The topographical model of MS: Empirical evaluation of the recapitulation hypothesis

Introduction

Earlier identification of multiple sclerosis (MS) progression has significant implications for patient care. Initially created to describe phenotypically different clinical entities of the disease, the diagnostic categories of relapsing–remitting multiple sclerosis (RRMS) and secondary progressive multiple sclerosis (SPMS) have been widely used for trial design, regulatory approval and, importantly, therapeutic decisions. However, while relapsing and progressive processes contribute to MS disease course^1,2 the relationship between them^3,4 has not been fully characterized.^2,3 These phenotypes focus on clinical manifestations captured by the Expanded Disability Status Scale (EDSS), and by design, they do not directly represent the totality of symptoms, relapse severity, and the pattern and manner of accumulation of disability. Furthermore, as we have previously shown in a long-term patient cohort, there is a delay in diagnostic and prognostic certainty in patients transitioning from RRMS to SPMS, ⁵ which reflects that there is often no definitive moment of change in diagnostic category. The phenotype categories dichotomize what is increasingly understood to be a continuous disease spectrum, ⁶ particularly taking into account pathological evidence of degeneration in early MS in some patients, ² and the observation that accelerated brain atrophy begins early in the disease course. ⁷ It is hypothesized that the accelerated loss of brain volume is attended by a loss of the compensatory mechanisms that constitute neurological reserve, and that progression may become clinically apparent after reserve is depleted. On this basis, the topographical model of MS was proposed as a clinical manifestation framework, to provide a unified visualization of MS clinical course across the spectrum of relapsing and progressive forms of the disease. ⁸ In this paper, we conduct the first empirical application of the model to the patient cohort in which we originally demonstrated diagnostic uncertainty in the transition from RRMS to SPMS.

The topographical model provides a dynamic visualization of long-term disease course (Figure 1). Like other conceptual depictions of clinical course, this model remains agnostic to specific pathophysiological and cellular mechanisms of the disease, but rather it seeks to encapsulate the distinct clinical paradigm through which MS clinical course manifests. ⁹ The central nervous system (CNS) is visualized as a pool with increasing levels of depth, with the spinal cord and optic nerves at the shallow end, the brainstem and cerebellum with intermediate depth, and the cerebral hemispheres comprising the deep end. The depth of the water in this visual model corresponds with the degree of functional reserve intrinsic to these different regions of the CNS.

OPEN IN VIEWER

Figure 1.

Visualization of the topographical model.

To depict disease activity, lesions rise as focal peaks emerging from the base of the pool; those that cross the surface of the water—the clinical threshold—cause demonstrable signs and symptoms of an MS relapse. Disease activity in the spinal cord and optic nerves, occurring in the shallow end, is predisposed to cause the hallmark clinical relapses of MS: symptoms referable to partial myelitis and optic neuritis. ⁹ In the model, relapse severity is indicated by the maximal height of a topographical peak and the degree of recovery is depicted as the extent to which the peak recedes back down toward the base of the pool, below the clinical threshold. Lesion localization yields the particular clinical symptoms and functional systems (FSs) affected. In this way, the model can be individualized to encapsulate a particular patient’s “disease topography”—the individual clinical pattern and severity of disease. ⁸

The topographical model operates not solely as a visual depiction, but rather posits that there is a distinct intersection between disease topography and the loss of functional reserve that drives MS clinical course. ⁹ The model depicts that as time passes and functional reserve (the water level) declines, progression clinically recapitulates a patient’s prior relapse symptoms, incrementally manifesting above the clinical threshold a patient’s underlying disease topography. ⁸ This “recapitulation hypothesis” proposes that a patient’s relapse signs/symptoms and underlying lesion localization could potentially foretell an individualized pattern of disability accumulation as reserve is lost and the disease progresses. This hypothesis is depicted in the model as topographical peaks, emerging above the clinical threshold both acutely upon initial formation (relapse), and later reappearing insidiously as the threshold itself declines and functional and structural resilience is lost (progression). ⁹ Progression is thus conceptualized as a dynamic clinical threshold, characterized by above- and below-threshold disease which, over time, declines to unmask a patient’s individual “disease topography.”

Further integrating two concepts into the conceptualization of clinical course may offer a more individualized and clinically nuanced depiction of disease course so as to allow for earlier, more precise identification of progression: first, the central concept of localization, which underlies much of the symptomatic and prognostic variability of the disease;^10–12 and, second, the loss of functional reserve, as the primary driver for progressive accumulation of disability. ⁹ Earlier and more individualized identification of progression could have implications for MS care on several levels, from counseling patients about prognosis and evaluating the relative merits of changing disease-modifying therapies, ⁹ to determining who may be the best candidates for emerging trials and interventions focusing on remyelination and repair.

