The changing landscape of neuroimaging in frontotemporal lobar degeneration: from group-level observations to single-subject data interpretation

ABSTRACT

Introduction

While the imaging signatures of frontotemporal lobar degeneration (FTLD) phenotypes and genotypes are well-characterized based on group-level descriptive analyses, the meaningful interpretation of single MRI scans remains challenging. Single-subject MRI classification frameworks rely on complex computational models and large training datasets to categorize individual patients into diagnostic subgroups based on distinguishing imaging features. Reliable individual subject data interpretation is hugely important in the clinical setting to expedite the diagnosis and classify individuals into relevant prognostic categories.

Areas covered

This article reviews (1) single-subject MRI classification strategies in symptomatic and pre-symptomatic FTLD, (2) practical clinical implications, and (3) the limitations of current single-subject data interpretation models.

Expert Opinion

Classification studies in FTLD have demonstrated the feasibility of categorizing individual subjects into diagnostic groups based on multiparametric imaging data. Preliminary data indicate that pre-symptomatic FTLD mutation carriers may also be reliably distinguished from controls. Despite momentous advances in the field, significant further improvements are needed before these models can be developed into viable clinical applications.

KEYWORDS:

• Frontotemporal dementia
• biomarkers
• machine-learning
• diagnosis
• magnetic resonance imaging
• neuroimaging
• classification models
• Dementia
• precision medicine

Article highlights

• The neuroimaging signatures of specific FTLD phenotypes and genotypes are relatively well defined.

• The meaningful interpretation of imaging data from individual patients, suspected patients, and pre-symptomatic mutation carriers is challenging.

• A multitude of classification models have been proposed to facilitate single-subject categorization, but early studies suffered from limited sample sizes, model overfitting, and suboptimal validation.

• The binary classification schemes of early machine-learning initiatives are increasingly superseded by robust multi-class models.

• MRI-based classification models distinguish established FTLD from healthy controls or AD with relatively good accuracy.

• Instead of using established cases with long symptom duration, model validation should ideally be undertaken with subjects with a suspected diagnosis or patients with short symptom duration.

• Preliminary data indicate that pre-symptomatic FTLD mutation carriers may also be differentiated from controls based on imaging data alone.

• Despite considerable advances in machine-learning in FTLD, existing models need further optimization and validation before they can be implemented in routine clinical practice.

• Machine-learning is likely to become an integral part of data interpretation in FTLD, expediting the diagnostic process and facilitating an earlier entry into clinical trials.

Abbreviations

AD – Alzheimer<apos;>s disease; AngleR – angle between cortical profile and principal diffusion direction; APOE4 – apolipoprotein E4; ALS – amyotrophic lateral sclerosis; ANN – artificial neural networks; ASL – arterial spin labeling; AUC – area under the curve; AxD – axial diffusivity; bvFTD – behavioral variant frontotemporal dementia; CBF – cerebral blood flow; CBS – corticalbasal syndrome; CDR – Clinical dementia rating scale; C9orf72 – chromosome 9 open reading frame 72; CSF – cerebrospinal fluid; CST – corticospinal tract; DFA – discriminant function analysis; DTI – diffusion tensor imaging; DLB – dementia with Lewy bodies; EOAD – early onset Alzheimer<apos;>s disease; FA – fractional anisotropy; FDG-PET – fluorodeoxyglucose positron emission tomography; fMRI – functional magnetic resonance imaging; FTD – frontotemporal dementia; FTLD – frontotemporal lobar degeneration; GAN – generative adversarial neural network; GM – gray matter; GMD – gray matter density; GRN – progranulin; KNN - K-nearest neighbor; LBD – Lewy body dementia; LOAD – late-onset Alzheimer<apos;>s disease; LoCo - Loss in connectivity; MAPT – microtubule-associated protein Tau; MCI – mild cognitive impairment; MD – mean diffusivity; ML – machine learning; MRI – magnetic resonance imaging; nfvPPA – non-fluent variant primary progressive aphasia; PCA – principal component analysis; PET – positron emission tomography; PPA – primary progressive aphasia; PSP – progressive supranuclear palsy; pTDP43 – phosphorylated 43-kDa TAR DNA-binding protein; RD – radial diffusivity; ROC – receiver operating characteristic curve; ROI – region of interest analysis; RUSBoost – Random undersampling boosting; rs-fMRI – resting state functional magnetic resonance imaging; STAND – structural abnormality in neurodegeneration; SuStaIn – Subtype and stage inference; SVD – singular value decomposition; SVM – support vector machine; svPPA – semantic variant primary progressive aphasia; TBM – tensor-based morphometry; TBSS – tract-based spatial statistics; VBM – voxel-based morphometry; VD – vascular dementia; VOI – volumes of interest; VOL – volumetry; WM – white matter; WMD – white matter density; WMH – white matter hyperintensity; wSDM – weighted symbolic dependence metric.

Declaration of interests

The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or conflict with the subject matter or materials discussed in this manuscript apart from those disclosed.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Additional information

Funding

The authors are supported by the Health Research Board (HRB EIA-2017-019), the Spastic Paraplegia Foundation (SPF), the EU Joint Programme – Neurodegenerative Disease Research (JPND), the Andrew Lydon scholarship, the Irish Institute of Clinical Neuroscience (IICN), and the Iris‘OBrien Foundation. The sponsors had no bearing on the opinions expressed herein.