DIAGNOSTIC ACCURACY OF CCTA

The June 13 issue of Arch in Internal medicine carried an article by Dr Benjamin Chow from University of Ottawa, Canada, studying the accuracy of CT coronary angiogram ( CCTA ), across different medical facilities. Are all CT angiogram interpretation reliable and the same? The answer is NO.
They studied 169 subjects of intermediate cardiac risk group from across 4 centers in Canada. These subjects all had their CCTA done and were agreeable to have their invasive coronary angiogram done within 10 days of their CCTA. The CCTA and coronary angiogram were read, in a blinded fashion by independent observers.
Well, the centers failed badly. Their sensitivity is only about 81.3%, specificity about 93.3%, positive predictive value was 91.6% and negative predictive value was 84.7%. All these numbers were way below those published by specialised centers in the initial papers. Whats even more upsetting is the spread of accuracy across different centers. See the table below.

Diagnostic accuracy ranges across centers

Statistical measure of CCTA accuracy	Lowest value achieved by any center (%)	Highest value achieved by any center (%)	p
Sensitivity	50	93.2	<0.001
Specificity	92	100	<0.001
Positive predictive value	84.6	100	<0.001
Negative predictive value	42.9	94.7	<0.001

The numbers were varying so much, how can one draw a conclusion. Obviously, the experience in individual centers vary alot. Therein lies the need for constant quality assurances from these centers. Afterall, the CCTAs are not cheap, and they carry a significant radiation risk, although I am told that the later machines, have much less radiation. Couple with these risk, if the CCTA varies so widely in their accuracy across centers, it may increase the need for invasive coronary angiogram, thereby negating the need for the CCTA isn the first place.
As more an more data comes out with the older 64slice MSCT, we know that that technology is just not good enough.
As for the newer 256 slice MSCT, or the dual source, or multi-detector MSCT, supposedly better, we need much more data. It looks like in all these Xray technics where we interpret a mixture of gray ( from white to black ), nothing is really black or white. It is all in different shades of gray, and so observer experience is of paramount importance.
As one of my old physician colleague use to say, when we discuss Xrays, the eye sees not, what the mind knows not.
In CCTA, nothing is really black and white. It is all a mixture of gray. Oh yes, you get significant radiation, not as much as Fukushima, but about 500 CXRays.