Archives of Academic Emergency Medicine. 2023; 11(1): e45 REV I EW ART I C L E The value of Coronary Artery Disease – Reporting and Data System (CAD-RADS) in Outcome Prediction of CAD Pa- tients; a Systematic Review and Meta-analysis Koohyar Ahmadzadeh1 a , Shayan Roshdi Dizaji1 a , Mohammad Kiah1, Mohamad Rashid2, Reza Miri3∗, Mahmoud Yousefifard1,3 † 1. Physiology Research Center, Iran University of Medical Sciences, Tehran, Iran. 2. Student Research Committee, Babol University of Medical Sciences, Babol, Iran. 3. Prevention of Cardiovascular Disease Research Center, Shahid Beheshti University of Medical Sciences, Tehran, Iran. Received: March 2023; Accepted: April 2023; Published online: 15 June 2023 Abstract: Introduction: Coronary computed tomographic angiography (CCTA) reporting has traditionally been operator- dependent, and no precise classification is broadly used for reporting Coronary Artery Disease (CAD) severity. The Coronary Artery Disease Reporting and Data Systems (CAD-RADS) was introduced to address the inconsistent CCTA re- ports. This systematic review with meta-analysis aimed to comprehensively appraise all available studies and draw con- clusions on the prognostic value of the CAD-RADS classification system in CAD patients. Methods: Online databases of PubMed, Embase, Scopus, and Web of Science were searched until September 19th, 2022, for studies on the value of CAD-RADS categorization for outcome prediction of CAD patients. Results: 16 articles were included in this system- atic review, 14 of which had assessed the value of CAD-RADS in the prediction of major adverse cardiovascular events (MACE) and 3 articles investigated the outcome of all-cause mortality. Our analysis demonstrated that all original CAD- RADS categories can be a predictor of MACE [Hazard ratios (HR) ranged from 3.39 to 8.63] and all categories, except CAD-RADS 1, can be a predictor of all-cause mortality (HRs ranged from 1.50 to 3.09). Moreover, higher CAD-RADS categories were associated with an increased hazard ratio for unfavorable outcomes among CAD patients (p for MACE = 0.007 and p for all-cause mortality = 0.018). Conclusion: The evidence demonstrated that the CAD-RADS classifica- tion system can be used to predict incidence of MACE and all-cause mortality. This indicates that the implementation of CAD-RADS into clinical practice, besides enhancing the communication between physicians and improving patient care, can also guide physicians in risk assessment of the patients and predicting their prognosis. Keywords: Coronary artery disease; Risk assessment; CAD-RADS; Reporting and Data System Cite this article as: Ahmadzadeh K, Roshdi Dizaji S, Kiah M, Rashid M, Miri R, Yousefifard M. The value of Coronary Artery Disease – Re- porting and Data System (CAD-RADS) in Outcome Prediction of CAD Patients; a Systematic Review and Meta-analysis. Arch Acad Emerg Med. 2023; 11(1): e45. https://doi.org/10.22037/aaem.v11i1.1997. 1. Introduction Coronary computed tomographic angiography (CCTA) is an accurate and non-invasive tool with a high negative predic- tive value, which is increasingly being used for the evalua- ∗Corresponding Author: Reza Miri: Prevention of Cardiovascular Dis- ease Research Center, Imam Hossein Hospital, Madani Avenue, Tehran, Iran. Phone/Fax: +982177582721; Email: dr.rezamiri1@gmail.com, ORCID: https://orcid.org/0000-0002-8568-9948. † Corresponding Author: Mahmoud Yousefifard: Physiology Research Center, Iran University of Medical Sciences, Hemmat Highway, P.O Box: 14665-354, Tehran, Iran; Phone/Fax: +982186704771; Email: yousefifard20@gmail.com / yousefifard.m@iums.ac.ir, ORCID: https://orcid.org/0000-0001-5181-4985. a: First and second authors have had equal contributions. tion of patients with stable angina and acute coronary artery disease (CAD) (1). CCTA provides physicians with utile infor- mation on the presence of atherosclerosis, and its character- istics such as the extent and location (2). CCTA reporting has traditionally been operator-dependent, and no precise classification has been broadly used for summarizing and categorizing CAD severity in this imaging modality (3). Previously, efforts had been made to design and implement uniform structuralized reporting frameworks for the inter- pretation of imaging assessments such as breast (BI-RADS), prostate (PI-RADS), liver (LI-RADS), and lung (Lung-RADS) imaging reporting and data systems (4). To this end, in 2016, the Coronary Artery Disease Reporting This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index K. Ahmadzadeh et al. 2 and Data Systems (CAD-RADS), a multi-society consensus reached by radiologists and cardiologists, was introduced to address the inconsistent CCTA reports (5). CAD-RADS has been intended to establish a common lexi- con between multiple disciplines involved in patient man- agement. Moreover, the acceptable intra and inter-observer variability of this scoring system contributed to the consis- tent and efficient patient clinical management and facilitated data gathering for registries with research purposes (6, 7). CAD-RADS provides a precise yet simple representation of CAD severity by classifying patients based on the most severe stenosis identified in CCTA (8). CAD-RADS categorizes the stenosis according to the degree of luminal diameter steno- sis, ranging from the absence of any occlusion or plaque (cat- egory 0) to total occlusion of at least one coronary artery (cat- egory 5). There are also modifiers including N, S, G, and V, which stand for non-diagnostic, stent, graft, and vulner- ability, respectively, providing additional details of the CCTA finding (Table 1) (5). In addition to the previously mentioned advantages of CAD- RADS, its clinical meaningfulness would be pronounced when patient categorization provides guidance on their ther- apeutic and preventive management measures. To achieve this aim, a preliminary step is to confirm the validity of CAD-RADS categorization for patient prognosis. Previously, several reports have investigated the predictive value of the CAD-RADS categorization system on patient outcomes, con- sisting of the risk for a major adverse cardiovascular event (MACE) and all-cause mortality. This systematic review with meta-analysis aimed to comprehensively appraise all avail- able studies and draw a conclusion on the predictive value of the CAD-RADS classification system in CAD patients. 