Here is HR on the subject of the fibroscan:
"With regard to the fibroscan, it can, if the ribcage is not very narrow and the patient not very obese, obain very precise information re the elasticity of the liver, which is excellently correlated to the amount of collagen fibers (fibrosis) that the liver contains. Some fibroscans fail, because the quality of the elastic wave that can be generated is suboptimal due to operater lack of skill, rib deflections causing wave disturbaces, obesity or (rarely) real wound scarring. The problem - and this comment goes to Jim- is that in the " standardized procedure" the judgement of acceptance of a particular individual elastometric shot or picture is left to the machine and its image processing algorithms. If one understands what the essence of the wave speed measurement is as reflected in the details of the individual image, it becommes painfully clear, that the machine algorith sometimes "accepts" individual single measurements that it should have discarded and also that the speed interpolation tha the machine automatically performs and superimposes on the individual image is not infrquently clearly wrong. This normally is weeded out because only the median of 12 measurements is finally used as "the' reflection of the true elasticity and is the only value that determines the final fibrosis judgement. In terms of repetitions of the "shots" or measurement of individual elastic values, the rule is simple: the more you do, the closer you get to the reality. Just like 10 liver biopsies would give you more info than one and 40 will really focus you into what is truly present - sampling error is increasingly reduced and the reality of degree of fibrosis comes into clearer focus the more "fibroshots" you do.. It is practical reasons that limit the " standardized protocol" and it is more a minimum protocol rather than the maximum precision that you can obtain.More mesurements will only reduce the uncertainty window that a particular median will finally represent. Biopsies are only coarsly quantitative in re to fibrosis - hence the "stages". So the more biopsies you would do( if you could, or do a "wedge biopsy", autsch) and if you would (as some French studies did) quantify the fibers with integrated area measurement of the stained fibrous areas and would in parallel do a number of fibroscan shots to the extent that the median now stays rock solid and you have discarded improper waves, then you would come to a very, very close agreement between the two. In the real world, the uncertainties of both methods will combine to widen the window of correlation. It is however easier to narrow the part that stems from the fibroscan due to its noninvasiive nature."
http://www.medhelp.org/posts/Hepatitis-C/No-Fibroscan---Jim-Help/show/326325
Note: "In the real world, the uncertainties of both methods will combine to widen the window of correlation." In other words, it is the variability of both methods that produces the variation in correlation between the biopsy and the fibroscan in predicting staging. It is much more reliable to compare a fibroscan to a fibroscan than a fibroscan to a biopsy. Same machine, same operator, same area of the liver over a unit of time = correlation.