Abstract
Objective: To assess the long-term predictive value of quantitative lesion load measurement on brain MRIs in patients after a 10-year follow-up who presented initially with a clinically isolated syndrome of the optic nerve, brainstem, or spinal cord.
Background: Quantitative MRI measurement is being used in treatment trials as a surrogate marker in MS, but there is a lack of long-term MRI follow-up data in assessing the natural course of the disease from the earliest stages.
Methods: Using a semiautomated threshold technique, the total lesion volume (TLV), the course of the disease, and disability were assessed in 58 patients at onset and after 5 and 10 years.
Results: The TLV at presentation correlated significantly (r = 0.81, p = 0.0001) with the TLV and also with the Expanded Disability Status Scale (EDSS) score (r = 0.45, p = 0.001) at 10-year follow-up. In contrast there was no correlation of the TLV at 5 years with subsequent change in EDSS score over the next 5 years (r = 0.18, p = 0.12). The change in TLV over the first 5 years in patients who developed clinically definite MS (CDMS) differed significantly according to the type of disease course (relapsing-remitting with disability, secondary progressive, or benign) manifesting at 10-year follow-up.
Conclusion: Quantification of changes detected by T2-weighted brain MRI at the earliest clinical stages is strongly predictive of the subsequent development of CDMS as well as the clinical course and level of disability 10 years later.
In the last decade MRI has been widely used as a diagnostic tool and as a surrogate marker in monitoring the treatment of MS.1,2 It has proven an excellent alternative outcome measure for assessing the efficacy of new agents in the therapy of MS.3,4 MRI has also been used in short-term exploratory trials as a primary outcome measure where the decrease of MRI brain lesion burden or a reduction of MRI activity was assessed.5-7 However, long-term natural history MRI data of homogenous patient groups are less extensive.
The potential role of quantitative MRI for providing information about the course of the disease is not well understood. There has been a disappointing lack of association between the total lesion load and disability in cross-sectional studies in established MS, and therapeutic trials in established MS have also revealed weak correlations between the MRI findings and clinical outcome.3,4,8,9
Although MS is diagnosed on clinical grounds by the demonstration of neurologic impairment disseminated in space and time, prospective studies of paraclinical tests in the diagnosis of MS have shown that MRI is the most sensitive evaluation for the demonstration of dissemination in space.2 It is also an appropriate tool for the demonstration of dissemination in time.10 MRI identified all the patients who were diagnosed as having laboratory-supported MS. Moreover, MRI also predicted clinically definite MS (CDMS) in 95% of these patients who later developed significant clinical events.11 Difficulties in assessing the relationship between the lesion load and disability in the course of the disease may arise from the lack of long-term MRI follow-up studies and the scarcity of knowledge and understanding of the development of the MRI natural history.
In patients with clinically isolated syndromes (CISs) suggestive of MS, T2-weighted brain MRI revealed multifocal asymptomatic lesions indistinguishable from MS in 50 to 70%.12-14 Such lesions at presentation were accompanied by a relatively high risk of progression to CDMS in the next 1 to 5 years.15,16 A more precise, semiautomated quantification of the lesion load showed a significant correlation between the T2 lesion load at presentation and disability at 5-year follow-up.17 The study suggested that in patients with CISs the assessment of the lesion load at presentation is useful in predicting the subsequent clinical course over the next 5 years. The data also suggested that quantitative MRI could be helpful in selecting patients with a high risk of developing MS who would benefit most from early treatment and therefore be most suitable for early clinical trials. To make such inferences at the beginning of the disease it is important to establish the longer-term follow-up with the evaluation of MRI and clinical data.
The results of a 10-year clinical follow-up of the National Hospital cohort showed that the progression to MS in patients with a CIS who had an abnormal brain MRI occurred in 83%.18 For those with a normal MRI, progression to CDMS occurred in only 11%.18 To determine the long-term relationship between the quantitative total lesion volume (TLV) assessment of T2-weighted brain MRIs in patients presenting with a CIS of the optic nerve, spinal cord, or brainstem and the subsequent clinical and MRI course of the disease, the patients’ MRIs were quantitatively reanalyzed. The aims of the study were to 1) evaluate the development of the MRI brain lesion load after 10 years from presentation and its relationship with the initial brain MRI abnormalities; 2) assess the development of disability in a long-term follow-up in relation to the initial lesion load volume; 3) evaluate the impact of the lesion load at presentation on the type of disease that developed after 10 years; and 4) compare the rate of the accumulation of MRI pathology and the progression of disability in the first and second 5-year periods of the follow-up.