2. Methods 2.1. Study design and setting This systematic review and meta-analysis was designed to evaluate the predictive value of CAD-RADS in the assessment of outcomes in CAD patients. In this study, PICO was defined as: Patients (P): patients with suspected or known coronary artery disease, Index test (I): CAD-RADS classification tool, Comparison (C): coronary artery disease patients not devel- oping the outcome of the study, Outcome (O): Major adverse cardiovascular event (MACE) and all-cause mortality. 2.2. Search strategy Appropriate keywords related to the aim of this study were chosen based on MeSH (Medline database) and Emtree (Em- base database) terms, a review of the related literature, and consultation with experts in the field. A systematic search was performed using four online databases of PubMed, Em- base, Scopus, and Web of Science until September 19th, 2022. The search strategy used for this study is provided in sup- plementary material 1. Google and Google Scholar search engines and references of the included articles were also re- viewed to retrieve any papers that might have been missed. 2.3. Selection criteria All articles evaluating the value of CAD-RADS for the predic- tion of outcomes in CAD patients were included in this study. The exclusion criteria were commentaries and editorials, re- view articles, case reports, case series, and articles not report- ing the data of interest. 2.4. Data collection Two researchers independently reviewed the titles and ab- stracts of the retrieved articles and full-text screening was performed for possibly relevant articles and appropriate ar- ticles were included in the study. The information reported in the included articles was summarized and compiled in a checklist designed according to the criteria of Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines. Any disagreements were resolved by consulting a third reviewer. The data checklist included ar- ticle characteristics (name of first author, year of publica- tion, and country of study), study design, studied population, number of patients, number of men, age, studied outcome and number of patients developing the outcome, follow-up duration, reported CAD-RADS category, and the relevant ef- fect size reported for each CAD-RADS category. The effect size of interest was chosen to be hazard ratio and the authors of any articles not providing required information were con- tacted by email, with a 1-week reminder in order to gain ac- cess to their results. Any disagreement between the two re- viewers was resolved by the third reviewer. 2.5. Quality and certainty of evidence assessment The quality of the articles was assessed using the guidelines provided by the Quality Assessment of Prognostic Accuracy Studies (QUAPAS) tool (9). Based on this guideline the arti- cles are assessed according to their risk of bias (in domains of participants, index test, outcome, flow and timing, and analysis) and their applicability (in domains of participants, index test, outcome, and flow and timing). The Grades of Recommendation, Assessment, Development, and Evalua- tion (GRADE) guidelines were used to evaluate the certainty of evidence (10). The certainty of evidence table was de- signed using GRADEpro online software (www.gradepro.org). 2.6. Statistical analysis Analyses were performed in the STATA 17.0 statistical soft- ware. The predictive value of CAD-RADS for outcomes of CAD patients was recorded as hazard ratio (HR) and 95% confidence interval (CI) and the data were analyzed using This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index 3 Archives of Academic Emergency Medicine. 2023; 11(1): e45 Table 1: CAD-RADS classification system Category Maximal Stenosis Interpretation CAD-RADS 0 0% No CAD CAD-RADS 1 1 - 24% Minimal nonobstructive CAD-RADS 2 25 - 49% Mild nonobstructive CAD-RADS 3 50 - 69% Moderate stenosis CAD-RADS 4A 70 - 99% Severe stenosis CAD-RADS 4B Left main >50% or 3-vessel ≥ 70% Severe stenosis CAD-RADS 5 100% Total coronary occlusion CAD-RADS: Coronary artery disease-reporting and data system; CAD: Coronary artery disease. Table 2: Characteristics of included studies Study, Year Design# Sample size Age* Male% (N) Follow-up Event% (N) Major Adverse Cardiac Event (MACE) Altay, 2021 (12) Retrospective 359 54.17 54.31 (195) 8 years 6.96 (25) Bittner, 2020(13) Retrospective 3840 60.4±8.2 48.64 (1868) 2.08 years 2.99 (115) Duguay, 2017(20) Retrospective 48 56 ± 10 60.4 (29) 1.6 year 29.16 (14) Faber, 2021(21) Retrospective 1615 59 66.62 (1076) 10.5 years 3.31 (51) Finck. 2019 (a)(14) Retrospective 2011 59 ± 11 66.03 (1328) 10 years 2.88 (58) Lee, 2021(15) Retrospective 1492 58±6.14 50.87 (759) 3 months 4.22 (63) 31.5 months 6.90 (103) Maclean, 2022(22) Retrospective 720 58 [IQR 19] 62.08 (447) 5.4 years 7.5 (54) Mangalesh, 2022(23) Prospective 366 62 70.76 (259) 2.56 years 16.39 (60) Senoner, 2020(16) Prospective 1430 57.9±11.1 55.59 (795) 10.55 years 3.98 (57) Tang, 2022(17) Prospective 511 61 [33-94] 75.92 (388) 1 year 6.65 (34) Van Rosendael, 2019(24) Prospective 2134 54.72 49.01 (1046) 3.6 years 0.06 (130) Williams, 2020(18) Retrospective 1769 58±10 56.35 (997) 4.7 years 2.31 (41) Xie, 2018(19) Retrospective 5039 59.97 63.74 (3212) 5 years 15.30 (771) Yamamoto, 2021(25) Prospective 133 67 ± 11 69.92 (93) 3.33 years 10.52 (14) All-cause mortality Finck, 2019 (b)(26) Retrospective 1913 58.97 66.54 (1273) 9.7 years 5.17 (99) Huang, 2021(27) Retrospective 9625 59.8±10.7 44.28 (4262) 4.3 years 5.61 (540) Senoner, 2020(16) Prospective 1430 57.9±11.1 55.59 (795) 10.55 years 7.41 (106) Xie, 2018(19) Retrospective 5039 59.97 63.74 (3212) 5 years 6.23 (314) *: Age is reported as mean ± SD or median [IQR]. #: All studies are observational. the “meta” package. The studies utilizing original CAD-RADS categories (1, 2, 3, 4A, 4B, and 5), with CAD-RADS category 0 as the reference, were included in the meta-analysis. The ex- periments with reports of combinations of the original CAD- RADS categories into a subset category or continuous vari- able were excluded from the meta-analysis and have been reported qualitatively. A meta-regression analysis was per- formed to evaluate the effect of the follow-up duration on the predictive value of CAD-RADS. The Heterogeneity between included studies was evaluated using I2 statistics and Chi- squared test. Publication bias assessment was not applicable since less than 10 articles were included in each meta-analysis (11). 