Plasma glial fibrillary acidic protein (GFAP) departed from the normal range earliest among the markers examined in sporadic frontotemporal lobar degeneration (FTLD) syndromes, followed by the Trail Making Test Part B, white matter lesion volume and plasma neurofilament light chain (NfL), according to a modelling study published in the August 2026 issue of Alzheimer's & Dementia. Alberto Benussi of the University of Trieste, Barbara Borroni of the University of Brescia and colleagues, working with laboratory groups in Gothenburg and Brescia, describe their result as the "first data-driven temporal cascade of multimodal biomarkers in sporadic FTLD-associated syndromes".

The model was built from 489 patients and 1,904 patient-visit observations: 211 with behavioural variant frontotemporal dementia, 129 with primary progressive aphasia, 71 with corticobasal syndrome, 66 with progressive supranuclear palsy and 12 with frontotemporal dementia associated with amyotrophic lateral sclerosis. Plasma, MRI and clinical measures were fitted as sigmoid trajectories to covariate-adjusted longitudinal data. Insular atrophy showed the steepest transition of the MRI measures, and the CDR Dementia Staging Instrument plus National Alzheimer's Coordinating Center behaviour and language domains sum of boxes declined most steeply overall. The sequence was not identical across syndromes: Trail Making Test Part B inflected earliest in the behavioural variant, while insular atrophy dominated in primary progressive aphasia.

Until now, ordered biomarker cascades in this spectrum came almost entirely from families carrying known mutations, where the approach of symptoms can be anticipated. Adam Staffaroni and colleagues modelled 796 carriers and 412 non-carrier relatives in Nature Medicine in 2022 and found the ordering differed by gene, and a 2021 study in the Journal of Neurology, Neurosurgery and Psychiatry modelled 56 presymptomatic and 35 symptomatic GRN carriers against 35 non-carriers. The new work addresses the far larger group of patients with no identified genetic cause, and the authors frame it as a basis for stage-specific trial endpoints across four syndromes usually studied separately.

The counterpoint is that this is an observational study of people who already carry a clinical diagnosis, so the cascade describes the symptomatic course rather than any presymptomatic phase, and the ordering is a group-level model output, not a prediction for one patient. GFAP's position at the head of the sequence also sits against other evidence: in the ALLFTD analysis by Udit Sheth and colleagues, published in Molecular Neurodegeneration in March 2025 and covering 702 participants with an FTD syndrome, 127 presymptomatic carriers, 67 people with mild changes and 161 controls, GFAP, unlike NfL, separated controls poorly from those with mild symptoms. The syndromes in the Italian model are defined clinically, and clinically similar syndromes can arise from different underlying pathologies.