Assessment of Hip Involvement in Patients With Ankylosing Spondylitis: The Reliability Between the HIMRISS and BASRI-Hip System

Background. The aim is to test the reliability and validity of the Hip Inammation MRI Scoring System (HIMRISS) in assessing hip involvement of AS patients at different stages of the bath ankylosing spondylitis radiology index (BASRI-hip) scoring system. Methods. Fifty-two outpatients with ankylosing spondylitis(AS) were included in this study. The subjects’ data includes demographics, clinical characteristics, disease activity score, and functional index. Based on the Harris hip scoring (HHS) of involved hip and BASRI-hip score, we devided these patients into no hip involvement group(HHS ≥ 80 and BASRI ≤ 1)(Group A), mild hip involvement subgroup (BASRI=2 or BASRI ≤ 1 and HHS ≤ 79) (Group B) and moderate to advanced hip involvement subgroup (BASRI ≥ 3) (Group C). Data was analyzed statistically by SPSS software. Results. In total of 44 patients (88 hips), group A consisted of 21 hips, group B consisted of 42 hips and group C consisted of 25 hips. The test-retest intraclass correlation coecients (ICCs) in four raters were 0.955~0.977 and interrater ICC was 0.993. HIMRISS correlated moderately with the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) (r=0.540, p<0.001), the Bath ankylosing spondylitis functional index (BASFI) (r=0.581, p<0.001), ASDAS-ESR (r=0.604, p<0.001) and Ankylosing Spondylitis Disease Activity Score(ASDAS)- C reactive protein (CRP) (r=0.575, p<0.001). HIMRISS in group B and C was signicantly higher than that in group A: 29.38 (cid:0) 17.00 (cid:0) 40.94 (cid:0) vs 14.50 (cid:0) 11.38 (cid:0) 22.25 (cid:0) , p=0.009; 38 (31.13 (cid:0) 64.38 (cid:0) 11.38 22.25), p<0.001. previous studies applied on hip OA 24–25 (cid:0) and SpA (cid:0) 26 (cid:0) et (cid:0) 26 (cid:0) reliability of HIMRISS from 0.67 to 0.90 after two training sessions in cases with The reliability of detecting femoral BML, acetabular BML and synovitis effusion good after the two exercises ICC 0.73, and 0.88, BASDAI BASFI ASDAS-ESR p NSAIDs:Non-steroidal PROs:Patient-reported SF-12:Short Form-12; SpA:Spondyloarthritis; THA:Total Hip Arthroplasty; T1 TSE:T1-Weighted Spin Echo.


Introduction
Hip involvement is common in patients with ankylosing spondylitis(AS) 1-3 and is associated with severe functional impairment, work disability, a compromised psychological status and quality of life 4-6 . When hip involvement progresses to an advanced symptomatic stage, total hip arthroplasty (THA) which is acknowledged as the well-accepted treatment currently will be conducted, and a number of authors have reported their results using cemented or cementless techniques 7-12 . Unfortunately, hip prostheses with various exclusive designs have a limited life span, and revision surgeries are often needed periodically. Consequently, it is extremely crucial to estalish a comprehensive and reliable system for accurate and early diagnosis of hip involvement in AS patients.
In previous studies, the diagnosis of hip involvement has been focused on clinical ndings and radiographic changes. The symptoms of hip involvement symptoms are insidious onset and there is even no symptoms in the early course of the disease. The bath ankylosing spondylitis radiology index (BASRI-hip) which has been proved to be an objective and reliable grading system is the most widely used radiographic classi cation system to assess severity and progression of hip involvement in patients with AS 13,14-16 . Conventional radiographs can display postin ammatory structural injuries and are unable to re ect some key histopathological changes, including subchondral bone marrow edema(BME) and synovitis [17][18] . As a result, radiographs may underestimate the frequency of hip involvement leading to a diagnosis delay.
Magnetic resonance imaging (MRI) provides excellent visualization of bones and soft tissues and is the only imaging tool which is capable of visualizing bone marrow in ammation, a hallmark of AS 19-20 . In previous studies, however, the application of MRI imaging is mainly focused on sacroiliac joints and spine of patients with AS. There is few data of the changes of hip MR imaging 21-23 . The Hip In ammation MRI Scoring System (HIMRISS) was developed by using the data from a trial of intraarticular steroid therapy for osteoarthritis(OA), which is based on the concept that scoring of hip OA most usefully emphasizes the evidence of active in ammation 24-25 . The feasibility and the reliability of this imaging system have been fully validated in the setting of hip OA. Zheng et al introduced HIMRISS into the evaluations of spondyloarthritis (SpA), and the reliability and the clinical association were preliminarily validated as being e cient 26 . However, this study failed to give a precise de nition of clinical or radiographic hip involvement to these patient population. This study was aim to test the reliability and validity of HIMRISS for assessing hip involvement of AS patients at different stages of BASRI-hip scoring system.

