Volume 206, Issue 2 , Pages 139-144, 15 February 2003
Clinical–MRI correlations in the secondary progressive phase of MS: lessons from the treatment trials
Article Outline
- Abstract
- 1. Introduction
- 2. Clinical–MRI correlations in SP-MS: conventional techniques
- 3. Clinical–MRI correlations in SP-MS: nonconventional techniques
- 4. Conclusions
- References
- Copyright
Abstract
Conventional MRI techniques are sensitive to detect MS lesions and their change overtime. In relapsing-remitting MS correlations with clinical measures are weak suggesting a pathological heterogeneity of these lesions. There are less data available in secondary progressive phase of the disease. The best source for clinical MRI correlations analysis is the placebo arm of the published interferon beta trials. This review presents the main clinical-MRI findings from these trials then focuses on recent promising observations obtained with non conventional MRI techniques in SP MS patients.
Keywords: Clinical–MRI correlation, Multiple sclerosis, Treatment trials
1. Introduction
In multiple sclerosis (MS), conventional MRI techniques have confirmed their sensitivity to detect lesions and their change over time. Many studies have shown significant associations between MRI and clinical measures, but these correlations are always modest, confirming that conventional MRI cannot be considered as a true surrogate marker of clinical disease [1]. Moreover, the majority of these studies refer to the relapsing–remitting phase of MS, in which the inflammatory aspects of the disease are prominent. MRI activity, as measured by conventional parameters, is more related to this inflammatory process than to the degenerative axonal loss acting in the secondary progressive phase of the disease, which is clinically defined as a sustained progression of disability [2]. New MRI techniques are more specific for the different pathological processes within the lesions. They differentiate more easily the actual mechanism underlying the progression of clinical disability and allow its follow-up. This review will present at first the clinical–conventional MRI correlations in secondary progressive MS. Unfortunately, there are not many data available, and their primary source are the last published interferon beta trials in SP-MS [3], [4], [5], [6], which will be reviewed. The major input of the recent nonconventional MRI techniques such as magnetisation transfer, spectroscopy diffusion imaging and functional MRI will be outlined in the last part.
2. Clinical–MRI correlations in SP-MS: conventional techniques
MRI has a role in both acute and chronic phase monitoring in MS. Acute phase monitoring refers to the inflammatory changes in disease activity, which is measured by the quantification of new or enlarging T2 and T1 Gd-enhancing lesions. Chronic phase monitoring refers to the total burden of the disease reflected in the total T2 and T1 (black holes) lesion volumes. In the last years, the measure of parenchymal atrophy, reflecting the degree of axonal loss, has been added to these parameters.
2.1. T2 lesion volume and clinical disability
The T2 lesion volume percentage change on annual MRI examinations in the interferon beta trials is presented in Fig. 1. With either interferon beta-1a or interferon beta-1b, the findings are quite similar: the T2 lesion volume increases slowly from baseline at the same rate during the trial in each placebo group (16% from baseline at year 3 with interferon beta-1b, 10% from baseline at year 3 with interferon beta-1a). A significant reduction of lesion load accumulation in all the interferon beta-treated groups confirmed the efficacy of this treatment on these parameters such as in the relapsing–remitting MS [7], [8] (−2% from baseline at year 3 with interferon beta-1b, −1.3% from baseline at year 3 with interferon beta-1a). A slight progression of the T2 lesion load is observed between years 1 and 3 (−4% versus −2% for interferon beta-1b, −3.5% versus −1.3% for interferon beta-1a).
Fig. 1.
This figure plots on the same picture T2 lesion volume percentage change on annual MRI examinations in two interferon beta trials in SP MS [4], [6]. In the European SP trial [4], each annually comparison between IFN beta-1b and placebo groups are significant (p<0.0001). In SPECTRIM, differences between placebo and both doses of interferon beta-1a are significant beginning 6 months after the start of treatment (p<0.0001). No comparison was performed between the two trials. They are presented in the same picture to confirm the results in a similar range.
However, in spite of this clear treatment impact on MRI lesions, a clinical beneficial effect was only found in the interferon beta-1b trial; although statistically significant, this treatment effect was moderate as a relative reduction of 21.7%, and reduction in the proportion of patients with progression was observed. In the interferon beta-1a trial, no difference was observed between the placebo patients and the patients treated with any dose. This discrepancy was also found in the US interferon beta-1b trial (not yet published), which did not demonstrate any treatment effect on the clinical measures [9].
