Unbiased data mining identifies cell cycle transcripts that predict non-indolent Gleason score 7 prostate cancer

BMC Urology

Table 2 Univariate logistic regression and ROC-AUC analyses for predicting BCR in patients with Gleason score 7 tumours

Variable	Odds Ratio (OR)	95% CI OR	P value OR	ROC-AUC	SE AUC	P value AUC	Youden cut-off	Sensitivity	Specificity	PPV	NPV
(up in tumour)
BUB1	1.31	0.908–1.889	1.48E-01	0.720	0.055	5.35E-05	1.22	0.92	0.58	0.16	0.99
TPX2	1.18	0.994–1.406	5.79E-02	0.732	0.066	4.53E-04	3.54	0.85	0.64	0.18	0.98
NCAPG	1.51	0.913–2.510	1.08E-01	0.711	0.065	1.08E-03	0.94	0.92	0.48	0.14	0.99
UBE2C	1.07	0.998–1.142	5.60E-02	0.713	0.067	1.42E-03	7.79	0.69	0.69	0.17	0.96
MELK	1.21	0.908–1.612	1.93E-01	0.687	0.065	3.97E-03	1.82	0.69	0.69	0.17	0.96
CCNA2	1.27	0.967–1.680	8.57E-02	0.694	0.069	5.04E-03	2.58	0.69	0.70	0.17	0.96
CCNB1	1.13	0.954–1.345	1.56E-01	0.685	0.066	5.20E-03	4.60	0.92	0.48	0.14	0.99
CDK1	1.07	0.969–1.181	1.81E-01	0.691	0.070	6.27E-03	4.97	0.77	0.62	0.15	0.97
E2F2	2.84	0.715–11.309	1.38E-01	0.682	0.068	7.55E-03	0.37	0.77	0.63	0.16	0.97
DLGAP5	1.24	0.766–2.014	3.80E-01	0.657	0.070	2.54E-02	0.65	0.85	0.46	0.12	0.97
TMEM206	1.15	0.999–1.321	5.10E-02	0.672	0.086	4.51E-02	9.11	0.85	0.52	0.14	0.97
CDKN3	1.14	0.936–1.387	1.93E-01	0.643	0.079	6.93E-02	2.04	0.92	0.39	0.12	0.98
SHMT2	1.01	0.990–1.038	2.52E-01	0.605	0.069	1.25E-01	56.64	0.69	0.57	0.13	0.95
(down in tumour)
SRD5A2	0.90	0.786–1.021	9.85E-02	0.674	0.079	2.70E-02	7.98	0.92	0.44	0.13	0.99
CSRP1	0.998	0.996–1.001	1.30E-01	0.628	0.080	1.11E-01	417.97	0.54	0.72	0.15	0.95
PGM5	0.99	0.974–1.006	2.04E-01	0.592	0.085	2.84E-01	22.95	0.39	0.81	0.15	0.94
NFIB	0.98	0.941–1.029	4.68E-01	0.584	0.092	3.60E-01	27.09	0.69	0.56	0.12	0.95

Data are from TCGA-PRAD (TCGA, 2015). n = 158. CI-confidence interval, SE-standard error. Youden cut-off is the optimal point to separate biochemical recurrence (BCR) from non-BCR. PPV-positive predictive value, NPV-negative predictive value

ISSN: 1471-2490