OBJECTIVES Alzheimer’s disease is the most frequent cause of degenerative dementia. Despite the available diagnostic criteria, improvement of diagnosic accuracy is still required. The aim of this prospective study was to assess in a large population of patients referred to a memory clinic the diagnostic value of the combination of medial temporal lobe atrophy on temporal oriented CT and decreased temporoparietal uptake on HMPAO single photon emission tomography (SPECT).
METHODS The study was conducted in 125 patients aged 51-93: 64 with probable Alzheimer’s disease (Mean (SD) mini mental state examination (MMSE)=18.34 (6.93)), duration of disease=6.48 (2.93) years, 13 possible Alzheimer’s disease (MMSE=21.58 (5.48), duration of disease=6.08 (2.56)), 48 patients with miscellaneous memory disorders (MMSE=21.98 (6.10), duration the disease = 6.85 (3.91)).
RESULTS For the diagnosis of probable Alzheimer’s disease, the sensitivity of this association was 0.56, the specificity 0.93, the positive predictive value 0.95, and the negative predictive value 0.45. The diagnosic accuracy was 0.68. Both medial temporal atrophy and parietotemporal decrease in uptake were present in four of 13 patients with possible Alzheimer’s disease and 11 of 48 with miscellaneous memory disorders. The association was absent in 27 of 29 patients with frontotemporal dementia. In mild stages (MMSE>18; n = 32), the sensitivity of the association was 0.34, the specificity 0.93, the positive predictive value 0.85, and the negative predictive value 0.57. The diagnosic accuracy was 0.53.
CONCLUSION This association, although not sensitive, helps to select patients with high probability of Alzheimer’s disease at an early stage which can be of interest for clinical and research purposes.
- computed tomography
- Alzheimer’s disease
Statistics from Altmetric.com
The longer life expectancy in western countries has increased the number of subjects with dementia. Although dementia can be caused by more than 60 disorders,1 dementia of Alzheimer’s disease accounts for roughly 50%-60% of cases. As the proportion of the elderly population increases, so does the prevalence of dementia. The observed prevalence of dementia varies between 1% and 20% with an average of about 6%.1 The early detection of brain impairment is crucial for the diagnosis, prognosis, and treatment of dementia. The differential diagnosis of Alzheimer’s disease is based on clinical criteria, as there is no biological marker of the disease available during life.
Computed tomography has been used to exclude other causes of dementia such as hydrocephalus, metastatic disease, or subdural haematoma, but less so to establish a diagnosis of Alzheimer’s disease. Since 1980 several reports have shown that generalised ventricular and sulcal enlargement correlate with the presence and severity of Alzheimer’s disease.2 Over the past decade, many studies have suggested that structural changes in the temporal lobe and hippocampus strongly support the diagnosis of Alzheimer’s disease.3-10
Using single photon emission computed tomography (SPECT) and (99 mTc)-d,l-hexamethyl propylene amine oxime (HMPAO), focal reductions of cerebral uptake can be shown in patients with a clinical diagnosis of Alzheimer’s disease.11-21
The combination of medial temporal lobe atrophy shown by CT and decreased uptake in the parietotemporal cortex shown by SPECT has already been reported in 86% of 51 patients with a clinical diagnosis of Alzheimer’s disease.22
The aim of our study was to assess the diagnostic value of the combination of medial temporal lobe atrophy on CT and parietotemporal uptake decrease on SPECT in a large population of patients referred to a memory clinic.
Patients and methods
From January 1991 to December 1994, temporal lobe oriented CT and HMPAO-SPECT were performed in 125 patients aged 51-93 and recruited in the Lille University memory clinic. Diagnoses were established after several evaluations by a senior staff neurologist, a psychogeriatrician, a neuropsychologist, a speech therapist, and by laboratory tests. Patients were assessed with a comprehensive neuropsychological test battery including the mini mental state examination (MMSE),23 the Mattis dementia rating scale,24 assessment of short and long term memory,25 26 aphasia, agnosia, and gestual and constructional apraxia. A semistructured interview with the patient and the family was conducted by a psychogeriatrician.
