Page 71 - ARNM-2-3

Advances in Radiotherapy
& Nuclear Medicine Association between PET-derived flows and angiography

corrected for magnification distortion and pincushion. We then ascertained the level of global MFR that would
Two trained technicians evaluated all QCA measurements optimally distinguish between the presence and absence
to reduce interobserver variability. A significant stenosis of significant (≥50%) stenosis in each vessel (binary
was defined as ≥50% of luminal obstruction, given the variable). A global MFR <1.7 had the optimal sensitivity
overall low number of events in the population. (66%) and specificity (56%) for significant LAD stenosis,
a global MFR <1.7 had the optimal sensitivity (57%) and
2.3. Statistical analysis specificity (55%) for significant RCA stenosis, and a global
Categorical variables were compared with Chi-square tests MFR <1.6 had the optimal sensitivity (65%) and specificity
or Fisher’s exact tests as appropriate. Heterogeneity was (65%) for significant LCX stenosis (Table S1, bottom). We
evaluated through the coefficient of variance (standard performed survival analysis based on whether patients had
deviation [SD]/mean). Continuous variables were MFRs above or below the coronary territory-specific MFR
analyzed using pairwise comparisons with paired t-tests cutoffs.
or Wilcoxon signed-rank tests as appropriate. A two-
tailed P < 0.05 was considered statistically significant. The 3. Results
association of impaired regional and global flow reserve 3.1. Patient characteristics
(MFR-cutoffs defined below) and significant stenosis
82
(≥50% on QCA) was first assessed as binary variables. A total of 304 patients underwent dynamic Rb PET MPI
The sensitivity, specificity, positive and negative predictive and were clinically referred to coronary angiography
value, accuracy, and positive and negative likelihood ratio by their treating clinicians during the study period.
were calculated for the ability of MFR impairment to A total of 115 patients were excluded from the analysis
predict ≥50% stenosis on QCA. The association between due to prior CABG precluding accurate assessment
per vessel degree of stenosis and flow reserve was assessed of QCA and MFR, poor angiographic films and/or
using the Spearman correlation of continuous variables. uninterpretable QCA, and/or non-regadenoson stress
The vital status of all patients was assessed through (Figure 1). Among the 189 patients that were included,
chart review in September 2022 (3 years after the last 43% (82/189) were women, the median age was 66 years
PET was done). The relationship between regional MFR (interquartile range [IQR]: 56, 74), and the mean BMI
2
impairment (defined below) and all-cause mortality was was 34.5 kg/m (SD: 8.8). 51% (97/189) of patients had
evaluated in unadjusted and adjusted survival analyses. a perfusion defect of ³10% of the left ventricle (LV).
Kaplan–Meier curves were compared with log-rank tests. The baseline characteristics of the included group
Cox proportional hazards regression was performed with are summarized in Table 1. There was no statistical
adjustment for age, sex, body mass index (BMI), diabetes difference in the baseline characteristics in the patients
history, CAD history, heart failure (HF) history, smoking, who did or did not have a baseline perfusion defect of
and need for subsequent revascularization. Subsequent ≥10% of the LV. The distribution of the percentage of
revascularization was coded as a binary variable based on perfusion defects is shown in Figure S1. Median regional
whether the patient underwent subsequent percutaneous MFR was reduced in all coronary territories: 1.75 (IQR:
coronary intervention (PCI) or CABG. Statistical 1.40, 2.19) in the LAD territory, 1.76 (IQR: 1.36, 2.27)
analysis was performed in R (R Foundation for Statistical in the RCA territory, and 1.71 (IQR: 1.30, 2.16) in the
Computing, Vienna, Austria). LCX territory. The median global MFR was 1.73 (IQR:
1.36, 2.20).
2.4. Determination of regional coronary territory
MFR cutoff values 3.2. Quantitative coronary analysis
We first ascertained the level of regional coronary territory Table 2 shows the characteristics of the vessels through
MFR that would optimally distinguish between the QCA. For the LAD, the median obstruction diameter was
presence and absence of significant (≥50%) stenosis in the 1.6 mm [1.21, 1.93] and the median percent stenosis was
corresponding vessel (binary variables): left descending 29.81% [24.61, 45.00]; for the RCA, the median obstruction
artery (LAD) MFR <1.7 had the optimal sensitivity diameter was 1.73 mm [1.27, 2.30] and median percent
(69%) and specificity (60%) for significant LAD stenosis, stenosis was 28.57% [21.53, 45.09]; and for the LCX, the
right coronary artery (RCA) MFR <1.4 had the optimal median obstruction diameter was 1.73 mm [1.30, 2.15]
sensitivity (46%) and specificity (80%) for significant RCA and median percent stenosis was 27.51% [23.02, 36.67].
stenosis, and left circumflex artery (LCX) MFR <1.5 had 17% (32/187) of patients had LAD stenosis >50%; 20%
the optimal sensitivity (65%) and specificity (66%) for (37/183) of patients had RCA stenosis >20%; 9% (17/186)
significant LCX stenosis (Table S1, top). of patients had LCX stenosis >50%.

Volume 2 Issue 3 (2024) 3 doi: 10.36922/arnm.3786