Aortic Valve Calcium Score as Measured by Native Vs. Contrast Enhanced Computer Tomography and the Implications for the Diagnosis of Severe Aortic Stenosis in TAVR Patients With Low Gradient Aortic Stenosis.


 Purposewe compared between patients with low gradient (LG) and high gradient (HG) severe aortic stenosis (AS) as regard the burden of aortic valve calcium (AVC) using different methodologies. Moreover, we evaluated the accuracy of published thresholds for the diagnosis of severe AS in both groups. Methodswe measured the calcium volume and score using Agatston methodology in non-contrast (n-c) CT and with modified and fixed 850 Hounsfield unit (HU) thresholds in contrast enhanced (ce) CT. ResultsThe medians (IQR) of Agatston score, score with 850 HU and modified thresholds were 1288 AU (750-1815), 101 (65-256), 701 (239-1632), respectively. The calcium volume in ceCT using fixed 850 HU thresholds is significantly lower than the assessed volume in ncCT or in ceCT using modifiable threshold. LG patients were more obese; BMI 31.2 (29.1-35.1) vs 27.6 (26-31) and presented more with coronary artery disease (71.4% vs 40%). AF was documented in 42% in LG-patients vs 30% in HG patients. LVEF was severely depressed (less than 30%) in 28.6% in LG-patients. LG patients were more symptomatic (NYHA ≥ III in 71.4% patients vs 42%).The LG patients had smaller anatomy: annulus diameter 23.5mm (21.5-27) vs 25mm (23-25.5), LVOT diameter 23mm (20-20) vs 25mm (23-26.7mm). The annulus geometry was more eccentric; eccentric index 0.23 (0.19-0.27) vs 0.11 (0.1-0.2). Agatston score and calcium volume were lower in patients with LG; 1641AU (1292-1990) vs 928AU (572-1284) and 1537mm³ (644-1860) vs 286mm³ (160-700), respectively. Only 20% of patients with LG had Agatston score less than the previously supposed AVC score threshold for the diagnosis of severe AS (>2000AU in men and >1200 in women). The elimination of ncCT from the protocol reduced significantly the radiation dose by 400.3 ± 140 mGy*cm and 2.4 ± 2.8mSv.ConclusionThe diagnosis of severe LGAS should not depend on a single parameter as calcium score. The measurement of calcium score in contrast CT underestimate the calcium load significantly.


Introduction
Calci c aortic stenosis (AS) is often, although not solely, an age-related condition in which scarring and degeneration of the aortic valve (AV) promotes deposition of calcium within it (Pawade, Newby et al. 2015). Computer tomography (CT) is a well-established method for the quanti cation of aortic valve calcium (AVC). Most prior studies have used non-contrast (nc) CT to assess AVC as part of the evaluation of AS severity. The cut-off values for this purpose differed between males and females (≥ Page 3/13 2000 or ≥ 3000 for men and ≥ 1200 or ≥ 1600 for women (Clavel, Pibarot et al. 2014;Zamorano, Gonçalves et al. 2016). The CT protocol used in these studies was similar to that used by Agatston for calcium scoring of coronaries (Agatston, Janowitz et al. 1990). AVC not only can aid in determining the severity of AS, but also can predict prognosis in AS patients. (Clavel, Pibarot et al. 2014). These cut off values are only validated in patients with high gradient AS (HG AS) (Clavel, Messika-Zeitoun et al. 2013). Discordant AS would encompass classical low ow low gradient (cLFLG), paradoxical LFLG, normal ow LG, and also patients with mean pressure gradient (PG) > 40 mm Hg and aortic valve area index (AVAi) > 0.6 cm². Clavel and colleagues have reported that only 50% of patients with discordant AS had higher values than the cut off value. Patients with discordant AS also had a lower AVC burden (Veulemans, Piayda et al. 2021), and this group of patients accounts for approximately30% of TAVR patients. The PG across the AV is determined not only by the AVarea, but also independently by ow in the LVOT, AVC-load, and systemic arterial compliance (Clavel, Messika-Zeitoun et al. 2013). Veulemans et al. showed that CT can diffrentiate the severity of LG AS only in men (Veulemans, Piayda et al. 2021). This study evaluated the AVC in contrast enhanced CT (ceCT) with xed threshld of 600 Houns eld (HU). This xed threshold could underestimate the calcium (Bittner, Arnold et al. 2016), especially in patients with discordant AS and low AVC burden. Even a multicenter trial only included few patients with discordant AS (161 from nearly thousand patients) (Pawade, Clavel et al. 2018). The evaluation of AVC using ceCT in patients with LG AS is therefore quite limited. Thus, we aimed to evaluate the AVC in both native and ceCT in patients with HGAS and in patients with LGAS.