Two major facts can be deduced from these data:
2.2. MRI activity, relapses and disability progression
The overall MRI activity of the disease in the two trials published in SP-MS patients is presented in Fig. 2, Fig. 3. The mean cumulative number of active lesions—i.e. new or enlarging lesions (T1 Gd-enhancing and new PD/T2 lesions)—observed annually in the IFN beta-1b study cohort is shown in Fig. 2. In IFN beta-1a study, MRI activity is reported as a mean combined unique active lesion per patient per scan—a parameter taking into account for T1 Gd-enhancing and active PD/T2 lesions without double counting (Fig. 3).
Fig. 2.
This figure shows the mean cumulative number of active lesions—i.e. new or enlarging lesions (T1 Gd-enhancing and new PD/T2 lesions) observed annually in the IFN beta-1b study cohort [4]. MRI scan at entry is a baseline. A significant reduction in the number of new or enlarging lesions in the interferon beta-1b group at annual time point compared with baseline was observed (p<0.0001).
Fig. 3.
This figure shows the median-combined unique active lesion per patient per scan—a parameter taking into account for T1 Gd-enhancing and active PD/T2 lesions without double-counting in the IFN beta-1a study [6]. This number was reduced by 78% for 22 μg and 89% for 44 μg compared with placebo (p=0.005 and p<0.0001) with a difference between doses (p=0.009).
Activity of the disease is still persistent in SP-MS patients as shown in the analysis of MRI activity in both placebo groups, although quantitatively less important than in RR-MS patients [10]. Treatment is associated with a substantial reduction of lesion activity in both trials (78% reduction at year 2 in the frequent cohort subgroup in IFN beta-1b trial, 89% reduction of median number of CU active lesions for 44 μg IFN beta-1a compared to placebo). These data are well-correlated with the relapses activity (Fig. 4): SP-MS patients are still relapsing, less than in the remitting phase, and IFN beta reduces the relapse rate at a rate similar in RR patients (30% with IFN beta-1b compared to placebo, 31% with IFN beta-1a compared to placebo).
Fig. 4.
This figure shows on the same picture the mean exacerbation/person/per year in two interferon beta trials in SP-MS Fig. 3, Fig. 5. Mean annual relapse rate was reduced by about 30% in the treatment group compared to placebo in interferon beta-1b study (p=0.002). A reduction in a similar range was observed in the interferon beta-1a study (22 μg versus placebo=RR=0.69 {0.56–0.84} p<0.001; 44 μg versus placebo=RR=0.69 {0.56–0.85} p<0.001).
There are differences in the treatment effect between the two trials if the progression of disability is analysed according to clinical activity: it still persits in the subgroup of patients without superimposed relapses [3] with IFN beta-1b and in patients without active lesions at baseline. Inversely, disability progression is reduced only in patients with superimposed relapses in IFN beta-1a study, with disappearance of any reduction in the overall cohort.
Therefore, if the dynamics of the degenerative process are dissociated with the activity of the disease, a persisting inflammatory activity can accelerate the disability progression and offers a treatment window in these SP patients.
2.3. T1 hypointense lesions and disability progression
T1 lesions have been shown to represent a subgroup of lesions in which more severe damage has occured [11]. Therefore, additional analysis of the T1 hypointense lesion load offers more information on the behaviour of the most severe lesions. The median percentage change in hypointense T1 lesion load from baseline in the IFN beta-1b is shown in Fig. 5. A 14% per year increase in T1 lesion load was observed in placebo-treated patients [12], similar to the increase of the natural course of 12% observed in a previous study [13]. IFN beta-1b significantly reduced this increase to 7.7% per year compared to placebo-treated patients.
Fig. 5.
This figure shows median percentage change in hypointense T1 lesion load from baseline in the IFN beta-1b trial [12]. If a linear increase is observed in lesion load across time in both arms, the rate of increase is significantly slower in the interferon beta-1b-treated patients compared to placebo patients (p=0.0003).
A positive correlation was found between absolute change in T1 lesion load and EDSS progression for the whole group of patients. However, this correlation was still significant in the IFN beta-1b arm, but was weak in the placebo group, suggesting that the degenerative process was better imaged when the inflammatory aspects of the lesions have been suppressed by the treatment.
Acute T1 hypointense lesions either become more or less hypointense over time depending on, at least, the demyelinative activity of the lesion and the extent of axonal loss: severe initial axonal loss appears to be a predictor for the future increase of T1 hypointensity such as completely demyelinated lesions in a longitudinal MRI-histopathological study [14].