Methods.
Patients.
Patients with a CIS presenting to the National Hospital, Queen Square, London, or the Physicians Clinic at Moorfield’s Eye Hospital, London, between 1984 and 1987 were recruited. The study was approved by the Medical Ethics Committee of both hospitals, and informed consent was obtained from each patient before entry into the study. The criteria for inclusion into this study were as follows: 1) the syndrome was acute or subacute in onset with symptoms reaching their peak within 14 days of their onset; 2) the patients were aged 10 to 50 years at presentation, with the upper limit set to reduce the likelihood of age-related nonspecific abnormalities in the MRI data; 3) appropriate investigations, including the initial brain MRI, did not reveal an alternative diagnosis. Patients with CISs of the spinal cord always underwent either spinal MRI or myelography, and some patients with optic neuritis had an orbital CT. Some patients with bilateral optic neuritis and spinal cord syndromes were treated initially with a short course of steroids. None of the patients received long-term immunomodulatory or cytostatic treatment. The initial cohort is described in detail in the shorter-term follow-up reports after 1 and 5 years.15,19
Patients were located after 10 years from reference to our previous records or if they had moved by using the National Health Service central register at The Office for National Statistics. Ethical approval was obtained for release of information regarding the health authority in which each patient was registered. Recruitment of patients is described elsewhere.18 The current study examines the quantitative MRI brain lesion load at presentation and the clinical outcome in 71 patients after a mean follow-up of 9.7 years. Fifty-eight patients had brain MRIs performed at all three time points, i.e., at presentation and at 5- and 10-year follow-ups. Because of the unavailability of initial data for quantitative assessment, problems in electronic data storage over the 10-year follow-up period, and patient dropouts, 26 patients were lost in the 10-year follow-up in comparison with the quantitative evaluation of the 84 patients at the 5-year follow-up.17
At the 10-year follow-up all patients were reexamined by one of us and defined according to the Poser criteria20 on clinical grounds solely. Disability was rated using Kurtzke’s Expanded Disability Status Scale (EDSS).21 CDMS was diagnosed when two separate attacks disseminated in time and place and clinical evidence of two separate lesions were verified. Clinically probable MS (CPMS) was diagnosed when there were either two separate attacks disseminated in time without new signs or one attack with evidence on examination of two lesions, the second having developed at least 6 months after the first. To minimize the chance of including patients with slowly evolving acute disseminated encephalomyelitis, a period of at least 6 months was required between the two clinical episodes. At the 10-year follow-up, in those patients with CDMS, the disease subtype was defined as either benign, relapsing-remitting (RR), or secondary progressive (SP). Benign patients were defined as having a typical RR course of the disease and an EDSS score ≤ 3 at 10-year follow-up. Patients with RRMS were defined as having a typical RR course with an EDSS score greater than 3 at follow-up. The SP type comprised patients with a history of RR disease that entered a phase of gradual deterioration of at least 6 months’ duration with or without superimposed relapses.
Magnetic resonance imaging.
All baseline and 5-year follow-up MRIs were performed on a Picker 0.5-T superconducting scanner (Picker, Cleveland, OH). A spin-echo 2,000/60 sequence was used in all cases with contiguous axial slices. Twenty-three of the 46 scans that were considered abnormal at baseline had 10-mm–thick slices; the remaining baseline MRIs and all 5-year follow-up scans had 5-mm–thick slices. In general, slice thickness affects volume measurements because of partial volume and the detection of smaller lesions with thinner slices, but considering the overall low TLV (and low lesion number) at baseline we did not expect this to be a confounding factor that would affect the overall results. At 10-year follow-up a GE Signa (Milwaukee, WI) 1.5-T scanner was used. A conventional dual-echo spin-echo sequence (2,000/30/90) with a slice thickness of 5 mm was performed. Brain MRI at presentation was classified as abnormal if there were one or more asymptomatic lesions compatible with demyelination. Scans were reported normal if they were completely normal or in the case of a brainstem syndrome if only the symptomatic lesion was seen. Lesion volume quantification was performed using a semiautomated, local thresholding technique. This method, provided by the DispImage package, is a thresholding technique based on the local environmental intensity of the lesion.22 Lesions were identified on hard copies and measured by one of us. Modification of a part of individual boundaries was sometimes necessary in poorly defined or confluent lesions. Only a few lesions could not be delineated at all with the contour technique. Intraobserver variability with this approach is approximately 3%.23 A correction for nonuniformities caused by the transmission and reception properties of the radio frequency (RF) coil24 was not performed in the 10-year follow-up MRIs because the correction of field inhomogeneity did not affect derived lesion volume or reproducibility, inferring that this process may not be essential for semiautomated lesion segmentation using a local thresholding approach. The volume was calculated automatically as the computed area multiplied by the slice thickness. For some statistical analyses, the patients’ initial lesion volumes were divided into three groups. Group 1 consisted of 26 patients with a normal brain MRI. Forty-five patients who showed MRI abnormalities suggestive of demyelination were subdivided according to the TLV and categorized as Group 2 with a low lesion volume (≤3.0 cm3) or Group 3 with a high lesion volume (>3.0 cm3, which corresponded to approximately 10 lesions at presentation, with 0.3 cm3 being the median size of a single lesion at presentation). The methods and results of the assessment of new lesions (lesion counting) is fully described elsewhere.18
Statistical analysis.
