conflict of Interest:
The authors have no financial interest to report.
Data availability
The data underlying this article will be shared on reasonable request to the corresponding author.
References
- 1.Ravelli A, Martini A. Juvenile idiopathic arthritis. Lancet. 2007;369(9563):767–78. [DOI] [PubMed] [Google Scholar]
- 2.Hinze C, Gohar F, Foell D. Management of juvenile idiopathic arthritis: Hitting the target. Nat Rev Rheumatol. 2015;11(5):290–300. [DOI] [PubMed] [Google Scholar]
- 3.Giannini EH, Ruperto N, Ravelli A, Lovell DJ, Felson DT, Martini A. Preliminary definition of improvement in juvenile arthritis. Arthritis Rheum. 1997;40(7):1202–9. [DOI] [PubMed] [Google Scholar]
- 4.Guzman J, Burgos-Vargas R, Duarte-Salazar C, Gomez-Mora P. Reliability of the articular examination in children with juvenile rheumatoid arthritis: Interobserver agreement and sources of disagreement. J Rheumatol. 1995;22(12):2331–6. [PubMed] [Google Scholar]
- 5.Mazzoni M, Pistorio A, Magnaguagno F, Viola S, Urru A, Magnano GM, et al. Predictive value of magnetic resonance imaging in patients with juvenile idiopathic arthritis in clinical remission. Arthritis Care Res. 2023;75(1):198–205. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Vega-Fernandez P, Ting TV., Oberle EJ, McCracken C, Figueroa J, Altaye M, et al. Musculoskeletal ultrasound in childhood arthritis limited examination: A comprehensive, reliable, time-efficient assessment of synovitis. Arthritis Care Res. 2023;75(2):401–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Vega-Fernandez P, Rogers K, Sproles A, Thornton S, Huggins J, Lovell DJ, et al. Diagnostic accuracy study of the pediatric-specific ultrasound scoring system for the knee joint in children with juvenile idiopathic arthritis. Arthritis Care Res (Hoboken). 2024;76(2):251–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Vega-Fernandez P, De Ranieri D, Oberle E, Clark M, Bukulmez H, Lin C, et al. Comprehensive and reliable sonographic assessment and scoring system for inflammatory lesions of the paediatric ankle. Rheumatol (United Kingdom). 2023;62(6):2239–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Petty RE, Southwood TR, Manners P, Baum J, Glass DN, He X, et al. International League of Associations for Rheumatology Classification of Juvenile Idiopathic Arthritis: Second Revision, Edmonton, 2001. J Rheumatol. 2004;31(2):390–2. [PubMed] [Google Scholar]
- 10.Mandl P, Bong D, Balint PV., Hammer HB, Miguel M, Naredo E, et al. Sonographic and Anatomic Description of the Subtalar Joint. Ultrasound Med Biol. 2018;44(1):119–23. [DOI] [PubMed] [Google Scholar]
- 11.Collado P, Martire MV, Lanni S, De Lucia O, Balint P, Guillaume-Czitrom S, et al. OMERACT International Consensus for ultrasound definitions of tenosynovitis in juvenile idiopathic arthritis: Systematic literature review and Delphi process. Arthritis Care Res. 2023;75(11):2277–84. [DOI] [PubMed] [Google Scholar]
- 12.Østergaard M, Peterfy C, Conaghan P, McQueen F, Bird P, Ejbjerg B, et al. OMERACT rheumatoid arthritis magnetic resonance imaging studies. Core set of MRI acquisitions, joint pathology definitions, and the OMERACT RA-MRI scoring system. J Rheumatol. 2003;30(6):1385–6. [PubMed] [Google Scholar]
- 13.Østergaard M, Peterfy CG, Bird P, Gandjbakhch F, Glinatsi D, Eshed I, et al. The OMERACT rheumatoid arthritis magnetic resonance imaging (MRI) scoring system: Updated recommendations by the OMERACT MRI in arthritis working group. J Rheumatol. 2017;44(11):1706–12. [DOI] [PubMed] [Google Scholar]
