Table 2 Results from univariable and multivariable binary logistic regression models to describe pathological ECE risk
From: Early biomarkers of extracapsular extension of prostate cancer using MRI-derived semantic features
Variable | Univariable Linear Regression Model | Multivariable Linear Regression Model | ||||||
|---|---|---|---|---|---|---|---|---|
Coef. | SE | OR (95% CI) | p-value | Coef. | SE | OR (95% CI) | p-value | |
Intercept | – | – | – | – | −2.701 | 0.773 | 0.067 (0.015, 0.305) | 0.000 |
Age at MRI | −0.026 | 0.025 | 0.975 (0.927, 1.025) | 0.317 | ||||
Prostate volume | −0.011 | 0.010 | 0.989 (0.970, 1.009) | 0.269 | −0.033 | 0.015 | 0.968 (0.940, 0.996) | 0.027 |
PSA | 0.053 | 0.040 | 1.055 (0.975, 1.141) | 0.186 | – | – | – | – |
PSAD* | 2.210 | 1.218 | 9.117 (0.838, 99.178) | 0.070 | – | – | – | – |
Index lesion lenght (ILL) | 0.152 | 0.034 | 1.164 (1.089, 1.244) | 0.000 | – | – | – | – |
Tumor capsular contact length (TCCL) | 0.179 | 0.032 | 1.196 (1.124, 1.273) | 0.000 | 0.093 | 0.047 | 1.097 (1.001, 1.203) | 0.048 |
Index lesion PI-RADSv2 | ||||||||
More Aggressive (≥4) | 1.205 | 1.069 | 3.336 (0.410, 27.115) | 0.260 | – | – | – | – |
Smooth capsular bulging | ||||||||
Yes | 2.006 | 0.508 | 7.434 (2.749, 20.106) | 0.000 | – | – | – | – |
Capsular disruption | ||||||||
Yes | 2.456 | 0.478 | 11.659 (4.573, 29.727) | 0.000 | 1.112 | 0.576 | 3.040 (0.983, 9.402) | 0.054 |
Unsharp margin | ||||||||
Yes | 1.810 | 0.431 | 6.113 (2.628, 14.220) | 0.000 | – | – | – | – |
Irregular contour | ||||||||
Yes | 2.303 | 0.424 | 10.000 (4.360, 22.935) | 0.000 | – | – | – | – |
Black strition periprostatic fat | ||||||||
Yes | 1.888 | 0.415 | 6.607 (4.360, 22.935) | 0.000 | – | – | – | – |
Measurable ECE | ||||||||
Yes | 3.937 | 0.774 | 51.250 (11.235, 233.774) | 0.000 | 1.652 | 0.991 | 5.217 (0.749, 36.357) | 0.095 |
Gleason score | ||||||||
More aggressive [7 (4 + 3), 8 (4 + 4), 9 (4 + 5), 9 (5 + 4)] | 2.481 | 0.408 | 11.955 (5.369, 26.620) | 0.000 | 1.823 | 0.518 | 6.188 (2.242, 17.081) | 0.000 |