The Mann-Whitney test was used to calculate differences in lesion volumes between groups. The reported changes represent median changes of the group at time B minus median at time point A. The relationship between lesion volume and clinical variables was expressed using the Spearman rank correlation coefficient (SRCC). The chi-square test was used to analyze differences between groups, and Kruskal-Wallis one-way analysis of variance was calculated where appropriate. An estimate of the likelihood of a patient developing CDMS and the proportion of patients for which this has had already occurred at a particular time were calculated with the Kaplan-Meier analysis.
Results.
Patients recruited for the 10-year follow-up had a similar frequency of CISs at presentation as the initial and 5-year follow-up cohorts. The number of abnormal scans at presentation (45 of 71; 63%) approximated the original patient group (80 of 132; 61%) and the 5-year follow-up cohort (57 of 89; 64%), suggesting that the present patients represent the original groups. The 58 patients who had MRIs at all three time points, i.e., at presentation and at 5- and 10-year follow-ups, did not differ in their clinical presentation or the number of abnormal scans from the total 10-year follow-up cohort (n = 71). Table 1 shows the patients’ characteristics and the number of abnormal scans at presentation. After a mean follow-up period of 9.7 years (SD 2.0), 40 patients (57%) had developed CDMS, and 4 patients were classified as having CPMS.
Patient characteristics
Quantitative lesion volume assessment at presentation and follow-up (MRI follow-up, n = 58).
Patients with a normal MRI at initial presentation had a lower lesion volume at 5-year (median 0 cm3, range 0 to 4.2) and at 10-year follow-ups (median 0 cm3, range 0 to 16.7) than did patients with MRI abnormalities, suggesting demyelinating lesions at baseline (median 1.3 cm3, range 0.15 to 55.1); the latter group had a median lesion load of 3.5 cm3 (range 0.58 to 36.6; p = 0.0001) at 5-year and 6.8 cm3 (range 0.55 to 88.6; p = 0.0001) at 10-year follow-ups. The former patient group had a much slower accumulation of MRI pathology over the first (median 0 cm3, range 0 to 4.2) and second 5-year follow-up periods (median 0 cm3, range −0.34 to 16.7). The latter group (patients with MRI abnormalities) had a median change of lesion volume of 2.0 cm3 (range −0.39 to 24.9; p = 0.0001) in the first and 3.0 cm3 (range −0.30 to 67.0; p = 0.0002) in the second 5 years of the follow-up.
The TLV at presentation correlated strongly with the TLV at 5 years (r = 0.87, p = 0.0001) and 10 years (r = 0.81, p = 0.0001). This strong predictive value of the initial lesion load for the subsequent lesion load was still present when patients with abnormal MRI at presentation were considered exclusively (10-year follow-up n = 45, r = 0.77, p = 0.0001).
Table 2 presents the qualitative MRI assessment of the disease load with the emphasis on the correlation between the quantitative TLV measurement and the detection of new lesions. At all time points there was a very strong correlation between the absolute TLV and total lesion number. Changes in TLV also correlated significantly with the number of new T2 lesions for each quinquennium, although the strength of this correlation was somewhat weaker between years 5 and 10.
