Skip to main content

Table 4 Features based predictors that discriminate aggressive grade (Gleason ≥7 Vs ≤ 6) prostate cancers a) univariate semantic predictors b) multivariable semantic predictors (up to 3 semantics) c) PIRADS based predictor

From: Radiological semantics discriminate clinically significant grade prostate cancer

  Features Error Sensitivity/Specificity PPV/NPV E [AUC],σ, [CI]
a) Single Predictive Semantic: ≥ GS 7 Vs =6 GS || Samples: 90 (57 Vs 33)
 1 F4:ADC-intensity 0.384 0.821/0.342 0.687/0.505 0.678 (0.114) [0.401,0.905]
 2 F8:enhanced-homogenicity-R 0.379 1/0 0.636/0 0.583 (0.099) [0.356,0.765]
 3 F7:early-enhancement-R 0.393 0.96/0.027 0.618/0.014 0.57 (0.116) [0.281,0.758]
 4 F11:extraprostatic-extension-R 0.378 1/0 0.628/0 0.542 (0.115) [0.314,0.749]
 5 F3:T2-intensity 0.396 0.965/0.016 0.637/0.005 0.549 (0.119) [0.254,0.751]
b) Multivariable Semantic Predictor: ≥ GS 7 Vs =6 GS || Samples: 90 (57 Vs 33)
 1 F3:T2-intensity;F4:ADC-intensity: F8:enhanced-homogenicity-R 0.333 0.793/0.465 0.719/0.576 0.701 (0.124), [0.426,0.94]
 2 F4:ADC-intensity;F8:enhanced-homogenicity-R:F11:extraprostatic-extension-R 0.329 0.799/0.464 0.73/0.557 0.687 (0.118), [0.46,0.94]
 3 F2:border;F4:ADC-intensity: F8:enhanced-homogenicity-R 0.334 0.819/0.412 0.681/0.6 0.706 (0.095), [0.474,0.886]
 4 F4:ADC-intensity;F5:homogeneity-R:F8:enhanced-homogenicity-R 0.341 0.808/0.418 0.712/0.553 0.698 (0.118), [0.457,0.948]
c) PIRADS as a Predictor: ≥ GS 7 Vs = 6 GS || Samples: 90 (57 Vs 33)
 1 PIRADS (> = 3 Cancer) 0.34 (0.096) 0.981/0.104 0.654/0.483 0.542 (0.06), [0.458,0.701]
 2 PIRADS (> = 4 Cancer) 0.296 (0.091) 0.981/0.22 0.686/0.717 0.6 (0.08), [0.462,0.752]
 3 PIRADS (> = 5 Cancer) 0.531 (0.101) 0.22/0.893 0.752/0.4 0.557 (0.083), [0.371,0.707]