Clinical impact of low-radiation computed tomography coronary angiography diagnosis for coronary artery stenosis

Abstract Background: The objective of this study aims to assess the clinic impact of low-radiation computed tomography coronary angiography (LR-CTCA) diagnosis for coronary artery stenosis (CAS). Methods: This study will comprehensively search the following electronic databases from inception to the present: PUBMED, EMBASE, Cochrane Library, PsycINFO, Web of Science, Google, Allied and Complementary Medicine Database, Chinese Biomedical Literature Database, VIP database, WANGFANG, and China National Knowledge Infrastructure. All these electronic databases will be searched without language restrictions. All case-controlled studies on assessing the clinical impact of LR-CTCA diagnosis for patients with CAS will be included. Quality Assessment of Diagnostic Accuracy Studies tool will be utilized to evaluate the methodological quality for each qualified studies. Results: We will assess the clinic impact of LR-CTCA diagnosis for CAS by measuring sensitivity, specificity, positive likelihood ratio, negative likelihood ratio, and diagnostic odds ratio. Conclusion: The results of this study will summarize the latest evidence of LR-CTCA diagnosis for CAS. Systematic review registration: PROSPERO CRD42019139336.


Study registration
The protocol of this study has been registered on PROSPERO (CRD42019139336). The study will be reported in accordance to the guideline of preferred reporting items for systematic reviews and meta-analysis (PRISMA) protocol statement. [36]

Ethics and dissemination
We will not analyze individual patient data, thus, no ethic approval is required. We will plan to publish results of this study via peer-reviewed journals or conference proceedings.
2.3. Eligibility criteria 2.3.1. Types of studies. All case-controlled studies (CCSs) on assessing clinical value of LR-CTCA diagnosis for CAS will be considered for inclusion.
2.3.2. Types of patients. We will include patients with invasive coronary angiography diagnosis of CAS in this study regardless the gender, age, and region.

Type of index test.
Index test: LR-CTCA diagnosis for CAS has been utilized in the intervention group.
Reference test: invasive coronary angiography diagnosis for CAS has been used in the control group.
2.3.4. Types of outcome measurements. In this study, we will measure sensitivity, specificity, positive likelihood ratio, negative likelihood ratio, and diagnostic odds ratio.
2.4. Search methods for identification of studies 2.4.1. Electronic searches. In this study, the databases of PUBMED, EMBASE, Cochrane Library, PsycINFO, Web of Science, Google, Allied and Complementary Medicine Database, Chinese Biomedical Literature Database, VIP database, WANG-FANG, and China National Knowledge Infrastructure will be comprehensively searched from their inception to the present regardless language restrictions. We will consider all CCSs of LR-CTCA diagnosis for CAS. A detailed example of search strategy for PUBMED is presented in Table 1. We will adopt similar strategies to all other electronic databases.

Other resources.
Other resources, such as dissertations, conference proceedings, and reference lists of relevant reviews will be included.

Data collection 2.5.1. Selection of studies.
Two independent authors will screen study eligibility according to the inclusion criteria. All divergences between 2 authors will be solved by consensus with a third author involved. The whole process consists of 2 stages. At first stage, all records will be scanned, and all duplicated and irrelevant studies will be excluded. At the second stage, full texts of the remaining studies will be read to further judge if they meet all eligibility criteria. All excluded studies will be recorded for specific reasons. The results of study selection will be shown in the PRISMA flow chart.
2.5.2. Data collection and management. Two independent authors will extract data from all eligible studies based on the previously developed data extraction form. A third author will help to resolve all disagreements between 2 authors via discussion. The following information will be extracted: (1) Study characteristics: first author, year of publication, region, design, study setting, sample size, and so on; (2) Patient characteristics: patient demographics, inclusion and exclusion criteria, and so on; (3) Interventional and reference tests: time, methods of samples, study period, and so on; (4) Outcomes: true positives, false positives, and so on.

2.5.3.
Dealing with missing data. Whenever there is insufficient or missing data, we will directly contact primary authors of original studies to require those data. If we cannot get back those data, we will analyze the available data only.

Methodological quality assessment
Two independent authors will assess methodological quality for each qualified study using the Quality Assessment of Diagnostic Accuracy Studies tool. [37] Any different opinions regarding the methodological quality assessment between 2 authors will be solved by a third experienced author through discussion.

Statistical analysis
RevMan V.5.3 and Stata 12.0 software will be used to carry out statistical analysis. We will calculate descriptive statistics and 95% confidence intervals. Additionally, we will also operate descriptive forest plot and a summary receiver operating characteristic plot.
2.7.1. Assessment of heterogeneity. We will use I 2 statistic to investigate heterogeneity among eligible studies. The value of I 2 50% means low heterogeneity. On the other hand, the value of I 2 > 50% means significant heterogeneity.

Data synthesis.
If there is low heterogeneity (I 2 50%), we will pool the data and carry out meta-analysis. If there is significant heterogeneity (I 2 > 50%), we will carry out subgroup analysis. We will pool the data and perform meta-analysis if there is low heterogeneity after subgroup analysis. Otherwise, we will not carry out meta-analysis if there is significant heterogeneity after subgroup analysis. In addition, we will carry out the bivariate random-effects regression to summarize the estimates of sensitivity and specificity.
2.8. Additional analysis 2.8.1. Subgroup analysis. We will conduct subgroup analysis based on the different study characteristics, and patients. Table 1 Search strategy for PUBMED database.

Number
Search terms 1 Coronary artery disease 2 CAD 3 Atherosclerotic heart disease 4 Coronary artery stenosis 5 Heart disease 6 Chest pain 7 Shortness of breath 8 Heart attack 9 Ischemic heart disease 10 Or 1-9 11 Computed tomography 12 Coronary computed tomography angiography 13 Coronary angiography 14 Low -radiation computed tomography  15  CT  16  CCTA  17  Sensitivity  18  Specificity  19 Or 11- 18 20 Case-control studies 21 Case -control  22  Studies  23  Trials  24  Case  25  Control  26 Or 20-25 27 10 and 19 and 26 Li and Zeng Medicine (2019) 98:46 Medicine 2.8.2. Sensitivity analysis. We will carry out sensitivity analysis to explore the stability and robustness of pooled outcome results by removing the low methodological quality studies.

Reporting bias.
We will perform funnel plots and relevant regression tests [38] to check if there is reporting bias in this study.

Discussion
Although many studies utilized LR-CTCA diagnosis for CAS, there is no comprehensive systematic review comparing the diagnosis accurate of LR-CTCA with other diagnosis tools. We hope this study will provide the most current available evidence to present whether LR-CTCA diagnosis is more accurate than other diagnosis tools in the patients diagnosed as CAS. The results of this study will provide helpful evidence for the LR-CTCA diagnosis in patients with CAS. However, this study may still have several limitations. First, although this study tries to search literature records comprehensively, this study may still miss some potential studies. Second, different methodological qualities may cause high heterogeneity.