coronary artery calcium scores from regular chest CTs

As mentioned in prior blogs (see below), i think the data are quite good for quantitative coronary artery calcium (CAC) as a marker for future clinical coronary artery disease (CAD). but the big issue to me is the excess radiation exposure. however, loads of patients are getting chest CTs for a variety of reasons, and a recent article found that CAC scoring on routine chest CTs is quite a good predictor of mortality (see CAD calcium score from chest CT jacccardiovasc2016 in dropbox, or J Am Coll Cardiol Img 2016;9:152). details:

--4,544 community-living people (mean age 68, 63% male) had "whole-body" CT scans, which included both EKG-gated CTs with 3mm cuts for their CAC scores and 6 mm chest CTs. All were done between 2000 and 2003, with mortality followup through 2009.
--157 people died, and these were matched with 494 controls by sex and age.
--cases and controls were well-matched for BMI, total cholesterol, HDL, and use of lipid-lowering agents. not so well-controlled for diabetes, hypertension, smoking, or family history of CAD, though on review of the numbers, the difference in hypertension was not significant
--results:
    ​--the Spearman correlation of CAC scores between the 2 different CT scans was very high, at 0.93 (p<0.001), with 1.0 being perfect (at least i am told this correlation is very high, per my son Jake, my local statistician masquerading as a medical student)
    --the CAC scores were lower in the 6mm vs 3mm CT scans (median of 22 vs 104 Agatston units), with p<0.001
    --adjusted for traditional CAD risk factors, for each standard deviation higher, the CAC score on the 6 mm CT scans was associated with a 50% higher odds of death [OR = 1.5 (1.2-1.9)], essentially the same as the odds ratio for the dedicated 3 mm CT scan [OR = 1.5 (1.1-1.9)]
    
so, this brings up a few points:
--CAC scores are a strong predictor of CAD events and all-cause mortality, adding predictive value to the Framingham Risk Score (and is usually cited as the best of the "non-traditional" risk factors, as per the Am Heart Assn)
--the JAMA study in the blog below was an assessment of the Framingham Study participants, which found a 1.6% incidence of CAD over 9.4 years in those with an Agatson score of “0”, even if they were felt to be "statin-eligible" by the quite aggressive 2013 ACC/AHA guidelines).
--though the EKG-gated CTs are not generally approved by insurances for coverage and only 600K are done annually in the US, there are >7.1 million chest CTs done annually
--it is interesting that regular chest CTs are so very strongly correlated with CT scans done specifically to assess CAC scores, since the latter are finer (3mm cuts vs 6mm) and are timed/gated to the EKG to minimize motion artifact from the beating heart
--many of the huge numbers of chest CTs done are for people who are already at high risk of CAD (eg, they smoke or have one or more of other CAD risk factors which lead to the chest CT). so, it may be really useful for the radiologist to comment on the CAC score, in order to help us risk stratify patients for either more or less rigorous clinical CAD prevention (eg, whether to use statins, or perhaps aspirin).
--the National Study of Lung Cancer, the study on LDCT which propelled the recommendations to check LDCTs in smokers (see blog below for details and critique), used large numbers of regular high dose CTs to check on suspicious lesions (over 3 years, of the 26,722 people in that study, 18,146 had false positive tests, and 58% went on to regular or PET-CT scans). so, with the routinization of low dose chest CTs (LDCTs) in smokers, there will be many many more regular chest CTs done that could be used to assess CAC scores
--one wonders if LDCTs themselves would allow an accurate predictive model for CAD risk assessment through CAC scores. i did speak with a radiologist who said that he could definitely evaluate coronary arteries for calcium on LDCTs. and, since there is such a strong correlation between regular CTs and those done specifically for CAC scoring, LDCT evaluation for CAC would likely be very useful if it turns out the the CAC scores for them matched well with those of regular chest CTs
--though in the study there was some increase in all-cause mortality in those with a "0" on either type of CT scan, the data were not granular enough to know what they actually died from. it would be really useful to know that: for example, since the 6 mm CT scan could conceivably miss small calcifications, a read of "0 Agatston units" may not have the same low rate of events as a "0" with the 3 mmEKG-gated  scans.
--and, there are some concerns about generalizability of the results, since the patients in this study were predominantly non-Hispanic white adults, many of whom self-referred for whole-body scans for "preventive care".
--so, the relative beauty of this study is that the data are already there, we just need to get the radiologists to read and report it (currently, CAC on chest CTs is reported only on 44% of chest CTs, and is not quantified). and there really could be very important changes in how aggressive we are in our CAD prevention approach, both increasing and decreasing its intensity depending on the CAC score

for more background, see http://blogs.bmj.com/ebm/2015/08/05/primary-care-corner-with-geoffrey-modest-md-comparison-of-the-2013-accaha-lipid-guidelines-to-atpiii/ which looks at the actual predictive value of CAC scores and the considerable overtreatment by the new 2013 Am Heart Assn guidelines

there have been several prior blogs on LDCT screening in smokers, for example, see http://blogs.bmj.com/ebm/2013/07/16/primary-care-corner-with-geoffrey-modest-lung-cancer-screening-with-low-dose-ct-ready-for-prime-time/