Foley, Thomas

Abstract Background To utilize high spatial resolution reconstructions for cardiac imaging at energy‐integrating detector CT (EID)‐CT with comparable noise to similar reconstructions at photon‐counting detector (PCD)‐CT, methods to control EID‐CT image noise are needed. Supervised convolutional neural networks (CNN) have shown promise for denoising, but a challenge remains to efficiently create high‐quality and unbiased estimates of noise without access to dedicated software or proprietary information, such that natural noise texture is retained in CNN‐denoised CT images. Purpose This study aims to develop and test image‐based noise estimation methods that can be used to train a CNN model, and to evaluate denoising performance and noise texture preservation for EID‐CT coronary CT angiography (cCTA) images reconstructed with high‐resolution kernels. Methods U‐net CNN models were trained for denoising. To supervise training, noise‐only images were estimated directly from high‐resolution kernel (Bv59) reconstructed EID‐CT (HR EID‐CT) patient images using two different methods: subtraction of low‐ and high‐strength iterative reconstruction (IR); subtraction of adjacent image slices with the same IR strength. The noise estimates from these methods contain differing noise texture and anatomical information. Networks were trained and validated separately for three data sets: the training data from each of the two noise‐estimation methods, and a 50%–50% partition of training data between the two methods. The trained models were applied to two sets of testing data: CT images of a uniform water phantom to measure noise power spectra (NPS), and an independent cohort of seven patient cCTA HR EID‐CT exams. The denoised patient images were compared to standard resolution EID‐CT reconstructions (Bv40). As a low‐noise reference, patient images acquired on the same day with a PCD‐CT and reconstructed using a similar kernel as HR EID‐CT were used for comparison. Results Models trained with each noise‐image estimation method denoised the HR EID‐CT images by 74%–79% to achieve a comparable noise magnitude to the HR PCD‐CT images. The peak, average, and 10% peak frequencies of the NPS of the input images (6.08, 6.24, and 12.0 cm −1 ) were better approximated by the model trained on adjacent slice subtraction (6.56, 5.87, and 11.5 cm −1 ) than by the model trained on subtraction of low‐ and high‐IR images (4.64, 5.44, and 11.3 cm −1 ). In cCTA images, the IR subtraction model images retained anatomic structures from input images but resulted in undesirable salt‐and‐pepper noise texture and CT number bias. The model trained on adjacent slice subtraction images had more natural texture and no significant bias, but the model sometimes removed small anatomic structures. The model trained on the mixed training dataset preserved both noise texture and anatomy from the model inputs and enabled visualization of small structures seen in PCD‐CT images that were previously unresolved by EID‐CT. Conclusions The noise texture and anatomical accuracy in CT images denoised with an image‐based supervised CNN are greatly influenced by the characteristics and partitioning of training data. With higher‐resolution reconstructions and noise texture‐preserving deep learning denoising, the quality of cCTA images from EID‐CT can be enhanced to enable resolvability of subtle anatomy similar to PCD‐CT.

Abstract Aims Aortic valve calcification (AVC) is prognostic in patients with aortic stenosis (AS). We assessed the AVC prognostic value in non-severe AS patients. Methods and results We conducted a retrospective study of 395 patients with non-severe AS, LVEF ≥ 50%. The Agatston method was used for CT AVC assessment. The log-rank test determined the best AVC cut-offs for survival under medical surveillance: 1185 arbitrary unit (AU) in men and 850 AU in women, lower than the established cut-offs for severe AS (2064 AU in men and 1274 AU in women). Patients were divided into 3 AVC groups based on these cut-offs: low (<1185 AU in men and <850 AU in women), sub-severe (1185–2064 AU in men and 850–1274 AU in women), and severe (>2064 AU in men and >1274 AU in women). Of 395 patients (mean age 73 ± 12 years, 60.5% men, aortic valve area 1.23 ± 0.30 cm2, mean pressure gradient 28 ± 8 mmHg), 218 underwent aortic valve intervention (AVI) and 158 deaths occurred during follow-up, 82 before AVI. Median survival time under medical surveillance was 2.1 (0.7–4.9) years. Compared with the low AVC group, both sub-severe and severe AVC groups had higher risk for all-cause death under medical surveillance after comprehensive adjustment including echocardiographic AS severity and coronary artery calcium score (all P ≤ 0.006); while mortality risk was similar between sub-severe and severe AVC groups (all P ≥ 0.2). This mortality risk pattern persisted in the overall survival analysis after adjustment for AVI. AVI was protective of all-cause death in the sub-severe and severe AVC (all P ≤ 0.01), but not in the low AVC groups. Conclusion Sub-severe AVC is a robust risk stratification parameter in patients with non-severe AS and may inform AVI timing.

Abstract OBJECTIVES In patients with aortic root aneurysm, the aortic basal ring is frequently dilated. It has been speculated that the muscular part of the basal ring dilates most. The purpose of this study was to analyse the segmental dilatation of the basal ring, comparing normal and dilated roots in patients with tricuspid aortic valves. METHODS Retrospective analysis of computed tomography studies in patients with normal and dilated aortic roots was performed. Lengths of segments of the basal ring corresponding to each of the 3 sinuses, and to the muscular and fibrous parts were measured. Fractions of these segments relative to the total basal ring perimeter were calculated. RESULTS We analysed 152 normal and 126 dilated aortic roots and 86 propensity-matched pairs. Basal ring dilatation was present in all segments of dilated aortic roots with subtle differences between the segments corresponding to the 3 sinuses. The muscular part of the basal ring dilated proportionately to its fibrous part, with no difference in fractions of measured muscular part in normal and dilated roots [42.2% (interquartile range 4.3%) vs 42.1% (interquartile range 6.3%)]. CONCLUSIONS Basal ring dilatation was present in all segments corresponding to the 3 sinuses in dilated aortic roots. Both muscular and fibrous parts dilated equally, supporting the need to stabilize the entire basal ring when performing aortic valve repair surgery.