CT Liver Fat Fraction Quantification

Non-Alcoholic Fatty Liver Disease Clinical Trial

Official title:

AI-based Quantification of Liver Fat Fraction Using Two-energy Ultra-low Dose CT Compared to MRI

Clinical Trial Details — Status: Enrolling by invitation

Administrative data

• NCT number: NCT05676177
• Other study ID #: 9785-22-SMC
• Status: Enrolling by invitation
• Phase: N/A
• Start date: January 29, 2023
• Est. completion date: December 31, 2025

Study information

• Verified date: April 2023
• Source: Sheba Medical Center
• Contact: n/a
• Is FDA regulated: No
• Study type: Interventional

Clinical Trial Summary

Our aim is to develop an AI based tool to use ultra-low dose CT in two separate energy levels using a single-energy CT machine to quantify liver fat in individuals at risk for having non-alcoholic fatty liver disease (NAFLD), compared to MRI which serves as the standard of reference.

Secondary aim of our study is to validate the developed artificial intelligence (AI)-based model on a second group of participants ("external validation").

Description:

Non-alcoholic fatty liver disease (NAFLD) is the most common cause of chronic liver disease. It affects 25% of the global population, with a higher proportion in Middle Eastern countries, especially in individuals with type 2 diabetes mellitus (T2DM), in whom NAFLD may be seen in up to 70% of patients. Studies have shown that in coming years, the disease is likely to become more prevalent, with increasing number of patients presenting with the more advanced disease form, nonalcoholic steatohepatitis (NASH), the latter gradually becoming a leading cause for liver transplantation, alongside viral hepatitis. NAFLD is characterized by excessive fat deposition (steatosis) in liver cells and can appear in both obese and non-obese individuals. Among individuals diagnosed with NAFLD, an estimated 12-14% have NASH, which can lead to liver fibrosis, cirrhosis and hepatocellular carcinoma (HCC). As NAFLD is associated with cardiometabolic disorders, including obesity, insulin resistance, T2DM, hypertension and atherogenic dyslipidemia, it increases the risk of cardiovascular events and death. When discovered early, NAFLD can be treated by both lifestyle modification and various drugs. Although the gold standard for detecting NAFLD and quantifying the fat contents in liver cells is a non-targeted liver biopsy, blood tests and non-invasive imaging can assist in early diagnosis of patients at risk for developing NASH and for follow-up after treatment. Ultrasound for detection and assessment of hepatic steatosis is limited by subjective assessment; and variable sensitivity and specificity (53-76% and 76-93%, respectively). US may fail in obese patients or those with ascites, and is highly operator- and platform-dependent (inter- and intra-reader agreement ~50%). Moreover, it has limited utility in fat fraction quantification. Novel US methods are being constantly developed for accurate quantification, but are yet to be agreed upon and used in daily clinical routine.

The most commonly used method for quantifying the amount of fat in the liver is MRI, and specifically chemical shift imaging sequences. However, MRI has limitations, including the cost of scans, limited availability worldwide and patient-specific limitations, including claustrophobia and implanted electronic devices which may be unsafe in the MRI magnetic field. Currently, single and dual energy CT have shown limited utility in diagnosing and quantifying liver steatosis, and although CT is readily available worldwide, currently CT cannot be used for liver fat fraction quantification or for early NAFLD diagnosis. Attempts to utilize dual-energy CT, with and without use of artificial intelligence (AI) has shown limited success. Moreover, dual-energy CT is not readily available in most medical centers.

A prior study the investigators performed has already shown that ultra-low dose chest CT can diagnose liver steatosis, but the investigators have not yet assessed its capabilities in quantifying the amount of liver fat. Therefore, the investigators' aim is to develop a novel methodology in which ultra-low dose abdominal CT could be used for both diagnosing NAFLD and quantifying liver fat contents.

Recruitment information / eligibility

• Status: Enrolling by invitation
• Enrollment: 150
• Est. completion date: December 31, 2025
• Est. primary completion date: December 31, 2024
• Accepts healthy volunteers: Accepts Healthy Volunteers
• Gender: All
• Age group: 18 Years and older

Eligibility

Inclusion Criteria:

Adult patients (age =18 years)

• At risk for hepatic steatosis (defined as at least one of the followings: age >50 years, over weight (BMI>25), impaired fasting glucose or impaired glucose tolerance, T2DM, gestational diabetes, hyperlipidemia, hypertension, elevated liver enzymes, family history of steatosis or cirrhosis, increased liver span per medical examination, increased ferritin levels and the patatin-like phospholipase domain-containing 3 polymorphism), as decided by the treating endocrinologist in our institute's Medical screening department. 12-14

• No history of malignancy involving the liver.

• No known risk factors for hepatic iron deposition (multiple prior blood transfusions, known hemochromatosis).

• Subjects able to understand study procedures and provide informed consent.

• Subjects able to hold their breath during CT and MRI scans.

Exclusion Criteria:

Patients younger than 18 years.

• Patients with risk factors from hepatic iron deposition (multiple prior blood transfusions, known hemochromatosis).

• Patients with known malignancy that involves the liver.

• Patients unable to hold their breath for both CT and MRI.

• Patients with severe claustrophobia.

• Patients with implanted devices of shrapnel.

• Pregnant people

Related Conditions & MeSH terms

• Fatty Liver
• Liver Diseases
• Non-alcoholic Fatty Liver Disease
• Steatosis of Liver

Intervention

Diagnostic Test:
• Ultra Low Dose CT
Two immediately consecutive scans with either one or two breath-holds First scan (ULD_DECT_1) 140 kilovolt peak (kVp) - fixed current 10 or 20 milliampere (mA) (if body mass index (BMI)>30), that is 5 or 10 mAs Second scan (ULD_DECT_2) 80 kVp - fixed current 20 or 40 mA (if BMI>30), that is 10 or 20 mAs
• Fat quantification MRI
Dual echo scans, as well as proton density fat fraction (PDFF) scans, will be performed to assess liver fat fraction

Locations

Country	Name	City	State
Israel	Prof. Noam Tau	Ramat Gan

Sponsors (1)

Lead Sponsor	Collaborator
Sheba Medical Center

Country where clinical trial is conducted

Israel, 

Outcome

Type	Measure	Description	Time frame	Safety issue
Primary	Developing AI model of liver fat fraction assessment on data obtained from ultra-low dose CT, using MRI data as a standard of reference	The investigators will develop an AI based tool to use ultra-low dose CT in two separate energy levels using a single-energy CT machine to quantify liver fat in individuals at risk for having NAFLD, compared to MRI which serves as the standard of reference. The MRI data will be extracted from the dual-echo scan, which can produce an MRI-based liver fat-fraction, and this data will be then used to create an AI CT model.; The AI model will be developed to be able to accurately produce an exact quantification of the liver fat fraction (exact percentage) on ultra-low dose CT.	Through study completion, up to 24 months
Secondary	External validation of the AI CT liver fat fraction model using a second participant group not included in the development of the AI-based CT model	After developing the AI-based CT liver fat fraction model, the model will be tested on participants who will undergo the same study procedures (CT and MRI), but these participants' MRI and CT scans are not used to develop the AI model, but rather to test and validate the model.; The AI-model based ultra-low dose CT fat fraction data obtained from these participants will be compared to the data from the MRI standard of reference, and a difference of up to 2% between CT liver fat fraction and MRI fat fraction will be considered a successful validation of the developed AI model.	Through study completion, up to 24 months