Functional Ensemble Survival Tree: Dynamic Prediction of Alzheimer’s Disease Progression Accommodating Multiple Time-Varying Covariates

Abstract

With the exponential growth in data collection, multiple time‐varying biomarkers are commonly encountered in clinical studies, along with a rich set of baseline covariates. This paper is motivated by addressing a critical issue in the field of Alzheimer’s disease (AD) in which we aim to predict the time for AD conversion in people with mild cognitive impairment to inform prevention and early treatment decisions. Conventional joint models of biomarker trajectory with time‐to‐event data rely heavily on model assumptions and may not be applicable when the number of covariates is large. This motivated us to consider a functional ensemble survival tree framework to characterize the joint effects of both functional and baseline covariates in predicting disease progression. The proposed framework incorporates multivariate functional principal component analysis to characterize the changing patterns of multiple time‐varying neurocognitive biomarker trajectories and then nest these features within an ensemble survival tree in predicting the progression of AD. We provide a fast implementation of the algorithm that accommodates personalized dynamic prediction that can be updated as new observations are gathered to reflect the patient’s latest prognosis. The algorithm is empirically shown to perform well in simulation studies and is illustrated through the analysis of data from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) (http://adni.loni.usc.edu/). We provide implementation of our proposed method in an R package funest.

Supervised Two-Dimensional Functional Principal Component Analysis with Time-to-Event Outcomes and Mammogram Imaging Data

“…, 2008; Jiang et al. , 2021) as shown in one of our simulation studies. Other radiomic feature‐based methods such as deep neural networks can also be adopted (Wu et al.…”

“…While we have proposed using the proportional hazards model in this article, our modeling framework can be directly extended to other types of survival setups. For instance, the features extracted from the proposed methods can be directly used under the random survival forest (Ishwaran et al, 2008;Jiang et al, 2021) as shown in one of our simulation studies. Other radiomic featurebased methods such as deep neural networks can also be adopted (Wu et al, 2019;McKinney et al, 2020).…”

Individual dynamic prediction of clinical endpoint from large dimensional longitudinal biomarker history: a landmark approach

“…Providing accurate predictions of health events that can exploit all the available individual information, even measured repeatedly over time, has become a major issue with the expansion of precise medicine. After the first proposals of dynamic predictions from repeated marker information [1,6], some authors have recently begun to tackle the problem of large dimension of longitudinal markers [18,19,22]. In comparison with this recent literature, our method has the advantage of (i) considering any nature of markers with measurement error while other considered only continuous outcomes [19], (ii) proposing the use of many summaries from the biomarkers as individual posterior computation from the longitudinal model (compared for instance to [22] who only include one or two summaries), (iii) exploiting the time-continuous information from survival data rather than discretized scale as in [22], and (iv) considering a vast variety of machine learning techniques as well as a superlearner rather than focusing only on one specific technique [18].…”

“…Computing dynamic predictions in the context of a large number of repeated markers is a very new topic in statistics, and only a few proposals have been made very recently. Zhao et al [18] and Jiang et al [19] focused on random forests. Using a landmark approach, Zhao et al transformed the survival data into pseudo-observations and incorporated in each tree the marker information at a randomly selected time.…”

Individual dynamic prediction of clinical endpoint from large dimensional longitudinal biomarker history: a landmark approach

“…Providing accurate predictions of health events that can exploit all the available individual information, even measured repeatedly over time, has become a major issue with the expansion of precise medicine. After the first proposals of dynamic predictions from repeated marker information [1,6], some authors have recently began to tackle the problem of large dimension of longitudinal markers [22,18,19]. In comparison with this recent literature, our method has the advantage of (i) considering any nature of markers with measurement error while other considered only continuous outcomes [19], (ii) proposing the use of many summaries from the biomarkers as individual posterior computation from the longitudinal model (compared for instance to [22] who only include one or two summaries), (iii) exploits the time-continuous information from survival data rather than discretized scale as in [22], and (iv) considering a vast variety of machine learning techniques as well as a superlearner rather than focusing only on one specific technique [18].…”