3. Results 3.1. Study characteristics The systematic search of online databases of PubMed, Em- base, Scopus, and Web of Science resulted in 346 non- duplicate records. 92 of these records were deemed to be eli- gible and upon further evaluation, 15 articles were chosen to be included in this study. Two articles were found via manual search, one of which was included. Finally, 16 articles (12-27) were included in this study (Figure 1). All included articles had assessed suspected or known coro- nary artery disease patients with a low to intermediate prob- ability of CAD. The articles had defined CAD-RADS cate- gories as introduced by Cury et al. (5) and only one article had a slight definition variation (5% difference in categories 1 and 2) (13). The results of analyses did not differ signifi- This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index K. Ahmadzadeh et al. 4 Table 3: Results of miscellaneous CAD-RADS reporting CAD-RADS category Study, Year Hazard Ratio (95% CI) Major Adverse Cardiovascular Event (MACE) 1 and 2 Lee, 2021 (15) 2.4 (0.8-6.9) 1.5 (0.7-3.2) 2 and 3 Van Rosendael, 2019 (24) 1.95 (1.19-3.2) ≥3 Duguay, 2017 (20) 3.12 (1.03-10.17) Yamamoto, 2021 (25) 1.3 (0.16-11.1) 4 (4A and 4B) Senoner, 2020 (16) 36.48 (4.94-269.67) Tang, 2022 (17) 13.49 (4.809-37.86) ≥4 Lee, 2021 (15) 1.7 (1.1-39.2) Lee, 2021 (15) 8.5 (3.7-15.8) Mangalesh, 2022 (23) 3.801 (1.58-9.145) Van Rosendael, 2019 (24) 2.68 (1.3-5.53) Yamamoto, 2021 (25) 1.3 (0.35-5.15) 4B+5 Bittner, 2020 (13) 21.84 (8.63-55.26) Yamamoto, 2021 (25) 2.6 (0.72-9.2) 1.97 (1.12-3.45) male Overall Faber, 2021 (21) 5.34 (2.42-11.8) female 2.34 (1.23-4.45) < 65 years 2.8 (1.46-5.35) ≥65 years Maclean, 2022 (22) 2.96 (2.2-4) All-cause mortality 4 (4A and 4B) Huang, 2021 (27) 2.761 (1.961-3.887) 4B and 5 Senoner, 2020 (16) 2.97 (1.59-5.57) Overall Finck, 2019 (b) (26) 2.03 (1.44-2.86) non-diabetic 1.72 (0.98-3.01) diabetic CAD-RADS: Coronary artery disease-reporting and data system; CI: confidence interval. cantly by the inclusion of this article and thus, the record was not excluded. The characteristics of the included studies are demonstrated in Table 2. 3.2. Value of CAD-RADS in prediction of MACE 14 out of the 16 included articles assessed the value of CAD- RADS categories for the prediction of MACE (12-25). 10 arti- cles had reports of combinations of the original CAD-RADS categories into a subset or overall category, which were not included in the meta-analysis and are reported separately (13, 15-17, 20-25). 8 articles (12-19) were included in the meta-analysis for the evaluation of the value of original CAD-RADS categoriza- tion for the prediction of MACE. The results of our analysis demonstrated that all the original CAD-RADS categories (1, 2, 3, 4A, 4B, 5) can predict MACE in CAD patients. The haz- ard ratios of CADS-RADS 1, 2, 3, 4a, 4b, and 5 for prediction of MACE were 3.39 (95% CI: 2.19-5.23), 4.19, (95% CI: 2.93-5.99), 5.99 (95% CI: 3.83-9.38), 7.29 (95% CI: 3.54-15.02), 6.27 (95% CI: 5.02-7.84), and 8.63 (95% CI: 4.67-15.95), respectively. All the analyses were statistically significant (p < 0.0001) (Fig- ure 2), with an increase in trend in the risk of MACE across CAD-RADS categories (Regression co-efficient = 0.066; 95% CI: 0.018-0.114; p = 0.007). The follow-up of the studies ranged between 3 months to 10 years. A meta-regression analysis was performed to assess the effect of follow-up duration on the predictive value of CAD-RADS for MACE. The results showed that the difference between follow-up durations had no significant effect on the predictive value of CAD-RADS in any of the categories (Sup- plementary Table 1). 3.3. Value of CAD-RADS in the prediction of all- cause mortality 4 out of the 16 included articles evaluated the value of CAD- RADS in the prediction of all-cause mortality (16, 19, 26, 27). One article (26) only reported the results for combinations of the original CAD-RADS categories into a subset or overall category, which was not included in the meta-analysis and is reported qualitatively. 3 articles (16, 19, 27) had assessed the value of the original CAD-RADS categories in the prediction of all-cause mortal- ity. The results of the analysis showed that all CAD-RADS cat- egories, except CAD-RADS 1, can be a predictor of all-cause mortality. The hazard ratios of CADS-RADS 1, 2, 3, 4a, 4b, and 5 for prediction of all-cause mortality were 1.50 (95% CI: 0.96- 2.34, p = 0.073), 1.85 (95% CI: 1.28-2.68, p = 0.001), 1.65 (95% CI: 1.31-2.07, p < 0.0001), 1.98 (95% CI: 1.26-3.09, p = 0.003), 2.78 (95% CI: 1.64-4.71, p < 0.0001), and 3.09 (95% CI: 1.91- 5.01, p < 0.0001), respectively (Figure 3). The analysis showed This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index 5 Archives of Academic Emergency Medicine. 