Skip to main content

Table 2 Statistical differences of radiomic features determined by using RF classifier between pseudoprogression and true progression

From: Differentiation of Pseudoprogression from True Progressionin Glioblastoma Patients after Standard Treatment: A Machine Learning Strategy Combinedwith Radiomics Features from T1-weighted Contrast-enhanced Imaging

Feature Gini Importance True progression Pseudoprogression p value
Median Interquartile range Median Interquartile Range
Feature1 3.73 0.998 0.995–0.999 0.996 0.993–0.999  < .001
Feature2 2.91 1.30 × 10–5 2.0 × 10–6–6.8 × 10–5 3.39 × 10–5 7.34 × 10–6–1.19 × 10–4  < .001
Feature3 2.08 3.0 × 10–13 1.04 × 10–14–4.2 × 10–12 5.59 × 10–13 1.26 × 10–13–7.91 × 10–12 .079
Feature4 2.08 − 0.20 − 1.21–0.83 − 0.58 − 1.79–1.09  < .001
Feature5 1.98 1.14 × 104 1725.0–72,802.4 2.03 × 104 9098.51–56,899.20 .015
Feature6 1.53 3.32 × 10–4 1.44 × 10–4–7.51 × 10–4 4.65 × 10–4 1.71 × 10–4–7.57 × 10–4  < .001
Feature7 1.45 16.22 1.20–241.05 37.06 11.14–254.88 .137
Feature8 1.42 221.32 14.89–5051.62 349.15 89.95–5227.03 .765
Feature9 1.39 5.25 × 10–6 2.45 × 10–7–2.04 × 10–5 6.44 × 10–7 2.43 × 10–7–8.2 × 10–6 .828
Feature10 1.32 5.35 × 108 2.20 × 107–1.64 × 1011 2.34 × 109 1.85 × 108–9.06 × 1010 .374
Feature11 1.25 4.84 × 10–5 1.3 × 10–5–1.96 × 10–4 7.3 × 10–5 5.74 × 10–6–1.75 × 10–4 .008
Feature12 1.25 14.49 1.08–342.81 35.44 2.47–189.35 .244
Feature13 1.24 − 2393.65 − 61,416.60–36,264.10 − 1.26 × 104 − 152 × 105–5.76 × 104 .015
Feature14 1.09 1.5 × 10–13 5.05 × 10–15–8.51 × 10–9 2.72 × 10–13 8.2 × 10–14–1.83 × 10–11 .445
Feature15 1.07 1.8 × 10–5 1.22 × 10–6–7.55 × 10–5 2.8 × 10–5 5.75 × 10–6–1.08 × 10–4 .005
Feature16 1.01 0.998 0.994–0.999 0.996 0.993–0.999  < .001
Feature17 0.93 3.27 × 10–5 − 3.14 × 10–4–7.18 × 10–4 1.47 × 10–4 − 5.47 × 10–4–4.69 × 10–4 .050
Feature18 0.93 − 744.67 − 1.68 × 104–1.03 × 104 748.24 − 11,634.40–18,560.10 .138
Feature19 0.89 0.12 1.91 × 10–4–8.93 0.14 5.95 × 10–4–5.43 .197
Feature20 0.82 0.55 0.35–0.66 0.56 0.50–0.73 .028
Feature21 0.81 1.3 × 10–13 1.20 × 10–14–2.97 × 10–12 2.52 × 10–13 5.05 × 10–14–2.07 × 10–9 .161
Feature22 0.76 1.83 × 109 6.18 × 107–7.80 × 1010 6.48 × 109 7.02 × 108–9.28 × 1010 .048
Feature23 0.75 0.998 0.994–0.999 0.997 0.994–0.998 .256
Feature24 0.74 0.998 0.994–1.000 0.998 0.997–0.999 .347