The fate of the new T2/T1 Gd-enhancing lesions was analysed in the IFN beta-1b study [15]. Hypointense lesions were more likely to form from large (>5 mm) to small (<5 mm) lesions (25% versus 9%, respectively). IFN beta-1b reduced the number of new enhancing lesions compared to placebo, but more efficiently for small lesions (70% decrease) than for large lesions (46% decrease). Therefore, the proportion of total Gd-enhancing lesions developing into hypointense lesions was not different in both placebo and treatment group. Large enhancing lesions are more prone to be demyelinated and more destructive. In this case, it appears that interferon beta fails to alter this course.
1.4. Cerebral atrophy and disability progression
Significant change of brain volume with progressive parenchymal atrophy was shown in less than 2 years in a group of 29 MS patients [16]. This finding, independent of the methodology used, has been confirmed in other studies [17] with an average rate of tissue loss reaching 1% per year. This progressing parenchymal atrophy is considered as an indirect measure of degree of axonal loss [18]. The percentage change in cerebral volume compared every 6 months with baseline in the SP-MS patients (European IFN beta-1b trial) is shown in Fig. 6 [19]. In the placebo-treated patients, a mean reduction of 3.86% compared to baseline was observed after 3 years, but 0.89% reduction was yet observed after 6 months. In the IFN beta-1b-treated patients, a mean reduction of 1.39% in cerebral volume was observed after 6 months, but the mean total reduction of cerebral volume after 3 years was 2.91%, the difference between the two groups was not significant at any time of the trial.
Fig. 6.
This figure shows mean percentage change in cerebral volume compared with baseline for all patients in the IFN beta-1b trial over the study duration [19]. A mean reduction of 3.86% was observed by M36 in the placebo group compared to baseline, and a reduction of 2.91% in the interferon beta-1b-treated patients compared to baseline. There was no significant effect of treatment (p=0.14).
A post hoc analysis of change of cerebral volume when patients were stratified according to MRI activity at baseline (Gd-enhancing versus Gd-non enhancing patients) showed a greater loss in mean cerebral volume after 3 years in the placebo Gd-non enhancing patients compared to placebo Gd-enhancing patients (5.1% versus 2.6%, respectively). However, in the IFN beta-1b-treated patients, the reduction in the mean cerebral volume was greater in the active patients at baseline compared with inactive patients (3.7% versus 1.8%, respectively).
Globally, there was a correlation between a higher rate of cerebral volume loss and confirmed EDSS progression; for placebo patients with EDSS confirmed progression, the mean reduction of cerebral volume was 4.95% compared with 2.97% in placebo-treated patients without progression. However, no longitudinal relationship was observed between the change in EDSS and the change in cerebral volume. Better correlations were observed between the residual brain volume and disability when it was measured by a more sensitive rating scale such as the MSFC [20].
The main observation of this study is that there is a steady progression of parenchymal atrophy in SP patients, which is more pronounced in patients with faster clinical progression. Overall, there is no treatment effect on this MRI measure, but the anti-inflammatory effect on atrophy progression in treated patients reflects complex-mixed mechanisms. In other trials in which this parameter was studied, an early decrease of cerebral volume secondary to resolution of inflammation, edema and resolution of the blood–brain barrier breakdown has been observed [18]. The magnitude and duration of this effect are difficult to appreciate. It must be reminded that in RR-MS patients, a reduction in the progression of atrophy is observed in patients treated with IFN beta-1a [17]. Such an effect, weak and obscured by the acute inflammatory lesions, is possible in SP progressive MS, as nonactive IFN beta-1b-treated MS patients had a reduced atrophy progression compared with placebo-treated MS patients [19].
3. Clinical–MRI correlations in SP-MS: nonconventional techniques
3.1. Magnetisation transfer in SP-MS
Measurement of magnetisation transfer provides an evaluation of the structural integrity of tissue. In MS, low MT ratios (MTR) values—MTR quantifies the exchange of magnetisation between protons in free water and those bound to the macromolecules—result from a reduction in the integrity of the macromolecules caused by myelin damage and axonal destruction rather than from macromolecules dilution by inflammatory edema, as shown in a postmortem study confirming a correlation between reduction of MTR, axonal loss and demyelination [21]. In a comparative study of MTR in MS and traumatic brain lesions, there was a gradual increase of the MTR values at points at increasing distance from the center of MS lesions, meaning a centrifugal process, but was also reduced in the surrounding normal-appearing white matter (NAWM), whereas the transition was abrupt between traumatic lesions and NAWM: MTR detects microscopic MS pathology and confirms NAWM is abnormal in MS [22].