2023; 11(1): e45 an increasing trend in the risk of all-cause mortality across CAD-RADS categories (Regression co-efficient = 0.046; 95% CI: 0.008-0.084; p = 0.018). Although it should be noted that the results of CAD-RADS categories 4a, 4b, and 5 should be interpreted with caution due to the limited number of studies in the respective analy- ses. The follow-up of the included studies varied between 4 to 10 years. A meta-regression analysis was performed to assess the effect of follow-up length on the predictive value of CAD- RADS for all-cause mortality. Due to the limited number of included studies, the analyses could only be performed on CAD-RADS categories 1, 2, and 3, none of which showed re- lations to the follow-up duration (Supplementary Table 2). 3.4. Miscellaneous CAD-RADS reporting 10 articles had reports of combinations of the original CAD- RADS categories for the prediction of MACE (13, 15-17, 20- 25). Lee et. al (15) demonstrated that a combined category of CAD-RADS 1 and 2 could not predict 3-month (HR = 2.4, 95% CI: 0.8-6.9) and 31.5-month (HR = 1.5, 95% CI: 0.7-3.2) MACE. Van Rosendael et al. (24) reported that moderate CAD (con- sisting of CAD-RADS 2 and 3) was not a predictor of MACE (HR = 1.95, 95% CI: 1.19-3.2). Two studies utilized a category of CAD-RADS ≥ 3 with con- flicting results. Duguay et al. (20) reported a hazard ratio of 3.12 (95% CI: 1.03-10.17) for this combined category in the prediction of MACE, while Yamamato et al. (25) did not re- port a predictive value of CAD-RADS for MACE (HR = 1.3, 95% CI: 0.16-11.1). CAD-RADS 4 (combined category of CAD-RADS 4A and 4B) was utilized by two studies (16, 17), and was shown to be a predictor of MACE (HR = 36.48, 95% CI: 4.94-269.67 and HR = 13.49, 95% CI: 4.809-37.86). 5 experiments had data on a combined category of CAD-RADS 4A, 4B and 5 (15, 23- 25). Four of these experiments (15, 23, 24) showed that this combined CAD-RADS category can predict MACE. Two stud- ies combined CAD-RADS categories 4b and 5 and reported conflicting results. Bittner et al. (13) reported a hazard ra- tio of 21.84 (95% CI: 8.63-55.26) while Yamamoto et al. (25) reported a hazard ratio of 2.6 (95% CI: 0.72-9.2) for the pre- dictive value of this subset category for MACE. Two studies reported results for the predictive value of overall CAD-RADS. Faber et al. (21) demonstrated that CAD-RADS as a continuous variable can be predictive of MACE in males (HR = 1.97, 95% CI: 1.12-3.45), females (HR = 5.34, 95% CI: 2.42-11.8), patients < 65 years of age (HR = 2.34, 95% CI: 1.23- 4.45) and patients ≥ 65 years of age (HR = 2.8, 95% CI: 1.46- 5.35). Maclean et al. (22) reported a hazard ratio of 2.96 (95% CI: 2.2-4) for the predictive value of overall CAD-RADS for the prediction of MACE. Three articles (16, 26, 27) had reported combinations of the original CAD-RADS categories for prediction of all-cause mortality. Senoner et. al (16) reported that CAD-RADS 4 (combined CAD-RADS 4A and 4B) can predict all-cause mor- tality (HR = 2.97, 95% CI: 1.59-5.57). Huang et. al (27) also reported that combined CAD-RADS 4B and 5 can be predic- tive of all-cause mortality (HR = 2.761, 95% CI: 1.961-3.887). Finck et al. (26) reported that while an overall CAD-RADS cat- egory can be predictive of all-cause mortality in non-diabetic patients (HR = 2.03, 95% CI: 1.44-2.86), it does not predict all-cause mortality in non-diabetic patients (HR = 1.72, 95% CI: 0.98-3.01). Although it should be noted that the diabetic group consisted of only 132 patients. The results of the anal- yses on the miscellaneous CAD-RADS reporting are demon- strated in table 3. 3.5. Quality assessment The risk of bias was evaluated according to the guidelines of QUAPAS. The risk of bias was assessed to be unclear in the domain of patient selection in four studies due to no report of the sampling method. Seven articles were judged to have an unclear risk of bias in the domain of index test due to no mention of the specialty of the assessor. Two articles were evaluated as unclear in risk of bias in the domain of outcome due to no report of MACE definition and the source of gath- ered data (registries, medical records, etc.). All articles were judged to have a high risk of bias in the flow and timing do- main. However, considering that according to guidelines the treatment of coronary artery disease patients varies depend- ing on its severity, clinical management of CAD patients can- not be identical and thus, the risk of bias in the domain of flow and timing was decided to be excluded from the judg- ment of overall risk of bias. One study was found to have a high risk of bias in the domain of analysis due to a high loss to follow-up rate. Studies were rated as low in all other do- mains of risk of bias assessment. One study was judged to have unclear applicability in the domain of outcome due to non-informative outcome definition. Studies were assessed to have no concerns in applicability in other domains (Table 4). 3.6. Certainty of evidence All included studies were observational studies and the base level of evidence was set as low. The level of evidence for the outcome of MACE was reduced by two grades due to consid- erable risk of bias and observed heterogeneity in the analysis. It was increased by two grades due to the observed large mag- nitude of effect (HR > 2) and possible dose-response gradient observed and thus, the level of evidence for the outcome of MACE was judged to be low. The level of evidence for the outcome of all-cause mortality was decreased by two grades due to the considerable risk of bias and imprecision (wide CIs) and increased by one due to the possible dose-response This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index K. Ahmadzadeh et al. 6 Table 4: Risk of bias assessment Study, year Risk of Bias Applicability Overall Patient selection Index test Outcome Flow and timing* Analysis Patient selection Index test Outcome Flow and timing Altay, 2021 (12) Unclear Low Unclear High Low Low Low Unclear Low Some concern Bittner, 2020 (13) Low Low Low High Low Low Low Low Low Low Duguay, 2017 (20) Unclear Low Low High Low Low Low Low Low Some concern Faber, 2021 (21) Low Unclear Low High Low Low Low Low Low Some concern Finck, 2019 (a) (14) Low Unclear Low High Low Low Low Low Low Some concern