MR Assessment of Perianal Crohn's Disease

Overview

Perianal Crohn's disease (pCD) is the commonest form of fistulising Crohn's disease, with up to 38% of patients affected and with 30% of them experiencing recurring disease symptoms. Presence of fistula can lead to major morbidity due to cutaneous perianal abscess formation or drainage. To date, it is very difficult to quantify inflammation in patients with pCD due to the absence of reliable disease activity measurements. In addition to this, optimising therapies for pCD is quite challenging and may have a major impact on quality of life. Magnetic resonance imaging of pelvic is a standard examination for the anatomical evaluation of pCD which is significant in terms of surgical therapy and progress. The overall hypothesis is that newer MRI techniques such as magnetization transfer (MT), diffusion weighted image ( DWI) and dynamic contract enhancement (DCE) are better suited to measuring the inflammatory vs fibrotic burden in pCD. The aim of this project is to measure disease activity within pCD and luminal CD using MRI sequences before and after biological therapy

Background: Inflammatory Bowel Disease has the United Kingdom prevalence of \~620,000 patients. PCD is the commonest form of fistulising Crohn's disease with 20% of patients affected and with 30% of them experiencing recurring disease symptoms. Durable clinical remission rate for complex fistulas is only 37% at the end of a 10-year follow-up. To this date, it is very difficult to reliably measure disease activity within the fistulas in pCD. Patients are sometimes needlessly left on long term medical therapy when their pCD has reached a chronic, fibrotic and irreversible stage. This practice needlessly exposes patients to clinical risk and health care systems to the financial expense of up to \~£20k/patient/year. More commonly, medical therapies are stopped prematurely, leading to prompt disease recurrence and chronic fistulisation with a major impact on quality of life. Importantly, it is very difficult to optimise therapies for pCD due to the absence of reliable disease measures. It is not possible to reliably correlate therapeutic drug monitoring to disease activity states. Due to all these limitations there has been an alarming paucity of well-designed randomised clinical trials. To this effect, both the 5th Scientific Workshop on pCD of the European Crohn's and Colitis Organisation and the global consensus on the classification, diagnosis and treatment of pCD by the international Organisation of Inflammatory Bowel disease, have highlighted the urgent need of more sensitive and specific MR-markers of disease activity in pCD. Once reliable MR measures of disease activity have been validated then these can be used to design appropriate, inexpensive, sensitive and non-invasive MR based clinical activity scores. Various clinical and radiological outcome measures have been investigated to quantify pCD activity. Clinical assessment is crude and participant to inter-operator-variability, with response loosely defined as a reduction of ≥50% in the number of draining fistulas and remission the absence of draining fistulas on two consecutive visits. This assessment has never been validated, and takes no account of the inflammatory state of the disease deep within the perineum. A Perineal Disease Activity Index (PDAI) was developed to assess perianal disease severity and response to therapy. It is based on the assessment of quality of life and disease severity. The PDAI score is 87% accurate and wrongly classifies patients' disease activity in 10% of cases, but is insensitive to change after medical therapy and its remission cut-off never been properly determined. The gold standard imaging modality for pCD is pelvic MRI. A recent meta-analysis indicated a sensitivity of MRI for fistula detection of 0.87 (95% CI: 0.63-0.96) but a low specificity of 0.69 (95% confidence interval (CI): 0.51-0.82). A T2-weighted MR sequence with fat-suppression sequences is the current advised technique for MR fistula imaging. This forms the basis of the Van Assche 1 score. This MR-based score was suggested to provide combined information of anatomical fistula description and features reflecting inflammatory activity. The score is insensitive to clinical response, with no significant difference shown in clinical responders to anti-tumor necrosis factor (anti-TNF) therapy before and 6 weeks after treatment, nor was there a difference in the score between clinical responders and non-responders. Longer term studies replicated these findings and also showed a dis-concordance of at least 12 months between clinical and radiological remission. T1-weighted gadolinium-enhanced images did not change significantly before or after treatment . No correlation was found between these MR-based measures of fistula activity and other known markers of disease activity like C-reactive protein and the PDAI. The majority of published studies have been carried out at 1.5T. The 3T scanners available at the Universities of Nottingham,London and Manchester will provide better resolution and sensitivity allowing us to design better clinical protocols incorporating quantitative MRI. 3T MR scanners are available in National Health Service(NHS) trusts but are not routinely used for the assessment of pCD. Our aim is to develop an optimized protocol for 3T imaging although we expect that our results will also be able to improve 1.5T imaging. This work will create a paradigm shift in the radiological