(2026) Hyska | Time-resolved PCCT estimates post-TAVI pressure gradients

Daniel Stromer

In 71 post-TAVI PCCT examinations from 63 patients, time-resolved cardiac PCCT enabled calculation of prosthetic-valve hemodynamics from left-ventricular volumetry and direct valve-orifice planimetry. CT-derived mean and maximum transvalvular gradients correlated strongly with echocardiography (ρ = 0.85 for both), while corrected valve area showed a correlation of ρ = 0.73.

The most clinically useful parameter was the mean transvalvular gradient (Pmean). Using echocardiographic Pmean ≥20 mmHg as the reference for potential valve dysfunction, PCCT achieved AUC 0.94, sensitivity 88.9%, specificity 88.6%, and overall accuracy 88.7%. The independently derived optimal PCCT cutoff was almost identical at 20.4 mmHg.

PCCT-derived Pmean showed much better agreement than Pmax: mean Pmean bias was only 1.6 mmHg, whereas Pmax was systematically higher and showed wider limits of agreement. This suggests that a post-TAVI PCCT already obtained for structural assessment could potentially provide opportunistic functional valve assessment, although the retrospective single-center workflow still requires external validation.

Access publication

Figure 1. Schematic overview of the CT-based workflow for post-TAVI hemodynamic assessment. Time-resolved cardiac CT was used to derive prosthetic valve hemodynamics after TAVI by combining multiphase left-ventricular volumetry, valve orifice area planimetry, quality control for excluding cases with severe mitral regurgitation and hemodynamic modeling of transvalvular flow.

Protocol

Detail	Value
Scanner	NAEOTOM Alpha
Scan area	Cardiac / post-TAVI prosthetic valve
Acquisition mode	Quantum HD
Scan mode	Retrospectively ECG-gated spiral
Tube voltage [kVp]	120
IQ level	60
Dose modulation	Automatic tube-current modulation; ECG-based modulation at 30–70% RR
CTDIvol [mGy]	44.1 ± 13.6
DLP [mGy·cm]	670.0 [476.0–849.5]
Effective dose [mSv]	17.4 [12.4–22.1]
Rotation time [ms]	250
Detector slice collimation	120 × 0.2 mm
Temporal reconstruction	Full cardiac cycle at 10% RR increments
Pressure-curve temporal sampling	Resampled to 66 ms
Volumetry reconstruction	Bv36; 0.8/0.6-mm thickness/increment; 512 × 512; QIR 3
Valve-orifice reconstruction	BMI-adapted Bv60/Bv64; 0.2/0.1-mm thickness/increment; 512 × 512; QIR 4
Valve phase	Automatically selected best systolic phase
Contrast	Iodinated contrast, 350–370 mg I/mL
Contrast volume	75 mL
Injection rate	3 mL/s
Bolus triggering	Ascending aorta
Post-processing	Multiphase LV volumetry + prosthetic-valve orifice planimetry
Hemodynamic model	Flow/velocity calculation followed by simplified Bernoulli equation
Severe mitral regurgitation exclusion	CT-derived mitral regurgitation fraction >40%

Key Results

Parameter	Main finding
Study population	71 scans / 63 patients
Pmean correlation with TTE	ρ = 0.85, p < 0.001
Pmean agreement	CT 20.2 [13.4–27.6] vs TTE 20.0 [16.0–24.5] mmHg, p = 0.080; bias 1.6 mmHg
Valve-dysfunction detection	AUC 0.94, sensitivity 88.9%, specificity 88.6%, accuracy 88.7%
Optimal PCCT Pmean cutoff	20.4 mmHg
Longitudinal follow-up	Pmean change correlated with TTE (ρ = 0.80); all 8/8 paired follow-ups concordantly classified around the 20-mmHg threshold

Source:
Hyska S, Osoria-Velasquez J, Hagar MT, et al. Photon-counting detector CT for prediction of valve dysfunction after transcatheter aortic valve implantation: Comparison with echocardiography. Journal of Cardiovascular Computed Tomography. Articles in Press. Published September 18, 2026.