AB0199 GENES PREDICTIVE ON THE EFFICACY OF INFLIXIMAB IN THE TREATMENT OF RHEUMATOID ARTHRITIS: A PROSPECTIVE, MULTI-CENTRE, CLINICAL PERFORMANCE EVALUATION STUDY FOR AN IN-VITRO DIAGNOSTICS MEDICAL DEVICE

Abstract

Approximately 30% of rheumatoid arthritis (RA) patients fail to respond to first biological therapy, thus treatment selection of biologic therapy for patients with RA is of high importance. The lack of biomarkers to predict specific biological treatment response, in the case of non-responder (NR) patients leads to unnecessary exposure, delay of adequate therapy, progression of the disease and therapy cost increase. Predicting the patient’s responsiveness to the first biological therapy is still an unmet need in the clinical setting. Predictive in vitro testing would have a significant effect on the administration of biological therapy, on the real life implementation of cost effective personalized therapy.The purpose of this in vitro diagnostic medical device study was to demonstrate that particular gene expression profiles as genomic biomarkers (i.e. the IVD medical device) predict therapeutic response to infliximab, discriminate between responders and non-responders to infliximab treatment. Responders were defined if they reached DAS target value DAS28≤3.2 at 6 month (M6).110 bionaive patients were enrolled with moderate-high activity RA (DAS28-CRP >3.2), who have responded inadequately to DMARDs (including methotrexate), after they have been assigned to infliximab treatment. All patients received commercially available infliximab, procured according to SmPC, local guidelines and regulations in this non-interventional clinical study. The clinical response was evaluated according to the change from baseline in disease activity at M6. Clinical characteristics (RA duration, smoke, steroid treatment, etc.) and serological parameters (RF, ACPA, aCVM) were collected. A 3rdvisit scheduled around week 22 (M6) and change of DAS28-CRP value from the baseline has been evaluated. Gene expression profiling was performed from blood samples taken at month 0 (M0); - just before the first infliximab infusion. Global gene expression profiling was performed to identify differentially expressing genes using RNA sequencing. The set of differentially expressing genes were further reduced with a combination of machine learning modelling and various feature elimination methods. The expression of the reduced gene set was confirmed and further analysed using reverse-transcription and quantitative real-time PCR.A total of 250 genes were identified by a combination of differential gene expression analyses, feature elimination techniques and various machine learning modelling methods of which 44 genes showed significant differences between NR and good responder groups. Preliminary interim analysis identified associations between gene expression and clinical response/ non-response to infliximab therapy.Table.Three models containing gene expression + clinical data sets illustrates some statistical characteristicsModell building_IDAccuracySensitivitySpecificityModell VerificationAccuracySensitivitySpecificity00232100.00100.00100.00002328888.8987.5000249 98.82 96.55100.00002498477.7887.5000270 98.82 96.55100.00002708877.7893.75Our preliminary analysis shows that this set of genes and selected clinical parameters are predictive markers for infliximab specific response in RA patients. Ongoing work involves the clinical validation of these results in an independent patient cohort (n=60). This approach provides the opportunity to develop an in vitro diagnostic test method for the prediction of infliximab treatment responsiveness in bionaive rheumatod arthritis patients, hence to personalize infliximab therapy for these patients.Emese Kiss Consultant of: EK has received consultancy fees from Egis., Gyula Poór Consultant of: GyP has received consultancy fees from Egis and he was the coordinating investigator in this study, Gábor Zahuczky Grant/research support from: Egis, Katalin Tauberné Jakab Employee of: Egis., Miklós Sebeszta Employee of: Egis., Tamás Ponyi Employee of: Egis., Zsolt Holló Employee of: Egis.