Background Rituximab (RTX) is an effective treatment of rheumatoid arthritis (RA) after failure or intolerance to TNF blockers. Nevertheless, still 30–50% of RA patients does not respond. Therefore, identifying patients with favourable response to RTX is important in order to further optimize treatment for the individual patient. In vivo imaging of biodistribution of radiolabelled RTX in RA by Positron Emission Tomography (PET) might provide a tool to fulfil this unmet clinical need.
Objectives To investigate the association of in vivo biodistribution of Zirconium-89 (89Zr-)RTX in RA patients starting with RTX treatment and clinical response at 24 weeks post injection (p.i.).
Methods We included 20 anti-B cell therapy naïve RA patients (female18/20; age 53±11; DAS28 5.4±1.2) who started RTX treatment according to standard clinical protocol (without methylprednisolone). The first RTX infusion (1000 mg) was directly followed by infusion of 10 mg-18 MBq 89Zr-RTX. Whole body and PET-CT images of wrists/hands (n=22 joints/patient) were performed three days p.i. After 24 weeks of therapy, response was defined according to the EULAR response criteria (DAS28 at endpoint ≤5.1 and/or deltaDAS28>0.6) (1). 89Zr-RTX uptake was assessed visually (blinded for clinical data) and quantitatively by drawing volumes of interest (VOI) on PET positive tissues/joints. Standardized uptake values (SUVs) were calculated and background ROIs were drawn on metacarpal bone to calculate target-to-background (T/B) ratios.
Results At 24 weeks 13/20 patients were moderate to good responders to RTX. Whole body PET images showed 89Zr-RTX distribution in liver, spleen, kidneys, blood pool, lymph nodes (LN) (9/20pts) and in shoulders/elbows (9/20pts), without statistically significant SUV differences between responders and non-responders. Detailed hand images showed PET positive joints in 18/20 patients (ranging 1–20 joints/patient). Clinical responders had significantly higher 89Zr-RTX uptake in visually positive joints than non-responders (SUV: 3.2±1.4 vs 2.1±1.1 [p=0.04]; T/B ratios 6.2±2.5 vs. 3.4±1.7 [p=0.015], respectively). Using a T/B cut-off value of 4.0, positive predictive value and negative predictive value for response were respectively 90% and 75%, at a sensitivity of 82% and specificity of 86% (Fig 1). In contrast, no association was found between any clinical disease or serological (including a-CCP/RF) parameters and clinical response at 24 weeks.
Conclusions 89Zr-RTX PET-CT clearly visualizes arthritic joints in RA patients. Significant differences in baseline quantitative 89Zr-RTX uptake in PET positive hand joints were found between responders and non-responders to RTX treatment at 24 weeks. Our data point at a promising clinical value of 89Zr-RTX PET to predict at baseline the therapeutic response to RTX and therefore has potential for stratification of RTX treatment in RA. Further validation is required in larger cohorts.
Van Gestel et al; Arthritis Rheum 1998
Acknowledgement This study was financially supported by Hoffmann-La-Roche,The Netherlands
Disclosure of Interest S. Bruijnen: None declared, M. Tsang-A-Sjoe: None declared, H. Raterman: None declared, T. Ramwadhdoebe: None declared, D. Vugts: None declared, G. Van Dongen: None declared, M. Huisman: None declared, O. Hoekstra: None declared, P. Tak Grant/research support from: P.P. Tak participated in this study from his position at the Academic Medical Center, Amsterdam, The Netherlands. GSK did neither sponsor this study nor supported this study otherwise. P.P.Tak also has affiliations at Cambridge and Ghent but his activities at these locations were not involved in the current study., A. Voskuyl: None declared, C. Van der Laken: None declared