Age and sex specific coronary plaque characteristics among South Asians: Insights from the DILWALE CTA study

Abstract

Background

South Asians (SA) face disproportionately high rates of premature coronary artery disease (CAD), often underestimated by traditional risk calculators. We aim to characterize CT angiography (CTA)-derived plaque composition, plaque burden, and association with risk factors among SA in the DILWALE registry.

Methods

Clinically indicated coronary CTAs from 341 patients in the Baylor Scott and White DILWALE registry were analyzed using Artificial intelligence enabled quantitative coronary plaque analysis (AIQCPA) to characterize age- and sex-specific plaque burden.

Results

Of 341 patients, 63% (n = 215) exhibited any plaque by AIQCPA. Non-calcified plaque (NCP) was the predominant plaque subtype across all age groups, but calcified plaque volumes increased with age. The median PAV was 3.5% (IQR 0–18.5%). Age, male sex, and statin use were significant predictors of the presence of any plaque, while age and male sex predicted the presence of low attenuation plaque ≥ 10 mm³.

Conclusion

This study represents one of the largest CTA based cohorts evaluating plaque characteristics in SA in the United States. Our findings highlight the value of AIQCPA CTA in revealing subclinical plaque, particularly non-calcified plaque in SA. Future studies should identify optimal imaging strategies for SA along with outcome-based validation of plaque characteristics.

1. Introduction

South Asian (SA) individuals, originating from India, Pakistan, Bangladesh, Sri Lanka, Nepal and Myanmar, face a significantly elevated risk of coronary artery disease (CAD) [1]. This increased susceptibility is not fully explained by traditional risk factors alone, and consequently, existing risk assessment tools often underestimate risk in SA adults [[2], [3], [4]]. Analysis of 457,473 adults from UK Biobank showed that SA individuals experienced >2-fold higher risk of atherosclerotic cardiovascular (CV) disease (ASCVD) events compared to European individuals despite similar 10-year risk predicted by Pooled Cohort equation and QRISK3 equations [4]. Recognizing this, the 2018 American College of Cardiology (ACC)/American Heart Association (AHA) guideline for the primary prevention of CVD designated SA ethnicity as a “risk-enhancing factor,” particularly when considering the initiation of statin therapy in individuals at borderline risk [5].

Coronary imaging has become an increasingly important tool in personalizing cardiovascular risk assessment, aiming to level the playing field by uncovering high-risk individuals who may be overlooked by traditional risk calculators [6]. Computed Tomography (CT) based coronary artery calcium (CAC) score has traditionally been the preferred modality to detect subclinical CAD and has been shown to improve CV risk prediction [7,8]. Our group recently published the largest available dataset on sex specific distribution of CAC scores in SA living in the US [8]. We demonstrated that SA individuals showed a significantly higher prevalence of CAC > 0 when compared with individuals from Chinese, African American and Hispanic groups from the Multi-ethnic Study of Atherosclerosis (MESA). However, prior studies have demonstrated limitations of CAC scoring in SA, showing a higher burden of non-calcified (NCP) and low attenuation plaque (LAP) when compared with Caucasians [9]. CTA (CT angiography) has emerged as a valuable tool in this setting, assessing non calcified and vulnerable plaque burden, thereby providing additional prognostic information beyond CAC scoring [10]. In 5007 outpatients with suspected CAD, coronary CTA plaque metrics provided additional prognostic value beyond CAC, increasing the c-statistic for major adverse cardiovascular events (MACE) from 0.71 with clinical risk factors to 0.82 with CAC and 0.93 with CTA [11].

The DIL Wellness and Arterial health Longitudinal Evaluation (DILWALE) registry includes one of the largest cohorts of SA individuals within the United States. In this analysis, we aimed to characterize CTA-derived plaque composition, plaque burden, and association with risk factors among SA in the DILWALE registry.

2. Methods

2.1. Study population

The DILWALE study is an electronic health record (EHR)-based retrospective registry of adult SA patients within the Baylor Scott & White Health (BSWH) system, the largest nonprofit healthcare system in Texas comprising of 51 hospitals. The details of this registry are described elsewhere [2]. The registry included 31,781 SA adults (age ≥18 years), identified using a validated name based algorithm, with at least one clinical encounter within the BSWH system. The analysis period spanned from August 2008 to December 2023.