In SP-MS patients, a first study observed a decrease in lesions MTR inversely correlated with increasing disability [23].
The measurements given by MTR histograms of the whole brain take into account all the pathological changes, either the macroscopic demyelinated lesions or the microscopic abnormalities in NAWM. In a study comparing different subgroups of MS patients and controls, disabled relapsing onset patients, including some SP patients, exhibited abnormalities in different histogram parameters analysed, although benign and nondisabled relapsing patients had no abnormalities [24].
In primary progressive MS patients, disability may be explained mainly by the NAWM damage, whereas the macroscopic and microscopic lesions are both contributive to neurological impairment in SP-MS patients [24], [25]. The contribution of macroscopic lesions to the MTR maps can be subtracted by superimposition of segmented T2 images onto coregistered maps, which allows the measurement of the normal-appearing brain tissue (NABT) [26]. The NABT–MTR ratio changes were more evident in SP-MS patients and were only partly correlated with the extent of macroscopic lesions and the severity of intrinsic lesion damage [27]. More information will be obtained with further studies in larger cohort of patients.
3.2. MR spectroscopy in SP-MS
MR spectroscopy (MRS) provides information on chemical compounds present in brain tissue. The main metabolites studied in MS are N-acetyl-aspartate (NAA), marker of axonal integrity, choline and phosphocreatine, chemicals correlate of inflammation and demyelinating changes. MRS may also depicts lipid resonances associated with lipid breakdown at short echo times [1].
A significant decrease in NAA/Cr ratio was found in SP-MS patients either in T2 macroscopic lesions or in NAWM when compared to RR patients [28]. Discriminant values between RR and SP patients were found in the NAWM (NAA/Cr=1.75), with a negative correlation to clinical disability, independent of the duration of the disease. A progressive decline of NAA/Cr ratio was observed in central white matter of MS patients, but the correlation of NAA/Cr was lower in SP patients than in RR patients: it was suggested that atrophy may attenuate the decrease of NAA density, which conceals the true loss of axons and the strong impact of spinal cord atrophy on EDSS, which can bias the correlation of NAA/EDSS [29].
3.3. Diffusion-weighted imaging in SP-MS
Diffusion-weighted (DW) MRI permits the measurement of water self-diffusivity and gives information about the size, shape and orientation of cellular brain structures in vivo. The diffusion coefficient, which is lower in biological tissues than in free water, is called apparent diffusion coefficient (ADC). Diffusion anisotropy in white matter is linked to the axonal fiber orientation and can be disturbed by the pathological lesions. DW-MRI can provide another way to quantify the severity of tissue damage [30].
Higher ADC can be measured in NAWM of MS patients compared to controls, and a significnt increase of ADC was found in T1 hypointense compared to T1 isointense lesions and in nonenhancing compared with enhancing lesions [31], [32], [33]. SP-MS patients' lesions showed 50% higher diffusivity than in RR lesions, and a significant whole brain diffusivity assessed by diffusion histograms was significantly increased in few SP-MS patients studied [34].This technique is very promising as it has been possible to show a steady and moderate increase in prelesion NAWM diffusivity before the acute inflammatory event marked by the Gd enhancement [33].
4. Conclusions
MRI is sensitive to MS pathology, which explains the accumulation of T2 lesion load in SP-MS patients. However, the correlations with the fraction of disability measured by EDSS remain poor. The T2 lesions are in fact heterogeneous and have no the same clinical impact. The new MRI techniques are more accurate to analyse these different pathological conditions and reveal that not only macroscopic lesions, but also microscopic lesions inside the NAWM contribute to the progression of clinical impairment of neurological functions in MS patients. It will probably be necessary to develop multiparametric MRI models using these different techniques which will correlate with disability more reliably [35]. However, a clinical scale with a better efficiency than EDSS may also be more sensitive to the subtile variations of MS pathology, which will be demonstrated by these additional MRI approaches.
