Quantification of identifying cognitive impairment using olfactory-stimulated functional near-infrared spectroscopy with machine learning: a post hoc analysis of a diagnostic trial and validation of an external additional trial

Table 2 C-statistic for the prediction model in the diagnosis of AD and MCI

Olfactory-stimulated oxygenation difference in the orbitofrontal cortex	Previous trial (n = 97)			Additional trial (n = 36)
	AUC (95% CI)	Sensitivity (%)	Specificity (%)	AUC (95% CI)	Sensitivity (%)	Specificity (%)
Classic prediction model for AD and MCI	0.873 (0.800 to 0.945)	88.1	81.8	0.639 (0.482 to 0.796)	60.0	68.8
Prediction model for AD and MCI using machine learning algorithm	0.925	88.1	80.0	0.825	65.0	81.3
Classic prediction model for MCI (excluded patients with AD)^a	0.852 (0.764 to 0.939)	84.6	81.8	0.688 (0.527 to 0.848)	68.8	68.8
Prediction model for MCI using machine learning algorithm (excluded patients with AD)^a	0.860	88.7	81.8	0.854	66.7	81.3

Abbreviations: AD Alzheimer’s disease, AUC Area under the receiver operating characteristic curve, CI Confidence interval, CN Cognitively normal, MCI Mild cognitive impairment
^aWe excluded 16 patients with AD; therefore, the sample size for this analysis was 112

ISSN: 1758-9193