Association of MT staining and clinicopathological factors – individual tumor types taken into account.

Heterogeneity of studies analyzed using Cochran's Q-test (p-value displayed) and using I2. Egger's two-tailed test used for publication bias analysis (p-value displayed). * effect measure is odds ratio, OR, except for survival analysis using hazard ratio, HR. No heterogeneity test and no model indicated when 1 study per factor analyzed. CI, confidence interval.