Table 1.
Comparison of previous meta-analyses with the current study.
| Parameters | This study | Soffer et al [7] | Mohan et al [8] |
| Number of included studies | 20 studies on gastrointestinal ulcers and 19 studies on gastrointestinal hemorrhage | 5 studies on gastrointestinal ulcers and 5 studies on gastrointestinal hemorrhage | 9 studies for the diagnosis of gastrointestinal ulcers or hemorrhage (did not perform separate analysis between ulcers and hemorrhage) |
| Main outcome | Separate diagnostic performance of CADa models for the gastrointestinal ulcers or hemorrhage using WCEb | Separate diagnostic performance of CAD models for the gastrointestinal ulcers or hemorrhage using WCE | Pooled diagnostic performance of CAD models for gastrointestinal ulcers and hemorrhage using WCE (not a meta-analysis with DTAc; lack of consideration for the prevalence of ulcers or hemorrhage in each study and thus no calculation of TPd, FPe, FNf, or TNg in each study) |
| Search strategy | Search of MEDLINE through PubMed, Web of Science, and the Cochrane Library (2 independent authors searched the databases) | Search of MEDLINE through PubMed (2 independent authors searched the database) | Search of ClinicalTrials.gov, Ovid EBMh Reviews, Ovid, Embase, Ovid MEDLINE, Scopus, and Web of Science (a single medical librarian searched all the databases) |
| Inaccurate calculation (coding) of TP/FP/FN/TN | N/Ai | Inaccurate calculation detected in the study’s figures | Not a meta-analysis with DTA; lack of consideration for the prevalence of ulcers or hemorrhage in each study and thus no calculation of TP, FP, FN, or TN in each study |
| Determination of the heterogeneity between studies | Correlation coefficient between the logarithm of the sensitivity and specificity, beta of HSROCj model, visual examination of the SROC curve | I2 statistics (DTA meta-analysis did not determine heterogeneity with I2 statistic) | I2 statistic (DTA meta-analysis did not determine heterogeneity with I2 statistics) |
| Quality assessment | QUADAS-2k | QUADAS-2 | Not assessed |
| Publication bias | Deeks funnel plot asymmetry test | Not assessed | Not assessed |
aCAD: computer-aided diagnosis.
bWCE: wireless capsule endoscopy.
cDTA: diagnostic test accuracy.
dTP: true positive.
eFP: false positive.
fFN: false negative.
gTN: true negative.
hEBM: evidence-based medicine.
iN/A: not applicable.
jHSROC: hierarchical summary receiver operating characteristic.
kQUADAS-2: Quality Assessment of Diagnostic Accuracy Studies second version.