Assessment of abnormalities in T2-weighted brain MRIs: Comparison of semiautomated lesion volume quantification and lesion counting (n = 58*)
Sixteen of the 45 patients with an abnormal MRI at presentation had infratentorial MRI lesions. The infratentorial lesion volume at presentation correlated moderately with the absolute TLV at both follow-up time points (5-year: SRCC = 0.5, p = 0.001, and 10-year: SRCC = 0.44, p = 0.006). Patients with infratentorial lesions at presentation had a significantly higher baseline TLV than did patients without infratentorial MRI lesions, and they had a greater increase in TLV than did patients with no infratentorial lesions. At 10-year follow-up the former patient group had a median change in TLV of 15.9 cm3 (range 2.7 to 40.3) in contrast to the latter group (no infratentorial MRI findings), who had a median change in TLV of 3.4 cm3 (range 0.09 to 78.6; p = 0.0064).
Quantitative lesion volume assessment and clinical status at follow-up (clinical follow-up, n = 71).
Type of disease. The progression of the disease as assessed by quantification of MRI-detected pathology was reflected by the clinical outcome. Of the 45 patients who had an abnormal MRI at presentation, 31 (68.9%) developed CDMS and 3 patients CPMS at the 5-year follow-up. After 10 years 37 (82%) had CDMS, and 2 were diagnosed with CPMS. In contrast, of the 26 patients with a normal brain MRI at first visit, only 1 (4%) had CDMS and 1 CPMS after 5 years. At 10-year follow-up, three (12%) had progressed to CDMS and two to CPMS. At 10-year follow-up the former group (abnormal MRI) had a median lesion volume of 6.7 cm3 (range 0.5 to 88.6), whereas in the latter (normal MRI) the median TLV was 0 cm3 (range 0 to 16.7) (p = 0.0001, Mann-Whitney). The clinical diagnoses of the total group of patients for the two follow-up time points are shown in table 3.
Clinical diagnoses at follow-up and the median time to development
The initial TLV was predictive for the total cohort of the type of disease manifesting at 5 years (CIS, CPMS, or CDMS) (p = 0.0001, Kruskal-Wallis) and predictive of the type that developed at 10 years (CIS, CPMS, or CDMS, including benign, RR, and SP; p = 0.0001, Kruskal-Wallis). Over the first 5-year follow-up period the change of lesion volume in the group developing CDMS (n = 32) was predictive of the subsequent type of disease course. MRI was performed in 31 of the 32 patients at 5 years. Thirteen patients were classified as having the benign type of MS at the 10-year follow-up. These patients had a lower accumulation of MRI pathology (median 2.0 cm3, range −0.39 to 22.8) over the first 5 years in comparison with the remaining RR patients who had acquired an EDSS score greater than 3 (n = 7) (median 4.5 cm3, range −0.38 to 15.6) or those who developed SP disease (n = 12, at 10-year MRI assessment [1 patient is missing] median 8.9 cm3, range 0.99 to 17.4; p = 0.005, Kruskal-Wallis) at 10-year follow-up. Patients who developed a RR or SP type of disease had the greatest change in TLV. Table 4 presents the relationship between the type of disease, TLV, and median lesion diameter at 10-year follow-up. Median lesion diameter was also larger at 10 years in the SP group.
Relationship of the type of disease at 10-year follow-up with the total lesion volume (TLV) at baseline and 10-year follow-up and lesion size at 10-year follow-up
Disability.
The TLV at presentation correlated with the EDSS score at 10-year follow-up (n = 71, SRCC = 0.45, p = 0.001). When patients with normal MRI were excluded, a persisting moderate correlation between the quantitative assessment of the total lesion load and the EDSS score was observed (n = 45, SRCC = 0.4, p = 0.007). The progression of the lesion volume increase differed, although not significantly, in the two follow-up periods, i.e., first and second 5 years (see table 2).
An evaluation was made of the relationship between the absolute TLV at baseline and 5 years and the change in EDSS score over the following 5 years. In the first follow-up period we found a significant correlation between the baseline TLV and the change in disability (EDSS score at 5 years) (r = 0.55, p = 0.001). The same applied for the relationship between the baseline TLV and EDSS score at 10 years (see above). The assessment of the absolute lesion volume at 5 years, however, did not correlate significantly with the change in EDSS score over the subsequent follow-up period (r = 0.18, p = 0.12).
In those with an abnormal scan at presentation, the initial infratentorial lesion volume also showed a significant correlation with disability at 10-year follow-up (n = 45, SRCC = 0.45, p = 0.002). Patients who had MRI evidence of a lesion in the infratentorial region at presentation that was suggestive of demyelination had a median EDSS score of 3.0 (range 0 to 10) at 10-year follow-up in contrast to patients who did not reveal any infratentorial MRI abnormalities (median EDSS score 1.2, range 0 to 6.5; p = 0.001, Mann-Whitney).
