OBJECTIVE: To investigate whether tau phosphorylated at Thr217 (p-tau T217) assay in CSF can distinguish patients with Alzheimer disease (AD) from patients with other dementias and healthy controls. METHODS: We developed and validated a novel Simoa immunoassay to detect p-tau T217 in CSF. There was a total of 190 participants from 3 cohorts with AD (n = 77) and other neurodegenerative diseases (n = 69) as well as healthy participants (n = 44). RESULTS: The p-tau T217 assay (cutoff 242 pg/mL) identified patients with AD with accuracy of 90%, with 78% positive predictive value (PPV), 97% negative predictive value (NPV), 93% sensitivity, and 88% specificity, compared favorably with p-tau T181 ELISA (52 pg/mL), showing 78% accuracy, 58% PPV, 98% NPV, 71% specificity, and 97% sensitivity. The assay distinguished patients with AD from age-matched healthy controls (cutoff 163 pg/mL, 98% sensitivity, 93% specificity), similarly to p-tau T181 ELISA (cutoff 60 pg/mL, 96% sensitivity, 86% specificity). In patients with AD, we found a strong correlation between p-tau T217 and p-tau T181, total tau and β-amyloid 40, but not β-amyloid 42. CONCLUSIONS: This study demonstrates that p-tau T217 displayed better diagnostic accuracy than p-tau T181. The data suggest that the new p-tau T217 assay has potential as an AD diagnostic test in clinical evaluation. CLASSIFICATION OF EVIDENCE: This study provides Class III evidence that a CSF immunoassay for p-tau T217 distinguishes patients with AD from patients with other dementias and healthy controls.