Table 4.

Areas under the curve (AUC), sensitivities, specificities, accuracies and p-values of the regression models

Combination of genes and clinical risk factors	Sensitivity	Specificity	Accuracy	AUC	ORIQR	p
SNORD99	0.667	0.625	0.654	0.681	0.384	0.0426
SNORD22	0.500	0.875	0.629	0.704	0.560	0.0375
SNORD26	0.761	0.667	0.729	0.677	0.573	0.0578
SNORA50C	0.778	0.667	0.744	0.735	0.210	0.0108
SNORD99 + OBS + HTN	0,900	0,417	0,743	0,693	0.385	0.0657
SNORD22 + OBS + HTN	0,791	0,667	0,746	0,733	0.570	0.0586
SNORD26 + OBS + HTN	0,628	0,750	0,672	0,720	0.541	0.0495
SNORA50C + OBS + HTN	0,840	0,625	0,770	0,766	0.165	0.0046

The expression values of the genes validated with qPCR were used in the regression models alone or in combination with obesity (OBS; BMI ≥ 30) and hypertension (HTN) as clinical risk factors. All models were adjusted for age and gender. The p-values of the top four results are those of the genes in the models (association with the response). The p-values of the bottom four results are those of the overall model fit (chi-squared test against the null model). P-values < 0.05 are shown in bold. Odds ratios (OR) for the risk of ccRCC presence were calculated on the basis of a gene expression change in the size of the interquartile range (IQR) of the respective gene