References
- . The role of MR techniques in understanding and managing MS. Brain. 1998;121:3–24
- Axonal transection in the lesions of MS. N. Engl. J. Med. 1998;338:278–285
- Final analysis of the European multicenter trial on IFN beta-1b in SP MS. Neurology. 2001;57:1969–1975
- Effect of IFN beta-1b on MRI outcomes in SP MS: results of a European multicenter, randomized, double-blind, placebo-controlled trial. Ann. Neurol. 1999;46:850–859
- . Randomized controlled trial of Interferon-beta-1a in SP MS: clinical results. Neurology. 2001;56:1496–1504
- Randomized controlled trial of Interferon-beta-1a in SP MS: MRI results. Neurology. 2001;56:1505–1513
- . Interferon beta-1b is effective in RR MS: 2. MRI analysis of a multicenter, randomized, double-blind, placebo-controlled trial. Neurology. 1993;43:662–667
- . Randomized double-blind placebo-controlled study of interferon beta-1a in RR MS. Lancet. 1998;352:1498–1504
- . Interferon beta-1b in SP MS: clinical and MRI results of a 3-year randomized controlled trial. Neurology. 2000;54:2352; [Abstract]
- The effect of interferon beta on BBB disruptions demonstrated by contrast enhanced MRI in RR MS. Ann. Neurol. 1995;37:611–619
- Histopathologic correlate of hypointense lesions on T1-weighted spin-echo MRI in MS. Neurology. 1998;50:1282–1288
- T1 hypointense lesions in SP MS: effect of interferon beta-1b treatment. Brain. 2001;124:1396–1402
- Accumulation of hypointense lesions (black holes) on T1 spin-echo MRI correlates with disease progression in MS. Neurology. 1996;47:1469–1476
- A longitudinal MRI study of histopathologically defined hypointense multiple sclerosis lesions. Ann. Neurol. 2001;49:793–796
- The effect of interferon beta-1b on the evolution of enhancing lesions in SP MS. Neurology. 2001;57:2185–2190
- Progressive cerebral atrophy in MS: a serial MRI study. Brain. 1996;119:2009–2019
- Use of the brain parenchymal fraction to measure whole brain atrophy in RR MS. Neurology. 1999;53:1698–1704
- Brain atrophy in relapsing MS: relationship to relapses, EDSS and treatment with interferon beta 1a. Mult. Scler. 2000;6:365–372
- The effect of interferon beta-1b treatment on MRI measures of cerebral atrophy in SP MS. Brain. 2000;123:2256–2263
- Optimizing the association between disability and biological markers in MS. Neurology. 2001;57:1253–1258
- Axonal loss in MS lesions: MRI insights into substrates of disability. Ann. Neurol. 1999;46:747–754
- Characterisation of white matter lesions in MS and traumatic brain injury as revealed by magnetisation transfer contours plots. AJNR Am. J. Neuroradiol. 1999;20:977–981
- Correlation of MTR with clinical disability in MS. Ann. Neurol. 1994;36:62–67
- Magnetisation transfer ratio histogram analysis of primary progressive and other MS subgroups. J. Neurol. Sci. 2001;185:11–17
- Comparison of MS clinical phenotypes using conventional and magnetisation transfer MRI. Neurology. 1999;52:588–594
- A magnetisation transfer histogram study of normal appearing brain tissue in MS. Neurology. 2000;54:186–193
- In vivo assessment of the brain and cervical cord pathology of patients with primary progressive MS. Brain. 2001;124:2540–2549
- Localised proton magnetic resonance spectroscopy in RR versus SP MS. Neurology. 1999;53:1091–1097
- . Magnetic resonance spectroscopy: imaging axonal damage in MS. J. Neuroimmunol. 1999;98:2–6
- . The value of new magnetic resonance techniques in MS. Curr. Opin. Neurol. 2000;13:249–254
- Apparent diffusion coefficients in benign and SP MS by nuclear MR. Magn. Reson. Med. 1996;36:393–400
- Comparison of MS clinical subgroups using navigated spin-echo diffusion weighted imaging. Magn. Reson. Imag. 1999;17:653–661
- The pathogenesis of lesions and NAWM changes in MS: a serial diffusion MRI study. Brain. 2000;123:1667–1676
- Whole-brain diffusion MR histograms differ between MS subtypes. Neurology. 2000;54:1421–1426
- Correlates of MS disability assessed in vivo using aggregates of MR quantities. Neurology. 2001;56:1331–1334
PII: S0022-510X(02)00351-9
© 2002 Elsevier Science B.V. All rights reserved.
Volume 206, Issue 2 , Pages 139-144, 15 February 2003
