LAUSR

Background: Currently MRI is the preferred imaging method to detect bone marrow edema (BME), the hallmark of sacroiliitis. MRI plays an important role in the early diagnosis of axial Spondyloarthritis (axSpA). Biological disease-modifying anti-rheumatic treatment has revolutionized the therapeutic armamentarium of axSpA. With drugs targeting TNFα, 50% of axSpA patients achieve a clinically important response. Therefore, we hypothesized that if we would be able to demonstrate in vivo expression of TNFα in sacroiliac joints by scintigraphy with 99mTc-radiolabeled certolizumab pegol (99mTc-CZP), this might lead to more ‘evidence-based biological therapy’. Objectives: To investigate the agreement between BME on MRI-SIJ and tracer uptake on immunoscintigraphy with 99mTc- CZP in the same location in patients with axSpA. Methods: CZP was conjugated with S-HYNIC and subsequently radiolabeled with approximately 740 MBq Tc99m and injected intravenously. Static images with single photon emission tomography (SPECT)/computed tomography (CT) of SIJ were acquired 4-6h post injection. Uptake of the tracer was scored semi-quantitatively, per quadrant of the SIJ: 0=no uptake, 1=faint uptake or 2=clear uptake. BME on MRI was scored per quadrant over 6 slices as absent or present, providing a maximum score of 6 per quadrant. Agreement between MRI-SIJ and immunoscintigraphy was calculated (kappa; percentage agreement) for all quadrants separately. To calculate the agreement a cut-off of ≥1 was used for MRI scores as well as immunoscintigraphy scores. In addition, depth and intensity of BME lesions on MRI-SIJ (as defined in the Spondyloarthritis Research Consortium of Canada (SPARCC) method) were assessed per slice per SIJ. Results: 7 axSpA patients (mean age 36±5.7 years) had both MRI-SIJ and immunoscintigraphy available. The mean score for BME lesions seen on MRI was 12.9±13.2 and 4.86±5.4 for tracer uptake observed on the immunoscintigraphy. In 2 out of the 7 patients there was no BME on MRI and in the same 2 patients there was no tracer uptake seen on scintigraphy. In table 1 kappa coefficients and percentage agreement for every quadrant are shown. The mean and median of agreement for all quadrants was k=0.80 and k=0.86, respectively. Clear tracer uptake (score 2) was correlated to deep BME lesions on MRI-SIJ (extending over the depth of at least 1cm from the articular surface); the observed Spearman’s rho correlations were 0.986 (p<0.00) and 0.956 (p<0.00) for left and right SIJs, respectively. Regarding intensity, no significant correlation with clear tracer uptake was found. Interestingly, in one additional patient with complete ankylosis of the SIJs on radiographs no tracer uptake could be detected, suggesting that in vivo detection of TNF does not correlate with bone formation.Table 1 Agreement on presence of BME on MRI-SIJ in a quadrant and tracer uptake on the immunoscintigraphy in that same quadrant. Location of the SIJ quadrant Kappa (CI) Percentage agreement Right upper iliac quadrant (RQ1) 0.720 (0.21-1) 85.7% Right upper sacral quadrant (RQ2) 0.696 (0.14-1) 85.7% Right lower sacral quadrant (RQ3) 0.417 (0-1) 71.4% Right lower iliac quadrant (RQ4) 1 (1-1) 100% Left upper iliac quadrant (LQ1) 1 (1-1) 100% Left upper sacral quadrant (LQ2) 0.588 (0-1) 85.7% Left lower sacral quadrant (LQ3) 1 (1-1) 100% Left lower iliac quadrant (LQ4) 1 (1-1) 100% Conclusion: Inflammation on MRI-SIJ can be detected with immunoscintigraphy with 99mTc-CZP. The immunoscintigraphy showed good correlation with BME lesions on MRI in patients with axSpA. Deep lesions, considered specific for axSpA, showed an almost perfect correlation. Disclosure of Interests: Philippe Carron: None declared, Manouk de Hooge: None declared, Thomas Renson: None declared, Bieke Lambert: None declared, Kathia De Man: None declared, Dirk Elewaut: None declared, Filip van den Bosch Consultant for: AbbVie, BMS, Galapagos, Janssen, Lilly, Merck, Novartis, Pfizer and UCB, Speakers bureau: AbbVie, BMS, Janssen, Lilly, Merck, Novartis, Pfizer and UCB.