Objective: The purpose of this study was to propose new criteria for differentiating Pittsburgh compound B (PiB)-negative from PiB-positive subcortical vascular dementia (SVaD) using clinical and MRI variables. Methods: We measured brain amyloid deposition using PiB-PET in 77 patients with SVaD. All patients met DSM-IV criteria for vascular dementia and had severe white matter hyperintensities on MRI, defined as a cap or band ≥10 mm as well as a deep white matter lesion ≥25 mm. Eleven models were considered to differentiate PiB(-) from PiB(+) SVaD using 4 variables, including age, number of lacunes, medial temporal atrophy (MTA), and APOE ε4. The ideal cutoff values in each of the 11 models were selected using the highest Youden index. Results: A total of 49 of 77 patients (63.6%) tested negative for PiB retention, while 28 (36.4%) tested positive for PiB retention. The ideal model for differentiating PiB(-) from PiB(+) SVaD was as follows: age ≤75 years, ≥5 lacunes, and MTA ≤3, which together yielded an accuracy of 67.5%. Conclusion: When patients meet the DSM-IV criteria for vascular dementia and also have severe white matter hyperintensities, younger age, greater number of lacunes, and lesser MTA, these are predictive of a PiB(-) scan in patients with SVaD. Classification of evidence: This study provides Class II evidence that the combination of younger age, greater number of lacunes, and lesser MTA identifies patients with SVaD at lower risk of Alzheimer disease pathology.