Skip to main content

Table 1 Comparison of SynTwin performance as measured by AUROC on various cancer datasets

From: Enhancing clinical outcome predictions through effective sample size evaluation in graph-based digital twin modeling

Approximate sample size (Mortality rate %)

Cancer site

Designa

A

B

C

D

E

F

90,000 (50%)

Breast [9]

mean

0.791

0.784

0.783

0.864

0.852

0.494

95% CI

[0.781, 0.800]

[0.774, 0.794]

[0.773, 0.793]

[0.857, 0.872]

[0.844, 0.860]

[0.482, 0.507]

30,000 (58.53%)

Ovary

mean

0.858

0.848

0.842

0.882

0.884

0.431

95% CI

[0.843,0.872]

[0.833,0.864]

[0.827,0.858]

[0.868,0.895]

[0.871,0.897]

[0.411,0.452]

20,000 (34.65%)

Cervix uteri

mean

0.720

0.708

0.702

0.835

0.828

0.496

95% CI

[0.700, 0.741]

[0.687, 0.729]

[0.680, 0.723]

[0.819, 0.851]

[0.811, 0.846]

[0.473, 0.519]

10,000 (38.49%)

Small intestine

mean

0.783

0.772

0.761

0.872

0.861

0.468

95% CI

[0.761,0.804]

[0.751,0.794]

[0.738,0.783]

[0.856,0.888]

[0.843,0.878]

[0.442,0.493]

5,000 (46.25%)

Hepatic flexure

mean

0.768

0.789

0.780

0.852

0.844

0.554

95% CI

[0.730,0.807]

[0.752,0.826]

[0.743,0.818]

[0.821,0.883]

[0.812,0.876]

[0.509,0.599]

3,000 (44.31%)

Nasopharynx

mean

0.597

0.594

0.578

0.708

0.709

0.552

95% CI

[0.540,0.654]

[0.541,0.648]

[0.524,0.632]

[0.657,0.759]

[0.660,0.759]

[0.493,0.612]

2,000 (56.83%)

Oropharynx

mean

0.671

0.683

0.664

0.746

0.768

0.438

95% CI

[0.608,0.733]

[0.620,0.746]

[0.599,0.728]

[0.687,0.804]

[0.712,0.820]

[0.368,0.508]

1,000 (59.56%)

Other urinary organs

mean

0.467

0.530

0.466

0.670

0.635

0.522

95% CI

[0.356,0.578]

[0.419,0.640]

[0.355,0.578]

[0.562,0.778]

[0.522,0.749]

[0.419,0.626]

  1. aReal patients (A), digital twins (B), real patients and digital twins (C), closest digital twins (D), real patients and closest digital twins (E), and real patients outside the community (F). Bolded metric values are significantly better than the others