A level of confidence: beta-hydroxybutyrate and myocardial glucose uptake suppression on ¹⁸F-FDG PET imaging

Myocardial ¹⁸F-fluorodeoxyglucose positron emission tomography (¹⁸F-FDG PET) has established a role in diagnosing and assessing various inflammatory and infectious cardiac diseases (e.g., sarcoidosis, endocarditis, and myocarditis) by highlighting pathologic glucose uptake. A key limitation of ¹⁸F-FDG PET is the need to achieve physiologic myocardial glucose uptake suppression (MGS) to obtain diagnostic imaging through patient preparation.¹ Since normal myocardium preferentially utilizes glucose during carbohydrate and insulin-rich states, inadequate preparation results in ¹⁸F-FDG uptake by the myocardium, obscuring pathologic uptake by inflammatory cells, which also metabolize glucose.¹ Accordingly, a substantial amount of research has explored strategies to achieve MGS via a metabolic shift of the normal myocardium away from glucose in patients undergoing myocardial ¹⁸F-FDG PET imaging. Current strategies vary across institutions; however, consensus documents recommend the implementation of a high fat and low (or no) carbohydrate diet for several meals before imaging in conjunction with a period of fasting to optimize MGS.^2,3 Yet, current preparation protocols still generally suffer from inadequate suppression in ~ 20% of myocardial ¹⁸F-FDG PET studies, resulting in non-diagnostic or even false positive results.^1,4

Serological biomarkers hold promise to identify successful MGS and facilitate diagnostic ¹⁸F-FDG PET studies. Initial research failed to identify relationships between several markers of glucose and fatty acid metabolism (e.g., insulin, glucose, and free fatty acids) and MGS.^1,5 While review of dietary logs is commonly performed and could theoretically substitute as an assessment for adequate myocardial preparation, concerningly, no strong relationship has been established between estimated intake of nutrients, such as carbohydrates and fats, and MGS.⁶ Recent data support a role for beta-hydroxybutyrate (BHB), a ketone produced by the liver via metabolism of fatty acids as a source of energy during insulinopenic states, in evaluating the adequacy of MGS in both healthy and clinical populations.^7,8 While more sophisticated metabolite analysis might hold similar promise, BHB importantly can be evaluated using point-of-care instruments or within one hour in major hospital laboratories, making it a clinically attractive biomarker for evaluating adequate MGS just prior to ¹⁸F-FDG injection.^7-10 However, key questions remain, including the need to define cutoff values of BHB that are indicative of MGS as well as the optimal timing of BHB measurement on the day of ¹⁸F-FDG PET imaging.

The study by Madamanchi, et al. in the current edition of the Journal of Nuclear Cardiology begins to address these knowledge gaps.⁹ The investigators studied 114 consecutive patients referred for myocardial ¹⁸F-FDG PET imaging to assess suspected endocarditis at the University of Michigan. All individuals underwent assessment of BHB and other metabolic biomarkers at three separate time points: approximately four hours before ¹⁸F-FDG injection, approximately ten minutes before ¹⁸F-FDG injection, and approximately 45 min after ¹⁸F-FDG injection. ¹⁸F-FDG PET imaging was assessed for MGS through both quantitative assessment of standardized uptake values and semiquantitative grading. BHB effectively predicted MGS, while several other metabolic biomarkers did not. Since BHB is a more dynamic biomarker than glucose or hemoglobin A1c during short-term dietary modification, this finding may not be surprising. A cutoff value of BHB of ≥ 0.35 mmol/L yielded a sensitivity of 88% and specificity of 61% for MGS, while a threshold of ≥ 0.95 mmol/L offered a sensitivity of 45% and specificity of 100%. Notably, the threshold of 0.35 mmol/L still misclassified 32% of patients who demonstrated MGS on imaging despite BHB levels below this value, and no threshold value was universally effective. BHB measured four hours prior to ¹⁸F-FDG injection performed as well or better than measurements performed closer to imaging, which may be useful during inpatient evaluations. The area under the curve for this timepoint was 0.87, which was comparable to that seen in prior studies using BHB to assess MGS.^7,8

