Why Should Neuroradiologists Study Patients with Smell Loss?

The practice of radiology is to understand human physiology and pathology hidden from ordinary clinical view. The accomplishments of radiology in this regard are legendary, beginning with initial observations of bone by Wilhelm K. Roentgen and continuing to the present with developments of CT, MR imaging, PET scanning, and various functional paradigms that not only elucidate anatomic structures but also detail functions of various organs and vital physiological processes. But why study olfaction?

Loss of smell (hyposmia) is a common but hidden problem. It is common in the sense that current estimates suggest that 19 million people in the United States have some form of chronic smell loss (1, 2). It is hidden in the sense that most patients exhibit no outwardly obvious handicap associated with this loss and usually do not exhibit any nasal cavity pathologic abnormality. Yet, as a symptom, this loss reflects specific abnormalities in the biochemistry of the body. Hyposmia is a harbinger, a symptom of biochemical abnormalities in multiple organ systems that involve, among others, oncologic, nutritional, metabolic, endocrine, infectious, genetic, and hematologic processes (2). Drugs induce hyposmia (3), as does head injury (4). Patients with this symptom are not at risk so much of death as they are of eating spoiled food or becoming exposed to toxic gas. More commonly, patients are desperately unhappy because they cannot obtain gratification from eating, drinking, or appreciating odors that give the rest of us such great pleasure and social enjoyment. Another major issue related to smell loss is distortion, and patients with hyposmia may also develop phantom smell.

But why should radiologists be interested in this common but not deadly problem? The findings shown on standard radiographs, CT scans, and MR images of the brain and sinuses of these patients are usually normal except for a relatively small number of cases with sinusitis, nasal polyps, tumors of the nasal cavity, neuroepithelium, or olfactory bulbs. To investigate the underlying pathology of most of these patients, quantitative and objective methods had to be developed to evaluate smell sensation and the ability to obtain information about odors. This initial effort took the form of complex psychophysical tests involving measurements of thresholds and magnitude estimation similar to those obtained in audiology. Although useful, these tests are cumbersome, time consuming, and not always objective. What role does neuroradiology play in this problem considering that commonly used neuroradiologic studies are of little diagnostic value for evaluating this common symptom?

Multiple methods are available through which some degree of resolution of this problem can be obtained. Neuroradiologic tools such as diffusion-weighted MR imaging techniques have shown neural tracts (5) and may be used to identify underlying olfactory pathways from bulbs to the CNS, although these structures are small and difficult to identify. Labeled agents such as manganese have been placed directly into the nose of animals and can be followed by MR imaging along peripheral neural pathways directly into the brain (6, 7); these techniques have not been used in human studies but may offer a useful approach to stimulus entry and follow-through into the entire olfactory system. PET scanning offers functional evaluation of metabolically active brain regions responding to olfactory stimuli (8); however, the anatomic location of major olfactory structures at the base of the brain makes PET scanning inaccurate and difficult to quantitate. MR spectroscopy has been used to determine abnormalities in metabolite concentration and neurotransmitter levels (9) in specific CNS regions in which olfactory activation takes place. We (10) and others (11–15) have used functional MR imaging of olfactory function in the CNS. This technique has proved helpful to determine regional brain activation in normal study participants (10) and in patients with various forms of hyposmia (16, 17). These studies showed quantitative data that distinguished normal participants from patients with hyposmia (16–24). In addition, by using patients as their own control participants, we have developed techniques in which treatment that restored smell function to normal in previously hyposmic patients was shown to increase regional brain activation quantitatively in relevant olfactory pathways and CNS regions (18).

These functional radiologic studies deal with a complex and difficult problem. Olfaction is not a simple sensory phenomenon in which a single stimulus-response paradigm is paramount. Multiple brain subsystems impinge on this sense, such that emotion, memory, language, vision, and other sensory and cognitive phenomena influence the sensory function and thereby functional MR imaging responses to olfactory stimuli. Results of functional MR imaging studies in general, regarding variability in single participants and across participants, have been questioned (25, 26), which makes this task even more difficult. The skull can introduce multiple artifacts into functional MR images because of susceptibility effects, especially at the base of the skull near the orbitofrontal cortex. Data processing requires great care, and its analysis and noise in the system can be difficult to determine. Stimulus presentation, if not performed in a most simple manner, can introduce such profound artifacts that the olfactory signal intensity is deformed and lost among image processing artifacts.

These technical problems are compounded by clinical problems that further increase the unreliability of these results. Deformation of CNS signal intensity output can occur in the patients whom we are most interested in studying, such as those with dementia secondary to Alzheimer disease and/or vascular disease. Olfactory responses in these patients may be totally unreliable because memory, language, or emotional lack may bias results. Olfactory signals may also be deformed by these phenomena in patients with head injury if postconcussion syndrome is associated with exacerbated emotional responses but inhibition of memory and language responsiveness. Age can influence both subjective and CNS responsiveness because of changes in both peripheral (ie, olfactory epithelium) and CNS physiology, changes that may be identified by functional MR imaging (27). These problems reemphasize the need for objective, reliable, quantitative techniques by which olfaction can be measured and defective sensory pathways and responses identified.

In our experience, patients with smell loss can be identified only by objective means, using a quantitative technique to define their pathologic abnormalities. Only through the use of objective techniques can patients with hyposmia be identified, their symptoms quantified, and their treatment followed. We think that functional techniques that are currently best understood and used by neuroradiologists are the methods that are best used to perform this task. Although currently available techniques will require further development, only through these methods can this important, common, but hidden clinical problem no longer be obscure. Just as development of blood sugar tests served to identify and characterize patients with diabetes, we think that the use of one or several of these techniques can serve to identify and characterize patients with hyposmia. Precise identification and localization of the abnormalities in the olfactory pathways will help to further define the nature of the pathologic abnormality and to evaluate the results of therapy. These tests are now available in the hands of qualified neuroradiologists and can be used to identify these patients. This effort opens a new and valuable large group of patients for neuroradiologists to study. Application of these new techniques will further the usefulness of neuroradiology in the practice of medical science.

• Copyright © American Society of Neuroradiology