Quantification supports amyloid-PET visual assessment of challenging cases: results from the AMYPAD-DPMS study

Abstract

ABSTRACTSeveral studies have demonstrated the high agreement between routine clinical visual assessment and quantification, suggesting that quantification approaches could support the assessment of less experienced readers and/or in challenging cases. However, all studies to date have implemented a retrospective case collection and challenging cases were generally underrepresented.MethodsIn this prospective study, we included all participants (N=741) from the AMYPAD Diagnostic and Patient Management Study (DPMS) with available baseline amyloid-PET quantification. Quantification was done with the PET-only AmyPype pipeline, providing global Centiloid (CL) and regional z-scores. Visual assessment was performed by local readers for the entire cohort. From the total cohort, we selected a subsample of 85 cases 1) for which the amyloid status based on the local reader’s visual assessment and CL classification (cut-off=21) was discordant and/or 2) that were assessed with a low confidence (i.e. ≤3 on a 5-point scale) by the local reader. In addition, concordant negative (N=8) and positive (N=8) scans across tracers were selected. In this sample, (N=101 cases: ([18F]flutemetamol,N=48; [18F]florbetaben,N=53) the visual assessments and corresponding confidence by 5 certified independent central readers were captured before and after disclosure of the quantification results.ResultsFor the AMYPAD-DPMS whole cohort, the overall assessment of local readers highly agreed with CL status (κ=0.85, 92.3% agreement). This was consistently observed within disease stages (SCD+: κ=0.82/92.3%; MCI: κ=0.80/89.8%; dementia: κ=0.87/94.6%). Across all central reader assessments in the challenging subsample, global CL and regional z-scores quantification were considered supportive of visual read in 70.3% and 49.3% of assessments, respectively. After disclosure of quantitative results, we observed an improvement in concordance between the 5 readers (κbaseline=0.65/65.3%; κpost-disclosure=0.74/73.3%) and a significant increase in reader confidence (Mbaseline=4.0 vs.Mpost-disclosure=4.34,W=101056,p<0.001).ConclusionIn this prospective study enriched for challenging amyloid-PET cases, we demonstrate the value of quantification to support visual assessment. After disclosure, both inter-reader agreement and confidence showed a significant improvement. These results are important considering the arrival of anti-amyloid therapies, which utilized the Centiloid metric for trial inclusion and target-engagement. Moreover, quantification could support determining Aβ status with high certainty, an important factor for treatment initiation.