To pursue this goal, here we evaluate the recapitulation hypothesis, that a patient’s “disease topography” could, in essence, predict the clinical pattern of disability accumulation. We empirically study this theoretical model with clinical data from patients diagnosed with SPMS and with a known relapse/lesion history to demonstrate the extent to which the clinical manifestations of disability observed in the progressive stage of the disease recapitulate the symptoms of a patient’s prior relapses. We then render patient clinical course utilizing the topographical model disease simulation software to look for insidious reemergence of prior relapse symptoms that can serve as heralding the early signs—the leading indicator—of progression.

Methods

Clinical course visualization

The topographical model of MS is a clinical manifestation framework designed to run as a stand-alone iOS application for iPad (Apple Inc., Cupertino, CA), with custom disease-course simulation software. By varying five factors (relapse occurrence, severity, recovery, topographical distribution, the baseline level of reserve, and the rate of threshold decline), this software can generate millions of hypothetical disease-course depictions as rendered video files. This study is the first utilization of the topographical model to depict specific individual patients.

For 8 of the 10 patients who had sufficient historical data, clinical course was rendered using this custom disease-course simulation software. Figure 2 shows the disease-course simulation interface. Using this interface, lesions were plotted with variable frequency throughout disease course, divided in 1-year intervals. The topographical peaks were rendered in three sizes: large, medium, or small, corresponding to range of relapse severity. To depict a range of relapse/lesion recovery (the degree to which each topographical peak recedes below the threshold) each topographical peak’s recovery capacity was also selected at three levels: high, moderate, or low. Baseline reserve volume (the water level) was selected at three levels, and independently, the rate of depletion was also set at one of three rates: fast, moderate, and slow.

OPEN IN VIEWER

Figure 2.

Disease-course simulation interface.

For each case, relapse symptoms, severity, recovery, and progression rate were time-matched to clinical course based on chart documentation. To plot clinical course in the topographical model software, we matched topographical peaks to their referable EDSS FS by lesion location and laterality, utilizing our previously presented Localization and Functional Systems map. ¹⁴ This map codifies our assumptions regarding localization as a determinant of clinical symptoms, and illustrates the set of approximations that allowed us to best map the placement of topographical peaks.

To account for the fact that lesions at different levels within the CNS can yield motor symptoms, the Localization and Functional Systems map in the model is designed to include “tracks” as laid out in Figure 2, such that motor symptoms can result from lesions in the motor pathways from the spinal cord, through the brainstem, and into the hemispheres. For the purposes of plotting these cases in the model, if brainstem signs were present, then the topographical peak was placed in the brainstem; if motor symptoms affected arm and leg without additional localizing features, it was placed in the cervical cord; and if such symptoms affected only the legs it was placed in the thoracic cord. These approximations were made on the basis of clinical course, noting that magnetic resonance images (MRIs) were not available for the vast majority of the early relapsing events. In select cases, we were able to consult available MRIs to confirm localization. In particular, in the case shown in Figure 3, the MRI was useful to confirm the presence of a brainstem lesion at the time of the 2005 relapse, and we highlight this lesion/topographical peak concordance.

OPEN IN VIEWER

Figure 3.

Depiction of an individual patient’s disease topography mapped utilizing the topographical model disease simulation software.

Once the relapse events/topographical peaks were placed in time and space in the model, we then could approximate reserve levels for each case in the disease-course simulation. Individualized reserve levels were necessary to ensure that each relapse occurrence, degree of recovery, and findings on the neurological exam at time of SPMS diagnosis matched the documented clinical history. In the model, the combined volume of above-threshold topographical peaks corresponds with the degree of accumulated disability. ⁸ For each case, we used the clinical features of above-threshold disability as documented at the most recent visit to benchmark the “end point” of depleted reserve in each simulation. We then worked backward from this end point by iteratively adjusting the level of baseline reserve (high, medium, or low) and the depletion rate (fast, medium, or slow) to establish the best-fit clinical trajectory for each case. To illustrate this process: once we approximated each case simulation, we tested the model by adjusting the reserve parameters. If for a given case we set baseline reserve too high, above-threshold progression did not become manifest over the course of the simulation. If we adjusted the reserve too low, relapses that were documented to have full recovery did not resolve below the threshold. In this way, we identified for each case the baseline level of reserve and the rate of threshold decline necessary to ensure that all clinical milestones in the simulation accurately corresponded with the clinical documentation.

The topographical model iOS application was used to output rendered video files depicting disease course. Congruent with the contemporary Lublin/Reingold definition of SPMS, ¹⁵ for each case we identified the gradual above-threshold emergence of topographical peaks as the leading indicator of progression.

Results

Ten patients from our long-term cohort followed by a single academic MS specialist had sufficient clinical documentation for inclusion in this pilot study. Our cohort was 80% female, with an average age of 31.6 ± 8.6 years at initial diagnosis. Patients were followed for an average of 23.8 ± 8.8 years, for a total of 238 patient-years of data reviewed. Patients experienced a mean of 3.1 relapses during the RRMS epoch. Duration of disease was 15.5 ± 8.5 years prior to SPMS classification, with a mean subsequent follow-up of 7.5 ± 2.6 years after SPMS diagnosis to most recent clinic visit recorded.

