Biotech Values

[iwfal]

Aetna considers calcium scoring (e.g., with ultrafast (electron beam) CT, spiral (helical) CT, and multislice CT) experimental and investigational for all other indications because the definitive value of calcium scoring for assessing coronary heart disease risk has not been established in the peer-reviewed published medical literature.

1) I don't know the date of the above and I do know that there has been a lot of recent movement in this area in California such that as of about 6 months ago all the major insurers but one had accepted calcium scans for at risk populations. (I actually came to know, for accidental reasons, one of the biggest publishers in this area and have asked him about it.)

2) Most of the high risk people I know (based on BMI), none of whom have had a heart attack, have had a calcium scan.

3) A particular comment on the above - it is absurd for Aetna to assert that it has not been established in peer-reviewed medical literature. There are literally hundreds of articles spanning, at this point, about 2 decades. Many with very large databases.

the score depends on what you're looking at - spatial distribution of calcified plaque, overall amount of calcium,

Absolutely - as I mentioned earlier in the thread one of the ingredients of the secret sauce was that the people working in this field have been very cooperative and have combed through vast amounts of data to find out what artifacts matter (i.e. standardizing on a way to count a calcium score that makes it most predictive). I speculated earlier in the thread that if the CIMT people did the same they could very very significantly improve the predictive power of CIMT. But to date they have not done so.

Second, it also depends on who you're looking at - age (specificity tends to decrease with advanced age), a/symptomatic, other risk factors like LDL-C etc.

Yep, one of the factors in interpreting calcium scans is age. Of course it also is in Framingham.

Overall, we are in agreement that the best thing to do is the non-radiative improved Framingham model.

Yep - but as a practical matter no one has standardized that improved Framingham. See my above comments on one of the reasons calcium score is so predictive compared to other tests like CIMT. For instance, my personal physician, who is about as active in this field as it is possible to be, does not use anything other than the Framingham model because it is so unstandardized. Thus as chooser you are left with a choice - a mediocre predictor wo radiation (C=0.70 is pretty mediocre), or a good predictor with radiation. And if you are known to be at >= medium risk via Framingham model your choice is starker because the stakes go up and the predictive ability of the calcium score go way up (c statistic near 0.9 is an outstandingly predictive test):

1) Start taking high dose statins and niacin and know that 50% (WAG - I could calc, but have run out of time) of the time you will be taking dangerous dosages of dangerous drugs when you don't need to.

2) Suffer a heart attack which, best case, damages your heart (worst case kills you)

3) Get a calcium scan to decide between the two choices above.

And FWIW, the risk of low dose radiation is far from clear. E.g. the "low dose" hiroshima population got about 100 times the dose of an EBCT (and around 5% extra risk of cancer). Even if you assume linearity of dose vs excess cancer rate that implies an excess cancer risk of 0.05%. And that linearity assumption is problematic once you start approaching background rates per year (background rates being about 6x an EBCT) since there is a lot of data that indicates that the body turns on repair mechanisms proportionate to the general rate of radiation (e.g. fascinating studies like irradiating mice with high but sublethal doses and then radiating them with a normally lethal dose and they don't die - or epidemiological data indicating that locales with high uniform natural background rates of radiation actually have lower rates of cancer.)