Skip to main content

Table 2 Summary information of the selected biomarker models (original measurements)

From: The accuracy and robustness of plasma biomarker models for amyloid PET positivity

 

Model

AIC

BIC

R2, unadjusted

R2, adjusted

AUC, 95% CI

LRT x2

P value

IP-MS for Aβ42/40

 All participants

AG

124.2

143.6

40.7%

37.5%

86.5% (79.7, 93.4)

53.3

< 0.0001*

 CU

A

62.6

76.7

33.5%

25.9%

82.3% (68.5, 96.1)

16.7

< 0.001*

 CI

AGP

59.2

77.0

58.7%

52.9%

93.5% (87.5, 99.5)

43.5

< 0.0001*

Simoa for Aβ42/40

 All participants

G

145.5

159.4

22.2%

19.4%

77.5% (68.9, 86.1)

28.0

< 0.0001*

 CU

G

71.9

81.4

11.0%

3.1%

72.8% (58.1, 87.6)

5.4

0.04

 CI

GP

71.7

85.0

42.2%

37.6%

87.1% (78.4, 95.9)

32.4

< 0.01*

Demographic

 All participants

155.6

166.7

12.7%

10.5%

70.7% (61.3, 80.0)

 CU

74.0

81.7

2.4%

− 4.0%

58.1% (41.9, 74.3)

 CI

78.9

87.8

27.8%

24.5%

81.3% (71.0, 91.5)

  1. Abbreviations: AIC Akaike information criterion, AUC area under the curve, BIC Bayesian information criterion, CU cognitively unimpaired, CI mild cognitive impairment, LRT likelihood ratio test
  2. *P value of the likelihood ratio test comparing the selected model with the demographic model on the respective sample group