Abstract
There is neuropathologic evidence that confluent MRI white matter lesions in the elderly reflect ischemic brain damage due to microangiopathy. The authors hypothesize that measuring changes in the progression of white matter lesions as shown by MRI may provide a surrogate marker in clinical trials on cerebral small-vessel disease in which the currently used primary outcomes are cognitive impairment and dementia. This hypothesis is based on evidence that confluent white matter lesions progress rapidly as shown in a recent follow-up study in community-dwelling subjects. The mean increase in lesion volume was 5.2 cm3 after 3 years. Based on these data in a clinical trial, 195 subjects with confluent lesions would be required per treatment arm to demonstrate a 20% reduction in the rate of disease progression over a 3-year period. Like any other MRI metric, the change in white matter lesion volume cannot be considered preferable to clinical outcomes unless it has been demonstrated that it matters to the patient in terms of function.
Cerebral small-vessel disease is an important cause of stroke, cognitive decline, and dementia. It causes about one-fourth of all strokes, both by causing lacunar infarction and by predisposing to intracerebral hemorrhages.1–5⇓⇓⇓⇓ Other clinical features include motor and cognitive dysexecutive slowing, dysarthria, mood changes, urinary problems, gait disturbances, and a higher incidence of falls in the elderly.6–8⇓⇓ Despite its importance on a population basis, no specific therapies are available. Treatment is limited to risk factor control and antiplatelet agents, based on trials that included all stroke subtypes. However, these trials fail to prevent both stroke recurrence and cognitive decline in many patients.
Only a small number of studies have examined therapy specifically in cerebral small-vessel disease. Currently, the clinical outcomes used in these studies are cognitive impairment and dementia. The lack of data may reflect the following: First, the clinical symptomatology of small-vessel disease-related cerebral tissue damage—apart from stroke—is only now becoming clear. Brain-imaging abnormalities often exist long before symptoms occur, and threshold effects that may lead to a temporal dissociation have been reported.9 Cognitive and other neuropsychiatric symptoms are often characterized by an insidious onset, and on an individual basis, it is difficult to separate whether these vascular changes are the sole cause for a patient’s symptoms, whether they add to another disease process like a degenerative disorder, or if they are even unrelated to the clinical symptomatology.
Second, small-vessel disease has been viewed as comprising almost exclusively hypertensive microangiopathy. Although hypertension is clearly the most prevalent risk factor for cerebral small-vessel disease, many hypertensive patients have normal brain MRI, whereas others with the same severity of hypertension have severe white matter disease. Factors modulating hypertensive damage and other pathogenic processes are important. These include genetic factors,10 endothelial dysfunction, 11 and low levels of free radical scavengers.12 In parallel, the recognition of genetically inherited microangiopathies like CADASIL (cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy) highlights the complexity of cerebral small-vessel diseases.13
Third, and most importantly, small-vessel disease per se is difficult to visualize in vivo. Although arteriosclerotic damage to medium-sized arteries can be visualized by conventional angiography and to some extent by high-quality MR angiography, both methods are unable to visualize small cerebral vessels with a diameter as small as 20 to 40 μm.14,15⇓
An alternative approach to visualizing the small vessels is to identify and monitor end-organ damage, that is, the brain consequences of small-vessel disease. This is now possible with MRI. We hypothesize that monitoring the progression of coalescent white matter lesions over time may be used as a surrogate in treatment trials with primary clinical endpoints.
We propose that MRI fulfills two major prerequisites for such an approach. These are 1) a reasonable confidence that observed tissue changes are indeed related to microangiopathy and 2) evidence for a rate of progression of these abnormalities that allows for observing possible therapeutic effects within a reasonable period of time.
White matter lesions: a marker for tissue damage due to cerebral small-vessel disease.
