Assessment of pericoronary adipose tissue attenuation

We read with interest the recent study by Sagris et al.¹ and congratulate the authors for the first systematic review and meta-analysis examining pericoronary adipose tissue (PCAT) and associations with major adverse cardiac events (MACE). Nonetheless, this field is perhaps more nuanced and immature than might be apparent from this study.

1. Variability of measurement techniques

   PCAT attenuation is obtained from image post-processing. While several studies have used the fat attenuation index (FAI), many techniques are available and there has been no validation of between-vendor differences. FAI is derived from an artificial intelligence algorithm that quantifies weighted 1 mm concentric layers of PCAT, in theory adjusting for CT scanners, scan parameters, and reconstruction algorithms. As studies included in this paper include those done with FAI and mean PCAT attenuation, there would be potential for significant differences in FAI and PCAT attenuation based on software, and measurement technique. We also note the omission of Goeller et al.² (which would fall within the search strategy dates), that reported increased MACE, and also used different software (Autoplaque) and mean PCAT attenuation.

2. Variability in the PCAT threshold

   Variable binary PCAT thresholds were also used, and as such, the aggregation of values from different measurement methods may provide non-generalizable pooled estimates. While a PCAT of ≥−70.1 HU has been purported as the ideal value for prognostication, there is need for derivation of a standardized threshold obtained from all data to facilitate clinical interpretation.

3. Variability of the site of PCAT measurement

   The included studies have predominantly used right coronary artery (RCA)-PCAT measurement with the presumption this is representative of pan-coronary inflammatory status. Emerging studies have demonstrated differences in PCAT at different sites along the vessel, between arteries and across grades of plaque severity.³ Additionally, sites closer to the aorta may be affected by contrast blooming; hence, a defined distance from the ostium is often used, but some of the included studies have not accounted for this. This is important given the analysis for stable vs. unstable plaques has included mainly studies that used RCA-PCAT or limited to proximal coronary PCAT, rather than lesion-specific values. The assumption of commonality of PCAT attenuation ignoring their location again clouds the accuracy of these results.

4. Cohort heterogeneity

   PCAT is variable amongst differing clinical cohorts.⁴ The authors have included a mix of stable and unstable CAD cohorts and it may not be appropriate to assume the inflammatory status (which is a dynamic process) can be distilled to a pooled value. Furthermore, the authors have included a non-contrast CT study.⁵ Whilst PCAT measurement methods remain yet to be standardized across vendor and scanner types, we suggest that further reviews may need to standardize for contrast-enhanced studies only.

The work from the authors is commendable, but there remain important limitations and caution must be taken with the current results. Subsequent systematic reviews might need to provide more granularity regarding potential confounding features.

Data availability

No new data were generated or analysed in support of this research.

References

Author notes