Coronary CTAs and risk factors documented in the registry were obtained for clinical indications (i.e., ambulatory patients referred for CTA by a clinician). All patients undergoing CTA also underwent CAC scoring using standard iterative threshold of 130 Hounsfield units (HU) at 120 kVp, followed by quantitative plaque analysis as described below. Traditional risk factors were collected from the EHR using measurements taken nearest to, but within 1 year of, the CTA date.

2.2. Coronary CTA protocol

Three CT platforms were utilized across the study cohort: the Siemens SOMATOM Definition Flash dual-source CT scanner, the Siemens SOMATOM FORCE dual-source CT scanner, and the GE Revolution Apex CT scanner. Despite the use of multiple scanners, acquisition parameters and reconstruction slice thickness were comparable across all platforms. All studies were reconstructed at a 0.6 mm slice thickness.

To optimize image quality, heart rate control with a target <65 bpm was achieved using oral or intravenous beta-blockers. Sublingual nitroglycerin was administered prior to image acquisition to ensure coronary vasodilation. Contrast enhancement was achieved using Omnipaque 350 (iohexol), administered at an injection rate of 4–6 ml/s adjusted based on patient body habitus. Image acquisition was triggered automatically using a region of interest (ROI) placed in the descending aorta. Radiation dose was minimized in accordance with As Low As Reasonably Possible (ALARA) principles, with dose modulation applied in retrospective acquisition protocols.

Quantification of Plaque and Stenosis: All CTAs were analyzed using an artificial intelligence (AI) - enabled quantitative coronary plaque analysis (AIQCPA) software (Heartflow. Inc, Mountain View, CA, first-generation algorithm, which received FDA 510 (k) clearance in October 2022) [12]. The software generates a 3D model of arterial lumen using a proprietary algorithm. A deep learning algorithm is applied to quantify and segment plaque utilizing vendor specific adaptive thresholds. Segmental NCP, calcified plaque (CP), LAP were quantified, and the sum of all segments was used to calculate total plaque volumes (TPV). Plaque volumes were normalized to the total per patient vessel volumes, (calculated as Plaque subtype volume/ vessel volume x 100) to calculate the Percent Atheroma Volume (PAV). Percent plaque composition was calculated as plaque subtype volume/total plaque volume *100. In order to reduce false positives, and in accordance with prior studies using AIQCPA for plaque volume analyses, a TPV < 20 mm³ was used as a threshold for normal findings [[13], [14], [15]]. The CTAs of patients with a CAC score of 0 and TPV < 20 mm³ underwent additional visual inspection by a CT reader (AK) to ensure that visually significant non calcified plaque was not excluded.

2.3. Statistical analysis

Descriptive statistics were used to summarize patient demographics and plaque characteristics which were reported by age, sex and CAC score. The age groups used were <40 years, 40–54 years, 55–65 years and > 65 years. Continuous variables were assessed for normality and presented as medians with interquartile ranges (IQRs), and categorical variables as counts and percentages. Differences between age, sex and CAC score groups were assessed using the Wilcoxon Two-sample t-test or the Kruskal-Wallis test for continuous variables and chi-square or Fisher’s exact tests for categorical variables. Multivariable logistic regression analysis was performed using age, sex, BMI, diabetes, hyperlipidemia, hypertension and statin use as covariates to evaluate the association with the presence of any plaque and the presence of LAP > 10 mm³; with results displayed as odds ratios (OR) with 95% confidence intervals (CI). We included an age and sex interaction term in the initial model but found this to be non-significant and thus was not retained in the final model. Statistical significance was defined as p < 0.05. All analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC).

3. Results

A total of 341 patients from the DILWALE registry (mean age: 55.7 ± 12.3 years; 68% [n = 234] men) were included in the analysis (Table 1). Hypertension (60%, n = 205) and hyperlipidemia (82%, n = 281) were the most prevalent comorbidities, while diabetes mellitus was present in 32% (n = 110) of participants. Fifty seven percent of patients were receiving statin therapy. (Central figure)

Table 1.

Baseline characteristics of patients undergoing coronary CTA.