All subjects fulfilled the diagnostic and statistical manual of mental disorders, third edition, revised; American Psychiatric Association, 1987 27 criteria for primary degenerative dementia. Dementia of Alzheimer type (probable Alzheimer’s disease=64, possible Alzheimer’s disease=13) was diagnosed according to NINCDS-ADRDA.28 Patients with memory impairment but no clinical diagnosis of Alzheimer’s disease were grouped together under the same heading of miscellaneous memory disorders. This group consisted of patients with frontotemporal dementia (n=29), cortical Lewy body disease (n=10), vascular dementia (n=5), Parkinson’s disease (n=1), corticobasal degeneration (n=1), posterior cortical atrophy (n=1), and progressive supranuclear palsy (n=1).
Dementia of frontotemporal type was diagnosed according to the criteria of Lund and the Manchester group for frontotemporal dementia.29 Patients with dementia of frontotemporal type had a history of behavioural disturbance predating dementia, with early personality changes, loss of social skills, dishibition, and apathy. Vascular dementia was diagnosed according to the report of the NINCDS-AIREN international workshop;30 31 senile dementia of Lewy body type (SLDT) according to the criteria of McKeith et al 32 validated by the same group;33 34corticobasal degeneration according to the indications provided by Riley et al;35 posterior cortical atrophy according to Benson et al;36 and the psychiatric syndromes according to the DSM III-R.27
Patients with other diagnoses, with an atypical history, or with associated systemic pathology were excluded. No patients received psychotropic medication in the month before the study. Table 1 shows the breakdown of the diagnoses and the characteristics of the population.
No CT or SPECT were performed in controls because the use of irradiation in controls is forbidden by French law.
Brain CT (Siemens, Somatron DR, Erlandger, Germany) was performed according to Jobst et al 7 Planes were 20° caudal to the orbitomeatal line in 2 mm contiguous slices through the posterior fossa. Left and right medial temporal lobe width (combined hippocampal formation and parahippocampal gyrus) were measured from the film, about midway through the brainstem, by a neuroradiologist, blinded to the clinical diagnosis.
We defined a medial temporal lobe thickness of 11.5 mm or less as indicative of atrophy according to the literature.7 10The first 50 scans were assessed by two independent neuroradiologists. Between rater deviation was <1 mm. The raters agreed in 92% of cases.
SPECT was performed with a Tomomatic 564 (Copenhagen) and intravenous HmPAO 99m Tc (555Mbq) by the same investigator blinded to clinical diagnosis. Data were collected from three slices centred 1, 5, and 9 cm above the canthomeatal plane. We studied the frontal and temporoparietal regions. SPECT was assessed blinded to the clinical diagnosis on a scale of 0 to 3, in which 0 represented no uptake decrease, 1 a minimal to mild decrease, 2 a moderate decrease, and 3 a severe uptake decrease. SPECT scores in the left, right, or both parietotemporal regions with a grading of 2 or more were considered to indicate a clear decrease in uptake for the purpose of subsequent analysis. SPECT was assessed by two neurologists. The two raters disagreed in nine cases, which were rated again to obtain a consensus. We have previously shown that this visual assessment of cerebral uptake has a good level of between rater reliability (κ=0.68).37
Neuropsychological assessments were performed within two months before or after the imaging assessment.
Table 2 shows the results.
We have compared patients with Alzheimer’s disease with patients with frontotemporal dementia to calculate the sensitivity, specificity, predictive values, and diagnosis accuracy of the test.
