Discordance between cerebrospinal fluid (CSF) biomarkers and amyloid PET imaging in Alzheimer's Disease (AD) is an area of active research. Disorders in CSF dynamics, however, are often not discussed in these studies. Features of abnormal CSF dynamics can be obtained from T1-weighted MRI. In particular, Disproportionately Enlarged Subarachnoid-space Hydrocephalus (DESH), characterized by tight CSF spaces at the high convexity, ventriculomegaly and enlarged Sylvian fissures, is a strong marker of abnormal CSF dynamics. We have an automated method to detect these imaging features on T1-weighted MRI; i.e. a “computational DESH” or CDESH score1 (Figure 1). Using CDESH as a surrogate measure of altered CSF dynamics we investigate if it helps explain discrepancies between CSF and PET biomarkers of amyloid in the ADNI cohort. CDESH scores for baseline MRI from 843 ADNI-G0/2 participants with CSF assays for aB1-42, pTau, tTau and AV45-SUVR were calculated. CDESH+ and CDESH- groups were defined (Figure 2). We tested the hypothesis that the group-wise distributions of each biomarker were consistent using an empirical Kolomogorov-Smirnov test. Baseline demographics and AD biomarker results, divided by diagnostic group, is shown in Table 1. Seven percent of participants were CDESH+. Distributions of CSF ab1-42, pTau, and tTau, were significantly inconsistent between groups (Figure 3). The distributions for AV45 were not significantly inconsistent between groups (Figure 4). In a subsample (N=143 participants) CSF Ab 1–40 and 1–42 values were available: the 42:40 ratio distributions are not significantly inconsistent between groups. The ratio of pTau to Ab1-42 was not significantly inconsistent between groups. Thirty percent of CDESH+ participants have low CSF Ab1-42 and low AV45 (i.e. discordant markers), different than the 10% of CDESH- participants (Yates-corrected ratio test p<0.001) (Figure 5). MMSE and CDR sum-of-boxes distributions did not differ between groups.