The most common neuropathologic correlate of white matter hypodensity on CT has been shown to be rarefaction of myelin, and therefore such CT abnormalities have been termed leukoaraiosis.16,17⇓ Equally important, these correlative studies confirmed microangiopathy as the major cause for this kind of tissue damage.16,17⇓ Subsequent observations using MRI, which detects many more white matter abnormalities than CT, led to uncertainties as to how to interpret such apparently asymptomatic cerebral tissue changes. The significance of such MRI observations was initially doubted because corresponding neuropathologic changes were not demonstrated for all abnormalities.18 It soon became clear that only larger MRI white matter lesions, defined as early confluent or confluent in early classification schemes, paralleled CT evidence of leukoaraiosis.19,20⇓ By contrast, punctate foci of signal hyperintensity on MRI can result from a number of other heterogeneous etiologies.19 Figure 1 presents an example of the neuropathologic correlates of confluent white matter lesions depicted on MRI. Pathologic studies have also shown that smooth periventricular hyperintensities, including caps around the ventricular horns, have to be differentiated from subcortical white matter abnormalities as they are related to disruption of the ependymal lining with subependymal widening of the extracellular space without representing a vascular etiology.19,21⇓
Figure 1. Examples of confluent white matter lesions: small cavities of tissue destruction within an area of extensive demyelination corresponding to a patchy hyperintensity on the premortem (A) and postmortem (B) MRI (straight arrows in [A], [B], and magnification of a part of the lesion in [C]). A large area of myelin pallor is also seen around the posterior horn (curved arrows). Kluver Barrera; original magnification, ×32. (Reprinted in part from Fazekas et al. Pathologic correlates of incidental MRI white matter signal hyperintensities. Neurology 1993;43:1683–1689.)
These pathologic data are complemented by a large body of evidence demonstrating that confluent white matter abnormalities are associated with symptomatic cerebrovascular disease. White matter lesions are most extensive and prevalent in patients with small-vessel strokes, and they are associated with other signs for small-vessel diseases including radiologic lacunar infarction and asymptomatic microhemorrhages.22,23⇓ In studies examining a sufficiently large population with a wide enough range of white matter changes, the severity of white matter lesions has been consistently associated with cerebrovascular risk factors.24,25⇓ Finally, extensive white matter damage is the hallmark lesion type of hereditary microangiopathies, in which the extent of lesions was shown to correlate with the severity of clinical symptoms.26 Thus, after excluding individuals with known white matter diseases such as multiple sclerosis (MS), the vast majority of white matter lesions encountered in the elderly can safely be assumed to be of ischemic, and more specifically of microangiopathic, origin. As has been pointed out in a recent review, there are many types of cerebral microangiopathy, but only a few occur commonly.27 Arteriolosclerosis is by far the most frequent form.27 Pathologic studies described arteriolosclerosis as an ubiquitous finding of persons above age 50.28 Severe arteriolosclerosis occurs predominantly in age groups above age 60, with more than half of individuals being affected. Numerous studies suggest that the severity of arteriolosclerosis corresponds to the extent of white matter damage.29–31⇓⇓
Progression of white matter lesions.
Five studies have reported longitudinal data on the progression of white matter lesions. The first of these studies was published in 1996 and described a mild increase in lesion extent over 5 years in 13 elderly subjects.32 The next study was done in a group of patients with heterogeneous clinical presentations and reported progression of lesions in 8 of 14 individuals over a period of 2 years.33 One year later, the Austrian Stroke Prevention Study reported the rate of white matter lesion progression over a 3-year observational period in a community-dwelling cohort of 273 middle-aged and elderly subjects free of signs or symptoms of neuropsychiatric disease.34 Based on the combined judgment of three independent raters, white matter lesions progressed in 17.9% of subjects, but only 8.1% showed marked lesion progression. A study on older people with gait dysfunction reported the first quantitative data over a period of 4 years.35 In this study, measurements were restricted to three consecutive slices, and volume estimation was done with a point grid. The investigators found an absolute increase in white matter hyperintensity volume of 1.1 cm3 after 4 years.
