Authors: Daria Pawlik, Antoine Leuzy, Olof Strandberg, Ruben Smith
Published: 2020-12-07
DOI: 10.1002/alz.041800
Source: Full article
AbstractBackgroundThe hippocampus is an important region to study in Alzheimer’s disease (AD) since it is an early site of tau accumulation. Accurate quantification of hippocampal tau signal using the most common tau‐PET tracer 18F‐flortaucipir is complicated, however, by off‐target binding in the adjacent choroid plexus. We here present a novel method for compensating for this off‐target choroid plexus signal.MethodIt has been shown that off‐target binding in the choroid plexus is higher using 18F‐flortaucipir compared to the recently developed tau tracer 18F‐RO948. Using data from 30 patients that underwent both 18F‐flortaucipir and 18F‐RO948 PET, we created a binary hippocampal mask where 18F‐flortaucipir signal was higher than 18F‐RO948 signal (Figure 1). This mask, representing voxels affected by off‐target binding from the choroid plexus, was then applied to 18F‐flortaucipir PET‐scans from 145 patients across the AD‐spectrum. In this way the off‐target component was excluded. Results were analysed with and without partial volume error correction (PVEc/non‐PVEc).ResultsWe found that unmasked hippocampal standardized uptake value ratios (SUVR) were significantly correlated to choroid plexus SUVRs using both non‐PVEc (r = 0.28, p < 0.001) and PVEc data (r = 0.18, p < 0.05) (Figure 2a‐b). After applying the mask these correlations disappeared (Figure 2c‐d). We next tested the diagnostic accuracy in separating cognitively impaired (CI) from cognitively unimpaired (CU) subjects using receiver operating characteristic (ROC) curve analysis. The area under the curve (AUC) from ROC analysis increased after masking from 0.79 (95% C.I., 0.72–0.87) to 0.84 (95% C.I., 0.77–0.91) for non‐PVEc data, and from 0.80 (95% C.I., 0.72–0.87) to 0.83 (95% C.I., 0.77–0.90) for PVEc data (Figure 3). Correlations between 18F‐flortaucipir hippocampal SUVR and MMSE improved significantly for masked vs unmasked data, both with (r = ‐0.50 vs. r = ‐0.45) and without (r = ‐0.50 vs. r = ‐0.44) PVEc. Similar results were found using the ADAS‐Cog Delayed Word Recall test.ConclusionCorrecting 18F‐flortaucipir hippocampal SUVR for spill‐in from the choroid plexus using an off‐target mask is feasible, increases the diagnostic accuracy of hippocampal SUVRs and improves correlation with cognitive measures. Importantly, the proposed mask provided better results than applying PVE correction alone.