Premium
P4‐338: MACHINE LEARNING‐BASED HISTOPATHOLOGICAL APPROACH TO TAU PATHOLOGY
Author(s) -
Signaevski Maxim,
Farrell Kurt W.,
Tabish Nabil,
Baldwin Elena,
Prastawa Marcel,
Koll John,
Fernandez Gerardo,
Zeineh Jack,
Purohit Dushyant,
Hanson Russell,
Crary John
Publication year - 2018
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1016/j.jalz.2018.07.161
Subject(s) - digital pathology , deep learning , artificial intelligence , convolutional neural network , senile plaques , pathology , computer science , neuroscience , alzheimer's disease , medicine , disease , biology
Background:Creutzfeldt-Jakob disease (CJD) is a rapidly progressive dementia with an illness duration averaging approximately 46 months. There are several challenges to conducting research and surveillance activities on CJD, including its rapid progression, difficulty to diagnosis, geographic dispersal, and rarity. Several studies have been conducted examining the utility of using teleneurology for research purposes of neurologic diseases. We conducted a feasibility study of using teleneurology for research and surveillance purposes inCJD.Methods:Subjectswere included in the study if theymet criteria for probable sporadic CJD (sCJD) or had a positive real-time quaking induced conversion (RT-QuIC) result. Subjects and a research partner were given a choice of in-person visit, teleneurology visit, or medical record review only. A standardized history and examination were collected as well as standardized instruments to measure cognition (Telephone Interview for Cognitive Status, TICS), functional status (MRC Prion Disease Rating Scale), and neuropsychiatric symptoms (Neuropsychiatric Inventory Questionnaire, NPI-Q). For teleneurology visits, subjects participated in the evaluation using secure software (Cisco Jabber) that could be installed on any internet connected device with a camera of their choice. Subjects were followed longitudinally on a monthly or bimonthly basis. Results:Over a 10-month period, the study received 81 referrals from 27 states. All but two enrolled subjects (95%) chose the teleneurology arm of the study and the majority participated from their homes (88%). Subjects and study partners expressed ease of use, convenience, and likelihood of recommending the teleneurology modality over other research modalities to potential participants. The quality of the examination depended on subject cooperation aswell as internet connectivity. All but one participant proceeded to autopsy (95% autopsy rate). Conclusions: This study demonstrates the feasibility of conducting research and surveillance activities of CJD subjects using teleneurology. Subjects preferred this modality, felt comfortable with its use, and may have been more likely to proceed to autopsy given involvement in the study. Further research should study the validity and reliability of instruments used to study this population remotely.