Recently, additional information was provided by 6-year follow-up data from the Austrian Stroke Prevention Study.36 In a total of 296 community-dwelling subjects, the total white matter lesion volume was quantified by outlining individual lesions over the whole brain on serial MRI scans, which were obtained at baseline and after 3 and 6 years. As in the previous studies, white matter lesions progressed over time in a linear fashion; that is, the volume change was almost identical during the first and the second triennial follow-up period. The total volume change of white matter lesions from baseline for the whole group was small. After 3 and 6 years, the median (interquartile range) volume increase was 0 (0; 0.3) and 0.1 (0; 0.7) cm3, with only 9.2 and 17.2% of study participants showing a progression that exceeded the magnitude of possible measurement error (1.81 cm3). Whereas these data might suggest a “benign” course of white matter lesions, stratification of data by the baseline grade of white matter abnormalities demonstrated that the small overall increase in lesion volume was due to the negligible rate of lesion progression in the large subset of study participants with no or only punctate abnormalities. By contrast, study participants with a baseline finding of early confluent or confluent changes showed a remarkably rapid increase in lesion volume. The median (interquartile range) volume increase over the 6-year period was 2.7 (0.5; 5.9) cm3 in subjects with early confluent lesions at baseline and 9.3 (7.1; 21.0) cm3 for individuals with confluent abnormalities. Almost two-thirds of study participants with early confluent, and all subjects with confluent, lesions demonstrated progression beyond the measurement error (1.81 cm3) over 6 years. In contrast, this was seen in none of the subjects with a normal baseline MRI scan and in only 14.6% of those with punctate foci. Table 1 gives a breakdown of the volume change of white matter lesion progression by strata of lesion grading. Examples of baseline white matter lesion grading and rapid progression of confluent abnormalities over 6 years of follow-up are shown in figure 2. Data are not available about the rate of progression in patients with symptomatic disease (either stroke or dementia) and extensive confluent white matter lesions. However, one would expect it to be at least as great as found in the confluent subset in the Austrian Stroke Prevention Study.
Table 1 Volume increase of white matter lesions by baseline lesion grade: 3- and 6-year data of Austrian Stroke Prevention Study
Figure 2. Baseline lesion grades (A to C) and example of rapid progression of confluent white matter lesions during 6-years of follow-up (D to F). Arrows show punctate (A), early confluent (B), and confluent lesions (C) and confluent white matter lesions at baseline (D). In the single slice shown, the volume of lesions increased by 2.6 cm3 between baseline and 3-year follow-up (E) and by 8.1 cm3 between baseline and 6-year follow-up (F). (Reprinted with permission from Elsevier [The Lancet 2003;361:2046–2047].)
Discussion.
Although clinical endpoints are the standard in therapeutic trials, reliable and specific laboratory or other markers may provide important information to our clinical impression. Simple examples are the treatment of bacterial infection or of diabetes mellitus in which the success of therapeutic intervention is first and best seen by a decrease in inflammatory markers or blood glucose levels. By analogy, based on the observation of a rapid progression of confluent white matter abnormalities, we suggest that trials on cerebral small-vessel disease with clinical primary endpoints could use the progression of lesion load as an outcome variable to evaluate therapeutic effects earlier and more readily than by the clinical observations alone. Research into MS faced a similar situation before it was recognized that the natural history of the disease on brain MRI is 5 to 10 times faster than that measured by clinical scales.37 Monitoring the accumulation of new lesions and the increase of total brain lesion volume allowed demonstration of therapeutic efficacy in a more sensitive manner. This has helped to establish beneficial treatments and to select promising new substances for larger trials with a primary clinical endpoint.38 In addition to high sensitivity, it is their objective and reproducible nature that makes morphologic abnormalities like white matter lesions on MRI attractive targets for therapeutic trials; data collection is not endangered by unblinding of the observer, and the extraction of information from stored images can be done repeatedly and flexibly.39 Prerequisites for successful implementation of such a strategy are the strict adherence to predefined imaging protocols and expert assessment of signal changes. Data indicate that this is feasible and best done using centralized rating procedures.20,40⇓