Relapse symptoms were categorized by FS, laterality, and topographical regions depicted in the model: spinal cord/optic nerve, brainstem/cerebellum, and cerebral hemispheres. Regarding the motor system, there were 11 instances of lower extremity weakness recorded during relapses, with 100% of these symptoms present at most recent visit matched by laterality; there were five instances of upper extremity weakness with 80% of these present at most recent visit matched by laterality. There were 12 such instances of sensory dysfunction in the lower extremities with 91.7% matched by laterality at most recent visit, and seven instances of sensory dysfunction in the upper extremities with 71.4% matched at most recent visit. There were seven instances of brainstem/cerebellar symptoms recorded during relapses, with 100% present at most recent visit. There were a small number of visual, cerebral, and bowel/bladder symptoms captured at relapsed time-points. In total, 83.3 ± 0.2% of prior acute relapse symptoms were noted at the time of SPMS diagnosis, and 91.0 ± 0.2% of symptoms were present at most recent subsequent clinical visit. These data demonstrate concordance between the topographical distribution of relapse symptoms and accumulated deficits from subsequent progression.

Based on the disease-course renderings mapped in the topographical model, above-threshold progressive disease became apparent an average of 7.75 years (range 4–16) earlier than the diagnosis of SPMS was applied in practice (Table 1). A representative case (Case A in Table 1) is depicted with associated neuroimaging utilizing the topographical model in Figure 3.

OPEN IN VIEWER

Table 1.

Duration of disease and time to onset of progression for each case.

Case	Onset Date	Recapitulation Date	SPMS Dx Date	Leading Indicator
A	1986	1997	2008	11 years
B	2004	2005	2011	6 years
C	1999	2004	2008	4 years
D	1977	2003	2008	5 years
E	1979	1986	2002	16 years
F	1988	1999	2005	6 years
G	1996	1998	2005	7 years
H	1994	2000	2007	7 years

Discussion

The topographical model of MS was designed to serve as a conceptual framework for depicting the clinical manifestation of MS disease course, and furthermore to set the stage for empirical research. ⁸ It has previously been used only to depict hypothetical disease-course archetypes. Here we employed patient-level data to evaluate this more clinically nuanced and individualized representation of disease course that might allow for earlier identification of progression in MS. These pilot data are the first empirical test of the recapitulation hypothesis, which posits that the accumulation of disability is the consequence of a progressive unmasking of underlying disease burden.

We first demonstrated the model's utility in depicting patients' disease topography as the loci of clinical progression. Indeed, the vast majority of relapse symptoms with associated laterality were present at the time of SPMS diagnosis and most recent visit. Our findings are congruent with studies that have shown that localization and severity of relapses, along with the extent of recovery, are important drivers of MS progression and future disability.^9,16–20 These data support our hypothesis that progression clinically recapitulates prior relapse symptoms and unmasks previously clinically silent lesions.

We then utilized the recapitulation hypothesis to identify leading indicators of progression in individual patients to allow for targeted, earlier recognition of progressive disease. Using this model, we found leading evidence of progression emerging above the clinical threshold an average of 7.75 years earlier than when SPMS was clinically diagnosed. Thus, identifying a patient’s individual disease topography via this model allows for mapping of the particular pattern of disability accumulation and may offer an earlier window of understanding into a patient’s risk and pattern of progression. We found an increasing concordance of prior relapse symptoms with those of progression as time continued to pass after SPMS diagnosis. This is further congruent with the declining clinical threshold proposed by the model, and helps to visualize that the loss of reserve may be the principal driver of disability accumulation as MS progresses. ²¹ This pilot retrospective evaluation of the recapitulation hypothesis does not yet allow for prognostication claims at the individual patient level, which indeed will require prospective testing of the model using larger cohorts to refine it.

Recent work by Keegan and colleagues ²² has shown that motor progression may result from a single critically located lesion, most commonly located in the cervical spinal cord or at the cervico-medullary junction. As with the Keegan et al. work, our study also utilized corticospinal tract signs as a marker of continued progression, making use of the quantifiable nature and high salience of the motor FS in real-world clinical documentation. While the motor system is an apt one to test the recapitulation hypothesis, the topographical model suggests that a patient’s pattern of progression may not be limited to motor system dysfunction alone. Our results are congruent with natural history studies that have found that disease activity in the early years of disease impacted the later onset of progression and degree of future disability, ¹³ though our cohort is too small, and the early relapse data not sufficient, to formally replicate those studies here. That the specific clinical features of a patient’s disability progression could be mapped to lesion location may be an important missing piece in our ability to model and predict disease course. ²³ While prospective validation of this model will be needed before it can be used to predict onset of progressive symptomatology in practice, the findings of our pilot study suggest that particular attention should be paid to patients’ disease topography as put forth by this model, and that clinicians should be particularly alert for motor worsening that reflects prior relapses as potentially heralding SPMS.