In another light: brain imaging

Rational Imaging: Suspected early dementia. Jonathan M Schott, Jason D Warren,  Frederik Barkhof, Martin N Rossor,  and Nick C Fox. Published in the BMJ, 20 September 2011. BMJ. 2011; 343: d5568.

---

In Another Light (Rapid response part I)
27^th September 2011
By Dr Peter & Dr Sian Gordon

Because we are not radiologists we cannot authoritatively contribute to the debate on the usefulness of imaging in the diagnosis of dementia. However the case presented here is unusual in the following ways:

(i) the patient is aged 58 years, which would make this early-onset dementia, and

(ii) the diagnosis is an unusual variant.

We therefore take issue with the statement that “all patients with suspected dementia should have structural brain imaging”.[1]This conclusion cannot be justified from this case.

It would seem entirely reasonable to us, as non-radiologists, to use imaging in this atypical illness in a young patient; it is not reasonable to generalise this as applying to an elderly population that is increasingly classified according to cognitive state.

We have been concerned for some time that significant harm is being done to our patients by overenthusiasm for “early diagnosis”, confusion regarding nomenclature, and inappropriate application of research protocols in everday clinical practice.

A summary of our concerns follows.
A ‘Position Paper’ by Dubois and colleagues suggests that certain biomarkers are now reliable enough to be employed clinically in the earliest possible diagnosis of Alzheimer’s disease[2]; however a follow- up Swedish study concluded the opposite[3] and a review in the Lancet considered the proposed ‘Dubois criteria’ as follows: “but they proceed as if such biomarkers are already available. Moreover, the presence of biomarkers must not be confused with the disease itself. Thus, this new framework for Alzheimer’s disease might represent a tower of Babel rather than a coherent lexicon.”[4]

At the same time as neuroscience is focussing research on the development of biomarkers for the “pre-clinical” detection of Alzheimer’s disease, there has been a general impetus, originating from UK health departments and voluntary organisations to encourage the early diagnosis of Alzheimer’s dementia. This was identified as an area for action given consistent evidence of the delayed diagnosis of dementia. It is perhaps understandable given that the nomenclature (“AD”) is often used inter- changeably for ‘Alzheimer’s disease’ and ‘Alzheimer’s dementia’ that these two imperatives risk becoming confused.

The National Institute on Ageing Preclinical Alzheimer’s disease Workgroup has classified “preclinical “Alzheimer’s disease based upon microscopic pathology. In this study, out of 126 cognitively intact patients with mean age 83.7 years at death, post-mortem found that 53/126 or 43% had pathology consistent with preclinical Alzheimer’s disease[5]. Worldwide epidemiological follow-up studies have found that the rates of transition between mild cognitive impairment (MCI) and Alzheimer’s dementia is about one-third to one-half over three years when looking at the type of MCI considered most likely to be ‘pre-Alzheimer’s.’[6] The InDDEx study, covering 4 years follow up found 5% per year of those with MCI progressed to dementia[7], whilst the Pala study, covering 2 – 6 years follow-up, found that 4.1% progressed to Alzheimer’s dementia[8]. Taken together, these findings suggest a large potential for false- positive diagnoses of Alzheimer’s disease, if diagnostic criteria are stretched to include identification of ‘pre-clinical’ states. It is vital that the clinical community heeds the warning of the research community that “at this point our recommendations are strictly for research purposes only.”[9]

A correct early diagnosis of Alzheimer’s disease may be clarifying and appreciated by patients even without disease-modifying treatment, and a diagnosis could be valuable since it allows informed planning for the future[10]. However the consequences in falsely diagnosed cases could be grave. Conferring a diagnostic label is far from a neutral act and is most potent when it is provided by those, such as doctors, deemed to have authority and dependability. Physicians should be aware that explaining to a patient that they have a chronic disease which will likely drastically reduce their quality of life over time, means altering the course of that patient’s life and that of his or her family whether or not they go on to develop dementia.

We must consider whether the earliest possible diagnosis of Alzheimer’s disease might result in an increased suicide risk and possible requests for doctor-assisted suicide[11]. Preserved awareness and the ability to perform planned actions, provide a potentially deadly combination.

It seems somewhat reckless to advocate for greater recognition of ‘early Alzheimer’s disease’ when knowledge is virtually non-existent on how patients would react to such a diagnosis[12], long before dementia may (or, indeed, may not) develop, and whilst research seems to point to large numbers of people with MCI not going on to develop dementia. Our evolving understanding of the pathologic changes that predate Alzheimer’s disease may one day revolutionize management. Better ways of identifying persons at risk of Alzheimer’s dementia will clearly be necessary for finding strategies to prevent clinical manifestation. But as we learn to identify risk, we need to do so in a way that is more likely to benefit patients than harm them.[13]

Competing interests: None declared

---

In Another Light (part II)
30^th September 2011
By Dr Peter & Dr Sian Gordon

We agree with Dr Jasvinder Singh that this is a welcome case-report but for a different reason. This report[14] exposes itself to criticism for its extrapolation from a narrow base to the widest possible: “all patients with suspected dementia should have structural brain imaging.”

The 2006 NICE guidelines[15] do indeed state: “Structural imaging should be used in the assessment of people with suspected dementia to exclude other cerebral pathologies and to help establish the subtype diagnosis.” However the same NICE guideline when it explains the evidence-base makes several counter statements:

• “Such brain imaging changes are not absolutely diagnostic in any individual but may be used to inform clinical judgment regarding diagnosis and likelihood of progression.”
• “Reported sensitivities and specificities in this regard vary depending on the method (that is, visual rating, volumetric or voxel based) and the anatomical area studied.”
• “Further work is needed to harmonize and standardize methods on analysis and reporting of clinical CT and MRI studies before such methods could be widely adopted clinically.”

Dr Jasvinder Singh cites just one research paper for the most prevalent dementia group: that of the Alzheimer’s type. In the study that Dr Singh cites[16], only nine of the study group went on to develop Alzheimer’s dementia: given such a small denominator we must be careful to conclude this as a guiding basis for brain imaging on a much wider scale.

As practising doctors, our personal determination is to make a timely diagnosis of dementia as early as we confidently can. The latest findings on cerebral pathology by Aging-Alzheimer’s Association workgroup[17] presents sobering reading; one is left, after reading this comprehensive pathological review, questioning our current definitions of brain disease. Meantime, it is important that we are not blinded, herd-like, from the truth that in today’s routine clinical setting, neuroimaging for dementia is at best supportive in the diagnostic process.