“…One study used both CNS‐originating EVs and general EVs for ROC analysis, and we did not exclude it (Yan et al., 2022). Among the studies using general EVs, thirty of those studies did not include an ROC analysis (Athauda et al., 2019; Bhattacharyya et al., 2022; Cerri et al., 2018; Chan et al., 2023; Chan et al., 2021; Chan et al., 2023; Chung et al., 2020; Elkouris et al., 2019; Fraser et al., 2016; Han et al., 2019; Herman et al., 2023; Ho et al., 2014; Hong et al., 2023; Kitamura et al., 2018; Lamontagne‐Proulx et al., 2019; Leng et al., 2020; Leng et al., 2022; Miyamoto et al., 2023; Sproviero et al., 2021; Sproviero et al., 2022; Taymans et al., 2023; Wang et al., 2017; Wang et al., 2018; Wang et al., 2020; Wang et al., 2022; Xia et al., 2019; Yang et al., 2015; Yang et al., 2018; Yeh et al., 2021; Zhao et al., 2020), six did not include sensitivity or specificity (Grossi et al., 2021; Gualerzi et al., 2019; He et al., 2021; Ohmichi et al., 2019; Ruf et al., 2021; Yao et al., 2018), and one study only performed the ROC analysis directly in serum instead of EVs (Citterio et al., 2023). In total, the meta‐analysis included 21 studies (Barbagallo et al., 2020; Cao et al., 2019; Cao et al., 2020; Cao et al., 2017; Chung et al., 2021a; Chung et al., 2021b; Dos Santos et al., 2018; Gui et al., 2015; Hadisurya et al., 2023; Hong et al., 2021; Lucien et al., 2022; Manna et al., 2021; Shim et al., 2021; Stuendl et al., 2016; Stuendl et al.,…”