Table 5.

Descriptive values of the ROC, sensitivity and specificity for environmental factors and GRSs

DatasetParametersAUC (95% CI)Sensitivity (%)Specificity (%)
Training setaGRS0.61 (0.58–0.66)83.333.1
Training setaAlcohol0.72 (0.69–0.76)7074.5
Training setaBetel0.86 (0.83–0.89)85.984.9
Training setaCigarette0.69 (0.67–0.73)89.650.4
Training setaAge0.54 (0.51–0.68)25.673.5
Training setaBetel + GRS0.89 (0.86–0.91)86.786
Validation setbGRS0.63 (0.57–0.68)61.666.9
Validation setbAlcohol0.78 (0.73–0.81)80.874.5
Validation setbBetel0.86 (0.83–0.89)85.686.7
Validation setbCigarette0.71 (0.68–0.74)91.150.4
Validation setbAge0.72 (0.69–0.77)67.873.5
Validation setbBetel + GRS0.88 (0.84–0.91)86.386.5
CombinedcGRS0.64 (0.57–0.71)67.5555.22
CombinedcAlcohol0.73 (0.68–0.79)74.0473.63
CombinedcBetel0.88 (0.84–0.92)85.884.9
CombinedcCigarette0.7 (0.68–0.73)83.8958.76
CombinedcAge0.56 (0.48–0.61)67.145.6
CombinedcBetel + GRS0.91 (0.87–0.95)88.786.7
  • NOTE: The optimal sensitivity and specificity were set with the maximum of AUC.

  • aOSCC and control were recruited from the southern Taiwan (n = 602).

  • bOSCC and control were recruited from mid-Taiwan (n = 425).

  • cCombined the training set and validation set.