Early MRI in Acute GallstonE Disease

Overview

This pilot study aims to determine the utility of direct Magnetic Resonance Cholangiopancreatography (MRCP) in the assessment of suspected acute gallstone disease presentations. This will take the form of a randomized controlled trial, the design of which is based on recent internal audit data which indicated that a high proportion of patients ultimately require MRCP prior to diagnosis, treatment and discharge, and suggested that early scan may improve overall costs and outcomes. Patients with symptoms and signs suggestive of suspected gallstone disease and deranged liver function tests/amylase (i.e. suggestive of a potentially obstructive gallstone disease), will be enrolled across the full range of 'front-door' locations for surgical presentations within the hospital (Emergency Department, Surgical Emergency Unit, Ambulatory Assessment Unit), and randomized to one of two diagnostic pathways which are both existing variations in current clinical care: 1. MRCP is used as the first mode of imaging; 2. following a standard care model (ultrasound then MRCP if deemed appropriate). An assessment will then be undertaken of the cost-effectiveness of early MRCP versus standard care, using the primary outcome measure of cost to diagnostic scan report calculated using hospital episode statistics (HES), with secondary outcome measures to assess the overall utility which include length of stay, time to diagnosis, overall cost of admission using HES, in-hospital complications, Patient Reported Outcome Measures (PROMs), readmission and re-attendance rates (ED/GP), and service/radiology costs.