Whether serial assessment of MRI white matter lesion volume could play a similar role in therapeutic trials in cerebral small-vessel disease is as yet undetermined. Undoubtedly, there are many striking similarities with MS. First, extensive cerebral white matter lesions reflect an important pathologic process in vascular dementia.6,41⇓ Second, the rate of progression of confluent white matter lesions closely matches that of the white matter abnormalities seen in MS. According to the 3- and 6-year results of the Austrian Stroke Prevention Study,36 the annual increase in white matter lesion volume ranges between 12.5 and 14.4% in subjects with early confluent lesion and between 17.3 and 25.0% in those with confluent abnormalities. A similar annual increase in lesion volume has been reported in MS patients with relapsing-remitting (32.2%) and secondary progressive (13.3%) MS taken from natural history studies or the placebo arms of treatment trials from six European centers.42 It was assumed that 101 patients with relapsing-remitting and 337 patients with secondary progressive MS would be needed per treatment arm to show a 30% treatment effect on MRI lesion load with sufficient statistical power in a 3-year therapeutic trial.41 As can be seen from the sample size calculations presented in table 2, similar numbers of patients would be needed to demonstrate treatment effects in patients with small-vessel disease-related white matter lesions. A total of 227 patients with early confluent and confluent lesions per treatment arm would be needed to show a 30% therapeutic effect in a 3-year study. Trials in stroke are often powered to detect ≤20% risk reductions. This would require a large number of 522 participants with early confluent and confluent lesions per treatment arm, but if one focuses on subjects with confluent abnormalities alone, which is the most likely scenario in a subcortical vascular dementia trial, a 30 and 20% therapeutic effect can be detected with 87 and 195 patients per arm, respectively. It is important to emphasize that these sample size estimates are based on observations in a community-dwelling cohort and not on a clinic-based cohort. This may have implications on the numbers of subjects that have to be included in a clinical trial. Another point that needs to be considered is that all sample size estimates are for trials including only subjects with marked pre-existing white matter changes. This group may have a different susceptibility to cerebral small-vessel disease, and this could limit the generalizability of findings from this group to groups without pre-existing white matter lesions. However, many patients with vascular dementia and vascular cognitive impairment have severe confluent white matter lesions on MRI. It is also important to realize that change in white matter lesion load, like in MS, 43 cannot presently be considered a validated surrogate endpoint. Formal validation of surrogacy requires that 1) the treatment is effective on the surrogate endpoint, 2) the treatment is effective on the clinical endpoint of interest, 3) the surrogate endpoint and the clinical endpoint are significantly correlated, and 4) the effect of the treatment on the clinical endpoint disappears when adjusting for the surrogate endpoint.44 Currently, data exist only as to the correlation between white matter lesion extent and cognitive dysfunction. In cross-sectional studies, a weak but significant correlation between cognitive functioning and white matter lesion load has been described17,22,45⇓⇓; a correlation in longitudinal studies remains to be established.34 Clearly, just because MRI has a better power to detect change of small-vessel disease-related brain lesions does not mean that a therapeutic effect on lesion load does also matter to the patient in terms of function including activities of daily living, quality of life, caregiver burden, and cost/benefit ratio. Further studies evaluating the effect of white matter lesion progression on functional abilities, particularly in subjects with confluent white matter changes, are needed. The other surrogacy criteria44 can be assessed only if one actually performs a clinical trial on cerebral small-vessel disease, which includes white matter lesion volume as secondary endpoint. It is desirable that such studies also include newer MRI techniques such as diffusion-tensor imaging and magnetization transfer imaging, which may allow more sensitive estimates of tissue damage.46–48⇓⇓ Recent studies have found that cognitive function correlates more strongly with diffusion tensor imaging parameters than T2 lesion volume.49 Pathologic and other correlative studies showing that confluent white matter changes represent small-vessel disease-related cerebral damage and recently published evidence demonstrating rapid progression of confluent white matter lesions have set the stage for such exploratory trials.
Table 2 Sample sizes per treatment arm according to lesion grade at baseline and effect size in 3-year interventional trial with outcome measure being change in lesion volume
- Received November 20, 2003.
- Accepted February 23, 2004.
References
Letters: Rapid online correspondence
REQUIREMENTS
If you are uploading a letter concerning an article:
You must have updated your disclosures within six months: http://submit.neurology.org
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.
