Prognostic impact of the extent of myocardial perfusion defect assessed by dynamic computed tomography perfusion

Abstract

Dynamic computed tomography perfusion (DynCTP) provides functional information complementary to coronary CT angiography (CCTA), enabling combined anatomical and functional non-invasive assessment of coronary artery disease (CAD). However, standardized cut-off values for grading the extent of perfusion defects and their clinical implications remain poorly defined.

Purpose

To evaluate clinically relevant cut-off values for grading myocardial perfusion defect extent on DynCTP in symptomatic patients with suspected CAD or prior chronic coronary syndrome (CCS).

Methods

Prospective single-center observational study including consecutive unselected patients (April 2021–June 2025) who underwent CCTA and DynCTP when CCTA showed CAD-RADS ≥3 or non-diagnostic coronary segments. Perfusion defect extent was quantified using a 16-segment myocardial model: no defect (0 segments), mild-to-moderate (1–4 segments; 1–25%), and extensive (≥5 segments; >25%). Cardiovascular outcomes (CV-MACE) during follow-up included coronary revascularization (percutaneous or surgical), acute coronary syndrome (ACS), and cardiovascular death.

Results

210 patients were included (mean follow-up 2.2 ± 1.2 years), and we excluded cases with missing perfusion quantification. 203 patients were analyzed and 133 (65.5%) had positive CTP. Mean age was 68 ± 10 years and 82% were male. Hypertension, dyslipidemia, and diabetes were present in 75%, 73%, and 30%, respectively. Half had prior CCS; and pretest CAD probability in those without CCS was 29 ± 13%. Baseline CCTA showed mean CAD-RADS of 3.9 ± 1.1, segment involvement score (SIS) of 8.3 ± 3.1, and segment stenosis score (SSS) of 9 ± 5.6.

There was a stepwise increase across perfusion extent categories in coronary revascularization (4,2%, 26,7%, and 63,9% for 0, 1–4, and ≥5 segments, respectively; p<0.001) and ACS (1,4%, 3,3%, and 11,1%; p=0,026). Two cardiovascular deaths occurred, only in patients with ≥5 segments (Table 1). Event-free survival differed significantly among groups (p<0.01) (Figure 1).

In multivariable Cox models adjusted for clinical risk factors and CCTA anatomical burden (CAD-RADS stenosis and SIS) perfusion extent remained an independent predictor of CV-MACE. Compared with patients with no hypoperfused segments, adjusted hazard ratios were 5.85 (95%CI 1.64–20.85;p=0.006) for 1–4 segments and 24.90 (95% CI 7.37–84.13;p<0.001) for ≥5 segments. Adding perfusion extent to clinical and anatomical model provided an incremental improvement in model fit (likelihood ratio χ²=62,19,p<0,001).

Conclusions

In this cohort, Dynamic CTP perfusion defect extent was an independent and incremental predictor of downstream invasive management and CV-MACE beyond clinical and CCTA anatomical variables. A threshold of ≥5 hypoperfused segments identified a high-risk group with increased likelihood of revascularization and ACS. Given the low number of cardiovascular deaths, prognostic value for hard outcomes should be confirmed.MACE in perfusion defect extension categoriesKaplan-Meier survival analysis