Cortical microstructure in primary progressive aphasia: a multicenter study

Ignacio Illán‐Gala, Víctor Montal, Sergi Borrego‐Écija, Maria Luisa Mandelli, Neus Falgàs, Ariane E. Welch, Jordi Pegueroles, Miguel Santos‐Santos, Alexandre Bejanin, Daniel Alcolea, Oriol Dols‐Icardo, Olivia Belbin, Ma Belén Sánchez‐Saudinós, Núria Bargalló, Sofía González‐Ortiz, Albert Lladó, Rafael Blesa, Bradford C. Dickerson, Howard J. Rosen, Bruce L. Miller, Alberto Lleó, Maria Luisa Gorno‐Tempini, Raquel Sánchez‐Valle, Juan ForteaAlzheimer 2022 · DOI

Abstract

Primary progressive aphasia (PPA) presents significant diagnostic challenges in its early stages. This multicenter study evaluates cortical mean diffusivity (cMD) as a sensitive MRI-based biomarker for neurodegeneration. Results indicate that cMD detects microstructural tissue disorganization more effectively than traditional cortical thickness measurements, offering a more robust tool for early diagnosis and clinical trial monitoring.

Introduction

Primary progressive aphasia (PPA) comprises a group of neurodegenerative disorders characterized by progressive speech and language impairment. Current clinical practice relies heavily on structural MRI to measure cortical thickness (CT)—the macroscopic shrinkage of brain tissue. However, CT often fails to detect early-stage neurodegeneration, as significant neuronal loss typically precedes visible atrophy.

This study investigates cortical mean diffusivity (cMD), a metric derived from diffusion-weighted imaging (DWI) that measures water diffusion within the cerebral cortex. Increased cMD reflects microstructural breakdown, such as loss of cellular membranes and dendrites, potentially providing a "leading edge" indicator of disease progression before macroscopic atrophy occurs.

Study Population and Methods

The researchers utilized a multicenter dataset comprising 209 participants from the University of California, San Francisco (UCSF) and the Catalan Frontotemporal Dementia Initiative (CATFI). The cohort included 120 PPA patients across three main variants—non-fluent/agrammatic (nfvPPA), semantic (svPPA), and logopenic (lvPPA)—as well as 89 age-matched healthy controls.

Flowchart of participant recruitment and final sample distribution across centers and PPA variants.

To ensure high precision, the study employed a surface-based processing pipeline. This approach maps diffusion data to the midpoint of the cortical ribbon, minimizing "partial volume effects" from adjacent cerebrospinal fluid or white matter. Technical variations across different 3T MRI scanners were corrected using the ComBat harmonization algorithm.

Comparative Sensitivity of cMD and Cortical Thickness

The primary finding of the study is that cMD is significantly more sensitive than cortical thickness in detecting neurodegenerative changes across all PPA variants. While both metrics identified damage in the "epicenters" of each variant (e.g., the left inferior frontal gyrus in nfvPPA or the left temporal pole in svPPA), cMD abnormalities were far more widespread.

Comparison of effect sizes (Cohen’s d) for cortical thickness (A) and cortical mean diffusivity (B) across PPA variants. Green regions indicate higher sensitivity of cMD.

In patients with very mild symptoms (Global CDR® 0.5), cortical thinning was often negligible or highly localized. In contrast, cMD already showed significant increases in the core disease regions. This suggests that microstructural disorganization is a precursor to macroscopic tissue loss.

Clinical Correlations and Implications

Beyond diagnostic sensitivity, cMD demonstrated a stronger correlation with clinical disease severity, as measured by the CDR® plus NACC FTLD-SB score. The relationship between increased cMD and cognitive decline was observed across broad frontal and temporal regions, whereas CT correlations were more restricted.

For clinical trials, these findings suggest that cMD could serve as a more responsive endpoint than atrophy. By capturing the ongoing spread of pathology into new cortical areas—avoiding the "floor effect" seen in severely atrophied regions—cMD provides a clearer picture of disease progression.

Conclusion

Cortical mean diffusivity represents a powerful addition to the neuroimaging toolkit for PPA. Its ability to detect early microstructural damage and its strong alignment with clinical severity make it a promising candidate for tracking disease and evaluating the efficacy of emerging therapies. Future longitudinal studies will be essential to confirm the temporal sequence of these microstructural changes relative to symptomatic onset.

Summary updated

Abstract

Cortical mean diffusivity shows promise as a sensitive biomarker for the study of the neurodegeneration-related microstructural changes in PPA.
BibTeX
@inproceedings{illangala2022cortical,
  author = {Ignacio Illán‐Gala and Víctor Montal and Sergi Borrego‐Écija and Maria Luisa Mandelli and Neus Falgàs and Ariane E. Welch and Jordi Pegueroles and Miguel Santos‐Santos and Alexandre Bejanin and Daniel Alcolea and Oriol Dols‐Icardo and Olivia Belbin and Ma Belén Sánchez‐Saudinós and Núria Bargalló and Sofía González‐Ortiz and Albert Lladó and Rafael Blesa and Bradford C. Dickerson and Howard J. Rosen and Bruce L. Miller and Alberto Lleó and Maria Luisa Gorno‐Tempini and Raquel Sánchez‐Valle and Juan Fortea},
  title = {{Cortical microstructure in primary progressive aphasia: a multicenter study}},
  booktitle = {Alzheimer},
  year = {2022},
  doi = {10.1186/s13195-022-00974-0},
}
Figure