Table 5.
Model | Original group | Predicted group membership | sensitivity | Specificity | |
---|---|---|---|---|---|
1 | Discrimination of ESCC/control: | control | ESCC | ||
control | 44 | 6 | 80.0% | 88.0% | |
ESSC | 10 | 40 | |||
Ungrouped cases | 54 | 46 | |||
2 | Discrimination of ESCD/control: | control | ESCC | ||
control | 38 | 12 | 64.0% | 76.0% | |
ESCC | 18 | 32 | |||
Ungrouped cases | 44 | 56 | |||
3 | Discrimination of ESCD/control: | control | ESCD | ||
control | 38 | 12 | 70.0% | 76.0% | |
ESCD | 15 | 35 | |||
Ungrouped cases | 27 | 73 | |||
4 | Discrimination of ESCD/control: | control | ESCD | ||
control | 39 | 11 | 64.0% | 76.0% | |
ESCD | 18 | 32 | |||
Ungrouped cases | 41 | 59 |
Note: The discrimination model 1 and model 3 included variables of the hTERT and EYA4 and the five traditional risk factors (sex, age, smoking, alcohol drinking and family history of esophageal cancer). The discrimination model 2 and model 4 only included the five traditional risk factors. ESCD, esophageal squamous cells dyspalsia; ESCC, esophageal squamous cells cancer.