Research Article

Diagnostic Accuracy of Machine Learning-Based Radiomics in Grading Gliomas: Systematic Review and Meta-Analysis

Table 2

Detailed QUADS-2 tool assessment after the two reviewers reached a consensus.

Authors and yearPatient selectionIndex testReference standardFlow and timing

Cho et al. 2018The experiment was designed to be a retrospective study and did not include a random sample. However, the study avoided inappropriate exclusion (high risk)The index test results were interpreted knowing the results of the reference standard. It was unclear whether a prespecified threshold was used (high risk)It was unclear whether the reference standard was likely to classify the target condition accurately. The reference standard results were interpreted knowing the results of the index test (low risk)It was unclear whether there was an appropriate interval between the index test and reference standard and whether all patients were included in the analysis. All patients received a reference standard, but it is unclear whether it was the same reference standard (unclear)

Tian et al. 2018The experiment was designed to be a retrospective study and did not include a random sample. However, the study avoided inappropriate exclusion (high risk)The index test results were interpreted knowing the results of the reference standard. However, a pre specified threshold was used (high risk)The reference standard was likely to classify the target condition accurately. The reference standard results were interpreted without knowing the results of the index test (low risk)It was clear whether there was an appropriate interval between the index test and reference standard. All patients received a reference standard, but it was unclear whether it was the same reference standard. Not all patients were included in the analysis (low risk)

Hashido et al. 2018The experiment was designed to be a retrospective study and did not include a random sample. However, the study avoided inappropriate exclusion (high risk)The index test results were interpreted without knowing the results of the reference standard. Furthermore, a pre specified threshold was used (low risk)The reference standard was likely to classify the target condition accurately. The reference standard results were interpreted without knowing the results of the index test (low risk)It was unclear whether there was an appropriate interval between the index test and reference standard. All patients received a reference standard, but it was unclear whether it was the same reference standard. Not all patients were included in the analysis (unclear)

Vamvakas et al. 2019The experiment was designed to be a retrospective study and did not include a random sample. In addition, it was unclear whether the study avoided inappropriate exclusion (high risk)The index test results were interpreted knowing the results of the reference standard. It was unclear whether a prespecified threshold was used (high risk)The reference standard was likely to classify the target condition accurately. The reference standard results were interpreted knowing the results of the index test (high risk)It was unclear whether there was an appropriate interval between the index test and reference standard and included all patients in the analysis. All patients received a reference standard, but it was unclear whether it was the same reference standard (unclear)

Zhao et al. 2020The experiment was designed to be a retrospective study and did not include a random sample. In addition, it was unclear whether the study avoided inappropriate exclusion (high risk)The index test results were interpreted without knowing the results of the reference standard. However, it was unclear whether a prespecified threshold was used (low risk)It was unclear whether the reference standard was likely to classify the target condition accurately. The reference standard results were interpreted without knowing the results of the index test (low risk)It was unclear whether there was an appropriate interval between the index test and reference standard. All patients received a reference standard, but it was unclear whether it was the same reference standard. Not all patients were included in the analysis (unclear)