Acute biliary disease comprises a significant proportion of surgical presentations in Accident and Emergency departments (A\&E), as well as of emergency surgical admissions to secondary care. Historically, this patient group first undergoes an abdominal ultrasound to confirm the presence of gallstones and look at ancillary features that may suggest cholecystitis or the presence of bile duct stones. The vast majority of patients who have abnormal liver biochemistry then proceed to Magnetic resonance cholangiopancreatography (MRCP), because the ultrasound scan is unable to clarify matters sufficiently to satisfy clinical need. There is often considerable delay in carrying out two sequential imaging investigations for each patient. A number of studies have raised questions regarding the utility of ultrasound in acute biliary disease. The prevalence of Common Bile Duct (CBD) stones in patients with cholelithiasis is significant at 5-15%. CBD diameter, often used in US reporting as a marker of the presence of choledocholithiasis, may not correspond to the presence of a CBD stone - in one study, only 37% of patients with CBD stones had CBD dilatation on ultrasound. Ultrasound is a user-dependent technology, where the seniority of the scanner can influence the reliability of the report, and a 2013 study by Barlow et al. suggested that ultrasound findings in the context of deranged LFTs are often inaccurate in predicting CBD stones in gallstone pancreatitis, suggesting that MRCP be used to improve diagnosis. A 2015 retrospective study led by Qiu et al. found the sensitivity of peri-operative ultrasound for CBD stones to be very low at 44.95% and supported the use of MRCP as a diagnostic test if CBD stones were suspected. They also suggested the potential for economic benefits of using MRCP over US in patients with potential choledocholithiasis. MRCP has been shown to be more cost-effective than US in selecting patients for Endoscopic Retrograde Cholangio-Pancreatography (ERCP) with suspected CBD stones, has a comparable accuracy to ERCP 6, and has been shown to reduce the number of unnecessary ERCPs performed in gallstone patients 7. Epelboym (2013) and Nebiker (2009) however disagree, characterising the use of MRCP in such patients as expensive and inefficient, and support the use of ultrasound as the primary diagnostic modality 8,9. In a recent paper, Milburn et al conducted a retrospective study examining the patient journeys corresponding to 234 inpatient MRCP scans over a 2-year period. They found that increasing access to MRCP led to further interventions in 22% of cases due to an increased detection of complications and alternate pathologies (e.g. malignancy), and therefore increased overall length of stay. The median duration from request to scan was 2 days, with a further 1 day to scan report (therefore 3 days overall from request to report). They suggested improved access and timely reporting of MRCP could potentially reduce length of stay, though it may also increase the number of interventions 10. Clearly, there is a need to examine whether proceeding directly with MRCP (without preceding ultrasound) would prove to be cost-effective for these patients, and with a growing body of evidence as to the utility of early diagnosis to hasten the patient journey and improve treatment outcomes, further research is needed to determine whether the use of direct MRCP may be beneficial in this regard. A recent audit of patients seen in the John Radcliffe Hospital ED over a period of six weeks revealed 46 patients admitted with gallstone disease, an incidence of approximately 1 admission per day. This correlates well with local audit of surgical admissions, which gives a referral rate of approximately 2 patients a day with 45% of patients referred from ED and 55% being referred from primary care. Despite the broad range of pathologies encountered in the ED and the considerable overlap in presentation, the diagnostic accuracy of emergency physicians with respect to gallstone disease was high, with the suggested specificity of an ED cholecystitis diagnosis calculated at around 85%, based largely on clinical findings and blood tests without radiological confirmation. This suggests that ED clinicians may be able to identify patients whose presenting symptoms are likely to be secondary to acute gallstone disease with a reasonable degree of accuracy on a clinical basis alone. Subsequent analysis of the continued inpatient diagnostic pathways suggested that whilst ultrasound remains the initial diagnostic radiological modality in the majority of patients with such presentations, approximately 59% (27/46) of patients undergo cross-sectional imaging during their admission, with around 48% (22/46) undergoing inpatient MRCP. Whilst access times to MRCP have improved in recent years, this pathway nevertheless may potentially result in an unnecessarily extended time to definitive diagnosis, and potentially unnecessary radiological expenditure. This raises the possibility that using MRCP as the initial diagnostic test in such patients may improve pathway efficiency, reduce length of stay and radiological burden, and hence reduce admission costs, notwithstanding the potential for improved diagnostic accuracy and detection of pathology such as common bile duct stones and/or pancreatitis. It may also improve patient satisfaction and clinical outcomes (e.g. earlier access to ERCP and stone removal in biliary obstruction patients). This can be assessed in a prospective randomised study to assess utility. This pilot study aims to determine the utility of direct Magnetic Resonance Cholangiopancreatography (MRCP) in the assessment of suspected acute gallstone disease presentations. This will take the form of a randomized controlled trial, the design of which is based on recent internal audit data which indicated that a high proportion of patients ultimately require MRCP prior to diagnosis, treatment and discharge, and suggested that early scan may improve overall costs and outcomes. Patients with symptoms and signs suggestive of suspected gallstone disease and deranged liver function tests/amylase (i.e. suggestive of a potentially obstructive gallstone disease), will be enrolled across the full range of 'front-door' locations for surgical presentations within the hospital (Emergency Department, Surgical Emergency Unit, Ambulatory Assessment Unit), and randomized to one of two diagnostic pathways which are both existing variations in current clinical care: 1. MRCP is used as the first mode of imaging; 2. following a standard care model (ultrasound then MRCP if deemed appropriate). An assessment will then be undertaken of the cost-effectiveness of early MRCP versus standard care, using the primary outcome measure of cost to diagnostic scan report calculated using hospital episode statistics (HES), with secondary outcome measures to assess the overall utility which include length of stay, time to diagnosis, overall cost of admission using HES, in-hospital complications, Patient Reported Outcome Measures (PROMs), readmission and re-attendance rates, and service/radiology costs. In addition to standard MRCP scan processing by the radiology department, data obtained through this study will be processed using the new quantitative MRCP+ technology (provided by Perspectum Diagnostics) to assess its diagnostic efficacy in detecting biliary tree dilatation, stones or strictures. Perspectum Diagnostics' quantitative MRCP technology is a software package which allows previously acquired MRCP data to be enhanced and quantitatively characterised using advanced image processing techniques (MRCP+). Conventional MRCP scans can be difficult to interpret since the 2D representation suffers from occlusion problems and a lack of depth information. This new technology can enhance the data without the need for contrast agent, enabling the visualization of smaller ducts. Quantitative characterisation of the biliary tree computes biliary tube diameter at each point along the duct, along with information about the cross-section orientation and branching topology. Combining true 3D rendering of the enhanced data with quantitative characterization facilitates clear mapping of the biliary tree, which can potentially improve surgical planning, facilitate detection of strictures, dilatations and gallstones in suspected biliary colic presentations and objectively stratify patients. As yet however this application of the MRCP+ has not been fully assessed in clinical practice, and this subsequent part of the study would aim to address this and provide data to enable assessment of its diagnostic utility.