Finck, 2019 (b) (26) Low Unclear Low High Low Low Low Low Low Some concern Huang, 2021 (27) Low Unclear Low High Low Low Low Low Low Some concern Lee, 2021 (15) Unclear Low Low High Low Low Low Low Low Some concern Maclean, 2022 (22) Low Unclear Low High Low Low Low Low Low Some concern Mangalesh, 2022 (23) Unclear Low Unclear High Low Low Low Low Low Some concern Senoner, 2020 (16) Low Low Low High Low Low Low Low Low Low Tang, 2022 (17) Low Low Low High High Low Low Low Low Some concern Van Rosendael, 2019 (24) Low Low Low High Low Low Low Low Low Low Williams, 2020 (18) Low Unclear Low High Low Low Low Low Low Some concern Xie, 2018 (19) Low Unclear Low High Low Low Low Low Low Some concern Yamamoto, 2021 (25) Low Low Low High Low Low Low Low Low Low *Flow and timing domain was judged to be of high risk of bias due to differences in treatment of coronary artery disease patients; However, considering that according to guidelines the treatment of coronary artery disease patients varies depending on its severity, the bias in flow and timing domain was excluded from the judgment of overall risk of bias. gradient. The level of evidence for the outcome of all-cause mortality was judged to be very low (Table 5). 4. Discussion This systematic review and meta-analysis examined the ef- fectiveness of hierarchical CAD-RADS categorization in pre- dicting MACE and all-cause mortality in CAD patients. Ac- cording to our analysis, all original CAD-RADS categories can be used as a predictor of MACE, and all categories, except CAD-RADS 1, can be a predictor of all-cause mortality. More- over, patients with higher CAD-RADS categories had a higher risk of unfavorable outcomes. Our results confirm that CAD- RADS categorization of CCTA findings is valid for prognosti- cation of CAD patients’ outcomes, which is consistent with other scores such as modified Duke index (28). The results of our analysis demonstrate that CAD-RADS cat- egories 1 and 2, which are representative of non-obstructive CAD patients (less than 50% stenosis), are associated with an increased risk of unfavorable outcomes. Studies have shown that non-obstructive CAD is attributable to MACE in acute coronary syndrome patients treated with percutaneous coro- nary intervention and that non-obstructive CAD in CCTA should be considered as a clinically important finding (29, 30). This implies that the previous concept of dichotomiza- tion of CAD patients into an obstructive and non-obstructive group, may not be informative enough for the prediction of patient outcomes, and preventative and treatment strategy decision-making should not solely rely on the degree of ob- struction and various variables such as plaque characteristics should also be taken into account. Contrary to the results of the meta-analysis of the original CAD-RADS categories, the results of the articles utilizing a combination of CAD-RADS categories were less conclusive on their predictive value due to having wider CIs and reports of insignificant predictive value for higher CAD-RADS cate- gories. However, it should be noted that the references for the analyses of the combined CAD-RADS categories differed among studies and the studies not demonstrating a predic- This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index 7 Archives of Academic Emergency Medicine. 2023; 11(1): e45 Table 5: Certainty of evidence Total knowledge score Hazard Ratio (95% CI) Certainty Number of studies Risk of bias Inconsistency Indirectness Imprecision Other considerations MACE (follow-up: range 1 month to 10 years) 8 studies Serious Seriousa Not serious Not serious Large magnitude of effect Possible dose response gradient CAD-RADS 1: 3.39 (95% CI: 2.19-5.23) CAD-RADS 2: 4.19 (95% CI: 2.93-5.99) CAD-RADS 3: 5.99 (95% CI: 3.83-9.38) CAD-RADS 4A: 7.29 (95% CI: 3.54-15.02) CAD-RADS 4B: 6.27 (95% CI: 5.02-7.84) CAD-RADS 5: 8.63 (95% CI: 4.67-15.95) ⊕⊕©© Low All-cause mortality (follow-up: range 6 months to 8 years) 3 studies Serious Not serious Not serious Seriousb Possible dose response gradient CAD-RADS 1: 1.50 (95% CI: 0.96-2.34) CAD-RADS 2: 1.85 (95% CI: 1.28-2.68) CAD-RADS 3: 1.65 (95% CI: 1.31-2.07) CAD-RADS 4A: 1.98 (95% CI: 1.26-3.09) CAD-RADS 4B: 2.78 (95% CI: 1.64-4.71) CAD-RADS 5: 3.09 (95% CI: 1.91-5.01) ⊕©©© Very low CAD-RADS: Coronary artery disease-reporting and data system; MACE: Major adverse cardiac event; CI: Confidence interval. a. There was considerable heterogeneity among the studies. b. Wide CIs tive value for the CAD-RADS category had smaller sample sizes. This finding implies that CAD-RADS might be better utilized as the original separate categories and attempts of combining these categories (even only combining categories 4A and 4B) could reduce the accuracy of the classification system. However, it should be kept in mind that few stud- ies with limited sample sizes had evaluated combinations of CAD-RADS categories. The included studies had different durations of follow-up, however, our analysis indicated that follow-up had no effect on the predictive value of CAD-RADS for MACE and all-cause mortality. Faber et al. (21) evaluated the predictive value of CAD-RADS for MACE in male and female populations and demonstrated that CAD-RADS as a continuous variable might be a bet- ter predictive tool in the female rather than the male pop- ulation. Articles have demonstrated that the female CAD population has higher mortality than the male counterparts, which could explain the higher HR reported for the female population (31). Further studies could shed more light on the differences in the predictive value of CAD-RADS in male and female populations. The fundamental parameter for the categorization of CCTA in CAD-RADS is the luminal diameter of the most stenotic vessel and post-CCTA recommendation on the further di- agnostic test and management depends on the mere CAD- RADS category. Unlike other scoring indices such as the Duke index, CAD-RADS ignores the number of involved vessels and the location