This study evaluated the diagnostic performance of the HR mode, the UHR mode, and the adjusted for coronary imaging UHR mode of PCD-CT used in pre-TAVI scans for the detection of CAD with QCA as the reference standard. In the per-patient analysis, our findings indicate that both the UHR and adjusted UHR modes show high sensitivity and negative predictive value in this population with a very high calcium score. In the per-vessel analysis, the adjusted UHR mode maintains a high sensitivity and negative predictive value compared to the standard UHR and HR modes.

Our findings align with those of a prior study by Hagar et al., which reported a sensitivity, specificity, and accuracy of 96%, 84%, and 88%, respectively, per patient using the UHR mode for CAD detection, despite including distal segments in their analysis that we excluded in our study [12]. However, it is important to highlight the disparity in scan parameters between Hagar et al.‘s study and our standard and adjusted UHR mode. Their protocol involved a retrospective ECG-gated dual-source helical scan with a collimation of 120 mm × 0.2 mm, and tube voltage settings of 120 kV or 140 kV, along with automated tube current modulation and dose modulation with ECG pulsing between 20% and 80% of the R-R interval. In contrast, our study employed prospective ECG triggering with a tube voltage of 90 kV to minimize radiation exposure. As a result, the adjusted mode in our study achieved a lower radiation dose (CTDIvol of 20.5 mGy [18.3–26.9] and dose length product of 306.5 mGy∙cm [262.5-419.8] in comparison to the threefold higher dose reported by Hagar et al. (CTDIvol of 67.7 mGy ± 19.2 and dose length product of 936 mGy∙cm ± 278).

In our study, the implementation of the adjusted UHR scan mode was associated with a modest increased CTDIvol compared to the standard UHR mode (standard: CTDIvol 18.6 mGy [14.8–22.2], adjusted UHR mode: 20.5 [18.3–26.9]). However, the sensitivity at the per-patient level was similar, at 100%, across both modes. Nevertheless, at the per-vessel level, the sensitivity of the adjusted UHR was higher compared to the standard UHR mode, though the differences in diagnostic accuracy between the scan protocols were not statistically significant, for both the per-patient and per-vessel analyses. It is important to acknowledge that our study may have been underpowered to detect differences between the scan modes due to its small sample size (n = 20 per group).

Given these findings, the adjusted UHR protocol may be favoured for pre-TAVI scans in a clinical setting, as it offers improved sensitivity for CAD detection with only a modest increase in radiation dose, which is acceptable given the relatively older TAVI patient population. Future research with larger sample sizes is necessary to comprehensively compare the three scan protocols and assess whether the adjusted UHR protocol significantly enhances the diagnostic evaluation of CAD in pre-TAVI patients and therefore justifies the additional dose associated with this scan mode.

Furthermore, while the diagnostic performance of the HR mode was lower compared to the UHR and adjusted UHR modes, the SNR and CNR were significantly higher for the HR mode. Despite the higher SNR and CNR in the HR mode, these metrics do not always translate to better visual assessment. This paradox can be explained by the fact that, although human readers prioritise higher contrast over lower noise when assessing images, the overall diagnostic performance also heavily depends on resolution and the ability to detect small details. The UHR and adjusted UHR modes, with their superior resolution, can reveal more subtle diagnostic features that are crucial for accurate assessments, thus making them more effective despite potentially lower SNR and CNR.

A meta-analysis conducted by Gatti et al. demonstrated that implementing routine CCTA in the pre-TAVI workup could save 41% of ICAs if a disease prevalence of 40% is assumed [2]. In our study we had a similar disease prevalence of 40% and CCTA correctly ruled out CAD in 19/36 patients, indicating that 53% of the patients could have been safely deferred from ICA. However, the proportion of patients who could avoid invasive coronary angiography might be underestimated, as a recent meta-analysis by Aarts et al. showed no significant difference in clinical outcomes between patients with concomitant significant CAD who were treated with TAVI with and without preceding PCI at both short- and long-term follow-up [15].

Recently, the integration of spectral information with UHR mode has become available [16]. This development might improve the accuracy of PCD-CT for detecting CAD in pre-TAVI patients [16,17,18]. TAVI patients often have extensive coronary calcifications, leading to blooming artefacts and overestimation of stenosis, which may account for the high false positive rate and relatively low specificity for CCTA [3]. Incorporating spectral information with UHR imaging, particularly using optimal virtual mono-energetic imaging and calcium subtraction algorithms, may address this limitation, potentially improving the specificity of the pre-TAVI CT scan [19, 20].

There are some limitations regarding our study. At our institute, TAVI patients are typically scanned using the HR mode, resulting in a limited sample size of patients scanned on the adjusted and standard UHR mode for this study. Therefore, a larger sample size could make our findings more robust. Additionally, the presence of severe aortic stenosis in our patient cohort precluded the use of beta-blockers or nitro-glycerine. Furthermore, this study does not include a comparison of the diagnostic accuracy of PCD-CT with energy-integrating detector CT. Moreover, in our study, we reduced the tube potential to 90 kV in the adjusted UHR mode to limit radiation dose. However, this reduction comes with a trade-off: the lower tube potential increases the risk of blooming artefacts, particularly in patients with severe calcifications or stents, which could compromise image quality and diagnostic accuracy for detecting CAD. Increasing the tube potential would improve image quality, but it would also lead to a higher radiation dose. This highlights the challenge of finding a balance between dose reduction and image quality in patients with high coronary calcium scores or stents. Lastly, our analysis was limited to assessing the degree of stenosis and did not include plaque characterization.

In conclusion, while both UHR and adjusted UHR modes of PCD-CT show potential for improved sensitivity and negative predictive value for the detection of CAD in patients undergoing pre-TAVI scans, no definitive difference from HR mode was observed.