About the CAC Percentile Calculator

Motivation

Patient classification for cardiovascular risk assessment following cardiac CT calcium scoring is currently done using either absolute Agatston calcium scoring (such as Agatston score ≥ 100 as a threshold) or more commonly using percentile score criteria. Currently, applications like Vitrea Vscore allow users to check the patient Agatston score as a percentile value against selected populations from common databases: Hoff [1], Raggi [2], and MESA [3].

These databases provide percentile information based on absolute Agatston score, and include age, gender, and in some cases, ethnicity categories [3] for a more precise risk prediction.

While the MESA study (6,110 subjects) [3] and Hoff study (35,246 subjects) [1] have considerable sample sizes, categories such as patients younger than 55 years old, women, and/or specific ethnic minorities are often underrepresented. For example, the MESA sample included only 719 Chinese subjects (with only 371 women), mostly from Los Angeles County and Chicago [4], across all age classes.

A relatively simple solution for a much larger and more representative database population for calcium scoring is to pool results from several calcium databases with diverse patient populations for more reliable percentile estimates. Recently, De Ronde et al. [5], provided percentile tables and curves based on such a pooled database. This included twelve databases from different geographic origins, pooling 134,336 Western subjects and 33,488 Asian subjects separately. This represents 37 times the size of the MESA study with much larger geographic coverage.

De Ronde et al. were unable to access the original raw data for all selected databases. Instead, they computed new pooled calcium percentile values as the weighted average of the Agatston score percentile values from the various published patient groups (Western vs. Asian, gender, and age classes). The weights in the weighted average correspond to the relative sample sizes of the different databases [5].

Methodology and Data Source

The core percentile data used in this calculator is based primarily on the pooled multi-ethnic percentile data published by De Ronde et al. in JACC: Cardiovascular Imaging (2020;14:414-420) .

Prior to integrating this De Ronde et al. (2020) pooled database, a review of the detailed results was conducted with the study authors for the following categories: Western Men, Western Women, Asian Men, and Asian Women. This review revealed some minor errors or inconsistencies between the published tables and the percentile curves presented in the original publication .

Correction Assumptions:

Corrections derived from this review were based on the following assumptions:

1. Non-Decreasing Percentiles with Age: Percentile curves for calcium Agatston score versus patient age are assumed to be non-decreasing, considering that coronary calcification generally increases with age. (Examples from the original study showing decreasing curves required adjustment based on this rule).
2. Non-Crossing Percentile Curves: Percentile curves for a given population group should not cross each other. (Examples from the original study showing crossing curves required adjustment).
3. Handling Local Decreases: In cases where a percentile value showed a temporary decrease between consecutive age classes (e.g., values like 1-0-2), the lowest value in that dip sequence is adopted to ensure the function remains non-decreasing (resulting in 0-0-2). (This conservative approach was applied, e.g., changing the 50th percentile for Western Men <40Y from 1 to 0).
4. Oldest Age Group (>74): The original review suggested the oldest age category (>74) might be excluded if results remained inconsistent or sample sizes were too low. (Note: For this calculator's dataset, the decision was made to include the >74 group but apply the non-decreasing rule rigorously, acknowledging potential limitations due to smaller sample sizes).

Ethnicity Categories Used

This calculator uses the ethnicity categories "Western" and "Asian". This choice directly reflects the pooling methodology and the final categories presented in the primary data source (De Ronde et al., 2020) . The study authors grouped data from various global studies (including USA, Europe, Brazil, Korea, Japan) into these two broad categories due to limitations in accessing uniform, fine-grained ethnicity data across all pooled studies and based on observed differences in calcium scores (notably lower scores in the Asian cohort). The "Western" group is primarily a mix of North American and European cohorts, while the "Asian" group is predominantly Korean and Japanese participants from the source studies.

This differs from calculators based solely on studies like MESA , which was designed to recruit distinct US-based ethnic groups (White, Black, Hispanic, Chinese American) and could therefore provide separate percentiles for them. The categories used here are accurate for the underlying De Ronde pooled dataset.

Summary of Applied Corrections

The following table highlights specific numerical corrections applied to the base data from Table 2 of De Ronde et al. , in addition to the systematic application of the non-decreasing rule across all percentiles and age groups.

45-49: 0

45-49: 2

<40: 1

<40: 0

>74: 1

>74: 16

40-44: 3

40-44: 1

70-74: 476

70-74: 514

References

1. Hoff JA, et al. 2001. Age and gender distributions of coronary artery calcium detected by electron beam tomography in 35,246 adults. Am J Cardiol. 2001 Jun 1;87(11):1335-9DOI | PubMed
2. Raggi P, et al. 2000. Identification of patients at increased risk of first unheralded acute myocardial infarction by electron-beam computed tomography. Circulation. 2000 Feb 29;101(8):850-5DOI | PubMed
3. MESA Study (McClelland RL, et al. 2006). Distribution of coronary artery calcium by race, gender, and age: results from the Multi-Ethnic Study of Atherosclerosis (MESA). Circulation. 2006 Jan 3;113(1):30-7DOI | PubMed
4. MESA Chinese Cohort Details. . Context: Referenced in relation to the underrepresentation of certain groups in older databases. Specific recruitment details (e.g., Los Angeles County, Chicago) are typically found within the main MESA study design or primary distribution paper [3].
5. De Ronde Pooled Analysis (De Ronde MWJ, et al. 2020). A Pooled-Analysis of Age- and Sex-Based Coronary Artery Calcium Score Percentiles. JACC Cardiovasc Imaging. 2020 May;13(5):1217-1230DOI | PubMed