of the culprit lesion. As a result, the CAD- RADS category cannot replace the CCTA report and should be interpreted in conjunction with the detailed CCTA re- port. Evidence indicates that along with dimensional param- eters of coronary atherosclerotic lesions, plaque characteris- tics also play a pivotal role in risk prediction of CAD patients (32, 33). CAD-RADS reflects plaque characterization through an extra modifier that assesses the presence of vulnerable or high-risk plaque. In our review, due to the unavailability of data, we could not perform an analysis on the effect of such This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index K. Ahmadzadeh et al. 8 additional modifiers on patients’ risk of MACE and mortality. It can be assumed that detailed categorization of CCTA find- ings in addition to CAD-RADS scoring may yield more prog- nostic information than what is presented in our study. Al- though the CAD-RADS was not designed to take the place of the complete descriptive reports of CCTA, the lack of incor- poration of such contributing factors may render CCTA’s rich data obsolete. Nevertheless, the structuralized reporting of CCTA using CAD-RADS and the discrete prognostic value of each category will lead to better patient care by consolidat- ing the communication between referring and interpreting physicians (34). To improve the representativeness of CAD- RADS categorization, CAD-RADS 2.0 (35) was introduced in 2022, which incorporates parameters reflecting plaque bur- den and ischemia in the form of modifiers added to stenosis assessment. This could improve the decision-making pro- cess for CAD patients by better representing the extent of CAD and the lesions’ characteristics. Yet, CAD-RADS’s short- coming remains due to the fact that it cannot be used to eval- uate the number of involved vessels. This reduces the effec- tiveness of the CAD-RADS classification system in predicting outcomes and progression in patients with multiple vessels involved (36). With the latest advancements in image processing, there is potential for the addition of other parameters related to atherosclerosis pathogenesis, including pericardial fat atten- uation (37), to enhance the precision of prognostic and di- agnostic performance of the CAD-RADS scoring system. In this regard, machine learning was shown to be promising in integrating multiplex CCTA parameters with both other per- formed test variables and patient characteristics. This would enable individualized and highly accurate risk prediction, which would significantly impact delivering optimal patient care (38, 39). Moreover, besides risk stratification, the CAD-RADS scor- ing system recommends further test studies and treatment plans for each CAD-RADS category. Although these recom- mendations are derived from expert consensus, since there currently are scarce evidence on the treatment strategies of CAD-RADS guidelines, the treating physician should imple- ment an individualized treatment plan for each patient and not solely rely on the recommendations provided by CAD- RADS guidelines. Future studies should investigate the effec- tiveness of these treatment plans on the disease progression of patients. 5. Limitations We acknowledge that our study has limitations. The defini- tion of MACE varied between the studies, which limits ac- curate comparison of the studies. A recent systematic re- view (40) has shown that only a limited number of studies match the conventional MACE definition of acute myocar- dial infarction, stroke, and cardiovascular death. Since the definitions vary slightly between studies, all reports of MACE were pooled and analyzed together. Also, not all the included studies specified whether the included patients were acute or chronic CAD patients or whether the patients were symp- tomatic or asymptomatic, which might lead to different prog- noses and treatment plans. Further studies should more ac- curately specify the patient population and compare the pre- dictive capabilities of CAD-RADS in acute and chronic CAD patients. Also, a major contributor to the outcome of CAD patients is the treatment strategies devised for the patients, which was not reported in the included articles and may vary depending on national and institutional guidelines. 6. Conclusion Low to very low levels of evidence demonstrated that the CAD-RADS classification system can be used to predict out- comes of MACE and all-cause mortality. This indicates that the implementation of CAD-RADS into clinical practice, besides enhancing the communication between physicians and improving patient care, can also guide physicians in risk assessment of the patients’ prognosis. Further well-designed clinical trials or prospective cohort studies are needed to pro- vide a high level of evidence for predicting the value of CAD- RADS in CAD patients. 7. Declarations 7.1. Acknowledgments Not applicable. 7.2. Conflict of interest The authors declare that they have no competing interests. 7.3. Fundings This study was funded by Shahid Beheshti University of Med- ical Sciences (Grant No. 43005143). 7.4. Authors’ contribution Study design: MY, RM Data gathering: MR, KA Analysis: MY, KA Interpretation of results: all authors Drafting and revising: all authors 7.5. Availability of data and materials The gathered data and checklist can be provided to qualified researchers with the intent of replicating the procedure and results. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index 9 Archives of Academic Emergency Medicine. 2023; 11(1): e45 References 1. Abdelrahman KM, Chen MY, Dey AK, Virmani R, Finn AV, Khamis RY, et al. Coronary Computed Tomography Angiography From Clinical Uses to Emerging Technolo- gies: JACC State-of-the-Art Review. J Am Coll Cardiol. 2020;76(10):1226-43. 2. Cordeiro MA, Lima JA. Atherosclerotic plaque character- ization by multidetector row computed tomography an- giography. J Am Coll Cardiol. 2006;47(8 Suppl):C40-7. 