MRI Detection of Active Sacroiliitis in First Degree Relatives of Ankylosing Spondylitis Patients with Clinical and Laboratory Correlations

Background. Detection of ankylosing spondylitis (AS) in the preclinical stage could help prevent long term morbidity in this patients’ population. The aim of this study was to examine the prevalence of active sacroiliitis in rst-degree relatives of AS patients using MRI with clinical and laboratory correlations as these patients may benet from MRI screening and early treatment. Methods. Seventeen rst-degree relatives of AS patients were recruited prospectively. AS screening questionnaires (Ankylosing Spondylitis Disease Activity Score, Bath Ankylosing Spondylitis Disease Activity Index & Visual Analogue Scale), blood tests (C-Reactive Protein, HLA-B27), and an MRI of the SIJs were taken. Two musculoskeletal radiologists interpreted the MRI scans, and two physiotherapists applied four symptom provocation tests (Gaenslen's test, posterior pelvic pain provocation test, Patrick's Faber (PF) test and palpation of the long dorsal SIJ ligament test), and two functional movement tests (active straight leg raise and Stork test). Cohen's A 2017 longitudinal study by Costantino et al. detected radiographic evidence of sacroiliitis in 68.5% of surveyed rst-degree relatives of SpA patients over a period of 15 years. 8 The prevalence of MRI conrmed active sacroiliitis in rst-degree relatives was previously estimated at 25.5%; observations were made from a single radiologist’s interpretation of a single coronal oblique STIR sequence, a signicant limitation. 9 This current study examines the prevalence of MRI conrmed active sacroiliitis in rst degree relative of AS patients using 4 MRI sequences (axial T2FS and T1 and coronal oblique T1 and STIR), which are independently interpreted by two musculoskeletal radiologists and correlated with clinical and laboratory tests. MRI this


Introduction
Ankylosing spondylitis (AS) is the prototypic form of spondyloarthritis (SpA) with a predominantly axial presentation and bilateral symmetric sacroiliitis. 1 Earlier identi cation of spondyloarthritis using diverse imaging, clinical and proteomic measures offer potential to initiate early therapy, reduce future disability and diagnostic costs. 2 In particular, utilization of anti-tumor necrosis factor (TNF) alpha biologic therapies has been shown to be effective in treating AS symptoms and associated with limitation in radiographic progression of the disease. 3 The Assessment of SpondyloArthritis International Society (ASAS) recently developed widely endorsed criteria for identifying and classifying axial spondyloarthritis (axSpA) that includes magnetic resonance imaging (MRI) to detect active in ammation in the sacroiliac joints. 4 Validated self-report symptom tools [Bath Ankylosing Spondylitis Disease Activity Index (BASDAI), Ankylosing Spondylitis Disease Activity Score (ASDAS), and spinal pain Visual Analogue Scale (VAS)] have been utilized in assessing disease activity in AS patients. 5 It has been well documented that biomarkers such as Human Leukocyte Antigen B27 (HLA-B27) status and C-reactive protein (CRP) have diagnostic, prognostic, and predictive value. 6,7 A 2017 longitudinal study by Costantino et al. detected radiographic evidence of sacroiliitis in 68.5% of surveyed rst-degree relatives of SpA patients over a period of 15 years. 8 The prevalence of MRI con rmed active sacroiliitis in rst-degree relatives was previously estimated at 25.5%; observations were made from a single radiologist's interpretation of a single coronal oblique STIR sequence, a signi cant limitation. 9 This current study examines the prevalence of MRI con rmed active sacroiliitis in rst degree relative of AS patients using 4 MRI sequences (axial T2FS and T1 and coronal oblique T1 and STIR), which are independently interpreted by two musculoskeletal radiologists and correlated with clinical and laboratory tests.

