More prognostic data for cardiac CT and the limitations of virtual histology

As a follow up to a post here in December 2009, another paper confirming the prognostic ability of cardiac CT was just published in Circulation Cardiovascular Imaging (Russo et al, May 11th).

In a series of 441 patients imaged using 16 detector CT, the best predictor of events was a combination of risk score + calcium score + CT angiogram findings. Follow-up up was 32 months on average, and the number of hard CV events was 44.

Significant CAD on CT confered an annual event rate of 8.1%, mild CAD 3.9% and normal coronary arteries on CT meant a risk level very similar to that following a normal perfusion scan - 0.9% per year.

In the same Journal, Thim and colleagues tackled a related coronary issue, that of the reliable identification of necrotic core size in atheroma. This is thought to be related to risk of future CV events. They used an intravascular approach, IVUS-VH (virtual histology) and compared the findings with direct histology in a porcine model of atherosclerosis.There was no relationship at all between the true histology and the virtual histology findings.

I've discussed IVUS-VH before, in terms of the PROSPECT study.

What are your views on these two papers? Should we move away from intravascular coronary assessment towards a non-invasive imaging paradigm? Is the pig model of atheroma a reliable surrogate for human coronary disease (the IVUS-VH algorithm was based on human coronary lesions).

Let us know your views in the comments please.