Table 3.

Comparison of one and two-cutoffs for DMT eligibility screening

Model	Cutoffs	Total N	Specificity	Sensitivity	NPV	PPV	Accuracy (95% CI)1	P-value2
	([P-tau217], pg/mL)	(Intermediate N)
Site-Specific Cutoffs
One-cutoff	0.27	50	0.67	0.97	0.89	0.9	0.90 (0.78, 0.96)	0.011
Two-cutoff (Inclusive)	0.273, 0.399	50 (6)3	0.92	0.97	0.92	0.97	0.96 (0.86, 1)	< 0.001
Two-cutoff (Exclusive)	0.273, 0.399	444	0.88	0.97	0.89	0.97	0.95 (0.84, 0.99)	0.003
BioFINDER-2 Cutoffs
One-cutoff	0.231	50	0.5	1	1	0.86	0.88 (0.76, 0.95)	0.028
Two-cutoff (Inclusive)	0.209, 0.254	50 (5)5	0.67	1	1	0.9	0.92 (0.81, 0.98)	0.003
Two-cutoff (Exclusive)	0.209, 0.254	456	0.5	1	1	0.9	0.9111 (0.79, 0.98)	0.079

¹Accuracy is reported with 95% CI

²P-value: accuracy of model prediction of cerebral Aβ status compared to the no information rate

³BioFINDER-2 cutoffs identified 1 Aβ positive and 4 Aβ negative participants in the intermediate “gray zone” (5/50 = 10%)

⁴Remaining sample size after removal of 5 intermediate cases

⁵Site-specific cutoffs identified 3 Aβ positive and 3 Aβ negative participants in the intermediate “gray zone” (6/50 = 12%)

⁶Remaining sample size after removal of 6 intermediate cases