Ethical considerations and recruitment
Ethical approval was provided by the University of Saskatchewan Biomedical Research Ethics Board  with operational approval granted by Saskatchewan Health Authority, Canada. Potential participants were invited to take part in this study using a variety of recruitment strategies, including recruitment posters and yers advertised in 3 local hospitals, private practice rheumatology clinics, paid advertisement through local newspapers, in-person invitations during out-patient rheumatology clinic visits within the university-a liated hospital, electronic mails to patients diagnosed with AS and several other media outlets. Informed consent was obtained from all individual participants included in the study.

Inclusion and Exclusion Criteria
Inclusion criteria were rst-degree relatives of clinically diagnosed AS patients, aged between 18 and 45 years, with no history of any previous diagnosis of SpA, spinal surgery, recent infection, or recent trauma and no contraindication for using MRI. All test participants were required to be capable of giving informed consent before inclusion in the study. Potential participants who advised they were receiving anti-TNF α biologic medications were excluded from the study.

MRI Imaging
All participants underwent a dedicated unenhanced spondyloarthropathy MRI scan protocol of the sacroiliac joints, which included coronal oblique T1 and STIR and axial T1 and T2FS sequences ( Table 1). All scans were performed on a Siemens 3 Tesla scanner (Siemens, Skyra Healthcare, Erlangen, Germany) using a spine coil. All images were independently read by two fellowship-trained musculoskeletal radiologists (with 18 years of experience each) on Philips PACS workstations and Barco monitors. Each scan was scored according to the ASAS criteria. 10 The sacroiliac joints were assessed for acute enthesitis, periarticular osteitis (bone edema), articular erosions, and joint effusions (Fig. 1). The two radiologists were blinded to each other's results, and an inter-observer agreement analysis was performed for reliability of results.

Measurement Tools
Three disease activity measurement tools were employed: 1. Ankylosing Spondylitis Disease Activity Score (ASDAS) The ASDAS is a composite index used to assess disease activity in AS patients. The score can be calculated with the erythrocyte sedimentation rate (ESR) or CRP values. 5 Scores are categorized in four groups as: 'inactive disease activity' when < 1.3, 'moderate disease activity' from 1.3 to < 2.1, 'high disease activity' from 2.1 to 3.5, and 'very high disease activity' when > 3.5. 5 For this study, we used the ASDAS-CRP score as recommended by ASAS. 11 2. Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) The Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) score consists of 10 cm visual analogue scales used to answer six questions pertaining to the ve major symptoms of AS [5]. Scores are categorized as: 'mild disease activity' for values 1-3, 'moderate disease activity' for values 4-7, 'severe disease activity' for values 8-10. 12

Visual Analogue Scale (VAS) for spinal pain
The VAS is a continuous scale comprised of a vertical line measuring 100 mm in length and anchored by two verbal descriptors (best and worst) describing extremes of pain. 13,14 Research timeframe per participant For each participant, data collection was carried out within a seven-day period, in two stages.

Stage 1 Blood work and MRI scan
On the day of recruitment, participants rst met with a research assistant to gather demographic data and complete the BASDAI, ASDAS, and VAS disease activity measurement tools. A phlebotomist then drew blood samples for CRP, ESR, and HLA-B27 serum analysis. Participants then underwent an MRI scan of the sacroiliac joints with a 3.0 Tesla MRI using a dedicated non-contrast spondyloarthropathy protocol.

Stage 2
Physical tests Two to seven days following blood draw and MRI, test participants were physically evaluated by two experienced musculoskeletal physiotherapists for independent and blinded physical evaluation. Prior to data collection, the two physiotherapists met with the research team to familiarise with the physical test protocol and questions to ask participants regarding LBP history within the past three months.
The six recommended clinical tests described in the European Guidelines for the physical diagnosis and treatment of pelvic girdle pain 15

Study Participants
Nineteen rst-degree relatives of known AS-diagnosed individuals, aged 18 to 45 years, with no history of clinically diagnosed spondyloarthritis, met eligibility criteria and were recruited into the study. All participants were males. Two participants dropped out due to scheduling con icts and travel limitations, leaving 17 participants.

