Introduction Primary progressive aphasia (PPA) is characterised by progressive erosion of the language network, and is most commonly associated with frontotemporal lobar degeneration. Few biomarkers exist for diagnosis and tracking of disease progression in PPA. Novel techniques such as diffusion tensor imaging (DTI) may allow early disease detection and monitoring, as well as improving our understanding of the trajectory of PPA.

Methods Thirty patients with PPA (13 with non-fluent variant, 11 with semantic variant, 6 with logopenic variant), and 20 age-matched healthy participants were assessed using serial DTI at baseline and on average 1.2 years later. Baseline and follow-up DTI scans were registered using a group-wise approach and a region-of-interest analysis performed for individual white matter tracts. Annualised rates of change for DTI metrics were calculated and compared between groups. Changes in neuro-psychometric scores were correlated with white matter structural changes.

Results In the non-fluent group, rates of change in fractional anisotropy were most significant within the right and left superior longitudinal fasciculus (right, −5.7%/year, left, −4.3%/year, p<0.001). In the semantic group, change was most significant in right uncinate fasciculus (−14.8%/year, p<0.001). In the logopenic group, change was most significant in only radial diffusivity, within the right inferior longitudinal fasciculus (6.8% per year increase, p<0.05). Using DTI as an outcome measure, sample size estimates for future treatment trials were generally lower across all syndromic groups compared with volumetric MRI.

Conclusion This study demonstrates the feasibility of longitudinal DTI, identifying rates of disease progression across the spectrum of PPA. Syndrome specific trajectories of disease progression emerged. This highlights the potential use of serial DTI as a disease biomarker in PPA, particular in the design of future treatment trials.