PET MPI stratifies risk, but clinical relevance uncertain

A large, multicenter observational study has found that PET myocardial perfusion imaging (MPI) is a powerful predictor of incremental risk of cardiac death in patients with known or suspected coronary artery disease (CAD). The study was published online Dec. 5 in the Journal of the American College of Cardiology. However, an accompanying editorial suggests that an additional prognostic tool is less valuable than a tool that helps to identify those patients who would most benefit from available treatments.

Sharmila Dorbala, MD, of Brigham and Women’s Hospital in Boston, and colleagues enrolled 7,061 patients who were clinically referred to four centers for a pharmacological stress rubidium-82 (Rb-82) MPI. Each patient underwent the study using a dedicated or hybrid PET/CT scanner with site-specific protocols. Patients were followed for a median duration of 2.2 years.

The primary outcome of the study was cardiac death, with all-cause death being the secondary outcome. Three of the four centers were able to ascertain the causes of death among study enrollees in their centers, and identified 169 cardiac deaths in a cohort of 6,037 patients. All-cause death was studied in the full cohort of 7,061 patients; there were 570 deaths among the cohort.

Using a multivariable Cox proportional hazard model, the researchers studied the increase in relative hazard for cardiac death and all-cause death by percent scarred and percent ischemic perfusion abnormalities. They reported that “for each 10 percent increase in percent myocardium ischemic and percent myocardium scarred, the hazard of cardiac death increased by 34 percent and 57 percent, respectively.”

The researchers separately looked at the PET/MPI’s value as a prognosticator for patients with rest left ventricular ejection fraction (LVEF). There were 2,102 such patients enrolled in the study. Dorbala et al found that a 10 percent higher rest LVEF was associated with a lower hazard of cardiac death (hazard ratio 0.57); among these patients, for each 10 percent of myocardium ischemic there was an 84 percent increase in the risk of cardiac death and for each 10 percent of myocardium scarred there was “a trend toward a 23 percent higher hazard of cardiac death,” the authors wrote.

The researchers asserted that considering the percent myocardium ischemic and the percent myocardium scarred with available clinical data resulted in significant clinical incremental value for predicting cardiac death. They found that an abnormal PET MPI along with clinical risk markers provided significant risk reclassification in 12 percent of patients, “confirming the clinical utility of PET MPI,” the authors concluded.

While noting that this study “significantly advances our understanding of the prognostic value and effectiveness of PET MPI,” editorialists Paul Schoenhagen, MD, and Rory Hachamovitch, MD, both of the Cleveland Clinic, queried whether “although giving a ‘correct’ answer, th[is] stud[y] by design may pose the wrong question.” They pointed out that the clinical relevance of prognostic tools in patient management decisions remains unclear.

Schoenhagen and Hachamovitch explained that this study, as well as others, had demonstrated that myocardium scarring and LVEF were more powerful predictors of cardiac death than extent of ischemia, yet studies of post-MPI clinical practice show that referral to catheterization and revascularization is based on the degree of the ischemia. “It appears there is a mismatch between the drivers of patient risk and the drivers of patient management,” they wrote.

The editorialists suggested that the availability of registries such as the one used in the current study offers opportunities for researchers to assess “whether a test can identify which patients will gain a benefit from a specific therapeutic approach. The role of testing is defined in the context of a specific intervention and whether the effectiveness of the intervention is improved by the use of the imaging study to identify optimal candidates for treatment.”