In the probable Alzheimer’s disease group (n=64), the sensitivity (Se) of the combination medial temporal lobe atrophy and temporoparietal decrease in uptake for the clinical diagnosis of Alzheimer’s disease was 0.56; the specificity (Sp) of this association was 0.93; its positive predictive value (PPV) was 0.95, and its negative predictive value (NPV) was 0.50. The diagnostic accuracy (DA) of the association was 0.68. We split the population into two groups according to the MMSE score: patients with MMSE⩾ 18 (n=32, mean MMSE=23.68) and patients with MMSE <18 (n=32, mean MMSE=11.30). Table 3 reports the diagnostic values of the test. Seventy eight per cent of patients with probable Alzheimer’s disease had a medial temporal lobe atrophy shown by CT, 70% had a temporoparietal decrease in uptake, 92.2% had either one or the other, and 56.2% had both signs. The Se, Sp, NPV, PPV, and DA of the test was much lower in patients with Alzheimer’s disease with MMSE ⩾18 than in patients with MMSE <18 (table 3). However, the PPV remained at 0.85 in those 32 patients with a mean MMSE score of 24. Thus in patients with mild dementia (mean MMSE=24), the probability of a clinical diagnosis of Alzheimer’s disease was 85% (PPV) when the test was positive. On the contrary, when the test was negative, the probability of having a clinical diagnosis of Alzheimer’s disease was only 43% (as the NPV is 0.57).
In 17 of 36 patients the combination of medial temporal lobe atrophy and decreased temporoparietal uptake was bilateral. A medial temporal lobe atrophy was associated with the temporoparietal decrease in uptake in 36 of 50 (72%) patients whereas a temporoparietal uptake decrease was associated with medial temporal lobe atrophy in 36 of 45 (80%) patients.
Fourteen patients in the probable Alzheimer’s disease group had no medial temporal lobe atrophy and 19 had no temporoparietal decrease in uptake. Of 64 patients, 50 had medial temporal lobe atrophy on CT. The comparison between patients with and without medial temporal atrophy did not show significant differences for age, MMSE score, and Mattis DRSE score. The duration of the disease was longer in patients with medial temporal lobe atrophy (table 4).
In the possible Alzheimer’s disease group, nine of 13 patients did not have the association of medial temporal lobe atrophy with decreased temporoparietal uptake. These patients would have been diagnosed as having Alzheimer’s disease but two had vascular findings, one had extrapyramidal signs suggesting a senile dementia of Lewy body type but did not fulfil the current criteria for this disease,32one had epilepsy, four had anxiety or depression, and one had a progressive cognitive impairment but did not yet fulfil the criteria for dementia. In this group 11 patients had either unilateral medial temporal lobe atrophy or decreased temporoparietal uptake, or both, and two had none of these signs.
In the frontotemporal dementia group, 27 of 29 patients did not have the combination of medial temporal lobe atrophy and decreased temporoparietal uptake (15 had medial temporal lobe atrophy or decreased temporoparietal uptake). Two patients had both signs: one had frontal atrophy and frontal decrease in uptake spread to the temporoparietal areas, and the other patient had no frontal decrease in uptake. In both patients, the disease started with a progressive aphasia.38
In the cortical Lewy body disease group, five of the 10 patients had medial temporal lobe atrophy and decrease in uptake. Six had an anterior decrease in uptake.
In our study, the high dementia of frontotemporal type: Alzheimer’s disease ratio is probably due to our research interest in dementia of frontotemporal type and the rarety of patients with vascular dementia due to the fact that these patients are seen in a separate stroke unit. Moreover, patients with suspected vascular dementia tend to have MRI but no SPECT imaging.
The findings on CT are consistent with previous data.7 10The relevance of SPECT in the diagnosis of Alzheimer’s disease is controversial,39 although most authors have reported temporoparietal decrease in uptake as the most typical pattern in Alzheimer’s disease,11 12 14 18 39 40 others have also reported a decreased uptake in the frontal lobes13 19-21 41-44 and some45 have considered temporoparietal decrease in uptake to be non-specific.
Few authors were interested in the combination of medial temporal lobe atrophy shown by CT and temporoparietal decrease in uptake shown by HMPAO-SPECT.
