Objective: Summarising the development of mathematical models to further HD research through characterisation and disease progression.

Background: There have been major advances in HD research in the last decade, aided by the development of mathematical models. Presented are a collection of models that serve as tools to facilitate the characterisation and progression of HD.(1–4)

Method: Data was analysed from Enroll-HD, PREDICT-HD, REGISTRY, and Track-HD. Characteristic models have been developed such as a new CAP score that determines age-of-onset, an online tool EHDPA that summarises HD phenotypes, and a logistic regression model that examines the frequency of comorbidities in PwHD.

Other models assess the progression of HD. Scaled coefficients from a Cox regression model created a normalised prognostic index (PIN) score for motor diagnosis in HD, to predict HD progression. A probabilistic ML model identified hidden patterns related to disease progression. The HD-ISS was developed with a consensus methodology, to standardise staging. Lastly, a joint model (JM) using TMS, SDMT and CAG expansion was developed to predict age of motor onset and a deviance residual between observed and model-based status.

Results: Results show that the new CAP score is suitable for use in modelling of clinical data. EHDPA summarises assessment measures, and comorbidities such as sleep disorders are reported more frequently for PwHD. Disease progression models such as the PIN score was calculated by combining weighted measures of UHDRS‑TMS, SDMT, and CAP score. Nine disease states were identified and suggest that subtle motor and cognitive changes precede a motor diagnosis. The HD‑ISS classifies 4 stages of HD, with excellent internal validity. Lastly, a JM can discriminate diagnosed from pre-diagnosed individuals.

Conclusion: These models provide robust tools to understand natural history studies, track disease progression and interpret investigational drug-related adverse events by characterising the fundamental features of HD, such as age-of-onset, comorbidities, and providing descriptive reports. Disease progression models can predict HD progression and identify cohorts for clinical trial recruitment.

This research was supported by statistics and modelling team collaborators and entities who contributed to the CHDI modelling efforts.

References: 1. Warner JH, Long JD, Mills JA, Langbehn DR, Ware J, Mohan A, et al. Standardizing the CAP Score in Huntington’s Disease by Predicting Age-at-Onset. J Huntingtons Dis. 2022;11(2):153–71.
2. Langbehn DR, Sathe SS, Loy C, Sampaio C, Mccusker EA. A Phenotypic Atlas for Huntington Disease Based on Data From the Enroll-HD Cohort Study. Neurol Genet. 2023 Dec;9(6).
3. Griffin BA, Booth MS, Busse M, Wild EJ, Setodji C, Warner JH, et al. Estimating the causal effects of modifiable, non-genetic factors on Huntington disease progression using propensity score weighting. Parkinsonism Relat Disord. 2021 Feb 1;83:56–62.
4. Tabrizi SJ, Schobel S, Gantman EC, Mansbach A, Borowsky B, Konstantinova P, et al. A biological classification of Huntington’s disease: the Integrated Staging System. Lancet Neurol [Internet]. 2022 Jul;21(7):632–44. Available from: http://www.ncbi.nlm.nih.gov/pubmed/35716693

« Back to 2024 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/the-use-of-modelling-in-huntingtons-disease/