3. Chandrashekhar Y, Min JK, Hecht H, Narula J. CAD- RADS: A Giant First Step Toward a Common Lexicon? JACC Cardiovasc Imaging. 2016;9(9):1125-9. 4. Kumar P, Bhatia M. Coronary Artery Disease Reporting and Data System: A Comprehensive Review. J Cardiovasc Imaging. 2022;30(1):1-24. 5. Cury RC, Abbara S, Achenbach S, Agatston A, Berman DS, Budoff MJ, et al. CAD-RADS(TM) Coronary Artery Disease - Reporting and Data System. An expert con- sensus document of the Society of Cardiovascular Com- puted Tomography (SCCT), the American College of Ra- diology (ACR) and the North American Society for Car- diovascular Imaging (NASCI). Endorsed by the Ameri- can College of Cardiology. J Cardiovasc Comput Tomogr. 2016;10(4):269-81. 6. Maroules CD, Hamilton-Craig C, Branch K, Lee J, Cury RC, Maurovich-Horvat P, et al. Coronary artery dis- ease reporting and data system (CAD-RADS(TM)): Inter- observer agreement for assessment categories and mod- ifiers. J Cardiovasc Comput Tomogr. 2018;12(2):125-30. 7. Abdel Razek AAK, Elrakhawy MM, Yossof MM, Nageb HM. Inter-observer agreement of the Coronary Artery Disease Reporting and Data System (CAD-RADS(TM)) in patients with stable chest pain. Pol J Radiol. 2018;83:e151-e9. 8. Foldyna B, Szilveszter B, Scholtz JE, Banerji D, Maurovich-Horvat P, Hoffmann U. CAD-RADS - a new clinical decision support tool for coronary computed to- mography angiography. Eur Radiol. 2018;28(4):1365-72. 9. Lee J, Mulder F, Leeflang M, Wolff R, Whiting P, Bossuyt PM. QUAPAS: An Adaptation of the QUADAS-2 Tool to Assess Prognostic Accuracy Studies. Ann Intern Med. 2022;175(7):1010-8. 10. Schunemann H, Higgins J, Vist G, Glasziou P, Akl E, Skoetz N, et al. Completing ‘Summary of findings’ ta- bles and grading the certainty of the evidence. Cochrane Handbook for Systematic Reviews of Interventions. 2019:375-402. 11. Page M, Higgins J, Sterne J. Assessing risk of bias due to missing results in a synthesis. Cochrane Handbook for Systematic Reviews of Interventions: Wiley; 2019. p. 349- 74. 12. Altay S. Prognostic Value of Standard Coronary Com- puted Tomography Angiography Reporting System (CAD-RADS). Indian J Radiol Imaging. 2021;31(1):37-42. 13. Bittner DO, Mayrhofer T, Budoff M, Szilveszter B, Fol- dyna B, Hallett TR, et al. Prognostic Value of Coronary CTA in Stable Chest Pain: CAD-RADS, CAC, and Cardio- vascular Events in PROMISE. JACC Cardiovasc Imaging. 2020;13(7):1534-45. 14. Finck T, Hardenberg J, Will A, Hendrich E, Haller B, Mar- tinoff S, et al. 10-Year Follow-Up After Coronary Com- puted Tomography Angiography in Patients With Sus- pected Coronary Artery Disease. JACC Cardiovasc Imag- ing. 2019 (a);12(7 Pt 2):1330-8. 15. Lee JW, Kim JY, Han K, Im DJ, Lee KH, Kim TH, et al. Coro- nary CT Angiography CAD-RADS versus Coronary Artery Calcium Score in Patients with Acute Chest Pain. Radiol- ogy. 2021;301(1):81-90. 16. Senoner T, Plank F, Barbieri F, Beyer C, Birkl K, Wid- mann G, et al. Added value of high-risk plaque criteria by coronary CTA for prediction of long-term outcomes. Atherosclerosis. 2020;300:26-33. 17. Tang CX, Qiao HY, Zhang XL, Di Jiang M, Schoepf UJ, Rudzinski PN, et al. Functional CAD-RADS using FFR(CT) on therapeutic management and prognosis in patients with coronary artery disease. Eur Radiol. 2022;32(8):5210-21. 18. Williams MC, Moss A, Dweck M, Hunter A, Pawade T, Adamson PD, et al. Standardized reporting systems for computed tomography coronary angiography and cal- cium scoring: A real-world validation of CAD-RADS and CAC-DRS in patients with stable chest pain. J Cardiovasc Comput Tomogr. 2020;14(1):3-11. 19. Xie JX, Cury RC, Leipsic J, Crim MT, Berman DS, Gransar H, et al. The Coronary Artery Disease-Reporting and Data System (CAD-RADS): Prognostic and Clinical Implica- tions Associated With Standardized Coronary Computed Tomography Angiography Reporting. JACC Cardiovasc Imaging. 2018;11(1):78-89. 20. Duguay TM, Tesche C, Vliegenthart R, De Cecco CN, Lin H, Albrecht MH, et al. Coronary Computed Tomographic Angiography-Derived Fractional Flow Reserve Based on Machine Learning for Risk Stratification of Non-Culprit Coronary Narrowings in Patients with Acute Coronary Syndrome. Am J Cardiol. 2017;120(8):1260-6. 21. Faber M, Will A, Hendrich E, Martinoff S, Hadamitzky M. Sex- and age-specific differences in the long-term prog- nostic value of morphological plaque features detected by coronary computed tomography angiography. J Car- diovasc Comput Tomogr. 2021;15(3):274-80. 22. Maclean E, Mahon C, Sehmi J, Kanaganayagam G, Ngee TT, Nicol ED. CT multivessel aggregate stenosis score: A novel point-of-care tool for predicting major This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index K. Ahmadzadeh et al. 10 adverse cardiac events. J Cardiovasc Comput Tomogr. 2022;16(4):350-4. 23. Mangalesh S, Yadav P, Dudani S, Mahesh NK. Athero- genic index of plasma predicts coronary artery dis- ease severity and major adverse cardiac events in ab- sence of conventional risk factors. Coron Artery Dis. 2022;33(7):523-30. 24. van Rosendael AR, Shaw LJ, Xie JX, Dimitriu-Leen AC, Smit JM, Scholte AJ, et al. Superior Risk Stratification With Coronary Computed Tomography Angiography Us- ing a Comprehensive Atherosclerotic Risk Score. JACC Cardiovasc Imaging. 2019;12(10):1987-97. 25. Yamamoto A, Nagao M, Ando K, Nakao R, Fukushima K, Matsuo Y, et al. Risk stratification in coronary artery disease using NH(3)-PET myocardial flow reserve and CAD-RADS on coronary CT angiography. Int J Cardiovasc Imaging. 2021;37(11):3335-42. 26. Finck T, Will A, Hendrich E, Martinoff S, Hadamitzky M. Coronary computed tomography angiography as a tool for long-term cardiovascular risk stratification in diabetic patients. Heart Vessels. 2019 (b);34(7):1086-95. 27. Huang Z, Zhang S, Jin N, Hu Y, Xiao J, Li Z, et al. Prog- nostic value of CAD-RADS classification by coronary CTA in patients with suspected CAD. BMC Cardiovasc Disord. 2021;21(1):476. 28. Min JK, Shaw LJ, Devereux RB, Okin PM, Weinsaft JW, Russo DJ, et al. Prognostic value of multidetec- tor coronary computed tomographic angiography for prediction of all-cause mortality. J Am Coll Cardiol. 