Feature25 0.73 5.1 × 10–12 1.53 × 1013–4.33 × 1010 1.44 × 10–11 1.17 × 10–12–3.16 × 10–8 .141
Feature26 0.72 1.73 × 10–4 − 2.99 × 10–4–7.68 × 10–4 1.69 × 10–4 3.13 × 10–6–7.69 × 10–4 .060
Feature27 0.71 3.72 × 10–4 2.19 × 10–4–1.06 × 10–3 4.43 × 10–4 3.16 × 10–4–1.04 × 10–3 .006
Feature28 0.70 7.47 × 103 999.18–41,102.90 1.16 × 104 1287.42–25,000.80 .111
Feature29 0.69 − 342.35 − 4559.64–8392.05 672.42 − 8078.63–28,881.70 .208
Feature30 0.68 − 1.02 × 103 − 5065.29–1823.32 − 600.96 − 2031.27–3107.47 .125
Feature31 0.66 6.80 × 108 2.59 × 107–1.87 × 1010 2.12 × 109 3.88 × 107–3.44 × 1011 .103
Feature32 0.65 843.33 160.59–1046.56 753.13 258.74–1333.93 .147
Feature33 0.62 − 9.8 × 10–5 − 5.9 × 10–4–3.22 × 10–4 − 8.60 × 10–5 − 3.36 × 10–4–1.4 × 10–4 .799
Feature34 0.62 967.43 69.37–6660.59 2441.03 149.65–10,040.5 .002
Feature35 0.60 5.83 × 10–6 1.51 × 10–6–1.89 × 10–5 8.91 × 10–6 2.43 × 10–6–2.95 × 10–5 .015
Feature36 0.58 1.69 × 104 9790.15–26,645.1 18,893.30 13,895.80–32,379.50 .008
Feature37 0.58 1.99 × 10–4 − 3.4 × 10–4–8.7 × 10–4 2.87 × 10–4 − 7.65 × 10–5–1.44 × 10–3 .060
Feature38 0.53 − 467.89 − 3.00 × 104–1.79 × 104 799.64 − 35,322.10–20,325.90 .575
Feature39 0.53 4.95 × 10–9 1.34 × 10–9–1.60 × 10–8 5.83 × 10–9 5.38 × 10–10–3.45 × 10–8 .037
Feature40 0.53 14.17 − 5.15 × 103–1.25 × 104 − 939.69 − 27,364.50–5113.09 .026
Feature41 0.52 − 259.96 − 16,902.50–9521.71 1264.01 − 10,087.90–6781.62 .121
Feature42 0.52 − 1.20 − 2.52–0.10 − 0.81 − 2.07–0.20 .043
Feature43 0.52 2.22 0.30–12.95 2.76 1.54–5.43 .536
Feature44 0.52 9.53 × 1010 1.10 × 1010–4.10 × 1012 2.30 × 1011 4.82 × 109–7.26 × 1012 .023
Feature45 0.52 1.04 × 105 2.21 × 104–5.96 × 105 1.40 × 105 37,793.10–516,907.00 .025
Feature46 0.52 3.30 × 105 1.23 × 105–2.56 × 106 4.29 × 105 1.27 × 105–1.39 × 106 .505
Feature47 0.51 9.1 × 10–14 9.09 × 10–14–9.42 × 10–15 1.96 × 1013 4.65 × 10–14–8.41 × 10–13 .110
Feature48 0.50 0.51 0.39–0.64 0.48 0.38–0.62 .074
Feature49 0.49 6.27 × 10–7 5.39 × 10–8–4.22 × 10–6 1.37 × 10–6 1.71 × 10–7–4.76 × 10–6 .005
Feature50 0.49 − 4.9 × 10–3 − 0.55–0.50 0.06 − 0.68–0.63 .414
  1. Feature relevance was assessed by using mean decrease in Gini index–based feature importance
  2. P values are adjusted for false-discovery rate by using Benjamini–Hochberg method. 1–50 features are the same as in Fig. 4