Prediction of Amyloid β and Tau Pathology using deep learning based artificial intelligence system...World Congress of Gerontechnology, October 22-26, 2022, Daegu, South Korea.

Purpose The pathological cascade of Alzheimer's disease (AD) begins decades before the development of clinical symptoms (Jack et al, 2010). Thus, predicting Alzheimer's before disease onset can minimize the socioeconomic burden and delay the cognitive impairments of patients. The Rey-complex figure...

Descripción completa

Detalles Bibliográficos
Publicado en:Gerontechnology Vol. 21; pp. 2 - 3
Autores principales: Lee, K. H., Park, J. Y., Seo, E. H., Won, S. H.
Formato: abstract proceedings research Journal Article
Publicado: International Society for Gerontechnology Oct2022
Acceso en línea:Ver este registro en EBSCOhost
fields @attributes:
  recordID: 1
pdfLink:
plink: https://search.ebscohost.com/login.aspx?direct=true&db=ccm&AN=161396103&site=ehost-live
header:
  @attributes:
    shortDbName: ccm
    uiTerm: 161396103
    longDbName: CINAHL Complete
    uiTag: AN
  controlInfo:
    bkinfo:
    dissinfo:
    jinfo:
      jid:
        15691101
        904R
      jtl: Gerontechnology
      issn: 15691101
      maglogo: N
    pubinfo:
      dt: Oct2022
      vid: 21
      pid: 54298
      pub: International Society for Gerontechnology
    artinfo:
      ui:
        161396103
        161396103
        161396103
        10.4017/gt.2022.21.s.821.2.sp3
        161396103
      ppf: 2
      ppct: 1
      formats:
      tig:
        atl: Prediction of Amyloid β and Tau Pathology using deep learning based artificial intelligence system...World Congress of Gerontechnology, October 22-26, 2022, Daegu, South Korea.
      aug:
        au:
          Lee, K. H.
          Park, J. Y.
          Seo, E. H.
          Won, S. H.
        affil: Department of Biomedical Science, Chosun University, Gwangju, Republic of Korea
      sug:
        subj:
          Deep Learning Utilization
          Amyloid beta-Peptides
          Tauopathies
          Predictive Validity
          Magnetic Resonance Imaging
          Positron-Emission Tomography
          Neurodegenerative Diseases Prognosis
          Congresses and Conferences South Korea
          South Korea
      ab: Purpose The pathological cascade of Alzheimer's disease (AD) begins decades before the development of clinical symptoms (Jack et al, 2010). Thus, predicting Alzheimer's before disease onset can minimize the socioeconomic burden and delay the cognitive impairments of patients. The Rey-complex figure task (RCFT) was revisited as the sensitive neurocognitive evaluator that is associated with cerebrospinal fluid amyloid β, tau levels and neurodegeneration observed by magnetic resonance imaging (MRI) (Seo et al, 2021). But the application of these findings is limited in real life. Assessing brain biomarker content with positron emission tomography (PET) is not readily accessible to ordinary patients, and the scoring system of RCFT is unstable due to the interrater gap in scores. To overcome this problem, we propose AI, deep-learning-based solutions for RCFT scoring, and MRI-based PET prognosis. Method We obtained 20,040 scanned RCFT images from Gwangju Alzheimer's and Related Dementia cohort in Korea. The images copy, immediate recall, and delayed recall) rated by experienced psychologists were fed into an input for the DL model. DenseNet (Huang et al, 2017) architecture was used as the backbone. Finally, we conducted an external validation with 150 images scored by five experienced psychologists. The PET prognosis platform, NeuroAI, computes PET positivity rate based on T1, T2 Flair image, Apoe4 type, and demographics of patients. Results and discussion Our model obtained mean absolute error (MAE) = 1.24 [points] and R-squared (R²) = 0.977 for 5-fold cross-validation. For the 150 independent test sets, the MAE and R² between our model and average scores by five human experts were 0.64 [points] and 0.994, respectively. Our results suggested no fundamental difference between the rating scores of human experts and those of our AI psychologists. Altogether, our work proves the potential of AI usage as a faster and more cost-effective to contribute to screening the early stages of AD.
      pubtype: Academic Journal
      doctype:
        abstract
        proceedings
        research
        Journal Article
      ougenre: Article
    language: English
    refInfo:
    holdings:
      @attributes:
        islocal: N