Development and validation of a 1-year survival prognosis estimation model for Amyotrophic Lateral Sclerosis using manifold learning algorithm UMAP

Abstract

Amyotrophic Lateral Sclerosis (ALS) is an inexorably progressive neurodegenerative condition with no effective disease modifying therapies. The development and validation of reliable prognostic models is a recognised research priority. We present a prognostic model for survival in ALS where result uncertainty is taken into account. Patient data were reduced and projected onto a 2D space using Uniform Manifold Approximation and Projection (UMAP), a novel non-linear dimension reduction technique. Information from 5,220 patients was included as development data originating from past clinical trials, and real-world population data as validation data. Predictors included age, gender, region of onset, symptom duration, weight at baseline, functional impairment, and estimated rate of functional loss. UMAP projection of patients shows an informative 2D data distribution. As limited data availability precluded complex model designs, the projection was divided into three zones with relevant survival rates. These rates were defined using confidence bounds: high, intermediate, and low 1-year survival rates at respectively ( ), ( ) and ( ). Predicted 1-year survival was estimated using zone membership. This approach requires a limited set of features, is easily updated, improves with additional patient data, and accounts for results uncertainty.

Pathological neural networks and artificial neural networks in ALS: diagnostic classification based on pathognomonic neuroimaging features

“…The utility of a variety of supervised and unsupervised machine-learning approaches have been explored in ALS, including support vector machines, regressionbased approaches, random forests, discriminant function analyses, dimension reduction frameworks, but these are seldom applied to imaging data [17,35,36] due to challenges associated with MRI scanning, quality control, preprocessing, data acquisition costs and data harmonisation. Advanced neural network architectures have been successfully trialled in other conditions, including multilayer 'deep-learning' learning models and generative adversarial networks (GAN) [37][38][39][40].…”

Frontotemporal Pathology in Motor Neuron Disease Phenotypes: Insights From Neuroimaging

“…High-field MRI generates better quality images and acceleration techniques enable shorter data acquisition that may be better tolerated by patients. Quantitative MRI analyses using validated computational pipelines and reliance on robust comparative, correlative, and classifier models enhance the clinical interpretation of vast imaging datasets (270). The advent of structural and functional connectivity studies have ignited interest in the concept of disease-specific selective network degeneration rather than the emphasis on focal pathology (271).…”

Prediction of survival in amyotrophic lateral sclerosis: a nationwide, Danish cohort study

“…In addition to these four studies, other groups have also attempted to create predictive models in ALS with progression rate as one of the predictors [16][17][18]. Of these, the study of Westeneng et al stands out by its multinational design and cohort size [16].…”