Patient demographics and enrollment
Fifty-two AS outpatients who were admitted to the department of rheumatology and the department of adult joint reconstruction surgery in our institute from February 2018 to June 2019 were enrolled in this study. The inclusion criteria were as follows: (1) diagnoses of AS were made according to the 1984 modi ed New York criteria 27 ; (2) the age of the patients ranged from 18 to 45 years. Exclusion criteria that any selected patients with any of the followings: (1) systemic diseases of the muscular of nervous systems, (2) history of congenital or childhood disease, surgery, deep infection, trauma, and tumor of hip; (3) lower extremity replacement or amputation other hip joint; (4) MRI contraindications (e.g. pacemaker, metal implants, pregnancy, claustrophobia).
The subjects' demographics and clinical characteristics included gender, body mass index (BMI), age at outpatient visit, age at onset of AS, duration of AS, diagnosis delay, family history, medication status and extra-articular manifestations (EAMs) (current or past) including uveitis, psoriasis and in ammatory bowel disease (IBD). Disease activity was assessed respectively using the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) functional index (BASFI) 30 . The patient-reported outcomes (PROs) were assessed by using the Ankylosing Spondylitis Quality of Life (ASQoL) scales 31 and the short form-12(SF-12) 32 . The use of the medications including NSAIDs and DMARDs was recorded, and the patients who had taken treatment agents for one year or longer were considered as sustained users. These data for clinical characteristics were collected and evaluated independently by two rheumatologists (L.H.C. and M.S.L.) who had not participated in radiographic evaluations conducted from a face-to-face questionnaire and medical records. The Harris hip scoring (HHS) system 33 was directly evaluated by orthopedic surgeons (Z.L. and H.Y.) at the time of outpatient consultation. On the 100-point scale, a score of 90 points or more is de ned as an excellent outcome; 80 to 89 points, a good outcome; 70 to 79 points, a fair outcome; and 70 points or less, a poor outcome.
Laboratory data such as human leukocyte antigen (HLA-B27) status, serum erythrocyte sedimentation rate (ESR), high sensitive C reactive protein (CRP) and CRP level were also measured at enrollment.
The results of these clinical and laboratory parameters are provided in Table 1.  The value of continuous variables was presented as median and quartile (25-75%) and the categorical variables were based on presented as number plus percentage  The value of continuous variables was presented as median and quartile (25-75%) and the categorical variables were based on presented as number plus percentage

