Skip to main content
. 2021 Mar-Apr;54(2):87–93. doi: 10.1590/0100-3984.2019.0135

Table 2.

Radiomic features of higher prognostic potential for lung cancer obtained from contrast-enhanced CT images in our sample (N = 101).

Radiomic feature (range) Values Number of deaths Overall survival (days)
Mean (95% CI)
Hazard ratio (95% CI) P-value
Mean of the Fourier transform (94.92 to 122.91) High
Low
18
10
1,465.4 (985,2-1,945.6)
2,164.8 (1,745.4 - 2,5 84.1)
2.12 (1.01-4.48)
-
0.048
Co-occurrence matrix prominence (35,997.37 to 3,885,382.04) Low
High
17
11
1,190.3 (877,9-1,502.7)
2,155.8 (1,756.1-2,555.5)
2.12 (0.99-4.52)
-
0.051
Co-occurrence matrix correlation (-0.049 to 0.635) Low
High
17
11
1,594.8 (1,106.7-2,082.9)
2,051.5 (1,608.9-2,494.1)
2.07 (0.98-4.40)
-
0.057
Co-occurrence matrix cluster shade (-37,575.62 to 4,516.65) High
Low
17
11
1,157.2 (840,3-1,474.1)
2,101.9 (1,669.8-2,534.0)
2.08 (0.98-4.43)
-
0.058
Highest value of the Fourier transform (189 to 238) High
Low
18
10
1,483.4 (1,014.4-1,952.4)
2,226.3 (1,847.1-2,605.4)
2.04 (0.97-4.31)
-
0.061