BMC Neurology (Aug 2023)

PRedicting the EVolution of SubjectIvE Cognitive Decline to Alzheimer’s Disease With machine learning: the PREVIEW study protocol

  • Salvatore Mazzeo,
  • Michael Lassi,
  • Sonia Padiglioni,
  • Alberto Arturo Vergani,
  • Valentina Moschini,
  • Maenia Scarpino,
  • Giulia Giacomucci,
  • Rachele Burali,
  • Carmen Morinelli,
  • Carlo Fabbiani,
  • Giulia Galdo,
  • Lorenzo Gaetano Amato,
  • Silvia Bagnoli,
  • Filippo Emiliani,
  • Assunta Ingannato,
  • Benedetta Nacmias,
  • Sandro Sorbi,
  • Antonello Grippo,
  • Alberto Mazzoni,
  • Valentina Bessi

DOI
https://doi.org/10.1186/s12883-023-03347-8
Journal volume & issue
Vol. 23, no. 1
pp. 1 – 13

Abstract

Read online

Abstract Background As disease-modifying therapies (DMTs) for Alzheimer's disease (AD) are becoming a reality, there is an urgent need to select cost-effective tools that can accurately identify patients in the earliest stages of the disease. Subjective Cognitive Decline (SCD) is a condition in which individuals complain of cognitive decline with normal performances on neuropsychological evaluation. Many studies demonstrated a higher prevalence of Alzheimer’s pathology in patients diagnosed with SCD as compared to the general population. Consequently, SCD was suggested as an early symptomatic phase of AD. We will describe the study protocol of a prospective cohort study (PREVIEW) that aim to identify features derived from easily accessible, cost-effective and non-invasive assessment to accurately detect SCD patients who will progress to AD dementia. Methods We will include patients who self-referred to our memory clinic and are diagnosed with SCD. Participants will undergo: clinical, neurologic and neuropsychological examination, estimation of cognitive reserve and depression, evaluation of personality traits, APOE and BDNF genotyping, electroencephalography and event-related potential recording, lumbar puncture for measurement of Aβ42, t-tau, and p-tau concentration and Aβ42/Aβ40 ratio. Recruited patients will have follow-up neuropsychological examinations every two years. Collected data will be used to train a machine learning algorithm to define the risk of being carriers of AD and progress to dementia in patients with SCD. Discussion This is the first study to investigate the application of machine learning to predict AD in patients with SCD. Since all the features we will consider can be derived from non-invasive and easily accessible assessments, our expected results may provide evidence for defining cost-effective and globally scalable tools to estimate the risk of AD and address the needs of patients with memory complaints. In the era of DMTs, this will have crucial implications for the early identification of patients suitable for treatment in the initial stages of AD. Trial registration number (TRN) NCT05569083.

Keywords