AI Science Tracker
EN

Automated spinal-lesion measurements retain an association with neurological recovery

Balgrist University Hospital; University of Zurich; Polytechnique Montréal · 5 Nov 2026 · medicine

Discovery

Automated spinal-lesion measurements retain an association with neurological recovery

Source check: Source checked. Not independent scientific confirmation.

Finding: Deep-learning segmentation enabled automatic MRI lesion measurements in 83 spinal-cord injury patients from eight sites. In the observed cohort, lesion width and preserved tissue bridges separated groups with different six-month recovery profiles. These are exploratory within-cohort associations, with no independently validated patient-level prognostic rule or treatment effect.

Technology: SCIsegV2 deep-learning segmentation; conditional-inference tree stratification

Publication: Peer-reviewed paper · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.

AI contribution
SCIsegV2 learned image segmentation supplied lesion and spinal-cord masks for automatic morphometrics; learned conditional-inference trees then identified imaging-defined subgroups associated with observed neurological outcomes.
Test and result
Retrospective eight-site study of 83 patients with clinical MRI and longitudinal neurological assessments. SCIsegV2 automatically segmented lesion and cord masks; geometric processing produced lesion length, width and tissue-bridge measures. Automatic/manual measures correlated at rho 0.540–0.821. Conditional-inference trees fit to 75 six-month outcomes selected lesion width and, for motor subgroups, tissue-bridge width; subsequent longitudinal models described different observed recovery profiles.
Limitation / still unknown
Outcome-derived thresholds and trajectory comparisons use the same cohort, not an untouched external prognostic test. Segmentation errors, partial-volume effects and imaging timing affect very small bridge estimates. Overlap with prior segmenter development was not independently reconciled; no held-out segmentation validation is claimed. Missing outcomes and heterogeneous care/rehabilitation may confound associations. The study does not establish causal tissue mechanisms, deployable cut-offs or improved treatment.
Evidence sources and annotation context

Evidence context (descriptions follow the cited sources; test type is a catalog classification):

Evidence sources: https://pmc.ncbi.nlm.nih.gov/articles/PMC13739854/

Source checked · editorial details

The source page was checked against this entry. Independent confirmation is not assessed by this label.

Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.

Review date: 21 Nov 2026. Review scope: Read actual primary named segmenter/automatic-measurement operation, clinical cohort/reference, conditional-inference and longitudinal-model Methods, prognostic-association Results and limitations. Prior segmenter training datasets and supplementary code were not independently inspected; no independent validation is inferred. Exact primary identity/date and fresh primary PubMed correction/retraction relationships checked..

No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.

Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.

Open in the full catalog for filters, charts and related discoveries. Browse medicine discoveries.