Table 2.

Survival-associated image features and eigengenes, identified by Kaplan-Meier estimator and log-rank test (p < 0.05). For each variable, patients were stratified into low and high groups using the median as cut-off point. For P/N, P means positive relation to survival (i.e., patients with high feature values have good prognosis), whereas N means negative relation to survival.

Feature	P value	P/N	Feature	P value	P/N
rMean_bin10	2.23e-5	N	gMean_entropy	0.0194	N
rMean_bin6	8.55e-5	P	ratio_std	0.0245	N
rMean_std	1.18e-4	N	rMean_kurtosis	0.0269	P
rMean_entropy	2.45e-4	N	ratio_bin8	0.0297	N
gMean_std	7.70e-4	N	ratio_bin9	0.0312	N
rMean_bin5	0.0010	P	area_std	0.0319	N
major_bin9	0.0022	N	ratio_bin5	0.0322	N
major_entropy	0.0028	N	ratio_mean	0.0324	N
area_bin5	0.0056	P	major_bin1	0.0333	N
major_bin4	0.0058	P	major_bin2	0.0337	N
ratio_bin6	0.0059	N	bMean_bin10	0.0338	N
major_bin8	0.0060	N	major_bin10	0.0366	N
major_std	0.0072	N	bMean_std	0.0407	N
area_bin7	0.0089	P	eigengene3	7.46e-6	P
rMean_bin9	0.0097	N	eigengene9	1.19e-4	N
major_bin5	0.0113	P	eigengene13	9.39e-4	P
gMean_bin10	0.0113	N	eigengene11	0.0013	N
area_bin6	0.0124	P	eigengene1	0.0217	N
bMean_entropy	0.0164	N	eigengene2	0.0237	N
ratio_bin7	0.0176	N