The Quantitative Imaging Biomarkers Alliance has set out technical standards for liver MR elastography covering acquisition, reconstruction, quality assurance and analysis across vendors.

Its central quantitative claim is that, with a within-subject coefficient of variation of about 7%, a change in liver stiffness of at least 19% is needed to say with 95% confidence that biology has changed rather than the measurement.

This matters for follow-up reporting: a stiffness that rises from 3.0 to 3.4 kPa is a 13% change and is within measurement noise. Reports should state the change as a percentage and say whether it exceeds the threshold, and follow-up scans should use the same protocol.

Report change in liver stiffness as a percentage between examinations.
Call a change real only when it reaches about 19% or more.
Use the same scanner protocol for serial MR elastography where possible.
Include image quality and region-of-interest details in the report.

Why it matters

It stops small, meaningless shifts being reported as fibrosis progression or regression.

The statistics, in plain English

A within-subject coefficient of variation of 7% describes how much repeat scans of the same unchanged liver differ. Multiplying by about 2.77 gives the smallest change that is 95% likely to be real: roughly 19%.