The intersection of human genomic analysis and in-depth neuroimaging characterization is an emerging, yet rapidly maturing, approach that is clearly data-rich and filled with promise in the research laboratory setting but does not yet have a clear clinical role.

This approach has already been applied successfully for the study of disease and healthy normal trait variation in human subjects. In those studies, genomic analyses have exclusively centered on common single nucleotide variants probed by the commercially available microarrays while the partnered neuroimaging work largely consists of structural MRI scans.

The genomic aspect of the data will expand in the very near future into whole genome analysis, and additional sequencing modalities will be incorporated. Currently, much of this work has existed in the basic science research realm, but the elucidation of informative findings bodes well for the future of this work and suggests possible avenues for clinical diagnostic, prognostic, or theranostic incorporation.

Using Alzheimer’s disease (AD) as a case study, one can demonstrate that the top genetic variant known to be associated with risk (the epsilon-4 allele of APOE) can also influence hippocampal size in healthy individuals. This begs the question – and perhaps illustrates the relatively untapped power of neuroimaging – of whether or not the effect of APOE on AD risk might have been identifiable using MRI scans and genomics alone in a cohort of healthy individuals. (There is, of course, very encouraging ongoing work in other diseases and regions of the brain, but I will not devote space to discuss that here.)

The lingering question is how this can influence clinical practice above and beyond the more general informative nature of these types of focused research studies. In many ways, genomics hasn’t yet altered neurologic disease clinical practice, although I would argue that it is still very early in the game because the human genome draft sequence is only 10 years old. With the further maturation of next generation sequencing, there continues to be a strong sense that an "affordable" genome sequence is in the very near future.

But what about the combination of neuroimaging and genomics? It is highly likely that the combined approach could aid in the early diagnosis of neurologic disease – either in the early stages of symptomatic disease or in presymptomatic, at-risk individuals. However, the use of imaging in presymptomatic individuals has significant hurdles associated with it, including methodological issues related to reproducibility and sensitivity, cost and risk of the scan, and the possibility of unanticipated findings.

If you believe some people in the genomics field, it won’t be too long before a significant portion of the population is walking around with their individual, analyzed genome sequences on their cell phone or readily accessible in cloud-based storage. What about partnering those data with well-visit brain scans? Is that feasible and would it help our approach to neurologic health and disease management? What happens as the imaging technology changes or improves, and does this negatively impact the potential power of the longitudinal data collected? While this combination of data-rich technologies is enticing, the hurdles mentioned above are significant, not the least of which relates to general feasibility based on simple infrastructure. For genome sequencing, a sample can be collected and processed off-site at a high throughput sequencing facility. For brain imaging, the equipment resources must reside locally, and they have a defined daily capacity that isn’t easily scalable.

Perhaps the tangible clinical or translational benefit in neuroimaging genomics lies in the cross-validation of discoveries. For example, an AD risk gene also associated with neuroimaging findings in at-risk individuals provides clear and powerful supportive evidence. Additionally, neuroimaging affords the geneticist an ability to study the brain of phenotypically "unique" individuals during life. Take for example a cohort of older individuals who are cognitively normal yet demonstrate significant beta-amyloid PET signal. Is this an artifact of the PET tracer or does such a group represent a unique opportunity to search for protective genomic and/or environmental factors? The ability to study such a cohort longitudinally is particularly powerful.

The collision of neuroimaging and genomics on a routine basis in healthy individuals is likely far down the road. However, the burgeoning collection of research studies using genomics and imaging information to make discoveries is rapidly growing and suggests that a strong synergy can exist between the two paradigms. The clinical genomics revolution will happen first – and in many people’s eyes, is already happening – and then the question remains open about the neuroimaging field. Data are powerful, and the future of clinical management must involve some sort of advanced longitudinal testing in well individuals. Such tests should include neuroimaging and genomics, as long as both paradigms are reduced to practice with an appropriate balance of risk and benefit and can become more affordable. To some extent, the maturing, prevention-focused clinical trials in neurologic disease stand to motivate further discussion of these issues. The future is bright for both neuroimaging and genomics, but it remains to be seen if their educated incorporation into clinical practice – especially in the case of healthy individuals – can be achieved.

Dr. Huentelman is head of the neurobehavioral research unit in the neurogenomics division of the Translational Genomics Research Institute, Phoenix. He has no relevant disclosures.

The Collision of Neuroimaging and Genomics