Skip to main content

Table 2 Performance of logistic regression models in predicting Aβ-PET status

From: Clinical performance of an antibody-free assay for plasma Aβ42/Aβ40 to detect early alterations of Alzheimer’s disease in individuals with subjective cognitive decline

Model

AUC

95% CI

Se

Sp

Acc

PPV

NPV

P valuea

Aβ42/Aβ40

0.87

0.80–0.93

86.1%

80.5%

81.5%

49.2%

96.4%

.1544

Aβ42/Aβ40, age, sex, APOE

0.89

0.83–0.94

86.1%

77.4%

79.0%

45.6%

96.2%

.0054

Age, sex, APOE

0.81

0.74–0.88

80.6%

69.5%

71.5%

36.7%

94.2%

-

  1. Abbreviations: Acc accuracy, APOE apolipoprotein E, AUC Area under the ROC curve, CI confidence interval, NPV negative predictive value, PPV positive predictive value, Se sensitivity, Sp specificity
  2. P values in bold correspond to statistically significant results
  3. a P values correspond to the comparison of the AUCs of the Aβ42/Aβ40 model and the full model (Aβ42/Aβ40, age, sex, APOE ɛ4 number of alleles) with the AUC of the model comprising age, sex and APOE ɛ4 number of alleles, using the DeLong test