Reliability of BBB water exchange rate assessed by diffusion‐prepared pseudo‐continuous ASL (DP‐pCASL): Preliminary results from the Cognitive Recovery After Elective Surgery (CREATES) Study

Abstract

AbstractBackgroundImpaired blood‐brain barrier (BBB) function is hypothesized to be an important mechanism in the development of perioperative neurocognitive disorders (postoperative delirium and postoperative cognitive impairment), as well as in the onset and progression of Alzheimer’s (AD) and cerebrovascular disease. Non‐contrast neuroimaging methods have shown promise for detecting subtle alterations in BBB permeability; however, test‐retest reliability has not been well‐established. Here we report preliminary results for BBB water exchange rate (kw ) values from a novel diffusion‐prepared 3D pseudo‐continuous arterial spin labeling (DP‐pCASL) protocol with two different post‐labeling delays (PLD), and compared test‐retest reliability results with the original approach with a single PLD.MethodThe DP‐pCASL pulse sequence (resolution=3.5x3.5x8mm3, 10 slices) has been previously described (Shao X, et al., 2019). kw was calculated from the ratio of intra‐ and extra‐vascular perfusion signals using the original protocol with 1‐PLD [1.8 sec]; and the new protocol with 2‐PLDs [1.8 and 2.1 sec]. Total scan time was 15 minutes.Participants were healthy older adult controls enrolled in Cognitive Recovery After Elective Surgery (CREATES) study, a longitudinal study of relationships between BBB dysfunction and perioperative neurocognitive disorders following major elective surgery. Seventeen participants (11 females, age=78.5±5.4 yrs) underwent test‐retest MRI scans (∼30 days apart) on a Siemens 3T Prisma Fit system (Erlangen, Germany). kw maps were normalized into the MNI space, and regional analysis was performed for cortical gray/white matter, frontal/temporal/parietal areas, and nine subcortical brain regions. Test‐retest reliability was evaluated by intra‐class correlation coefficients (ICC).ResultWhole brain and regional kw values were calculated for the two protocols (Table 1). Test‐retest reliability increased using the 2‐PLD protocol. Whole brain kw ICC was 0.7 (good reliability; reference range, 0.5‐0.74) vs. 0.87 (excellent reliability; reference range, ≥ 0.75) for 1‐PLD and 2‐PLD protocols, respectively (Figure 1). ICC values increased across all brain regions with the 2‐PLD protocol (Figure 2).ConclusionIn a modest sample of cognitively healthy older adults, we found that measurement reliability of whole brain and regional kw values was notably improved by using the 2‐PLD DP‐pCASL protocol. These preliminary findings warrant confirmation in larger cohorts with and without late‐life cognitive disorders.