2007;50(12):1161-70. 29. Wang ZJ, Zhang LL, Elmariah S, Han HY, Zhou YJ. Preva- lence and Prognosis of Nonobstructive Coronary Artery Disease in Patients Undergoing Coronary Angiography or Coronary Computed Tomography Angiography: A Meta-Analysis. Mayo Clin Proc. 2017;92(3):329-46. 30. Bittencourt MS, Hulten E, Ghoshhajra B, O’Leary D, Christman MP, Montana P, et al. Prognostic value of nonobstructive and obstructive coronary artery disease detected by coronary computed tomography angiogra- phy to identify cardiovascular events. Circ Cardiovasc Imaging. 2014;7(2):282-91. 31. Group EUCCS, Regitz-Zagrosek V, Oertelt-Prigione S, Prescott E, Franconi F, Gerdts E, et al. Gender in car- diovascular diseases: impact on clinical manifestations, management, and outcomes. Eur Heart J. 2016;37(1):24- 34. 32. Gossl M, Versari D, Hildebrandt H, Mannheim D, Olson ML, Lerman LO, et al. Vulnerable plaque: detection and management. Med Clin North Am. 2007;91(4):573-601. 33. Finn AV, Nakano M, Narula J, Kolodgie FD, Virmani R. Concept of vulnerable/unstable plaque. Arterioscler Thromb Vasc Biol. 2010;30(7):1282-92. 34. Canan A, Ranganath P, Goerne H, Abbara S, Landeras L, Rajiah P. CAD-RADS: Pushing the Limits. Radiographics. 2020;40(3):629-52. 35. Cury RC, Leipsic J, Abbara S, Achenbach S, Berman D, Bittencourt M, et al. CAD-RADS 2.0 - 2022 Coronary Artery Disease-Reporting and Data System: An Expert Consensus Document of the Society of Cardiovascular Computed Tomography (SCCT), the American College of Cardiology (ACC), the American College of Radiol- ogy (ACR), and the North America Society of Cardiovas- cular Imaging (NASCI). J Cardiovasc Comput Tomogr. 2022;16(6):536-57. 36. Szilveszter B, Vattay B, Bossoussou M, Vecsey-Nagy M, Simon J, Merkely B, et al. CAD-RADS may underesti- mate coronary plaque progression as detected by se- rial CT angiography. Eur Heart J Cardiovasc Imaging. 2022;23(11):1530-9. 37. Antonopoulos AS, Sanna F, Sabharwal N, Thomas S, Oikonomou EK, Herdman L, et al. Detecting human coronary inflammation by imaging perivascular fat. Sci Transl Med. 2017;9(398):eaal2658. 38. Cormode DP, Roessl E, Thran A, Skajaa T, Gordon RE, Schlomka JP, et al. Atherosclerotic plaque composition: analysis with multicolor CT and targeted gold nanoparti- cles. Radiology. 2010;256(3):774-82. 39. Kolossvary M, Karady J, Szilveszter B, Kitslaar P, Hoff- mann U, Merkely B, et al. Radiomic Features Are Superior to Conventional Quantitative Computed Tomographic Metrics to Identify Coronary Plaques With Napkin-Ring Sign. Circ Cardiovasc Imaging. 2017;10(12):e006843. 40. Bosco E, Hsueh L, McConeghy KW, Gravenstein S, Saade E. Major adverse cardiovascular event definitions used in observational analysis of administrative databases: a sys- tematic review. BMC Med Res Methodol. 2021;21(1):241. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index 11 Archives of Academic Emergency Medicine. 2023; 11(1): e45 Figure 1: Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) diagram. Supplementary Material 1: Search strategy for the “Evaluation of the value of Coronary Artery Disease – Reporting and Data System (CAD- RADS) in outcome prediction of coronary artery disease patients” (Septemeber 19t h , 2022) PubMed: CAD-RADS [tiab] OR [tiab] CAD RADS[tiab] OR CADRADS[tiab] OR Coronary Artery Disease Reporting and Data System[tiab] OR coronary artery disease reporting[tiab] Embase: ‘Coronary Artery Disease Reporting and Data System’/exp OR ‘CAD-RADS’:ab,ti OR ‘CAD RADS’:ab,ti OR ‘CADRADS’:ab,ti OR ‘Coro- nary Artery Disease Reporting’:ab,ti Web of Science: (ALL=("Coronary Artery Disease Reporting and Data System" OR "CAD-RADS" OR "CAD RADS" OR "CADRADS" OR "Coronary Artery Disease Reporting")) Scopus: TITLE-ABS-KEY ("Coronary Artery Disease Reporting and Data System" OR "CAD-RADS" OR "CAD RADS" OR "CADRADS" OR "Coro- nary Artery Disease Reporting") Supplementary Table 1: Meta-regression for evaluation of effect of follow-up on predictive performance of CAD-RADS for major adverse cardiovascular event (MACE) Variable Meta-regression coefficient 95% Confidence interval P value CAD-RADS 1 1.00975 0.99771, 1.02193 0.113 CAD-RADS 2 1.00461 0.99264, 1.01671 0.452 CAD-RADS 3 1.00492 0.99291, 1.01708 0.423 CAD-RADS 4a 1.00490 0.97994, 1.03050 0.703 CAD-RADS 4b 1.00441 0.99411, 1.01482 0.402 CAD-RADS 5 1.01510 0.99035, 1.04047 0.234 CAD-RADS: Coronary artery disease-reporting and data system. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index K. Ahmadzadeh et al. 12 Figure 2: Value of CAD-RADS classification system in prediction of major adverse cardiovascular events in coronary artery disease patients. CAD-RADS: Coronary artery disease-reporting and data system; CI: Confidence interval; HR: Hazard ratio. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index 13 Archives of Academic Emergency Medicine. 2023; 11(1): e45 Figure 3: Value of CAD-RADS classification system in prediction of all-cause mortality in coronary artery disease patients. CAD-RADS: Coro- nary artery disease-reporting and data system; CI: Confidence interval; HR: Hazard ratio. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index K. Ahmadzadeh et al. 14 Supplementary Table 2: Meta-regression for evaluation of effect of follow-up on predictive performance of CAD-RADS for all-cause mortality Variable Meta-regression coefficient 95% Confidence interval P value CAD-RADS 1 1.00573 0.98776, 1.02402 0.534 CAD-RADS 2 1.00615 0.99350, 1.01897 0.342 CAD-RADS 3 1.00551 0.99266, 1.01853 0.402 CAD-RADS 4a Not assessable due to limited number of studies CAD-RADS 4b Not assessable due to limited number of studies CAD-RADS 5 Not assessable due to limited number of studies CAD-RADS: Coronary artery disease-reporting and data system. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index Introduction Methods Results Discussion Limitations Conclusion Declarations References