From: Projecting COVID-19 disease severity in cancer patients using purposefully-designed machine learning
Clinical variable | Severe-late | Severe-early | Non-severe | |
---|---|---|---|---|
Basic patient information | # of patients | 71 | 71 | 206 |
Male (%) | 49.3 | 49.3 | 50.0 | |
Age (years) | 62.6 ± 13.9 | 69.8 ± 12.9 | 60.3 ± 16.9 | |
Labs | Absolute Lymphocyte Counts (K/mcL) | 3.57 ± 17.04 | 1.30 ± 2.02 | 2.45 ± 14.00 |
D-Dimer (mcg/mL) | 6.58 ± 7.26 | 3.52 ± 2.90 | 2.16 ± 1.75 | |
Interleukin 6 (pg/mL) | 469.60 ± 816.95 | 116.50 ± 90.25 | 62.62 ± 47.35 | |
Ferritin (ng/mL) | 1016.36 ± 1045.78 | 983.09 ± 1279.56 | 467.76 ± 541.13 | |
Platelets (K/mcL) | 171.31 ± 114.04 | 196.04 ± 112.71 | 218.33 ± 118.85 | |
Cancer-related | Lymphoma (%) | 11.3 | 5.6 | 1.9 |
Lung (%) | 5.6 | 21.1 | 6.8 | |
Leukemia (%) | 19.7 | 8.5 | 5.3 | |
Diagnosis (ICD) | I1 - Hypertensive diseases (%) | 52.1 | 71.8 | 56.8 |
I4 - Cardiac disorders (%) | 31.0 | 35.2 | 26.2 | |
J4 - Chronic lower respiratory disease (%) | 19.7 | 28.2 | 17.5 | |
Radiology | Retic. Opacities (%) | 16.9 | 29.6 | 14.6 |
Effusions (%) | 12.7 | 15.5 | 7.8 | |
Airspace Opacity (%) | 32.4 | 81.7 | 38.3 | |
Performance | AUROC (Our method) | 0.704 | 0.829 | 0.710 |
AUROC (Yan et al.) | 0.456 | 0.634 | 0.499 | |
AUROC (Huang et al.) | 0.600 | 0.638 | 0.604 | |
Avg. Precision (Our method) | 0.366 | 0.578 | 0.772 | |
Avg. Precision (Yan et al.) | 0.190 | 0.315 | 0.588 | |
Avg. Precision (Huang et al.) | 0.291 | 0.326 | 0.691 |