Auraptene (AUR) and umbelliprenin (UMB) are naturally occurring prenylated coumarins that have demonstrated promising anticancer efects across various human cancer cell lines. This metaanalysis aimed to systematically assess, compare, and quantify the anticancer efcacy of AUR and UMB by synthesizing evidence from in vitro studies. A comprehensive literature search identifed 27 eligible studies investigating AUR or UMB against cancer cells. Mixed-efects models revealed signifcant negative associations between coumarin dose and viability for AUR (est.= − 2.27) and UMB (est. = − 3.990), underscoring their dose-dependent cytotoxicity. Meta-regression indicated slightly higher potency for UMB over AUR, potentially due to increased lipophilicity imparted by additional isoprenyl units. Machine learning approaches identifed coumarin dose and cancer type as the most infuential determinants of toxicity, while treatment duration and the specifc coumarin displayed weaker efects. Moderate (AUR) to substantial (UMB) between-study heterogeneity was detected, although the fndings proved robust. In summary, this meta-analysis establishes AUR and UMB as promising natural anticancer candidates with clear dose-toxicity relationships across diverse malignancies. The structural insights and quantifcations of anticancer efcacy can inform forthcoming eforts assessing therapeutic potential in pre-clinical models and human trials.

Keywords