Descriptive analysis
A descriptive overview of the demographic and clinical variables of the 17 male participants is presented in Table 2. Mean values and standard deviation are summarized for age, ASDAS-CRP, BASDAI, and VAS scores. Frequencies and percentages are also described for participants who tested positive for HLA-B27, CRP, and MRI; as well as participants reporting 'yes' for having had at least one episode of LBP in the last three months. Eight of 17 participants (47%) tested positive for HLA-B27 antigen; 2 (12%) tested positive for elevated CRP; while seven (41%) were positive for an MRI diagnosis of active sacroiliitis (Fig. 1). A descriptive overview of clinical participants (with LBP or HLA-B27), with MRI positive versus MRI negative results, is presented in Table 3. Two of the 7 (29%) participants who had an MRI con rmed active sacroiliitis had no history of LBP in the past three months. HLA-B27 was negative in 2 of the 7 (29%) participants who demonstrated active sacroiliitis on MRI. One participant who had no history of LBP and negative HLA-B27 demonstrated active sacroiliitis on MRI.  Table 3 Descriptive

Reliability assessments
The Cohen's Kappa score for the interobserver agreement between the two radiologists was 1.00 (p-value < 0.0001). While interobserver physiotherapist agreement for the six physical tests ranged from fair to perfect agreement (Kappa 0.26 to 1.00) for symptom provocation or movement analysis, no physical test, either singularly or in combination with others, demonstrated any signi cant predictive association with a positive MRI.

Discussion
First degree relatives of AS are considered high-risk population for developing SpA 21 ; however, these individuals may remain in the preclinical phase of the disease for a number of years before presenting with clinical and imaging features of SpA. 8, 9,22 Imaging plays a key role in the detection and monitoring of SpA, and MRI is considered the gold standard for identifying active sacroiliitis. 4,23 The results of this study showed that eight of 17 participants ( The 41% prevalence of MRI con rmed sacroiliitis in this study is higher than previously reported prevalence rate of 25.5% by Turina et al. 9 This maybe related to the fact that, in their study, the authors imaged the sacroiliac joints with a single coronal oblique STIR sequence, while in this current study, the MRI protocol included four sequences (axial T2FS and T1 and coronal oblique T1 and STIR), adding more anatomical and pathological details to the imaging data. In addition, the interpretation of the MRI scans in this current study was carried out by two fellowship-trained musculoskeletal radiologists compared to a single radiologist's read in the previous study. 9 There was perfect interobserver agreement between the two radiologists in detecting active sacroiliitis on MRI in this study. The physiotherapist inter-rater reliability results demonstrate fair to substantial agreement for both positive and negative responses to all physical test procedures 20 ; however, it is possible that these clinical tests were not su ciently sensitive for discriminating low back symptom reproduction in a small non-clinical sample where only 9 out of 17 participants described the presence of at least one episode of LBP in the past three months. This argument is supported by the BASDAI, ASDAS, and VAS scores observed in Table 1, which do not approach threshold values that would be clinically indicative of any active disease process. 11,12,26 It is therefore unclear whether these therapist evaluations would relate any more strongly to MRI ndings in a clinically recruited sample of patients with in ammatory LBP related to AS. It was noted that participants with negative MRI for active disease, were negative responses for pain provocation tests (Gaenslen's, PPPT, PFT, and ASLR) as well as detection of pelvic movement asymmetry on the Stork test.
Our study has several limitations. Despite a variety of advertising and recruitment strategies over a 2-year period, we were unable to identify more than 19 participants. A number of potential participants were unable to attend due to travel restrictions and personal factors, and a number of contacted patients with AS did not have rst-degree relatives who met the inclusion and exclusion criteria. Thus, at the completion of the study, we were only able to recruit 17 participants. A limitation of our sampling approach was all rst-degree relatives who entered the study were male subjects. While all 17 participants were asked for a recollection of occurrences of LBP in the past three months, none were undergoing any process of clinical investigation or clinical management. Thus, screening procedures were applied to participants who had not yet chosen to enter a clinical pathway. Since the physical tests applied in this study are clearly described as recommended clinical tests for the physical diagnosis and treatment of pelvic girdle pain, it is possible the procedures may not have been su ciently sensitive for use as a screening procedure in a non-clinical population. A further limitation is that recruitment in this study relied on volunteer participation by rst-degree relatives. This procedure to select study participants raises the possibility of selection bias, as individuals with greater concerns regarding a family history of back pain and spondyloarthritis may have been more likely to inquire and agree to participate in the study but may not be fully representative of the population with latent axial SpA.