Baseline Characteristics	n (%) or median (IQR)
Total population	341
Age (Years)	54 (46, 65)
Sex (Female)	107 (31%)
BMI (Kg/m2)	26 (24–29)
Hypertension	205 (60%)
Systolic blood pressure (mmHg)	126 (115–137)
Diastolic blood pressure (mmHg)	74 (66 – 82)
Diabetes	110 (32%)
Hyperlipidemia	281 (82%)
Hemoglobin A1c (%)	5.8 (5.4 −6.5)
Total cholesterol level (mg/dL)	173 (141 −213)
LDL-C level (mg/dL)	99 (71 −135)
HDL-C level (mg/dL)	46 (39 −55)
Triglyceride level (mg/dL)	123 (94–164)
Current Smoker	11 (3%)
Aspirin use (no)	217 (63%)
Statin use (no)	196 (57%)
CAC score 0	37%
CAC score 1–99	29.5%
CAC score > 100	33.5%

CAC: Coronary artery calcium.

Data reported as median (IQR) or n (%).

Central Illustration.

Central Figure.

DILWALE CT Sub study of South Asians in the US

Abbreviations: SAUS: South Asians in the US

AIQCPA: Artificial intelligence enabled quantitative coronary plaque analysis

CAC: coronary artery calcium score

CP: calcified plaque

NCP: non calcified plaque

PAV: Percent atheroma volume.

3.1. Plaque burden by age and sex

Among 341 patients, 63% (n = 215) demonstrated the presence of any coronary plaque by AIQCPA. Men showed a statistically significantly higher prevalence of any plaque when compared to women (69.2% vs 49.5%, p = 0.0005). (Fig. 1). The median CAC score was 19.5 Agaston Units (IQR 0–230). The mean TPV by AIQCPA was 237.4 ± 371.84 mm³ and median TPV was 72 mm³ (IQR 0–322 mm³). (Fig. 2) The mean and median PAV were 11.3% (SD 15.72) and 3.5% (IQR 0–18.5%). (Fig. 3) Women in general had lower TPV [median 15 mm³ (IQR 0–149 mm³) vs 94 mm³ (IQR 2- 350 mm³)], and all plaque subtypes compared to men (Fig. 2).The median TPV was statistically significantly lower in women compared to men in the 40–54 year age group [TPV 0 mm³ (IQR 0,10 mm³) vs 35 mm³ (IQR 0150 mm³), p < 0.0001] and in the 55–65 year age group [TPV 44 mm³ (IQR 1–129 mm³) vs 126 mm³ (47- 532 mm³), p = 0.005 ] (Table S2). Women aged > 65 years demonstrated a significantly higher median TPV compared to women in the 55 - 65 year age groups [Median TPV 255 mm³ (IQR 40.0, 637.0 mm³) vs 44 mm³ (IQR 1.0, 129.0 mm³)]. (Table S2)

Fig. 1.

Presence of any plaque by AIQCPA in the cohort.

Abbreviations: CAC: coronary artery calcium

AIQCPA: Artificial intelligence (AI) - enabled quantitative coronary plaque analysis.

Fig. 2.

Plaque volumes and distribution by age and sex.

Legend: Box plot of distribution of quantitative plaque volumes by age and sex.

Median plaque volume is in mm³.

Fig. 3.

Percent atheroma volumes (PAV) for total plaque and by plaque subtypes in the cohort.

Legend: PAV = plaque subtype volume/vessel volume * 100

Vessel volume = lumen volume + total plaque volume

Abbreviations: TP = Total Plaque; NCP = Non-Calcified Plaque; CP= Calcified plaque.

Detection of any plaque was more common in patients older than 65 years (88.9%), and plaque volume increased with age. However, AIQCPA still identified plaque in 48.7% (77) of patients aged 40–54 years. (Fig. 1) Age-based TPV, and the distribution of CP, NCP, and LAP by age, sex and CAC scores are summarized in Fig. 2, and Table 2. NCP predominated across all age groups with an overall percent NCP composition of 86%. Younger individuals had a greater predominance of NCP whereas CP volumes increased with age (Fig. 4). LAP ≥ 10 mm³ was present in 42 patients. Total Plaque and CP volumes were highest in the left anterior descending (LAD) territory. [Median TPV 27 mm³ (IQR 0–127 mm³). (Table S1).

Table 2.

Distribution of plaque volumes by Age, Sex and CAC score.

Plaque volumes are described as Median ± IQR (mm³)

CP: calcified plaque; NCP: non calcified plaque; LAP: low attenuation plaque.

