Diagnostic accuracy of CTseg segmentation software for dementia in Pacific peoples living in Aotearoa New Zealand
Journal of Brain Health · 05.06.2026 · 1 (1)
DOI: 10.65104/001c.162254 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Brain Health, online erschienen am 05.06.2026. Verfasst von Mukish Yelanchezian und 6 weiteren Autor:innen. Die Arbeit nennt 38 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Themen: Demenz
Abstract (Originalsprache)
Background
Aotearoa New Zealand’s Pacific population represents one of the largest diasporas globally, yet experiences a disproportionate burden of dementia– with prevalence rates nearly double that observed in New Zealand (NZ) European populations. Furthermore, the onset is often significantly earlier, compromising opportunities for timely diagnosis and intervention. Automated CT-based segmentation tools, such as CTseg, hold significant promise for identifying undetected dementia in Pacific peoples.
Objective
We aimed to rigorously evaluate the diagnostic accuracy of CTseg-derived volumetric markers—specifically total brain volume and hippocampal volume—for detecting dementia in Pacific patients receiving care in NZ memory services.
Methods
De-identified health and imaging data were analyzed from 204 older Pacific patients living in NZ. Logistic regression models adjusted for age and total intracranial volume (TIV) were used to estimate diagnostic accuracy metrics using Youden index, sensitivity, specificity, and area under the receiver operating characteristic curve (AUC).
Results
The diagnostic accuracy of CTseg-derived total brain volume was moderate with an AUC of 0.78 (95% CI: 0.72-0.85), sensitivity of 66.9% (95% CI: 58.3%-74.5%), and specificity of 79.2% (95% CI: 68.9%-74.5%). For hippocampal volume, the AUC was 0.75 (95% CI: 0.68-0.82), sensitivity 55.9% (95% CI: 47.2%-64.2%), and specificity 83.1% (95% CI: 73.2%-89.9%).
Discussion
CTseg produced volumetric estimates in a sample of Pacific Island patients that were comparable to those previously reported for a sample of European patients from a real-world memory service. Accessibility and scalability suggest potential utility for CT-based dementia detection across diverse NZ populations, facilitating equitable access to timely diagnosis and care.
Zitieren
APA
Yelanchezian, M., Ma'u, E., Oulghan, B., Yates, S., Dobbie, G., Cheung, G., & Cullum, S. J. (2026). Diagnostic accuracy of CTseg segmentation software for dementia in Pacific peoples living in Aotearoa New Zealand. Journal of Brain Health, 1(1). https://doi.org/10.65104/001c.162254
Vancouver
Yelanchezian M, Ma'u E, Oulghan B, Yates S, Dobbie G, Cheung G, et al. Diagnostic accuracy of CTseg segmentation software for dementia in Pacific peoples living in Aotearoa New Zealand. Journal of Brain Health. 2026;1(1). doi:10.65104/001c.162254
BibTeX
@article{Yelanchezian2026,
author = {Yelanchezian, Mukish and Ma'u, Etuini and Oulghan, Bede and Yates, Susan and Dobbie, Gillian and Cheung, Gary and Cullum, Sarah J},
title = {Diagnostic accuracy of CTseg segmentation software for dementia in Pacific peoples living in Aotearoa New Zealand},
journal = {Journal of Brain Health},
year = {2026},
volume = {1},
number = {1},
doi = {10.65104/001c.162254}
}
EUVIMED