Machine learning-based radiomics for amyotrophic lateral sclerosis diagnosis - Article abstract

Aug 3, 2026

Upload by Eden

Timely diagnosis and accurate phenotyping of amyotrophic lateral sclerosis (ALS) are critical for clinical management. The study evaluates radiomics analysis on T1-weighted MRI to build a machine learning classification pipeline. Data include 53 controls and 84 ALS patients from three different scanners, with dataset harmonization, feature selection, and multiple ML algorithms. A combined LASSO with SVM achieves 81.1% accuracy for ALS vs controls, while mRMR with SVM reaches 92.9% accuracy for “Classic” vs “non-Classical” motor phenotypes.

LanguageEnglish

CategoryHealthcare

FormatPDF

Pages9

Reading Time23 min

AccessFree to download

Introduction

Diagnostic complexity and criteria

Importance of early diagnosis and clinical trials

ALS phenotypic heterogeneity and classification systems

Expert Systems With Applications 240 (2024) 122585 Contents lists available at ScienceDirect Expert Systems With Applications journal [homepage: www.elsevier.com/locate/eswa](homepage: www.elsevier.com/locate/eswa) | Machine learning-based radiomics for amyotrophic lateral sclerosis diagnosis
Benedetta Tafuria, b, *, Giammarco Milellaa, b, Marco Filardia, b, Alessia Giugno b, d, Stefano Zoccolellab, c, Ludovica Tamburrino a, b, Valentina Gnonib, Daniele Ursob, Roberto De Blasib, Salvatore Nigrob, 1, Giancarlo Logroscinoa, b, 1
a Department of Translational Biomedicine and Neuroscience (DiBraiN), University of Bari Aldo Moro, Bari, Italy
b Center for Neurodegenerative Diseases and the Aging Brain, University of Bari Aldo Moro at Pia Fondazione “Card. G. Panico”, Tricase, Italy c Neurology Unit, San Paolo Hospital, Azienda Sanitaria Locale (ASL) Bari, Bari, Italy
d Department of Medical and Surgical Sciences, Institute of Neurology, Magna Graecia University, Catanzaro, Italy | | | | --- | --- | --- | | A R T I C L E I N F O | | A B S T R A C T | | Keywords:
Amyotrophic lateral sclerosis Magnetic resonance imaging Radiomics
Machine learning | | Timely diagnosis and accurate phenotyping of amyotrophic lateral sclerosis (ALS) is of paramount importance for the clinical management of patients. Magnetic Resonance Imaging (MRI) plays a key role in the clinical work-up of ALS. In this study we investigated the usefulness of radiomics analysis on T1-weighted MRI to define a machine learning-based classification pipeline.
We collected 53 controls and 84 patients with ALS from three different scanners. Following dataset harmonization, radiomics analysis was conducted using different features selection and machine learning algorithms to identify the best combination in distinguishing ALS patients from controls and “Classic” from “non-Classical” ALS motor phenotypes.
The combined Least Absolute Shrinkage and Selection Operator with Support Vector Machine (SVM) algorithm classified ALS patients with an accuracy of 81.1%. The Maximum Relevance Minimum Redundancy with SVM pipeline was able to distinguish “Classic” from “non-Classical” motor phenotypes with 92.9% accuracy.
Radiomics is a promising approach to characterize brain abnormalities in patients with ALS. Radiomics could help to improve diagnosis and may prove useful to assess disease severity and longitudinally monitor ALS patients along the disease course. | 1. Introduction Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disease showing a progressive degeneration of motor neurons in the cortex, brainstem and spinal cord (Kiernan et al., 2011). Although first described over 150 years ago, the diagnostic landscape for ALS remains complex. Therefore, various diagnostic criteria for ALS have been proposed over time: while some emphasize clinical evaluations of upper motor neuron (UMN) and lower motor neuron (LMN) signs (Brooks, 1994; Brooks et al., 2000; Hannaford et al., 2021), others incorporate neurophysiological or instrumental exams for increased diagnostic precision (de Carvalho et al., 2008). Achieving an early diagnosis is critical as it allows timely intervention and facilitates patient enrollment in clinical trials. Thanks to the introduction of new classification systems and through many population-based studies, the understanding of ALS epidemiology has progressively advanced. For instance, the EURALS consortium, which collected data from nearly 24 million individuals across Europe, reported an ALS incidence of 2.2 per 100,000 personyears. In contrast, studies have found the lowest incidences in East Asia and South Asia, at 0.89 and 0.79 per 100,000 person-years, respectively (Logroscino & Piccininni, 2019). Simultaneous degeneration of UMN and LMN not only represents a hallmark of the diagnosis of ALS disease but also explains the extreme phenotypic heterogeneity among ALS patients. In this context, over the past

FAQs

Why is timely diagnosis of amyotrophic lateral sclerosis (ALS) important?

Early diagnosis enables timely intervention and supports patient enrollment in clinical trials, improving clinical management and study participation.

What MRI input and overall method does the radiomics pipeline use?

Radiomics features are extracted from T1-weighted MRI, followed by dataset harmonization, feature selection, and machine learning classification algorithms.

How accurate are the proposed models for ALS classification and phenotype separation?

The LASSO-SVM pipeline distinguishes ALS patients from controls with 81.1% accuracy, and the mRMR-SVM pipeline separates “Classic” from “non-Classical” phenotypes with 92.9% accuracy.

Aug 3, 2026

Upload by Eden