BioTel Research Blog

November 07, 2019

Including Imaging Methods in your NASH Clinical Trial

Probably the best imaging technique for measuring fat in your liver is MRI proton density fat fraction (PDFF). MRI-PDFF was developed by researchers at UCSD and the University of Wisconsin. They each came up with two slightly different techniques that obtain the same answer within the liver. The complex-based technique was developed at the University of Wisconsin, and the magnitude-based technique at UCSD. Both utilize the chemical shift properties of water and fat, meaning the hydrogen protons that are in water spin slightly faster than the hydrogen protons in fat. Both take into account the multiple types of fat you have in the liver, whether it's unsaturated or saturated. Both also correct for iron in the liver.

What's the difference between complex and magnitude-based MRI-PDFF?

Some MRI scanners have the complex-based technique built in - meaning the site can buy it. This option may only be available at a limited number of sites. For all other scanners that don't have this option, BioTel Research can acquire magnitude-based images, which are just normal images acquired off an MRI scanner, and calculate the magnitude-based MRI-PDFF.

Should you consider using phantoms?  

Phantoms are helpful because they provide a kind of quality check, which is why we use them. Here at BioTel Research we currently use a belt with four phantoms, and will shortly have one with five different fat fractions in vials. They are placed under the patient and imaged with every scan. When fat fraction is calculated for the patient, it can be compared to the benchmark phantom values. This process provides great confidence that the values in the liver are accurate. 

What endpoints are you going to use?  

Recent publications suggest that anything greater than a 29% reduction in PDFF leads to a histological change. This was actually confirmed in the Madrigal MGL-3196 phase two study, which showed a 30% change was required to be a histological responder. Still another publication showed that a 25% change led to histological changes. This range is where there may be some clinical outcome associated with a reduction in proton density fat fraction.

Are you going to use proton density fat fraction for eligibility?

You're going to have a certain screen fail rate depending on the cutoff value you use. We've seen eligibility or cutoff criteria anywhere from 6% up to 15%. Based on our data, if you set your cutoff at 10%, you can expect about a 28% screen fail rate. If you use ultrasound for pre-screening, or have a cap threshold to use for FibroScan, you may be able to lower that rate. For PDFF, you're forced to conduct this process centrally for several reasons. The MRI scanner will produce a fat fraction map, so it will tell you what the fat fraction is at every location. Still someone actually has to go in and say, "Okay, we'll measure it here." And the question is, what's here? Are you going to measure it over the entire liver? Are you going to measure it over the different Couinaud segments? Are you going to measure it over the anterior/posterior part? The left lobe/right lobe? How are you going to do that? And what software are you going to use? Hence it's very difficult to perform PDFF for eligibility locally.

Lastly, I’ll make the case for using BioTel Research for you central reads. We performed the first multi-site PDFF study back in 2010 (published in 2013) to determine whether PDFF could actually be used in a multi-site clinical trial. So we have years of experience. We've performed over 10,000 analyses and worked with over 460 sites in 30 countries. We've implemented PDFF on every magnet for the three largest MRI manufacturers. And we can calculate PDFF from the raw images, thereby performing the magnitude-based PDFF - meaning we can implement PDFF on almost any MRI scanner. This is helpful because it doesn't limit your choice of clinical sites. And finally, we utilize phantoms to ensure that the scans are accurate.




Catch Up and Listen to our Webinar:

Liver Image Webinar

Written by Jon Riek, Ph.D

BioTel Research, One Preserve Parkway, Suite #600, Rockville, MD 20852