monitoring of disease activity in pCD. Recent work has highlighted the use of magnetisation transfer (MT) MRI techniques7 in quantifying inflammation and differentiating this from fibrosis. MT generates contrast that is determined by the fraction of large macromolecules and the immobilised phospholipid cell membranes in tissues. Particularly when combined with T2 measurements, MT may be helpful in differentiating and quantifying inflammation by the relative absence of collagen, and differentiating inflamed perianal fistulas that need treatment from other chronic fibrotic and irreversible fistulas that would not be responsive to treatment - better selecting patients for expensive medical or surgical treatment. DCE has shown marked enhancement in patients who later develop perianal abscesses needing surgical treatment or needed medical escalation of therapy. Changes to the timing of how the tissue takes up the contrast agent may also provide information on whether the tissue is still actively inflamed. Early pilot data on fistula volumetrics and diffusion-weighted MRI in luminal Crohn's disease have shown good correlation to disease activity both in pCD and in luminal disease. In diffusion-weighted MRI, the source of the contrast is the movement of the increased number of water molecules found in inflamed, oedematous tissues. We suggest that new MR techniques will make it possible to more accurately quantify the inflammatory burden within pCD. This will allow for better patient follow-up and more effective clinical decision making. Better medical management with more timely therapy onset, optimisation and cessation of therapy is needed. Moreover, although early data on therapeutic drug monitoring is emerging in luminal Crohn's disease, similar data is lacking in pCD due to the lack of appropriate measures of disease activity. Improving MR methodology, by employing quantitative MRI to separate inflammation from fibrotic tissues, and to determine fistula size better will allow for the design and validation of a sensitive clinical scoring system. We will use the improved sensitivity of 3T to develop optimal imaging approaches and then use these results to tailor a convenient clinical score. We will determine whether this score can be used at both 3T and 1.5T. Aims and hypothesis: The overall hypothesis is that newer MRI techniques such as MT, DWI and DCE are better suited to measuring the inflammatory vs fibrotic burden in pCD The aim of this project is to measure disease activity within pCD and luminal CD using MRI sequences before and after biological therapy. Materials and Methods: Patients This multicentre prospective cohort study will be conducted at three sites in Nottingham (Nottingham University Hospitals), London (Northwick Park NHS Foundation Trust) and Manchester (Salford Royal NHS Foundation Trust). A total of 25 patients (age 16-75 years) with pCD as deemed by a PDAI of \> 4 and a clinician assessment, prior to the onset of biological therapy. All eligible patients will receive infliximab or adalimumab or Ustekinumab. Patients will have infliximab 5 mg/kg i.v. at weeks 0, 2, 6 and 8 weekly thereafter, or Adalimumab subcutaneously (s.c.) at week 0 (160 mg), 2 (80mg), 40 mg every 2 weeks thereafter or Ustekinumab 6mg/kg i.v. at week 0 and a weight-based s.c. every 12 weeks thereafter. Patients will be excluded from the study if they are due to have surgery in the near future (12 weeks). Study Design: Recruited participants will have no more than three visits. Visit 1: This will include screening to ensure participants meet the inclusion criteria and possess none of the exclusion criteria. The MRI evaluation (visit 2) at 1.5 and then 3T will be performed for all those recruited before starting biological therapy at week 0. The second MRI scanning session will be 12 weeks after biological therapies onset (visit 3). The period between scanning (MRI) (visit 2) and biological therapy onset will be a maximum of 4 weeks MR Imaging Technique: Patient will be placed in a supine position (feet first) for scanning at both 1.5T and 3T, with the pelvis centred within a torso phased-array surface coil unless this position proves too uncomfortable.The MRI scanning will be undertaken on 1.5 Tesla (2 Philips and 1 GE) and 3T (Philips) scanners at the Sir Peter Mansfield Imaging Centre at the University of Nottingham, the Clinical Sciences Building at the Salford Royal NHS Foundation Trust , Northwick park NHS trust (1.5T) and University College London (UCL). Statistical analysis All data will be assumed to be non-parametric (due to the small sample size). Data will be presented as median +/- interquartile range with all comparisons of single and multiple categorical and continuous matched and non-matched variables done accordingly. Correlation between the clinical and MR endpoints will be carried out with a Spearman test. A univariate analysis will be undertaken of the MR endpoints at 12 weeks for responders and non-responders. For all candidate parameters, κ reliability values based on X2 probability testing will be calculated. We aim to undertake a multiple regression model in order to create the best-fit weighted scoring system. Double-weighting will be applied for inflammatory scores versus anatomical scores. However, this may not be possible at this present stage but this work will provide the pilot data to properly power a downstream study. P values of less than 0.05 will be deemed significant. All analyses will be carried out with GraphPad Prism.