Abstract
Paediatric-onset multiple sclerosis (POMS) causes white-matter lesion accumulation during a critical period of brain maturation, yet conventional MRI summaries reduce lesion burden largely to count and volume and therefore discard three-dimensional surface geometry. We investigate whether spectral shape descriptors provide complementary lesion-level information. Each lesion surface was encoded by its Laplace–Beltrami eigenspectrum (Shape-DNA) together with curvature and classical morphometric descriptors. Applied to 940 lesions from 25 longitudinal 3D FLAIR scans of nine POMS patients, $k$-means clustering in a 19-dimensional feature space identified two highly reproducible morphotypes (379 compact/focal and 561 larger/diffuse lesions; leave-one-patient- out mean Adjusted Rand Index $0.997\pm0.008$). Internal validity indices disagreed on the optimal number of clusters, whereas bootstrap stability favoured $k{=}2$ (mean Jaccard $0.984$ versus $0.814$ for $k{=}3$). Importantly, volume alone produced a higher Silhouette score than the full representation ($0.967$ versus $0.482$), confirming lesion size as the dominant within-POMS clustering axis. In a separate grouped POMS-versus-adult-MS discrimination experiment, however, volume was near chance (AUC $0.516$), Shape-DNA alone reached AUC $0.837$, and Shape-DNA plus morphometry reached $0.914$. Scale-normalised spectral-shape clustering also produced an orthogonal partition (ARI $=-0.026$ relative to the size-driven solution), and normalised spectral entropy remained different after volume matching ($r{=}0.278$). Cross-diagnosis analysis also showed lower spectral entropy in multiphasic disseminated encephalomyelitis than POMS ($r{=}0.645$, $p{=}0.0007$), while clinically isolated syndrome lesions were indistinguishable ($p{=}0.856$). These exploratory results indicate that spectral lesion geometry complements, rather than replaces, volumetric measures and motivates validation in larger multi-site paediatric cohorts with lesion-activity metadata.
Links to Paper and Supplementary Materials
Main Paper (Open Access Version): https://papers.miccai.org/miccai-2026-sat/paper/PIPPI_003.pdf
SharedIt Link: Not yet available
SpringerLink (DOI): Not yet available
Supplementary Material: Not Submitted
Link to Open Review
BibTex
@InProceedings{MitPra_ShapeDNA_MICCAISAT2026,
author = { Mittal, Prateek AND Singh, Prabal Pratap AND Chauhan, Joohi},
title = { { Shape-DNA Spectral Analysis of Lesion Morphotypes in Paediatric-Onset Multiple Sclerosis } },
booktitle = {Medical Image Computing and Computer Assisted Intervention -- MICCAI 2026 Workshops and Challenges},
year = {2026},
publisher = {Springer Nature Switzerland},
volume = {LNCS 17257},
month = {pending},
page = {pending}
}