This study leverages several unique advantages by assessing patients with suspected endocarditis. First, this population yields minimal ambiguity as to whether myocardial glucose uptake represents pathology or poor suppression since non-paravalvular uptake would not be expected. Another notable feature of this population was the high rate of inpatients included (87%), a population that is notoriously challenging to prepare for myocardial ¹⁸F-FDG PET imaging. This provided a larger than expected cohort of individuals with inadequate suppression (28.1%) and, therefore, greater power to compare BHB values among those with and without MGS.

Given these findings, the authors proposed a potential clinical protocol by which BHB could be implemented for assessing MGS. In this construct, patients would undergo BHB testing after confirmation of adequate dietary preparation at the time of arrival for imaging. If an individual was found to have a BHB value < 0.35 mmol/L, they would undergo a more critical assessment of their preparation prior to ¹⁸F-FDG injection. If strict adherence were to be confirmed, the authors would consider it reasonable to proceed with ¹⁸F-FDG PET imaging given the relatively high fraction of individuals with MGS despite having a low BHB value. A schematic of such a protocol implementing BHB is shown in Figure 1.

Figure 1.

Potential protocol by which measurement of beta-hydroxybutyrate could be incorporated into the assessment of patients prior to ¹⁸F-fluorodeoxyglucose positron emission tomography (¹⁸F-FDG PET) imaging for the assessment of myocardial inflammation.

The current study also has several limitations. Only patients with suspected endocarditis at a single institution using a specific preparation protocol were included, and many of these subjects were inpatients. Assessing laboratory BHB levels four hours prior to injection is feasible among inpatients, though not generally with outpatients. Therefore, using BHB levels upon arrival to outpatient imaging is more convenient in this population with reasonable testing characteristics found at this time interval. Additionally, there was inadequate statistical power to perform important subgroup analyses of individuals with diabetes mellitus and cardiomyopathy, conditions known to impact myocardial metabolism.¹¹ Accordingly, the current findings may not be broadly applicable to all individuals undergoing ¹⁸F-FDG PET for myocardial inflammation.

Regardless, this study prompts several important future areas of research to further define the role of BHB in the evaluation of MGS. First and foremost, it remains unclear whether unique thresholds of BHB are needed to evaluate MGS in different patient populations and with different preparation strategies. As such, a prospective multi-center study that includes a diverse group of patients with various indications (especially cardiac sarcoidosis) and preparation strategies is needed. Although initial evidence from a separate study supports the effectiveness of BHB in assessing MGS in patients with diabetes mellitus and cardiomyopathy, this prospective study should include an adequate number of individuals with these underlying conditions to confirm that the observed relationships between BHB and MGS persist in these populations.⁸ Additionally, the optimal strategy for integrating measurements of BHB into the assessment of patients prior to ¹⁸F-FDG injection and the interpretation of ¹⁸F-FDG PET imaging for myocardial inflammation requires clarification. This is underscored by the fact that current protocols employing dietary review alone may not yield greater certainty.⁶ Thus, the advantages of ketone measurements beyond current protocols must be defined, since BHB is an imperfect biomarker for identifying patients with adequate MGS. Finally, assessing other ketones should be considered to determine if they outperform BHB in the assessment of MGS.¹⁰

Although further investigation is needed before broad clinical implementation, BHB offers promise as a readily available serological biomarker that provides an additional level of confidence in the assessment of the adequacy of MGS prior to myocardial ¹⁸F-FDG PET inflammation imaging. Such a biomarker might improve the diagnostic yield of myocardial ¹⁸F-FDG PET for inflammation and reduce patient radiation and imaging costs by decreasing the number of non-diagnostic and false positive results.