- 14.Altman DG. Practical Statistics for Medical Research [Internet]. First edit. New York: Chapman and Hall/CRC; 1990. Available from: https://www.taylorfrancis.com/books/9781000228816 [Google Scholar]
- 15.Mukaka MM. Statistics corner: A guide to appropriate use of correlation coefficient in medical research. Malawi Med J. 2012;24(3):69–71. [PMC free article] [PubMed] [Google Scholar]
- 16.Lanni S, Marafon DP, Civino A, Alongi A, Proverbio E, Agostoni C, et al. Comparison Between Clinical and Ultrasound Assessment of the Ankle Region in Children With Juvenile Idiopathic Arthritis. Arthritis Care Res. 2021;73(8):1180–6. [DOI] [PubMed] [Google Scholar]
- 17.Licciardi F, Petraz M, Covizzi C, Santarelli F, Cirone C, Mulatero R, et al. Discordance between Clinical and Ultrasound Examinations in Juvenile Idiopathic Arthritis: An Experimental Approach. Children. 2022;9(3). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Pascoli L, Wright S, McAllister C, Rooney M. Prospective evaluation of clinical and ultrasound findings in ankle disease in juvenile idiopathic arthritis: Importance of ankle ultrasound. J Rheumatol. 2010;37(11):2409–14. [DOI] [PubMed] [Google Scholar]
- 19.Lanni S, Bovis F, Ravelli A, Viola S, Magnaguagno F, Pistorio A, et al. Delineating the Application of Ultrasound in Detecting Synovial Abnormalities of the Subtalar Joint in Juvenile Idiopathic Arthritis. Arthritis Care Res. 2016;68(9):1346–53. [DOI] [PubMed] [Google Scholar]
- 20.Rooney ME, McAllister C, Burns JFT. Ankle disease in juvenile idiopathic arthritis: Ultrasound findings in clinically swollen ankles. J Rheumatol. 2009;36(8):1725–9. [DOI] [PubMed] [Google Scholar]
- 21.Colebatch-Bourn AN, Edwards CJ, Collado P, D’Agostinor MA, Hemke R, Jousse-Joulin S, et al. EULAR-PReS points to consider for the use of imaging in the diagnosis and management of juvenile idiopathic arthritis in clinical practice. Ann Rheum Dis. 2015;74(11):1946–57. [DOI] [PubMed] [Google Scholar]
- 22.Malattia C, Damasio MB, Pistorio A, Ioseliani M, Vilca I, Valle M, et al. Development and preliminary validation of a paediatric-targeted MRI scoring system for the assessment of disease activity and damage in juvenile idiopathic arthritis. Ann Rheum Dis. 2011;70(3):440–6. [DOI] [PubMed] [Google Scholar]
- 23.Wakefield RJ, Freeston JE, O’Connor P, Reay N, Budgen A, Hensor EMA, et al. The optimal assessment of the rheumatoid arthritis hindfoot: A comparative study of clinical examination, ultrasound and high field MRI. Ann Rheum Dis. 2008;67(12):1678–82. [DOI] [PubMed] [Google Scholar]
- 24.Magni-Manzoni S, Epis O, Ravelli A, Klersy C, Visconti C, Lanni S, et al. Comparison of clinical versus ultrasound-determined synovitis in juvenile idiopathic arthritis. Arthritis Care Res. 2009;61(11):1497–504. [DOI] [PubMed] [Google Scholar]
- 25.Al-Matar MJ, Petty RE, Tucker LB, Malleson PN, Schroeder ML, Cabral DA. The early pattern of joint involvement predicts disease progression in children with oligoarticular (pauciarticular) juvenile rheumatoid arthritis. Arthritis Rheum. 2002;46(10):2708–15. [DOI] [PubMed] [Google Scholar]
- 26.Minden K, Kiessling U, Listing J, Niewerth M, Doring E, Meincke J, et al. Prognosis of patients with juvenile chronic arthritis and juvenile spondyloarthropathy. J Rheumatol. 2000;27(9):2256–63. [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data underlying this article will be shared on reasonable request to the corresponding author.