Table 3.
Results of meta regression analysis to assess causes of heterogeneity.
| Variable | HE4 | CA125 | ROMA (ALL) | ROMA (pre-menopause) |
ROMA (post-menopause) |
|---|---|---|---|---|---|
| Analysis of diagnostic odds ratio (DOR) | |||||
| Constant (SE) | 3.27 (0.24)* | 2.64 (0.19)* | 3.35 (0.21)* | 2.91 (0.20)* | 3.85 (0.18)* |
| Slope (SE) | − 0.29 (0.11)* | − 0.06 (0.16) | − 0.10 (0.17) | − 0.17 (0.17) | − 0.17 (0.13) |
| Tau-square | 0.50 | 0.42 | 0.75 | 0.57 | 0.40 |
| Meta regression with cutoff value as covariate# | |||||
| Tau-square | 0.36 | 0.45 | 0.84 | 0.64 | 0.46 |
| Constant (SE) | 1.31 (0.65) | 2.76 (0.35)* | 2.18 (1.60) | 2.49 (0.88)* | 3.67 (1.81) |
| S coefficient (SE) | − 0.14 (0.12) | − 0.14 (0.16) | − 0.03 (0.20) | − 0.12 (0.18) | − 0.16 (0.14) |
| Cutoff (SE) | 0.02 (0.01)* | − 0.002 (0.01) | 0.06 (0.08) | 0.04 (0.08) | 0.007 (0.06) |
| Meta regression with age as covariate# | |||||
| Tau-square | 0.81 | 0.51 | 1.06 | Could not be carried out due to non-availability of Age details separately for pre-menopause and post-menopause status | |
| Constant (SE) | 2.57 (1.79) | 1.60 (1.51) | 1.63 (2.18) | ||
| S coefficient (SE) | − 0.31 (0.16) | − 0.24 (0.19) | − 0.24 (0.25) | ||
| Age (SE) | 0.01 (0.03) | 0.03 (0.03) | 0.03 (0.04) | ||
| Meta regression with cutoff value and age as covariates# | |||||
| Tau-square | 0.49 | 0.78 | 1.30 | Could not be carried out due to non-availability of Age details separately for pre-menopause and post-menopause status | |
| Constant (SE) | 1.01 (1.57) | 1.62 (2.93) | 1.21 (2.65) | ||
| S coefficient (SE) | − 0.18 (0.15) | − 0.30 (0.25) | − 0.15 (0.35) | ||
| Cutoff (SE) | 0.03 (0.01)* | 0.02 (0.09) | 0.05 (0.12) | ||
| Age (SE) | − 0.003 (0.03) | 0.013 (0.04) | 0.02 (0.05) | ||
*Coefficients are statistically significant at P < 0.05, Human epididymis secretory protein 4 (HE4), Cancer antigen (CA125) and Risk of Ovarian Malignancy Algorithm (ROMA). Standard error (SE).
#Meta-regression was weighted by the inverse of the variance of diagnostic Odds Ratio and between study variance was estimated by Random Effect Model (REML).