This report presents three cases of atypical degenerative dementias in order to illustrate challenges associated with the use of biologic markers of Alzheimer's disease (AD) for diagnosis and management. Clinical diagnostic methods followed the NINCDS-ADRDA criteria for AD. Additional diagnostic studies included serial neurocognitive testing, MRI, neuroSPECT, ApoE genotyping, and a CSF assay of tau and beta-amyloid42. For patient 1, both the clinical and biologic markers were consistent with AD. The patient was diagnosed with AD with a high degree of confidence, even though the base rate of nonfamilial AD at this age group (<55 years) is exceedingly rare. This case argues favorably for the use of biologic markers to aid in confirming a diagnosis in an atypical dementia. Patient 2 met the NINCDS-ADRDA criteria for AD, although with less confidence. Neurocognitive data indicated a progressive right hemispheric syndrome, insight was preserved, and ApoE was 3/3, but tau concentrations and beta-amyloid42 were highly consistent with cut-offs for AD; the differential fell on the Pick's disease/frontotemporal degeneration spectrum. Patient 3 had no clinical evidence of the disease, even when evaluated via extensive neurocognitive testing over a 2-year interval. However, ApoE was 4/4, and CSF assay of tau and beta-amyloid42 were within the AD range. Therefore, while the CSF assay of tau and beta-amyloid42 markers was confirmatory of AD, each case was highly atypical. Results illustrate the lack of normative data available when using biologic markers for highly atypical cases, calling into question their usefulness for such patients. These cases illustrate the interplay between neuropsychological and biological markers in establishing neurodegenerative diagnoses.