This study developed a machine learning classifier to analyze early frame amyloid PET scans and measure neurodegeneration related to Alzheimer's disease. The classifier produced scores that strongly correlated with scores from an FDG PET classifier, indicating it can detect disease progression. Early frame amyloid scans had lower cortical but higher subcortical signals compared to FDG PET. Nonetheless, the amyloid PET classifier achieved similar performance to FDG PET at differentiating disease stages. This suggests early frame amyloid PET can provide a measure of neurodegeneration without an additional scan, useful for clinical trials and diagnosis.