Plaque Volume (mm3)	Age (Years)					Sex		Calcium score
	All	<40	40–54	55–65	> 65	Female	Male	CAC= 0	CAC 1–99	CAC>100
	(N = 341)	(N = 20)	(N = 158)	(N = 82)	(N = 81)	(N = 107)	(N = 234)	(N = 119)	(N = 95)	(N = 108)
Total Plaque	72 (0, 322)	0 (0, 7.5)	12 (0, 112)	98.5 (20, 378)	307 (120, 765)	15 (0, 149)	94 (2, 350)	0 (0, 1)	79 (29, 114)	506 (302, 869.5)
CP	6 (0, 50)	0 (0, 0)	0.0 (0, 11)	17.5 (0, 87)	56 (21, 164)	0.0 (0, 23)	10.5 (0, 60)	0 (0, 0)	6 (2, 15)	98.5 (41, 170)
NCP	61 (0, 265)	0 (0, 7.5)	12 (0, 95)	86 (19, 306)	259 (109, 592)	14 (0, 122)	87.0 (2, 298)	0 (0, 1)	66 (27, 95)	378 (254, 682.5)
LAP	1.0 (0.0, 4.0)	0.0 (0.0, 0.0)	0.0 (0.0, 2.0)	1.5 (0.0, 4.0)	3.0 (1.0, 8.0)
	1 (0, 4)	0 (0 – 0)	0 (0, 0)	0 (0, 2)	1.5 (0, 4)	0 (0)	1.0 (0, 5)	0 (0, 0)	1 (0, 2)	5 (3, 11.5)

Fig. 4.

Percentage of non calcified and calcified plaque stratified by age and sex.

Legend: Percent NCP = NCP volume/total plaque volume *100

Percent CP = CP volume/total plaque volume *100

Abbreviations: CP: calcified plaque; NCP: non calcified plaque.

3.2. Plaque burden among individuals with a coronary artery calcium score of zero

Thirty four percent of patients (n = 119) in this cohort had a CAC score of 0. The distribution of CAC scores is represented in Table 1. Visual inspection of the 23 patients with a CAC 0 and TPV < 20 mm³ by AIQCPA (those considered as having no plaque) was performed to ensure low volume NCP was not being missed. One out of 23 patients had minimal NCP in the ostial right coronary artery. This patient was counted as having plaque, changing the prevalence of any plaque in patients with CAC 0 from 9% (n = 11) by AIQCPA to 10% (n = 12) with a median TPV of 0 mm³ (IQR 0–1 mm³). (Fig. 1, Figure S1). Only hypertension was found to be a significant predictor of the presence of any plaque in patients with a CAC score of 0. No patients in the < 40 year age group with CAC 0 had plaque. LAP ≥ 10 mm³ was seen in 1 patient with a CAC 0.

3.3. Association of traditional ascvd risk factors with plaque burden and the presence of low attenuation plaque

In multivariable regression analyses, older age was independently associated with higher odds of both any plaque [OR 1.10 (95% CI: 1.06 – 1.13), p < 0.0001)] and LAP [OR 1.11 (95% CI: 1.06 – 1.16), p < 0.0001]. Women had lower odds than men of both any plaque [OR 0.28 (95% CI: 0.15 – 0.52), p < 0.0001] and LAP [OR 0.11 (95% CI: 0.04 – 0.36), p < 0.0003]. A trend for a possible effect was noted for hypertension [OR 1.67 (95% CI 0.96, 2.92), p = 0.07] and any plaque. In univariable analyses, diabetes mellitus, hyperlipidemia, statin use and hypertension were associated with the presence of any plaque and LAP ≥ 10 mm³. After multivariable adjustment, only statin use remained a significant predictor for any plaque [1.99 (1.11 – 3.57 p = 0.021] but a trend was noted for BMI as a possible predictor of LAP [CI 1.10 (1.00, 1.22)], p = 0.05] (Table 3, Table 4).

Table 3.

Association of traditional ASCVD risk factors with the presence of any plaque.

	Univariable		Multivariable
	Odds Ratio (95% CI)	p-value	Odds Ratio (95% CI)	p-value
Age	1.10 (1.07, 1.13)	< 0.0001	1.11 (1.07, 1.14)	< 0.0001
Female sex	0.45 (0.28, 0.71)	0.0007	0.22 (0.12, 0.42)	< 0.0001
Body-mass-index	1.02 (0.97, 1.07)	0.5285	1.04 (0.98, 1.10)	0.2033
Current smoker	1.83 (0.58, 5.79)	0.3064	2.41 (0.63, 9.21)	0.1983
Diabetes Mellitus	2.20 (1.33, 3.63)	0.0021	1.42 (0.76, 2.64)	0.2749
Hyperlipidemia	2.06 (1.17, 3.62)	0.0119	0.74 (0.34, 1.60)	0.4483
Hypertension	2.75 (1.75, 4.34)	< 0.0001	1.62 (0.93, 2.81)	0.0888
Statin use	2.81 (1.79, 4.42)	< 0.0001	1.99 (1.11, 3.57)	0.021

Multivariable analysis: Adjusted for Age, sex, current smoker status, diabetes mellitus, hyperlipidemia, hypertension, statin use.