Radiographic Classi cation System
The anteroposterior (AP) radiographs of the pelvis and MRI imaging were obtained on the same day on an outpatient basis. We excluded 8 patients whose radiographs were of inadequate quality and kept a total of 44 patients for review and analysis.
The BASRI hip system was adopted to assess the severity of radiological involvement in the hip joint 13 , and it classi ed the status of the hip joints into a ve-point scale from 0 to 4 (0 = normal, no change; 1 = suspicious, possible focal joint space narrowing; 2 = minimal, circumferential joint space narrowing > 2 mm; 3 = moderate, circumferential joint space narrowing ≤ 2mm, or bone-on-bone apposition of ≤ 2 cm; 4 = severe, bone deformity or bone-on-bone apposition of < 2 cm or total hip re-placement).
MRI scans on enrolled patients' both hips were performed on a 3.0 T MR system (Siemens MAGNETOM Skyra, Siemens Company) using the following imaging parameters: a transverse T1-Weighted Spin Echo (T1 TSE) sequence, a transverse T2 TSE + Fat-Sat sequence, a coronal T1 TSE sequence, a coronal T2 TSE + Fat-Sat sequence, and trans-verse and coronal fat-saturated. All sequences used a 380mm eld of view, a 3.5mm slice thickness, a 0.7mm slice interval, a 384 448 matrix, and several excitations (AVERAGE) of 2.
HIMRISS has three features: bone marrow lesions (BMLs), synovitis and effusion. BML is de ned as an area of hyper-intensity within the bone in STIR sequence. The BML scoring on either femoral head or acetabulum side was graded according to the method which was described by Maksymowych et al 34 . The range of BML scoring is 0-100. Effusion and synovitis are scored together depending on the maximum depth of the uid (0 = 0-1.9 mm, 1 = 2-3.9 mm, 2 = ≥ 4 mm) on the same central, anterior and posterior slices. The range of total effusion scoring is 0-30. Therefore, the nal HIMRISS scoring range is 0-130.

Reading exercises
Considering the possible effect of the raters' experiences on evaluating the hips, we chose four raters with different levels of training and experiences.
A senior rheumatologist (L.H.C.) who is an attending physician with 16 years of clinical experience, an adult hip surgeon (Z.L.) who is a vice director with 15 years of clinical experience, and two musculoskeletal radiologists (Z.J., X.P.) who are residents with 4 years and 9 years of clinical experience respectively participated in grading the MRI imaging.
All raters were not informed about patient demographics and clinical parameters. Firstly, they learned the scoring rule through a PowerPoint le of Outcome Measures in Rheumatology (OMERACT) 11 24 individually. Then a training session was held where they agreed on a criteria for radiographic evaluation that was based on ve MRI images.
Next, another 88 MRI images from the PACS (Picture Archiving and Communication System) workstation which were not included in the evaluation were collected by the rst author (M.S.L.) and were sent to the raters. Then they graded the MRI images in random sequences in different workrooms (exercise 1). Three months later, they repeated their works (exercise 2) without knowing the previous results to assess the reliability of test-retest.

Statistical analysis
Data were statistically analyzed using SPSS software for Windows (version 23.0; IBM, Armonk, NY, USA). Descriptive analyses for categorical variables were shown as percentages and frequencies and for continuous variables they were based on mean and standard deviation (SD) or median and quartile (25-75 %) if the data were skewed. The values of femoral BML, acetabular BML, synovitis effusion were summed up as the HIMRISS values for the single hip, and the mean values by all readers obtained in exercise 2 were taken as the nal HIMRISS scores. The inter-rater and intra-rater reliability of HIMRISS were calculated using intraclass correlation coe cient (ICC). The correlations of HIMRISS with clinical continuous variables and with ordinal variables were determined by correlation coe cient (r) in Pearson correlation analysis and Spearman rank correlation analysis respectively. We classi ed hips into no hip involvement group(HHS ≥ 80 and BASRI ≤ 1)(Group A), mild hip involvement subgroup (BASRI = 2 or BASRI ≤ 1 and HHS ≤ 79) (Group B) and moderate to advanced hip involvement subgroup (BASRI ≥ 3) (Group C) based on HHS of involved hip and BASRIhip score. Demographic features, clinical characteristics and radiographic parameters were compared using ANOVA (including post hoc analysis) and nonparametric Kruskal-Wallis test among these subgroups. Bonferroni method was taken to adjust the signi cance level in multiple comparisons. All reported P values were two-tailed with an alpha of 0.05.