Conclusion
MRI remains the gold standard for the diagnosis of active sacroiliitis. The study has identi ed active sacroiliitis in 41% (n = 7/17) of the rst-degree relatives of patients with ankylosing spondylitis. MRI detected active sacroiliitis in 2 participants who had no history of LBP, 2 participants with negative HLA-B27, and one participant with neither a history of back pain nor positive HLA-B27. The ndings of this study provide some support for MRI screening of this high-risk population for active sacroiliitis in order to help guide informed patient management, minimize long-term disease burden, reduce disability, and improve patients' quality of life; however, further studies are needed to explore the cost effectiveness and long-term bene ts of this imaging strategy moving forward. Catherine Trask has contributed to the (1) conception and design, or acquisition of data, or analysis and interpretation of data, (2) drafting the article or revising it critically for important intellectual content, (3) nal approval of the version to be published, and (4) agree to be accountable for all aspects of the work if questions arise related to its accuracy or integrity.
David Leswick has contributed to the (1) conception and design, or acquisition of data, or analysis and interpretation of data, (2) drafting the article or revising it critically for important intellectual content, (3) nal approval of the version to be published, and (4) agree to be accountable for all aspects of the work if questions arise related to its accuracy or integrity.
Melanie D. Bussey has contributed to the (1) conception and design, or acquisition of data, or analysis and interpretation of data, (2) drafting the article or revising it critically for important intellectual content, (3) nal approval of the version to be published, and (4) agree to be accountable for all aspects of the work if questions arise related to its accuracy or integrity.
George Katselis has contributed to the (1) conception and design, or acquisition of data, or analysis and interpretation of data, (2) drafting the article or revising it critically for important intellectual content, (3) nal approval of the version to be published, and (4) agree to be accountable for all aspects of the work if questions arise related to its accuracy or integrity.
Rhonda Loeppky has contributed to the (1) conception and design, or acquisition of data, or analysis and interpretation of data, (2) drafting the article or revising it critically for important intellectual content, (3) nal approval of the version to be published, and (4) agree to be accountable for all aspects of the work if questions arise related to its accuracy or integrity.
Stacey Lovo has contributed to the (1) conception and design, or acquisition of data, or analysis and interpretation of data, (2) drafting the article or revising it critically for important intellectual content, (3) nal approval of the version to be published, and (4) agree to be accountable for all aspects of the work if questions arise related to its accuracy or integrity.
Regina Taylor-Gjevre has contributed to the (1) conception and design, or acquisition of data, or analysis and interpretation of data, (2) drafting the article or revising it critically for important intellectual content, (3) nal approval of the version to be published, and (4) agree to be accountable for all aspects of the work if questions arise related to its accuracy or integrity.
Bindu Nair has contributed to the (1) conception and design, or acquisition of data, or analysis and interpretation of data, (2) drafting the article or revising it critically for important intellectual content, (3) nal approval of the version to be published, and (4) agree to be accountable for all aspects of the work if questions arise related to its accuracy or integrity.
. Data availability: The data related to this manuscript is available from the corresponding author upon request. Figure 1 A 29-year-old male rst-degree relative of ankylosing spondylitis patient. Axial STIR (a) and coronal oblique STIR (b) MRI sequences demonstrating bilateral periarticular osseous oedema (arrows) in keeping with active bilateral sacroiliitis.