Table 4.

Association of traditional ASCVD risk factors with the presence of low attenuation plaque (LAP= 0 mm³ vs LAP ≥ 10 mm³).

	Univariable		Multivariable
	Odds Ratio (95% CI)	p-value	Odds Ratio (95% CI)	p-value
Age	1.09 (1.06, 1.13)	< 0.0001	1.11 (1.06, 1.16)	< 0.0001
Female sex	0.35 (0.15, 0.81)	0.0142	0.11 (0.04, 0.36)	0.0003
BMI	1.04 (0.97, 1.12)	0.2962	1.10 (1.00, 1.22)	0.0586
Current smoker	1.09 (0.22, 5.47)	0.9139	0.88 (0.13, 5.90)	0.8984
Diabetes	2.18 (1.07, 4.47)	0.0325	1.26 (0.52, 3.03)	0.6094
Hyperlipidemia	3.91 (1.14, 13.39)	0.0298	1.61 (0.37, 7.00)	0.5231
Hypertension	2.97 (1.37, 6.44)	0.0059	1.76 (0.68, 4.59)	0.2446
Statin use	2.41 (1.17, 4.96)	0.0175	1.61 (0.62, 4.22)	0.3304

Multivariable analysis: Adjusted for Age, sex, current smoker status, diabetes mellitus, hyperlipidemia, hypertension, statin use.

4. Discussion

This analysis of the DILWALE registry revealed several insights into coronary artery plaque burden among SA individuals, a population at uniquely high risk for ASCVD but seldom characterized using advanced coronary imaging. First, 63% of participants were identified as having coronary plaque on CTA by AIQCPA, including 49% of those 45–54-years-old. After excluding a TPV < 20 mm³ to reduce false positives, any plaque was noted in 15% of patients under 40 years of age. While direct comparison is limited by sample size and differences in assessment, a similar prevalence (19%) has been reported in the CONFIRM registry among symptomatic individuals < 45 years of age [16]. Our findings on the distribution of plaque in the different coronary territories align with findings from the PARADIGM registry [17]. Second, male patients develop plaque earlier and have greater plaque burden than female patients across all age groups, consistent with previously reported trends. However, it is worth noting that women have higher CV risk at similar plaque volumes when compared to men [18]. Women in our cohort also demonstrate a significant increase in plaque volumes after age 65 years, consistent with the catch up increase in CV risk seen in women post menopause [19]. Third, a higher prevalence of NCP was noted compared to CP at all age groups, consistent with a prior large international clinical cohort using the same AIQCPA software [12]. NCP is particularly important due to its greater propensity for rupture, leading to adverse acute coronary events [20]. LAP was seen in 42 patients, and its prevalence increased with age. Secondary analysis of SCOT-HEART showed that LAP, which represents lipid rich or vulnerable plaque, is a strong predictor of fatal and non-fatal myocardial infarction (MI), irrespective of CAC score or coronary stenosis, and doubling of LAP conferred ∼60% higher MI risk [20]. Finally, age, male sex, and statin use were significant predictors of the presence of any plaque.

The increasing utilization of AIQCPA has allowed rapid quantification of plaque burden with increasing evidence for the prognostic value of TPV [21,22]. Outcomes with targeted use of preventive therapies based on AIQCPA findings are being studied [23]. Ours is one of the largest databases of CTA based AIQCPA in South Asians in the US leveraging the use of plaque analysis to gain insights into CVD burden in this population. Prior studies looking at CAC score (a density based score), have shown similar CAC scores in SA and Caucasians, however, this finding does not explain the increased CV risk seen among SA individuals [24]. Evaluating plaque burden and composition have shown additional benefit in refining CV risk and prognostication [[25], [26], [27]]. Two prior studies have shown conflicting results on whether SA have an increased total plaque burden when compared with Caucasians. A cross-sectional analysis of 165 SA adults with 1:1 matched non-Hispanic White individuals showed higher percent atheroma volume (Total PAV 11.9% vs. 6.2%), NCP (5.1% vs. 3.8%), and CP (4.9% vs. 2.4%) in SA adults [28]. Another study showed a similar total plaque burden but higher percent NCP composition in SA adults compared to Caucasians [9]. Similarly, studies in SA individuals with diabetes have shown a higher number of vessels with >50% stenosis and a higher prevalence of triple vessel disease [29,30]. However, these studies did not use AIQCPA and did not stratify based on age groups or CAC scores.

