Subtle blood‐brain barrier leakage rate and spatial extent: Considerations for dynamic contrast‐enhanced <scp>MRI</scp>

<jats:sec><jats:title>Purpose</jats:title><jats:p>Dynamic contrast‐enhanced (<jats:styled-content style="fixed-case">DCE</jats:styled-content>) <jats:styled-content style="fixed-case">MRI</jats:styled-content> can be used to measure blood‐brain barrier (<jats:styled-content style="fixed-case">BBB</jats:styled-content>) leakage. In neurodegenerative disorders such as small vessel disease and dementia, the leakage can be very subtle and the corresponding signal can be rather noisy. For these reasons, an optimized <jats:styled-content style="fixed-case">DCE</jats:styled-content>‐<jats:styled-content style="fixed-case">MRI</jats:styled-content> measurement and study design is required. To this end, a new measure indicative of the spatial extent of leakage is introduced and the effects of scan time and sample size are explored.</jats:p></jats:sec><jats:sec><jats:title>Methods</jats:title><jats:p>Dual‐time resolution <jats:styled-content style="fixed-case">DCE</jats:styled-content>‐<jats:styled-content style="fixed-case">MRI</jats:styled-content> was performed in 16 patients with early Alzheimer's disease (<jats:styled-content style="fixed-case">AD</jats:styled-content>) and 17 healthy controls. The leakage rate (K<jats:sub>i</jats:sub>) and volume fraction of detectable leaking tissue (<jats:styled-content style="fixed-case">v<jats:sub>L</jats:sub></jats:styled-content>) to quantify the spatial extent of <jats:styled-content style="fixed-case">BBB</jats:styled-content> leakage were calculated in cortical gray matter and white matter using noise‐corrected histogram analysis of leakage maps. Computer simulations utilizing realistic K<jats:sub>i</jats:sub> histograms, mimicking the strong effect of noise and variation in K<jats:sub>i</jats:sub> values, were performed to understand the influence of scan time on the estimated leakage.</jats:p></jats:sec><jats:sec><jats:title>Results</jats:title><jats:p>The mean K<jats:sub>i</jats:sub> was very low (order of 10<jats:sup>−4</jats:sup> min<jats:sup>−1</jats:sup>) and highly influenced by noise, causing the K<jats:sub>i</jats:sub> to be increasingly overestimated at shorter scan times. In the white matter, the K<jats:sub>i</jats:sub> was not different between patients with early <jats:styled-content style="fixed-case">AD</jats:styled-content> and controls, but was higher in the cortex for patients, reaching significance after 14.5 min of scan time. To detect group differences, <jats:styled-content style="fixed-case">v<jats:sub>L</jats:sub></jats:styled-content> proved more suitable, showing significantly higher values for patients compared with controls in the cortex after 8 minutes of scan time, and in white matter after 15.5 min.</jats:p></jats:sec><jats:sec><jats:title>Conclusions</jats:title><jats:p>Several ways to improve the sensitivity of a <jats:styled-content style="fixed-case">DCE</jats:styled-content>‐<jats:styled-content style="fixed-case">MRI</jats:styled-content> experiment to subtle <jats:styled-content style="fixed-case">BBB</jats:styled-content> leakage were presented. We have provided <jats:styled-content style="fixed-case">v<jats:sub>L</jats:sub></jats:styled-content> as an attractive and potentially more time‐efficient alternative to detect group differences in subtle and widespread blood‐brain barrier leakage compared with leakage rate K<jats:sub>i</jats:sub>. Recommendations on group size and scan time are made based on statistical power calculations to aid future research.</jats:p></jats:sec>