Table 5 shows the relationship between the change in brain lesion volume and the change in disability. We found a significant positive correlation between the MRI progression of disease assessed as lesion volume and the change in disability over the two 5-year follow-up periods. However, the magnitude of the relationship diminished with time.
Correlation between the change in disability and the change in lesion volume
In patients who had developed CDMS or CPMS at 5 years (n = 36) the EDSS score at 5-year follow-up was predictive of the subsequent progression toward disability. Only 1 of 23 patients (4%) with an EDSS score less than 3 reached an EDSS score of 6 at 10-year follow-up. In contrast, patients with an EDSS score greater than 3 (n = 13) at 5 years had a more unfavorable course, with 10 of 13 (77%) having an EDSS score of 6 or more at 10 years (chi-square, p = 0.015). The former group also had significantly lower median lesion load at 5-year follow-up (4.0 cm3, range 0.5 to 36.6) compared with the latter group (median lesion load at 5 years of 18.2 cm3 [range 1.5 to 114.3; p = 0.014, Mann-Whitney]).
Time to CDMS.
The TLV at presentation had a significant impact on the time between the first symptom (i.e., the presenting symptom) and a possible subsequent relapse that usually lead to a clinically definite diagnosis. For further analysis patients were divided into the following three groups to demonstrate the relationship between TLV at presentation and its predictive role for the subsequent development of CDMS: Group 1 (26 patients), normal brain MRI at presentation; Group 2 (34 patients), abnormal scan and TLV ≤3 cm3 at presentation; and Group 3 (11 patients), TLV >3 cm3 at presentation.
The relationship between the initial TLV and time to development of CDMS is shown in the figure (see table 6 for outcomes of the three groups in terms of disability). All 11 patients with a high TLV (>3 cm3) at presentation had developed MS before the 5-year follow-up; in contrast, 22% with the low TLV (≤3 cm3) had not developed CDMS after 10 years. The best prognosis, however, was clearly shown for patients with a normal brain scan at presentation. For the total patient group there was a strong negative correlation between the TLV at presentation and time to development of CDMS (SRCC = −0.64, p = 0.001). The presence of or the volume of infratentorial lesions at baseline had no influence on the time to development of CDMS.
Figure. Time to the development of clinically definite MS (CDMS) based on the total lesion volume at presentation.
Development of disease after a 10-year follow-up according to the lesion load at presentation
Clinical presentation and quantitative lesion volume assessment.
Patients with optic neuritis, a brainstem syndrome, or a spinal cord syndrome had a comparable TLV at presentation that did not differ significantly between the groups (p = 0.37, Kruskal-Wallis). A particular clinical syndrome was not predictive of subsequent TLV at the 5- and 10-year follow-ups or of the progression of accumulation of MRI pathology over time.
Presenting syndrome and clinical status at follow-up.
The type of clinical syndrome was predictive of the development of disability only in the first 5 years of the follow-up. Patients who had an optic neuritis as the presenting symptom had a median EDSS score of 1.0 (range 0 to 7.5), whereas patients presenting with a brainstem syndrome had a median EDSS score of 1.5 (range 0 to 8.5) as did those presenting with a spinal cord syndrome (range 0 to 6.5) (p = 0.039, Kruskal-Wallis). At 10-year follow-up the median EDSS score for the group that presented with an optic neuritis was 1.5 (range 0 to 7.5) in contrast to the other two groups, who had a median EDSS score of 2.5 (range: brainstem syndrome, 0 to 10; spinal cord syndrome, 0 to 6.5), although this difference did not reach statistical significance.
Discussion.
In this study we investigated the contribution of the quantitative assessment of MRI brain lesion load derived from the initial T2-weighted brain MRI for the subsequent course of the disease. Patients who presented with a CIS of the optic nerve, brainstem, or spinal cord were restudied at the 10-year follow-up. The main aim was to investigate the predictive value of objectively quantified TLV for the development of clinical MS, the type of disease course, and the degree of the disability over this relatively long period of follow-up.
To achieve reliable quantitative measurements of lesion volume, we used a semiautomatic local threshold contour technique that has been proven to have low inter- and intrameasurement errors.23 The absolute measurements obtained were somewhat higher than those we reported in a 5-year follow-up of the same cohort using a global thresholding technique.17 However, subsequent experience has shown that volumes measured using the global threshold are highly dependent of the threshold chosen, whereas measures using the local threshold contour technique provide absolute lesion volumes very similar to those achieved from manual outlining of lesions by an experienced observer.23 We therefore consider the contour method to be superior and accordingly applied it in the current study.