Our study in a clinical cohort of SA reveals that nearly 63% of the cohort had the presence of coronary plaque with a plaque volume > 20 mm³ by AIQCPA, a prevalence lower than reported in a recent assessment of plaque burden in an asymptomatic US based cohort (81%), though directed comparison is limited given different populations studied [14]. Further, a different AIQCPA software was used for the prior study, and inter-vendor comparison has known limitations [13]. Total plaque and NCP volumes are also lower in our cohort when compared to prior clinical cohorts of SA with similar risk factor burden [9,28]. The median PAV of 3.5% is also lower than the PAV of 9.1% (Table S4, Tzimas et al.) in a large multicenter AIQCPA based clinical cohort using the same software [12]. While we are limited in terms of outcomes of this population, the lack of an increased plaque, NCP or LAP burden is noteworthy, despite the slightly younger population in our cohort (mean age 55 vs 62 years). Our findings could reflect referral bias leading to more intensive CTA based testing prompted by broad awareness of increased CV risk in SA. However, a median PAV of 3.5% may still be clinically significant in this population. A prior study has shown that a PAV cutoff of ≥ 2.6% is associated with increased rates of acute coronary syndrome [31]. We also noted that the >65-year age group had a median PAV of 20%, higher than previously reported in SA and other populations [12,28]. If confirmed in larger cohorts, our findings may indicate that clinical events could occur at lower plaque volumes in SA, and with more rapid plaque progression as individuals age. The presence of plaque among individuals <40 years, along with an NCP predominance also needs further characterization in larger cohorts to inform screening practices especially in younger and potentially higher-risk SA. Prior work has also postulated that smaller coronary artery volumes in SA may increase the risk of thrombotic occlusion at similar plaque volumes, a hypothesis that warrants future study [32]. Furthermore, other contributors may underlie the observed increased risk of clinical ASCVD in SA, including potential differences in healthcare-seeking behavior.

While the prevalence of any plaque increased with an increase in CAC score, 10% of SA with a CAC score of 0 had any plaque. LAP was seen in only 1 patient with a CAC score of 0. These results highlight the low plaque burden observed among individuals with a CAC score of 0, with a plaque prevalence similar to other CTA based cohorts [14]. However, the detection of non-calcified plaque could lead to earlier initiation of preventive therapies and potentially change the trajectory of disease, especially in younger SA.

Multivariable modeling confirmed older age and male sex as strong predictors of any plaque and LAP. While hypertension, diabetes, hyperlipidemia and statin use predicted the presence of any plaque in univariable analyses, only statin use remained a significant predictor of the presence of plaque in adjusted models. This observation likely reflects confounding by treatment with high-risk individuals preferentially being started on a statin by their clinicians, rather than a true atherogenic effect. Future studies should further explore the relationship between traditional ASCVD risk factors and the presence of any coronary plaque and LAP.

Our study has several limitations. First, the cross-sectional design without longitudinal follow up precludes causality and association with clinical outcomes. Second, although we adjusted for known variables, residual confounding by unmeasured factors such as diet, physical activity, and socioeconomic status cannot be excluded. The unavailability of menopausal status in women remains an additional limitation given well characterized changes in CV risk post menopause. Third, our study did not control for risk factor exposure duration in terms of LDL-C, hemoglobin A1c and Lipoprotein (a) levels. Fourth, despite confirmation by visual inspection in patients with CAC =0, the exclusion of a total plaque volume <20 mm³ could potentially have missed some early plaque. This may be a missed opportunity for prevention in younger SA who may have lower plaque burden with predominantly non calcified plaque. Fifth, plaque quantification is known to be affected by the AIQCPA software vendor, image quality and spatial resolution. While we used the same AIQCPA software vendor for all studies, the use of CTA data from different scanners introduces the possibility of variability in plaque quantification and limits external generalizability to studies performed in other settings. Finally, since this is a real-world clinical cohort, there is a possibility of referral bias and early detection of plaque due to increased awareness. The generalizability of our findings to asymptomatic community dwelling SA and SA globally may be limited. Large-scale, multicenter, longitudinal studies are necessary, particularly in elucidating the interplay of genetic, lifestyle, and social factors in determining plaque burden and CV risk in SA.

