ObjectiveTo improve the timely recognition of patients with treatment‐responsive causes of rapidly progressive dementia (RPD).MethodsTwo‐hundred twenty‐six adult patients with suspected RPD were enrolled in a prospective observational study and followed for up to two years. Diseases associated with RPD were characterized as potentially treatment‐responsive or non‐responsive referencing clinical literature. Disease progression was measured using Clinical Dementia Rating® Sum‐of‐Box scores. Clinical and paraclinical features associated with treatment responsiveness were assessed using multivariable logistic regression. Findings informed the development of a clinical criterion optimized to recognize patients with potentially treatment‐responsive causes of RPD early in the diagnostic evaluation.ResultsOne‐hundred fifty‐five patients met defined RPD criteria, of whom 86 patients (55.5%) had potentially treatment‐responsive causes. Median (range) age‐at‐symptom onset in patients with RPD was 68.9 years (22.0‐90.7), with a similar number of males and females. Seizures, tumor (disease‐associated), MRI suggestive of autoimmune encephalitis, mania, movement abnormalities, and pleocytosis (≥10 cells/mm3 in CSF) at presentation were independently associated with treatment‐responsive causes of RPD after controlling for age and sex. Those features at presentation, as well as age‐at‐symptom onset < 50 years (i.e., STAM3P), captured 83/86 (96.4%) cases of treatment‐responsive RPD. The presence of ≥3 STAM3P features had a positive predictive value of 100%.InterpretationSelected features at presentation reliably identified patients with potentially treatment‐responsive causes of RPD. Adaptation of the STAM3P screening score in clinical practice may minimize diagnostic delays and missed opportunities for treatment in patients with suspected RPD.This article is protected by copyright. All rights reserved.