Project Summary/Abstract
Therapeutic goals in pediatric Crohn’s Disease (CD) have shifted from clinical improvement or remission to
endoscopic healing (EH) by ileocolonoscopy and transmural healing (TH) by magnetic resonance enterography
(MRE). This shift happened because patients who achieve complete healing (CH, TH and EH) experience lower
rates of subsequent hospitalization, therapy escalation, or surgery than those with EH alone or no healing. We
hypothesize that specific pre-treatment clinical, radiologic, genomic, and microbial factors along with attainment
of targeted anti-TNF biologic exposure will be associated with the primary endpoint, CH, and the major secondary
endpoints, EH and TH, 52 weeks after anti-TNF start.. We will test this hypothesis in a prospective cohort study
of 570 newly diagnosed pediatric-onset CD subjects who initiate treatment with anti-TNF medication within 6
months of diagnosis. We will administer personalized anti-TNF biologic therapy guided by therapeutic drug
monitoring (TDM) using a novel dosing algorithm which we developed.
Aim 1. Evaluate putative associations and explore novel associations between CH and baseline
measures of clinical and radiologic severity. We hypothesize that pre-treatment nutritional status,
antimicrobial serologies, and MRE findings will be associated with CH 52 weeks after anti-TNF start. Formal
hypothesis tests will be carried out to confirm the predictive power of a set of pre-specified measures using a
logistic regression model. We will also conduct exploratory analyses of novel predictors, identified via machine
learning methods, to assess their relationship with CH after adjusting for the confirmed primary predictors.
Aim 2. Evaluate putative associations and explore novel associations between CH and baseline host and
microbial genomic and metabolic factors. We hypothesize that pre-treatment gene expression signatures
and microbial factors will be associated with year 1 CH. We will characterize the host genotype, longitudinal
microbial taxonomic and metabolomic profiles, and ileal and colon host epigenome and transcriptome at baseline
and at 52 weeks after anti-TNF start.
Aim 3. Use a k-fold cross-validation procedure to determine the optimal predictive model of year 1 CH.
We hypothesize that a model which includes host gene signatures and microbes will improve prediction of CH
beyond one based on clinical and imaging factors alone. The model will include significant clinical and imaging
predictors from Aim 1 and the subset of baseline host and microbial characteristics found to be potentially
explanatory in Aim 2.
Impact. The proposed inception cohort study, CAMEO, will be unique in providing a robust, novel platform to
study factors that contribute to healing in pediatric CD that can then be immediately translated into clinical
practice, as well as guiding future therapies targeting the microbiota and host immune responses in patients
unlikely to achieve healing with current approaches.