乌鸦传媒

CHICAGO -- Quantitative myocardial blood flow assessed on PET scans caught some multivessel coronary artery disease missed with standard PET perfusion imaging, a study showed.

Qualitative myocardial perfusion imaging underestimated the extent of disease compared with gold standard angiography, calling multivessel disease single vessel or no disease in 45% of cases,, of the Cleveland Clinic, and colleagues found.

The newer quantitative analysis method significantly improved on that performance both in net reclassification and overall prediction combined with perfusion and other clinical variables, the researchers reported here at the American Society of Nuclear Cardiology meeting.

"Stress myocardial blood flow could form the basis of a risk prediction score beyond visual myocardial perfusion imaging, which could then be integrated along with other adjunct markers into clinical software for detection of multivessel coronary artery disease," Lou said.

The retrospective study included 202 consecutive patients who underwent a rubidium-82 PET/CT during rest and stress with regadenoson (Lexiscan) from December 2010 through August 2013.

Patients had to have a coronary angiogram for comparison within 6 months of the scans but no prior coronary artery bypass surgery or recent stenting, severe aortic stenosis, inflammatory cardiomyopathies, or transplantation.

Those angiograms, interpreted by readers who were blinded to the results, indicated multivessel disease in 54% of the cohort.

Compared with the angiogram, qualitative myocardial perfusion imaging on PET underestimated the extent of coronary artery disease for 27% of patients and overestimated it for 21%.

For multivessel disease, the visual myocardial perfusion imaging underestimated extent of disease for 45%.

But quantitative myocardial blood flow tracked as a continuous variable along with severity of angiographic disease, with significantly greater flow in normal versus one vessel disease (P<0.0001), one versus two vessel disease (P<0.007), and two versus three vessel disease (P<0.01).

The researchers also looked at a cutoff of 1.6 ml/min/g for an abnormal myocardial flow result, determined as the lower 95%confidence interval for patients in the cohort with normal angiograms.

When used that way, flow correctly reclassified 81% of the patients (34 of 42) that qualitative perfusion imaging called one vessel disease, but angiography confirmed as multivessel disease. The study technique wrongly reclassified 38% (18 of 37) as abnormal in cases where the other modalities found single-vessel disease or no disease.

The net reclassification benefit was significant at P=0.003.

Predictive ability improved when myocardial blood flow was added atop perfusion imaging in a multivariate model for multivessel disease identification. The C-statistic went from 0.787 to 0.828 and the total chi-square value went from 37.72 to 45.1.

Limitations included the retrospective design at a tertiary center with its attendant referral bias and likely selection bias in which patients received both PET and coronary angiography.

Without data from normal patients or those with some risk factors, the results could have been more a reflection of endothelial dysfunction, suggested , director of nuclear cardiology at Westchester Medical Center in Valhalla, N.Y.

"It may have reduced specificity if the study had included a large arm of normal subject," he said from the discussant panel at the session.

Comment