Purpose:
Coronary assessment in pre-procedural CT angiography for transcatheter aortic valve implantation (TAVI) planning has been shown to be, in principle, feasible, but often leads to inconclusive results due to high calcium load. Novel photon-counting detector (PCD)-CT technology offers substantial benefits over conventional energy-integrating detector CT in cardiovascular imaging. This includes, in particular, an improved spatial resolution. This early real-world study investigates the diagnostic performance of PCD-CT in patients undergoing TAVI workup for detection of significant coronary artery disease (CAD) in a non-selected patient cohort.

Methods:
Retrospective analysis of 32 consecutive patients with severe aortic valve stenosis who underwent contrast-enhanced, retrospectively ECG-gated CT angiography acquired on a PCD-CT system and invasive coronary angiography (ICA) as routine TAVI workup. Patients did not receive any nitroglycerin or additional beta-blockade. Diagnostic accuracy of CT for detecting significant CAD was compared with that of ICA applying a simplified 7-segment model including the proximal and mid coronary artery segments (LM, proximal and mid LAD, RCX and RCA). In both tests, significant coronary artery stenosis was defined as ≥50% luminal diameter narrowing in segments ≥2 mm.

Results:
Mean age was 79±8 years (59% males; mean BMI 27±5 kg/m2; 34% atrial fibrillation, mean heart rate 72±12 bpm). The mean number of days between CT and ICA was 4 days (range 0-28 days). Mean dose-length product was 919±410 mGy*cm, mean contrast volume was 60 ml. Image quality was rated as excellent in 40%, good in 50% and poor in 7% of the segments. Nine out of 224 segments (4%) were uninterpretable due to motion artefacts (6), severe calcification (2) and beam hardening artefacts in a stented segment (1). Per-patient and per-segment analysis is presented in table 1. In summary, CT correctly identified significant CAD in 59% of the patients. Due to severe calcification, 6% of the patients were falsely diagnosed with obstructive CAD. In one patient, CT underestimated a long-distance severe stenosis in the proximal LAD which was confirmed by invasive FFR. In 9% of the patients, no clear obstructive CAD was found but they had at least one non-evaluable coronary segment. When considering uninterpretable segments as a positive result, obstructive CAD was present in 69% of the patients. In one patient with prior stenting, CT correctly identified an in-stent re-stenosis in the proximal LAD and excluded relevant re-stenoses in the LM and proximal RCX. Though, due to motion and beam hardening effects, the proximal stented RCA segment was not interpretable.
In this real-world cohort without additional prior medical heart rate control, sensitivity of CT angiography for detection of obstructive CAD in the proximal and mid coronary segments was 95% and specificity was 84%.

Conclusion:
CT angiography in patients undergoing TAVI workup using novel PCD technology is feasible for CAD assessment in the proximal and mid coronary artery segments with a good sensitivity and specificity. This early, small study included a real-world, non-selected TAVI cohort with no additional heart rate control. The majority of segments with stents was interpretable but the accuracy of stent evaluation with PCD-CT has to be evaluated in a larger patient group.