Table 3. Osteopontin and survival in distinct clinical specimens.
| Number of studies | Concordance | P-value | |
|---|---|---|---|
| Specimen | |||
| (A) All tumors combined | |||
| Tumor | 31 | 0.825 | <0.0001 |
| Plasma | 14 | 1 | <0.0001 |
| Serum | 3 | 1 | 0.04 |
| Cancer type | |||
| (B) Plasma, individual cancers | |||
| Lung | 1 | 1 | 0.00 |
| Prostate | 2 | 1 | 0.08 |
| Breast | 2 | 1 | 0.08 |
| Head and neck | 3 | 1 | 0.13 |
| Cancer type | |||
| (C) Tumor, individual cancers | |||
| Liver | 7 | 0.857 | 0.06 |
| Breast | 5 | 0.800 | 0.19 |
| Esophageal | 2 | 1 | 0.24 |
| Head and neck | 2 | 1 | 0.25 |
| Lung | 2 | 1 | 0.25 |
| Cervical | 2 | 1 | 0.25 |
| Gastric | 2 | 0.714 | 0.28 |
The concordance and probability of error were calculated for the null-hypothesis that Osteopontin levels are not correlated with high risk for short survival. (A) All tumors combined in distinct types of clinical samples. (B) Plasma Osteopontin in individual cancers (for serum Osteopontin see main text). (C) Tumor Osteopontin in individual cancers. Bold values indicate P<0.1