To our knowledge, our study is the first prospective one conducted in a large population of patients. The sensitivity was higher in the study by Jobst et al 22 (86% as compared with 56% in our study). Jobst et al did not provide information about of the severity of the disease, except that the MMSE was <24 and it is therefore difficult to draw comparisons. Of the 12 patients who died in the study of Jobst et al, 10 fulfilled histopathological criteria for Alzheimer’s disease, nine of them having a clinical diagnosis of probable Alzheimer’s disease, and all had shown a combination of hippocampal atrophy and temporoparietal decrease in uptake.7
All the patients of the present study are still alive. However, 15 patients followed up in the Lille memory clinic died and had necropsy performed. All the clinical diagnosis were confirmed. The usual anatomoclinical correlation is between 80% and 90%.46 47
Eagger et al 6 studied the right and left medial temporal lobe separately in 35 patients with probable Alzheimer’s disease, with HMPAO-SPECT and CT. They reported temporal atrophy to be associated with temporal decrease in uptake on the right side in 20 patients and on the left side in 18 of 35 patients. Similarly, Wyper et al 48 compared 14 patients with histologically confirmed Alzheimer’s disease with 14 controls. They studied separately the two hemispheres and calculated the frequency of atrophy and decrease in uptake in several regions of interest in each hemisphere. Brain CT was performed parallel to the orbitomeatal line. Twenty temporal lobes were atrophied on CT in the Alzheimer’s disease group and eight in the control group. The combination of atrophy and decrease in uptake was present in 13 of 28 patients with Alzheimer’s disease and one of 28 controls. Methodological differences make comparisons with our study difficult. Moreover, these studies failed to report how many patients had the combination atrophy and decrease in uptake at least on one side. Reanalysis of the data of Wyper et al 48suggests that if the combination of medial temporal atrophy and temporoparietal decrease in uptake occurred in 13 patients then the association was present in 92% of patients with Alzheimer’s disease. On the other hand, if bilateral atrophy and decrease in uptake were present in all patients then the association would be present in 46% of patients, a proportion similar to that in our study (17 of 36=47.22%).
Pearlson et al 49 using MRI in nine patients with possible or probable Alzheimer’s disease and eight controls concluded that the volume of the left amydala and entorhinal cortex and a relative left tempoparietal decrease in uptake on SPECT discriminated best between patients and controls. The combination of the two measurements gave 100% discrimination. However, it was a very small group of patients and the complex methodology is not easily applicable.
In our study, 36 patients with probable Alzheimer’s disease had the combination of medial temporal lobe atrophy shown by CT and temporoparietal decrease in uptake shown by SPECT but 92.2% had medial temporal lobe atrophy or decreased temporoparietal uptake. In the dementia of frontotemporal type group, 27 of 29 did not have this combination. In the two patients with this combination the disease started with progressive aphasia.
Pasquier et al 50 reported a patient with probable Alzheimer’s disease without medial temporal lobe atrophy on CT: the diagnosis was histologically confirmed. The density of the neuropathological changes was less severe in the hippocampal region than in the frontal and parietal cortex, in agreement with the imaging data.
What is the relevance of the decrease in temporoparietal uptake ? It is seen in most cases with atrophy, but can be present in the absence of it. Wyper et al 48 found that atrophy and decrease in uptake are more frequent in patients with Alzheimer’s disease than in controls but in 50% of the patients a decrease in uptake was present in the absence of atrophy; thus atrophy is not the only cause of decrease in uptake.49 In our study, 20% of the patients had a decrease in uptake without atrophy and 28% of them had an atrophy without decrease in uptake.
The relevance of haemodynamic and metabolic modifications is still not clear and it is not known if they are a cause or a consequence of the neurodegenerative process. An interruption of the cholinergic connections between the neocortex and the Meynert basal nuclei has been suggested.51
The combination—medial temporal atrophy and temporoparietal uptake decrease—could correspond to the subgroup of Alzheimer’s disease type I of the classification of Blennow et al 52characterised by memory and parietal lobe deficits with both early and senile onset.
Although the association of medial temporal lobe atrophy and temporoparietal decrease in uptake may not have high sensitivity and specificity for Alzheimer’s disease, it may be more valuable in differentiating Alzheimer’s disease from frontotemporal dementia and may help to select patients in the early stage of Alzheimer’s disease.
We thank Jean-Pierre Pruvo and Michèle Hamon for their contribution in the CT scan study and Marc Steinling and Chantal Lamy-Lhuillier for having performed the SPECT