Search in the Abstract Database

Abstracts Search 2019

P198 Multi-alleles predict primary non-response to infliximab therapies in Crohn’s disease: a simple and practicable model

J. Tang*1, C. Zhang2, X. Wang2, X. Gao1

1The Sixth Affiliated Hospital of Sun Yat-sen University, Department of Colorectal Surgery, Guangzhou, China, 2Institute of Clinical Pharmacology, School of Pharmaceutical Sciences, Sun Yat-sen University, Guangzhou, China

Background

Infliximab (IFX), a rather expensive medicine, is first-line treatment of Crohn’s disease (CD) patients. The threptic efficacy of IFX has noticeable individual differences. Single-gene polymorphism is inadequate to predict primary non-response (PNR). In this study, we aimed at identifying genetic factors associated with PNR and predict patient primary response to IFX by develop multi-genetic prediction model.

Methods

A retrospective study was performed and patients with IFX therapy were recruited. Primary response was evaluated at 14th week according to simple endoscopic score for CD. Ninety tag single-nucleotide polymorphism (SNPs) within 27 genes were genotyped by MassARRAY Analyser system. Multi-variate prediction model was established to predict PNR. Area under the receiver-operating characteristic curve (AUROC) was applied to evaluate the performance of multi-variate model.

Results

Of 206 patients, 42 (20.4%) experienced PNR. Nine SNPs were associated with PNR (p < 0.05). The genetic prediction model included 5 SNPs, AUROC on representative training dataset and testing dataset is 0.794 (p = 6.00 × 10−6) and 0.812 (p = 7.90 × 10−5), respectively.

The ROC curve of SNPs on training and testing dataset.

The combined genetic-clinical prediction model, comprised 5 SNPs and one clinical indicator, is superior to genetic model, AUROC on representative training dataset and testing dataset is 0. 818 (p = 9.36 × 10−7) and 0.888 (p = 9.52 × 10−7), respectively.

Abstract P198

Combined predictive effect on endoscopic response in Nomography.

The sensitivity and specify is 86.9% and 72.0%, respectively. On 100 training datasets and 100 testing datasets that obtained from 100 splitting process, the mean AUROC difference between them is only 0.02.

Conclusion

Genetic polymorphisms can predict PNR to IFX therapy in CD, the genetic-clinical prediction model is stable and not overfitted.