A tensor based morphometry study of longitudinal gray matter contraction in FTD

Simona M. Brambati, Natasha C. Renda, Katherine P. Rankin, Howard J. Rosen, William W. Seeley, John Ashburner, Michael W. Weiner, Bruce L. Miller, Maria Luisa Gorno‐TempiniNeuroImage 2007 · DOI

TL;DR

This study uses tensor-based morphometry (TBM) to track gray matter loss in patients with frontotemporal dementia (FTD) over a one-year period. The findings reveal significant atrophy in limbic and paralimbic regions, suggesting TBM is a sensitive tool for monitoring disease progression in FTD and potentially aiding in clinical diagnosis and treatment evaluation.

What Problem Does This Paper Address?

Frontotemporal dementia (FTD) is a neurodegenerative disease characterized by progressive changes in personality and behavior, stemming from atrophy in the frontal and temporal lobes. While cross-sectional studies have identified structural abnormalities in FTD, there's a significant knowledge gap regarding the longitudinal progression of this atrophy. Understanding how the brain changes over time is crucial for accurately diagnosing FTD, differentiating it from other dementias, and evaluating the effectiveness of potential treatments. Previous longitudinal studies in FTD have been limited, and the authors aim to use a more advanced neuroimaging technique, tensor-based morphometry (TBM), to precisely map and quantify these changes over a one-year period.

Methods

The core technique employed is Tensor-Based Morphometry (TBM), implemented using Statistical Parametric Mapping (SPM2) software. TBM works by analyzing the deformation fields required to "warp" a subject's later MRI scan to match their earlier scan. The Jacobian determinant of this deformation field quantifies local volume changes at each voxel.

Here's a breakdown of their approach:

  • Image Pre-processing: T1-weighted MRI scans from two time points (baseline and 1-year follow-up) were pre-processed. A bias correction was applied to the later scan to ensure comparability.
  • High-Dimensional Warping: A high-dimensional deformation field was used to precisely align the corrected later scan to the earlier scan within each subject.
  • Jacobian Determinants: The determinant of the deformation gradient at each voxel was calculated. This value represents the local volume change.
  • Gray Matter Volume Change Quantification: The Jacobian determinant was multiplied by the segmented gray matter image from the first scan. This resulted in a map representing gray matter-specific volume change.
  • Normalization and Smoothing: A study-specific template was created, and normalization parameters were estimated. The resulting images were smoothed using a 12 mm isotropic Gaussian kernel and masked to include only gray matter.
  • Statistical Analysis: A "single-subject conditions and covariate" statistical model was used. This model identified brain regions showing greater gray matter atrophy in FTD patients compared to controls. Nuisance covariates like sex, age, and total intracranial volume were included.
  • Two Analysis Approaches:
    1. Whole-Brain Analysis: A conservative significance level of p<0.05, corrected for multiple comparisons (Family-Wise Error - FWE), was used to identify widespread changes.
  • Region of Interest (ROI) Analysis: A more sensitive threshold of p<0.001 (uncorrected) was applied to specific regions previously implicated in FTD (medial/orbital frontal, anterior cingulate, insula, amygdala, hippocampus).

Datasets

The study utilized MRI scans from two groups:

  • FTD Patients: 12 patients with mild to moderate FTD.
  • Healthy Controls: 12 age- and sex-matched healthy individuals.

All subjects underwent two T1-weighted MRI scans approximately 12 months apart. No new benchmarks or datasets were introduced in this paper.

Key Insights and Ideas

  • Limbic System Vulnerability: The study strongly suggests that limbic and paralimbic regions are particularly susceptible to atrophy in FTD. This aligns with the known role of these areas in regulating emotion, behavior, and social cognition, which are core deficits in FTD.
  • TBM as a Sensitive Tool: The findings highlight TBM's ability to detect subtle, longitudinal anatomical changes in the brain, even over a relatively short period (one year) and with a moderate sample size. This suggests TBM could be a valuable tool for tracking disease progression.
  • Discrepancy Between Imaging and Behavior: While imaging showed clear atrophy progression, the behavioral measures (NPI) did not show commensurate significant changes in this small group. This points to the potential for automated brain analysis to be more sensitive than clinical assessments in early stages or with small sample sizes, and raises questions about the timeline of behavioral manifestation relative to structural changes.
  • FTD Network Specificity: The observed atrophy pattern in FTD is distinct from that seen in Alzheimer's disease, which tends to affect temporoparietal regions. This reinforces the idea of distinct neurodegenerative pathways for different dementia types.

Key Results

  • Whole-Brain Analysis: The only region showing statistically significant gray matter atrophy progression in FTD compared to controls, after stringent correction for multiple comparisons (p<0.05 FWE), was the bilateral anterior cingulate/paracingulate gyri (BA 32).
  • Region of Interest (ROI) Analysis: With a more lenient threshold (p<0.001 uncorrected), several other regions showed significant atrophy:
  • Left ventro-medial frontal cortex (BA 11)
    • Right medial superior frontal gyrus (BA 32)
    • Insulae
    • Left medial temporal lobe, including the amygdala and hippocampus.
  • Functional Changes: While not the primary focus, the study noted significant progression in the CDR Sum of Box scores and some cognitive domains (working memory, learning) in FTD patients over the year. NPI scores showed worsening in aberrant motor behavior.

Limitations and Future Work

  • Sample Size: The study included only 12 FTD patients. The authors acknowledge that larger groups would increase statistical power and allow for more robust conclusions.
  • Disease Stage: The patients were at mild to moderate stages of FTD. Future studies could explore progression in earlier or later stages.
  • Medication Effects: Four FTD patients were on SSRIs, and two were on SSRIs and atypical neuroleptics. While controls were not on medication, the potential influence of these drugs on brain structure or atrophy rates is not fully explored.
  • Behavioral Correlation: The authors suggest that future investigations with larger cohorts and more sensitive behavioral tasks would be beneficial to directly correlate regional atrophy progression with declines in behavioral measures.
  • Specificity of TBM: The authors propose that further research should evaluate TBM's anatomical specificity by comparing it across different neurodegenerative diseases to confirm its utility in distinguishing disease patterns. They also suggest TBM could be a useful technique for assessing the effects of disease-modifying treatments in clinical trials.

Summary updated

BibTeX
@inproceedings{brambati2007tensor,
  author = {Simona M. Brambati and Natasha C. Renda and Katherine P. Rankin and Howard J. Rosen and William W. Seeley and John Ashburner and Michael W. Weiner and Bruce L. Miller and Maria Luisa Gorno‐Tempini},
  title = {{A tensor based morphometry study of longitudinal gray matter contraction in FTD}},
  booktitle = {NeuroImage},
  year = {2007},
  doi = {10.1016/j.neuroimage.2007.01.028},
}
Figure