Figure 9.

Leave-one-out sensitivity analyses assessing the influence of individual studies on the pooled standardized mean difference (SMD) estimates for cytotoxicity outcomes at different time intervals. Each plot illustrates the recalculated pooled effect size following sequential omission of a single study under the random-effects DerSimonian–Laird model. The consistency of effect estimates across exclusion scenarios indicates that the overall results were robust and not driven by any single study (95% CI).