AI - Thanks for posting that link. I had not seen the actual numbers. Table 3 indicates that my assumption in my original post was correct. If you look at the median and STD (standard deviation) values at pre- and post separately, you can see that the STD relative to the median value is much higher (X 4) for LAP than for any other measure. This means a broader band of error around the median, which results in the distributions across treatment groups overlapping more. This makes it harder to show statistical significance even if the medians are visually quite different.
Another thing I noted is that the V group (IPE) randomly happened to have higher baseline values across all plaque measures, although none were significant. This could complicate interpretation of changes over the trial (less room to move up/ more room to move down in the V group). On the rightmost column where it says ANCOVA, this may indicate they were adjusting for baseline differences, so the "adj p" values in the right column are likely the most accurate.
I am not a cardiologist, so I am unsure of the relationship between LAP and the other measures, but based on doing a little reading most did not seem to be dependent on each other. I agree however that LAP appears to be a fraction of total plaque, just as the other measures are all part of the total.
Stats geek note - This slide presents the data the wrong way. You use standard deviations (STD) to express variability for mean values, and interquartile range for median values. They presented a mish-mash of median and STD stats that do not really go together. I am sure this will be fixed on publication.
News 
Market Data 
Discover 
