Prediction of conversion to Alzheimer’s disease using deep survival analysis of MRI images

Abstract

The prediction of the conversion of healthy individuals and those with mild cognitive impairment to the status of active Alzheimer’s disease is a challenging task. Recently, a survival analysis based upon deep learning was developed to enable predictions regarding the timing of an event in a dataset containing censored data. Here, we investigated whether a deep survival analysis could similarly predict the conversion to Alzheimer’s disease. We selected individuals with mild cognitive impairment and cognitively normal subjects and used the grey matter volumes of brain regions in these subjects as predictive features. We then compared the prediction performances of the traditional standard Cox proportional-hazard model, the DeepHit model and our deep survival model based on a Weibull distribution. Our model achieved a maximum concordance index of 0.835, which was higher than that yielded by the Cox model and comparable to that of the DeepHit model. To our best knowledge, this is the first report to describe the application of a deep survival model to brain magnetic resonance imaging data. Our results demonstrate that this type of analysis could successfully predict the time of an individual’s conversion to Alzheimer’s disease.

A comparison of machine learning methods for survival analysis of high-dimensional clinical data for dementia prediction

“…Li et al 24 developed a multi-task learning-based survival analysis framework that handles block-wise missing data. Several authors [25][26][27] have developed neural networks for survival analysis on data from the ADNI. However, a comprehensive survey paper on the topic of machine learning for survival analysis does not include any studies employing these techniques for the prediction of Alzheimer's disease 28 .…”

Multilevel Survival Modeling With Structured Penalties for Disease Prediction From Imaging Genetics Data

“…We compared our method to various variations of the Cox model with different penalties and different types of input data. We also compared it to a recently proposed deep learning approach (deep survival analysis) [8]. The most striking result is that the use of MRI (either in isolation or when combined with genetics) resulted in a strong improvement when compared to genetic, age or gender (from 0.12 to 0.21 points of iAUC).…”

“…• the Cox Proportional Hazards model using only gender and age as input variables • the Cox Proportional Hazards model with a lasso penalty using only genetic data as input variables • the Cox Proportional Hazards model with a lasso penalty using genetic data, gender and age as input variables • the Cox Proportional Hazards model (either without penalty or with the ridge penalty) using only imaging data • a deep learning approach (deep survival analysis) recently proposed by Nakagawa et al [8] using only imaging data • the Cox Proportional Hazards using both genetic and imaging data which are combined in an additive framework. In this case, the hazard function is given by:…”

“…Deep learning approaches have also been proposed for survival analysis of progression to AD. This includes deep survival analysis which extends a regularized multi-layer perceptron to account for censored data [8] and temporal convolutional networks [9]. However, these different techniques do not combine imaging and genetic data.…”

Deep learning prediction of mild cognitive impairment conversion to Alzheimer’s disease at 3 years after diagnosis using longitudinal and whole-brain 3D MRI

“…Future studies should also combine imaging with non-imaging data, such as neurocognitive scores ( Nagaraj & Duong, 2021 ) in predictive models using ML. Nakagawa et al used deep survival analysis to model the prediction of conversion from either MCI or NC subjects to AD using volumetric data from MRI ( Nakagawa et al, 2020 ). A future extension of this analysis should investigate the use of data from the CNN models, both single-channel and longitudinal, using features extracted at the end of the convolutional layers.…”