BioMed Research International / 2018 / Article / Tab 5 / Research Article
Predicting Advanced Prostate Cancer from Modeling Early Indications in Biopsy and Prostatectomy Samples via Transductive Semi-Supervised Survival Analysis Table 5 Weights of features in models with multimodal characteristics.
ā Cox Model Cox Model with Transduction SVRc SVRc with Transduction Feature Original Weight Normalized Weight Original Weight Normalized Weight Original Weight Normalized Weight Original Weight Normalized Weight Clinical Stage 0.475 0.408 0.535 0.444 -3.083 -0.174 -6.658 -0.320 PSA 0.700 0.601 0.893 0.741 -9.763 -0.552 -15.587 -0.749 Dominant Biopsy Gleason Grade -0.094 -0.081 0.039 0.032 -4.867 -0.275 -6.679 -0.321 Biopsy Gleason Sum 1.164 1.000 1.205 1.000 -16.801 -0.949 -20.076 -0.965 Dominant Prostatectomy Gleason Grade 0.455 0.391 0.643 0.534 -7.005 -0.396 -7.460 -0.358 Prostatectomy Gleason Sum 0.260 0.223 0.181 0.150 -14.282 -0.807 -7.011 -0.337 Seminal Vesicle Invasion 0.842 0.723 0.846 0.703 -20.542 -1.161 -24.166 -1.161 Surgical Margin Status 0.371 0.319 0.414 0.344 -4.788 -0.271 -6.757 -0.325 Extra Capsular Extension Status 0.802 0.689 0.892 0.740 -6.217 -0.351 -6.485 -0.312 AR+ Area -1.099 -0.944 -1.026 -0.852 15.277 0.863 17.347 0.833 Cytoplasm Texture -1.195 -1.026 -0.931 -0.773 17.699 1.000 20.814 1.000 Luminal Area -0.467 -0.401 -0.424 -0.352 10.264 0.580 12.092 0.581 Glandular Size 0.475 0.408 0.535 0.444 -3.083 -0.174 -6.658 -0.320