Ethics and registration
All procedures involving human participants carried out in the studies were in accordance with the ethical standards of the institutional and/or national research committee and the 1964 Helsinki declaration and its later amendments or comparable ethical standards. This study was approved by the Beijing Jishuitan Hospital Institutional Review Board (project number S-305/2007), and the informed consent was obtained from each participant before the enrollment of this study.

Patient demographics and clinical parameters
The patient demographics and clinical parameters were displayed in

Reliability of HIMRISS scores
The test-retest and interrater ICCs were provided in Table 2. The interrater ICCs were calculated by four raters in the second scoring exercise.

Correlation analysis and intergroup comparisons
The results of correlations between HIMRISS and the clinical and laboratory parameters were shown in Table 3.   Group A represents the hips with no hip involvement (HHS ≥ 80 and BASRI ≤ 1), group B represents the hips with mild hip involvement (BASRI = 2 or BASRI ≤ 1 and HHS ≤ 79) and Group C represents the hips with advanced hip involvement (BASRI ≥ 3) The value of continuous variables was presented as mean ± standard deviation and the categorical variables were based on presented as number plus percentage The results of intergroup comparisons were provided in Table 1 HIMRISS as a quantitative imaging method possess outstanding advantages in the assessment of early hip involvement in AS compared with the traditional MRI descriptive diagnosis. As far as we know, this was the rst study which introduced the HIMRISS into this eld.
Our study showed excellent reliability among four raters in AS patients. The study conducted by Zheng et al 26 showed that it was more notable to see the correlations between disease activity and HIMRISS from bilateral rather than unilateral hip joint scores. The sum HIMRISS correlated moderately with ASDAS-CRP (r = 0.67, p < 0.001) and CRP (r = 0.55, p < 0.01) and weakly with ASDAS-ESR (r = 0.48, p < 0.001). Unfortunately, the study did not discriminate the degree of radiographic hip involvement and as a result failed to con rm the clinical advantage of HIMRISS in detecting early hip in ammation lesions in SpA patients. In our series, there was a signi cant correlation of HIMRISS with BASDAI (r = 0.540, p < 0.001), BASFI (r = 0.581, p < 0.001), ASDAS-ESR (r = 0.604, p < 0.001), ASDAS-CRP (r = 0.575, p < 0.001), ASQOL (r = 0.478, p < 0.001) and HHS (r=-0.468, p < 0.001). These results were also consistent with the research done by Zheng et al 26 .
Moreover, we classi ed hips into no, mild and moderate to advanced hip involvement subgroups according to HHS of involved hip and BASRI-hip score. The mild hip involvement subgroup had a signi cantly higher HIMRISS than no hip involvement subgroup (29.38 vs 14.50, p = 0.009). This result can be considered as further proof that HIMRISS is a reliable imaging tool to detect early hip structural damages in AS patients.
The main limitation to our study came from the nature of its single-center cross-sectional research. Relatively few participants took part in the study, which may have a negative impact on the evaluation of the rate and the degree of hip involvement in AS patients.
In conclusion, HIMRISS applied in patients with AS demonstrated a satisfying reliability, and it was in signi cant clinical association with a series of AS speci ed clinical and laboratory parameters. And there was a particular concern that the hips with mild hip involvement had a signi cantly higher HIMRISS than the ones without hip involvement. HIMRISS is a reliable quantitive assessment tool for evaluating early hip involvement in patients with AS.

Conclusions
The application of HIMRISS in patients with AS demonstrated a satisfactory reliability, which means it can be a reliable quantitive assessment tool for evaluating early hip involvement in patients with AS.

Declarations
Ethics approval and consent to participate All procedures involving human participants carried out in the studies were in accordance with the ethical standards of the institutional and/or national research committee and the 1964 Helsinki declaration and its later amendments or comparable ethical standards. This study was approved by the Beijing Jishuitan Hospital Institutional Review Board (project number S-305/2007), and the informed consent was obtained from each participant before the enrollment of this study.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
Competing interests