Table 2.
Estimated overall effect size and p values for continuous scores on molecular, histopathological, and clinical variables from fixed effects model
| Variable | Continuous score | Effect size | p value |
|---|---|---|---|
| CMS1 subtype | PCSS1 | 0.82 | 3E-55 |
| PCSS2 | − 2.55 | 1E-129 | |
| CMS2 subtype | PCSS1 | − 2.00 | 2E-156 |
| PCSS2 | 0.76 | 7E-60 | |
| CMS3 subtype | PCSS1 | − 0.57 | 2E-28 |
| PCSS2 | − 0.75 | 6E-56 | |
| CMS4 subtype | PCSS1 | 1.72 | 2E-130 |
| PCSS2 | 1.75 | 4E-106 | |
| MSI | PCSS1 | 0.76 | 1E-31 |
| PCSS2 | − 1.68 | 5E-71 | |
| Right location | PCSS1 | 0.087 | 0.09 |
| PCSS2 | − 0.23 | 1E-04 | |
| Late stage | PCSS1 | 0.16 | 0.002 |
| PCSS2 | 0.24 | 6E-06 | |
| High grade | PCSS1 | 0.33 | 3E-05 |
| PCSS2 | − 0.30 | 7E-05 | |
| Disease recurrence or death | PCSS1 | 0.23 | 5E-05 |
| PCSS2 | 0.19 | 0.001 |
The effect size statistic is log hazard ratio for disease recurrence or death and log odds ratio for all other variables. These estimates are not sensitive to fixed vs random effects modeling (Additional file 5: Table S4). Statistics for individual datasets, including I2 statistics, are also provided in Additional file 5: Table S4