We demonstrated that the quantified lesion load is strongly predictive of the risk of developing CDMS and disability in patients presenting with a CIS (see table 6). Of the 34 patients with an abnormal scan but a low initial lesion load (<3 cm3), 22% did not develop CDMS after a 10-year follow-up period, and only 18% had an EDSS score greater than 6. All patients with a lesion volume >3 cm3 at presentation were diagnosed having CDMS after 10 years, and 45% had an EDSS score greater than 6. The initial TLV correlated significantly with the elapsed time to developing CDMS. Moreover, we observed a significant correlation between the quantitative assessment of the lesion load and type of the disease that developed after 10 years.
After dividing the 10-year follow-up into two 5-year periods we found a different rate of conversion to CDMS between the first and second 5-year follow-up duration. In the first follow-up period 36 of the 71 patients (50.7%) were diagnosed with CDMS, whereas in the second follow-up period only 8 of the 35 remaining CIS patients (23%) developed MS. We also observed a decrease of the rate of accumulation of disability using the EDSS score in the later disease course, although this may reflect the nature of the EDSS scale, with some patients spending longer at some levels than others. We likewise observed a substantially lower rate of progression to CDMS.25
We observed a proportionately greater increase in TLV during the second 5-year follow-up period when compared with the generally small numbers of the new lesions (median 0) during the same time period. The correlation between the change in T2 abnormalities and change in EDSS score in the second 5-year period was also lower when we employed lesion counting (r = 0.2918) rather than TLV measurement (r = 0.37). From these observations, one could infer that the increase of the lesion volume in the second phase of the follow-up may be partly due to enlargement of the already existing lesions and not predominantly due to the appearance of new lesion sites. This would be in agreement with previous cross-sectional studies that demonstrated larger or more confluent lesions, or both, in patients with SP disease than in those with benign and primary progressive disease.26,27 It is also consistent with our finding of a larger median lesion size in patients with SP disease than in those with RR disease. The separation of TLV and lesion number changes during the second 5 years would also be consistent with the emergence of new pathogenetic mechanisms as the disease evolves, although it may partly reflect the greater difficulty in identifying new lesions in the presence of larger TLVs.
Our long-term follow-up observations corroborate earlier, shorter studies that described a favorable outcome of a normal initial brain MRI in patients with a CIS.15,28-30 This study demonstrates also that the predictive value of the quantified T2 lesion load for the subsequent development of disease becomes weaker as the disease progresses. For patients who had developed MS by the 5-year follow-up, the 5-year absolute TLV did not predict increase in disability over the subsequent 5 years. Furthermore, although still significant, the relationship between the change in lesion volume and the change in EDSS score over that period was weaker than it was for the first 5 years. Given the large sample size, with prospective follow-up and no obvious biases as the result of patient dropouts, we think that this weakening is probably of biological significance. In particular, these results emphasize the importance of the early MRI findings as predictors of late disability. A crucial question regarding the pathogenesis of MS is whether pathologic changes that occur early in the course of the disease contribute directly to later clinical progression. The extent of early pathologic changes may be more important than later pathologic evolution, at least when monitored with T2-weighted brain MRIs. Extensive early inflammation and demyelination might provide the template for the occurrence of subsequent axonal degeneration.31 The current data point to a particularly important contribution of early pathologic change to later disability.
Despite the limitations of the 10-year follow-up data that include a different scanner for the initial images (0.5 T),32,33 this long-term follow-up study demonstrates that TLV measurement on conventional T2-weighted brain MRIs at presentation is a powerful tool for predicting the subsequent course of the disease including the time to the development of CDMS and the degree of disability. The data will be valuable in planning and interpreting long-term therapeutic trials of agents in patients with a CIS; in such patients, T2-weighted lesion volume is clearly a relevant measure.
Nevertheless, the contribution of different pathologic processes such as permanent loss of tissue (myelin or axons, or both), demonstrated by T1-weighted scans34 or magnetization transfer35,36 spectroscopy,37 or atrophy38 will undoubtedly contribute in an important way to the accumulation of disability and should also be implemented in future studies. Different pathologic processes and pathogenetic mechanisms may emerge in the later course of the disease and may partly explain the weaker correlation between the TLV and disability in the second 5-year follow-up.
- Received July 8, 1998.
- Accepted October 24, 1998.
References
Letters: Rapid online correspondence
REQUIREMENTS
You must ensure that your Disclosures have been updated within the previous six months. Please go to our Submission Site to add or update your Disclosure information.
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.
