Table 2.

Modified QUADAS-2 tool

QUADAS domain	Risk of bias questions
Domain 1: Patient selection	Study design (retrospective vs prospective) Patient selection Clear description of selection criteria Case description
Domain 2: Index test	Image acquisition description Image analysis description Region of interest and data extraction
Doman 3: Reference test	Tumor classification quality Treatment details description
Domain 4: Flow and timing	Timing of treatment and imaging described
Domain 5: Data analysis, processing, and reporting	Withdrawals and exclusions explained Statistical tests reported and sufficient Data availability

The quality of the reporting for each study was assessed in five domains. For each domain, one to three aspects were scored as having low, medium, or high risk of bias. A detailed explanation is provided in Supplementary Materials