Study population:
Multi-slice CT (MSCT) images from all patients who underwent TAVR in 2020 at our institution (Zentralklinik Bad Berka) were reviewed. The CT data was available and interpretable in 311out of 340 patients. We excluded patients with valve-in-valve TAVR.

Pre-procedure cardiac CT angiography
For all patients, we analysed the MSCT images, which were performed as standard-of-care pre-TAVR .
Patients were evaluated using a Siemens Somatom De nition Edge scanner (Siemens Medical Solutions) using collimation of 0.6 mm at a xed pitch of 0.2 with an injection of 70 ml of iopamidol (Ultravist-370; Bayer Vital Pharma). A dedicated protocol was formulated, with kV and tube current modi ed according to the patient's size. Image acquisition for the heart was performed with retrospective ECG gating. CT Digital Imaging and Communications in Medicine (DICOM) data were analysed using Siemens syngo software, Syngo Via, for TAVR Planning.

Measurement of calcium volume
The calcium score and volume of the aortic valve and each cusp were evaluated by the specialist (ME) using three different methodologies: 1. in ncCT imaging using a threshold of 130 Houns eld (HU), 2. in ceCT scans using a modi able threshold and a xed threshold of 850 and 600. The modi cation of Houns eld was used; thus, 100 HU were added to the luminal attenuation HU (Bettinger, Khalique et al. 2017), Figure 1. In patients with LG AS the cardiac output was measured to assure that stroke volume index is ≤ 35ml/m 2 .

Assessment of radiation dose
Radiation exposure was measured according to the methods previously described by Shnayien and colleagues. The dose-length product (DLP) was obtained from an automatically generated protocolthat was based upon the CT dose index (CTDI)and was measured in mGy*cm. The effective dose (E) was measured in mSv and was derived from the DLP as suggested by the European Guidelines on Quality Criteria for Computed Tomography. Thus, we used a conversion coe cient (k) of 0.017 and the following formula: E = k × DLP. The size-speci c dose estimates (SSDE) is given in mGy, and was determined by .. . Eby multiplying conversion coe cients as a function of the sum of the lateral and anteroposterior dimensions with CTDI. (Shnayien, Bressem et al. 2020).

Statistical analysis
Continuous variables were tested for normality of distribution by using the Shapiro-Wilk test. Normally distributed variables were expressed as mean ± standard deviation. For non-normally distributed variables, the median and inter-quartile range (IQR) were calculated and tested for statistical significance with the Mann-Whitney U test. Categorical variables were compared by chi-square statistics. Statistical analyses were performed with SPSS (version 24.0; IBM Corporation, Armonk, NY). A two sided P < .05 was considered statistically signi cant.
LG patients had smaller dimensions in terms of their cardiac anatomy: annulus diameter 23.5 mm (21.5-27 mm) vs 25 mm (23-25.5 mm), LVOT diameter 23 mm (20-24 mm) vs 25 mm (23-26.7 mm). The annulus geometry of the LGAS group was more eccentric than in patients with HG AS, with an eccentricity index of 0.23 (0.19-0.27) vs 0.11 (0.1-0.2). The LVOT diameter to annulus diameter of the LGAS patients was also more conical in shape. Calcium volume and Agatston score was both lower in the LG patients; 1641 AU (1292-1990 AU) vs 928 AU (572-1284 AU) and 1537 mm³ (644-1860 mm³) vs 286 mm³ (160-700 mm³), respectively. Only 20% of patients with LG had an Agatston score higher than the published AVC score threshold for the diagnosis of severe AS.

