Table 3.

Summary and comparison of fold change, ∆Ct values, and PCR efficiency

	Pilot phase	Validation phase
miRNA	Median fold change relapse/control	Relapse (∆Ct − median)	Control (∆Ct − median)	Median fold change relapse/control	PCR efficiency	Raw p value	Corrected p value
hsa-miR-660	0.77	5.13	7.31	4.53	2.00	<0.001	<0.001
hsa-miR-375	2.97	3.24	4.43	2.28	1.89	0.001	0.008
hsa-miR-125a-3p	1.26	6.79	6.91	1.09	2.02	0.448	1.000
hsa-miR-362-3p	0.80	7.25	7.82	1.48	1.98	0.093	0.744
hsa-miR-210	0.51	6.15	5.21	0.52	1.95	0.121	1.000
hsa-miR-223	0.58	2.81	3.13	1.25	1.98	0.806	1.000
hsa-miR-487b	1.22	6.93	7.30	1.29	2.10	0.373	1.000
hsa-miR-532-3p	0.90	6.22	6.30	1.05	2.15	0.679	1.000

Eight candidate miRs were selected in the pilot phase and further analyzed in an independent cohort by RT-qPCR. Both in the pilot and the validation cohort, the levels of hsa-miR-375 were significantly higher in the relapse group (raw p value = 0.001, corrected p value = 0.008). For calculation of the relative miR expression, the Ct values of the reference gene were subtracted from the Ct values of the target miR. The fold change was estimated with the ∆∆Ct method as described by Livak. Correlations were tested for statistical significance with the non-parametric Mann-Whitney test, p values were corrected for multiple testing according to Bonferroni