5. Conclusions

Our study is one of the largest cohorts of US- based SA, originating from the well characterized DILWALE registry. Plaque burden in this referred SA cohort appears comparable to previously reported AIQCPA cohorts, though direct comparisons are limited. The study highlights the value of AI-enabled CTA in revealing subclinical plaque, particularly non-calcified plaque in SA. Our findings also underscore the need for larger population-specific studies to identify optimal imaging strategies for SA and future outcome-based validation.

Disclosures / conflicts of interest

AK reports non – financial support from HeartFlow for plaque analysis in the registry. MSK reports consulting fees or research grant support from Bayer, Novartis, Merck and Boehringer Ingelheim. AA reports consulting fees and research grants from HeartFlow, and Cleery. JL reports consulting fees and stock options from HeartFlow. MB has received support from HeartFlow and grant support from Cleerly and General Electric. Other authors report they have no conflicts of interest.

Statement of authorship

Priyanka Satish and Anandita Kulkarni designed the study. Priyanka Satish, Anandita Kulkarni and Aashna Vajramani wrote the manuscript. Tsung-wei Ma, Tanushree Agarwal and Swapnil Gupta helped with data curation, performed statistical analysis and contributed to the manuscript. Muhammad Shahzeb Khan, Matthew Budoff, Dinesh Kalra, Jonathon Leipsic, Amro Alsaid, and Suvasini Lakshmanan helped with study methodology and provided review and editing for the manuscript. All authors provided final review of the manuscript.

Author agreement

We the undersigned declare that this manuscript is original, has not been published before and is not currently being considered for publication elsewhere. We confirm that the manuscript has been read and approved by all named authors and that there are no other persons who satisfied the criteria for authorship but are not listed. We further confirm that the order of authors listed in the manuscript has been approved by all of us. We understand that the Corresponding Author is the sole contact for the Editorial process. He/she is responsible for communicating with the other authors about progress, submissions of revisions and final approval of proofs.

CRediT authorship contribution statement

Priyanka Satish: Writing – original draft, Conceptualization. Aashna Vajramani: Writing – original draft. Swapnil Gupta: Writing – original draft, Formal analysis, Data curation. Tsung-wei Ma: Writing – original draft, Formal analysis, Data curation. Tanushree Prasad: Writing – original draft, Formal analysis, Data curation. Suvasini Lakshmanan: Writing – review & editing, Methodology. Matthew Budoff: Writing – review & editing, Methodology. Dinesh K. Kalra: Writing – review & editing, Methodology. Muhammad Shahzeb Khan: Writing – review & editing, Methodology. Jonathon Leipsic: Writing – review & editing, Methodology. Amro Alsaid: Writing – review & editing, Methodology. Anandita Kulkarni: Writing – original draft, Methodology, Conceptualization.

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Anandita Kulkarni MD reports equipment, drugs, or supplies was provided by HeartFlow Inc. Muhammad Shahzeb Khan reports a relationship with Bayer Pharma AG that includes consulting or advisory and funding grants. Muhammad Shahzeb Khan reports a relationship with Novartis Pharmaceuticals Corporation that includes consulting or advisory and funding grants. Muhammad Shahzeb Khan reports a relationship with Merck & Co Inc that includes consulting or advisory and funding grants. Muhammad Shahzeb Khan reports a relationship with Boehringer Ingelheim Canada Ltd that includes: consulting or advisory and funding grants. Amro Alsaid reports a relationship with HeartFlow Inc that includes: consulting or advisory and funding grants. Amro Alsaid reports a relationship with Cleerly Inc that includes: consulting or advisory and funding grants. Jonathon Leipsic reports a relationship with HeartFlow Inc that includes: consulting or advisory and equity or stocks. Matthew Budoff reports a relationship with HeartFlow Inc that includes support. Matthew Budoff reports a relationship with Cleerly Inc that includes: funding grants. Matthew Budoff reports a relationship with General Electric Company that includes: funding grants. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.