Discussion
We report here that only 20% of severe AS patients with LG had a higher AVC volume when evaluated by MSCT than the published cutoff recommended for the diagnosis of severe AS. We also note that the mean AVC burden in this study was lower than in previously reported studies (Bittner, Arnold et al. 2016;Pawade, Clavel et al. 2018;Veulemans, Piayda et al. 2021). We included more patients with LGAS; however, even in patients with high gradient AS the calcium load was lower than in prior studies. Factors known to affect AVC burden include hyperlipidemia, diabetes, chronic kidney disease, male sex, and the presence of a LG. We included high risk patients with dyslipidemia (58%), diabetes (56.3%), and chronic kidney disease (17.6%). Thus, the risk pro le in ourstudy is comparable to that of prior studies, such as that of of Aggarwal et al. In that study, the patients included had hyperlipidemia (67%), coronary artery disease (42%), and diabetes mellitus (24%), and the reported predictors of an increased AVC burden age, male sex, higher velocity, smaller aortic valve area, and smaller LVOT. Likewise, the mean aortic annulus diameter (23.5 mm) and mean aortic annulus area (485.5mm²) in our study were comparable to what Bittner, et al reported (24.2mm and 462mm 2 , respectively)..The AVC loadin our patients was similar to the one in female patients with LG and in patients with AVA more than 1cm in the study of Veulemans et al, 1007 (521/1547). We found that the distribution of calci cation was uneven, being more prominent in the NCC and the RCC and less so in theLCC. This is consistent with Veulemans, et al who found that in AS of all severities and and etiologies AS entities, the NCC was the most calci ed. Cheng et al also reported that the NCC had the most calci cation and the LCC had the least calci cation. (Cheng, Chang et al. 2017). In previous CT studies of TAVR, the incidence of LG AS was between 33-56% (Pawade, Clavel et al. 2018;Veulemans, Piayda et al. 2021), which is consistent with the 41.2% incidence we found. This high incidence of LG AS in TAVR patients highlights the importance of studying this entity. We demonstrate here that previously published CT-AVC thresholds are not applicable in LGAS. In keeping with our results, Pawade et al found, in a lmulticentr study with 918 patients, that CT-AVC thresholds (women 1377 Agatston units, men 2062 Agatston units) were accurate only in patients with HG AS but not in patients with LG AS.. Furthermore, Veulemans and colleagues found that the AVC load thresholds were only useful for differentiating between moderate and severe AS in both males and females when the AS was HG. In general, the assessment of calcium volume using the 850 HU threshold underestimates the volume of calcium. The accurate measurement of AVC is valuable as certain calcium distribution patterns may increase the risk for the need of permanent pacemaker implantation. (REF: Fujita, B 2016 Eur Heart J Cardiovasc Imaging) Thus, this Limitations: Some of the limitations of this study are that it is a single-center study and that its design is retrospective and observational.

Conclusion
The published CT AVC threshold for the diagnosis of severe AS I suboptimal for LG AS. We need to standardize the methodology used of calcium measurement in CT before TAVR in order to determine an appropriate for the threshold for AS in patients with LG. LG: low gradient. Con icts of interest/Competing interests: the authors report no nancial relationships or con icts of interest regarding the content herein.

Abbreviations
Availability of data and material: all data are available on request at the Department of Cardiology, Zentralklinik Bad Berka, Germany.
Code availability: not available.
Authors' contributions: MG collected the data and did the provisional drafting of the manuscript. MG, EC and TO did the statistical analysis and the interpretation of data and helped in the drafting. PL, TK and HL revised the manuscript critically and give the nal approval. All authors have read and approved the manuscript.
Ethics approval and consent to participate : our analysis looked retrospectively at outcomes for a cohort of patients treated as part of routine care and is in no way an add-on for purposes of research. This was done internally as part of an audit/evaluation, to improve our quality of care.
Consent for publication Not applicable, as this study was conducted retrospectively and was done internally as part of an audit/evaluation, to improve our quality of care. This is in line with the European regulations; https://ec.europa.eu/research/participants/data/ref/fp7/89807/informed-consent_en.pdf.    calcium score and volume measument use different HU threshold leads to change the score and volume signi cantly (to the right with 850 HU threshold and to the left with modi able threshold).