Skip to main content
. 2020 May 8;15(5):e0226461. doi: 10.1371/journal.pone.0226461

Table 3. DNN model performance in test set.

CANCER ORIGIN SPECIFICITY SENSITIVITY (RECALL) PPV (PRECISION) NPV
AG 0.9993 0.9787 0.9787 0.9993
BLADDER 0.9986 0.9878 0.9759 0.9993
BRAIN 1.0000 1.0000 1.0000 1.0000
BREAST 0.9977 1.0000 0.9810 1.0000
COLORECTAL 1.0000 0.9861 1.0000 0.9993
ESOPHAGUS 0.9909 0.7410 0.7579 0.9902
HN 0.9971 0.9099 0.9619 0.9927
KIDNEY 0.9993 1.0000 0.9925 1.0000
LIVER 0.9993 0.9851 0.9851 0.9993
LUNG 0.9984 0.9740 0.9894 0.9961
PANCREAS 0.9979 1.0000 0.9167 1.0000
PROSTATE 1.0000 1.0000 1.0000 1.0000
SKIN 1.0000 1.0000 1.0000 1.0000
SOFT TISSUE 0.9993 0.9825 0.9825 0.9993
STOMACH 0.9921 0.9375 0.8721 0.9964
TESTIS 1.0000 1.0000 1.0000 1.0000
THYMUS 1.0000 0.8889 1.0000 0.9979
THYROID 1.0000 1.0000 1.0000 1.0000
OVERALL 0.9983 0.9595 0.9608 0.9983

PPV: positive predictive value; NPV: negative predictive value; AG: Adrenal Gland; HN: Head and Neck