Endoscopic Ultrasound Shear Wave Elastography in Patients With Non-alcoholic Fatty Liver Disease

Introduction: Non-alcoholic fatty liver disease (NAFLD) is the most widespread chronic liver condition across the world. NAFLD over time can progress to its more severe form, non-alcoholic steatohepatitis (NASH), which is associated with increased speed of progression to advanced fibrosis and cirrhosis. Identification of patients with suspected NAFLD/NASH and early stage fibrosis is crucial for prompt clinical management in weight loss and optimization of metabolic disorders for reversal of fibrosis, but the gold standard of liver biopsy can be impractical due to high costs and risk of sampling error and procedural complications. Elastography is a non-invasive imaging modality that provides information about tissue elasticity and is often used to measure liver stiffness as a correlation to fibrosis severity (F0-4). Elastography has been heralded as a promising alternative to liver biopsies given its non-invasive nature and improved cost effectiveness in the general population. Many elastography modalities exist, including Fibroscan, magnetic resonance elastography (MRE), and shear wave elastography (SWE). Unfortunately, despite being highly sensitive and specific, routine MRE for NAFLD/NASH fibrosis screening can also be cost prohibitive. Fibroscan and SWE are more cost effective and are routinely performed via a transabdominal probe for liver fibrosis screening in patients with NAFLD/NASH, but their accuracy may be impeded in patients with obesity. Endoscopic ultrasound (EUS) approaches to elastography may hold several technical advantages over transabdominal in patients with obesity due to EUS-SWE's ability to visualize the liver parenchyma under a thin sub-centimeter gastric or duodenal wall, unlike transabdominal approaches in which the quality may be impacted by a thick abdominal wall. The main disadvantage to EUS-SWE, however, is that it is a fairly novel technology with no established stiffness cutoffs for the various fibrosis grades, unlike VCTE of which fibrosis grade cutoffs are well studied and standardized for patients with NAFLD/ NASH. Specific aims/hypotheses: The investigators aim to determine the efficacy of EUS-SWE compared to other non-invasive scoring systems, such as the FIB-4 score and Fibroscan in patients with elevated body mass index for purposes of liver fibrosis screening. The investigators would subsequently establish optimal stiffness (kPa) cutoffs for liver fibrosis grading in reference to liver biopsies for EUS-SWE for this patient population, as no standard cutoffs currently exist. The investigators hypothesize that EUS-SWE will be more accurate than conventional non-invasive screening tools for liver fibrosis grading. Study design: The investigators are performing a cross-sectional study with prospectively collected data on patients with elevated body mass index and suspected NAFLD/NASH who are referred for upper endoscopy and EUS for any indication and liver biopsy either obtained during the same endoscopic session or available within 6 months of the endoscopy. Study participants are recruited from two sites, Brigham and Women's Hospital and Brigham and Women's Faulkner Hospital (same healthcare system and electronic medical record). Once identified and consented, patients undergo clinically indicated EUS followed by the shear wave elastography portion which in general takes 2-3 minutes and does not expose the patient to any additional anticipated risk other than minimally increased sedation time. The EUS-SWE device being used is a linear EUS endoscope with SWE features (Aloka Arietta 850, Olympus America, Center Valley, PA), an FDA-approved device. Patient demographics, laboratory findings, past medical history, elastography, and liver biopsy results are then collected prospectively and then entered into a patient registry. The patient registry results are placed in a secure cloud server maintained by the Mass General Brigham healthcare system with detailed procedures in place to maintain patient privacy. Statistical Analysis: Variables will be reported as mean ± standard deviation or median \[range\] if continuous and as proportions if categorical. As all comparisons will be paired, analyses between continuous variables will be performed using paired student's t-test and between categorical variables using McNemar's test. Accuracy will be measured by area under the curve (AUC) analysis. Optimal stiffness cutoffs will be performed based on calculations for Youden's index, along with thresholds with minimal sensitivity and specificity of 0.9. Statistical analysis will be performed using SAS 9.4 (Cary, NC).