Positive impact of cladribine tablets on reducing brain atrophy in patients with relapsing-remitting multiple sclerosis: A longitudinal study

Abstract

Background:

Measuring brain volume changes over time is an objective and dependable surrogate marker for the pathological processes that damage the brain in relapsing-remitting multiple sclerosis (RRMS). These measures are particularly valuable for monitoring the long-term impact of immunomodulatory treatments such as cladribine.

Objectives:

To evaluate the long-term impact of oral cladribine treatment on brain volume loss in patients with RRMS.

Methods:

This real-world study processed magnetic resonance imaging (MRI) scans using FreeSurfer’s recon-all-clinical pipeline leveraging SynthSeg for brain segmentation. Piecewise linear regression was used to analyze brain atrophy changes over 4.5 years before and after cladribine treatment and estimate the time breakpoint of atrophy rate change.

Results:

A total of 448 MRI exams from 102 RRMS patients were analyzed. Before the initiation of cladribine treatment, brain atrophy rates were significantly steep with an α1 slope between −1.27 and −0.62 for the Thalamus, DGM, Subcortical GM, Cerebral WM, and BP. Over 2 years after treatment, breakpoints marked a shift in atrophy rates, with post-breakpoint slopes (α2) becoming non-significant, reflecting stabilization of brain atrophy.

Conclusions:

Cladribine treatment in highly active RRMS patients protects the brain from atrophy, with stabilization occurring over 2 years after initiation. The extended observation period highlights its sustained benefits compared with shorter clinical trials.

Keywords

Brain volume atrophy multiple sclerosis cladribine FreeSurfer

Introduction

Brain volume loss, as measured by magnetic resonance imaging (MRI) is widely recognized as an objective and reliable surrogate marker for the underlying pathologic processes in multiple sclerosis (MS).^1–3 Brain volume loss is reported to occur from the earliest stages of MS and accumulates progressively throughout the disease course.⁴ Specifically, brain atrophy, both gray and white matter loss, is associated with versatile aspects of MS-related symptoms and permanent impairments, including cognitive decline,⁵ fatigue,⁶ mood disorders,⁷ and neurologic disability.⁸ Measuring brain volume changes over time can offer insights into therapeutic interventions’ efficacy and guide clinical management decisions in MS.

Cladribine (2-chloro-2′-deoxyadenosine [2-CdA]) is a purine analog that selectively targets B and T cells and suppresses autoreactive lymphocytes.^9,10 Treatment is administered for patients with relapsing-remitting multiple sclerosis (RRMS) as short oral courses at the beginning of Years 1 and 2, with no requirement for further courses of treatment. Cladribine treatment proved efficient in treating RRMS,¹¹ as evidenced by phase III trials (CLARITY and ORACLE) showing a significant reduction in relapse rates, disability progression, and MRI measures of disease activity.^12–14 Notably, the CLARITY study demonstrated that cladribine treatment significantly reduced brain atrophy compared with placebo, with the percentage of brain volume changes over 2 years in treated patients approximating physiological rates.¹⁵ A subsequent post hoc analysis of the CLARITY study, spanning 6–24 months, indicated that gray matter volume loss was significantly lower in cladribine-treated patients than in those receiving placebo.¹⁶ Despite these findings, there is a lack of real-world data on the long-term impact of cladribine on brain atrophy. A 2-year follow-up period appears too brief to detect the treatment effect on brain atrophy. A more extended observation period could make brain volume changes a more significant indicator of neurodegeneration. Therefore, gaining insights into how cladribine treatment preserves brain integrity and prevents cortical loss over time is essential. The current study provides long-term data on brain atrophy 4.5 years before and after initiation of oral cladribine treatment. This knowledge could enhance our understanding of treatment benefits and inform decision-making in treatment strategies.

Methods

Study design

We conducted a retrospective long-term exploratory analysis using anonymized Sheba Multiple Sclerosis Center Computerized Data Registry data. The analysis focused on MRI outcomes among highly active MS patients who completed the 2-year, 2-dose oral cladribine protocol, receiving a cumulative dose of 3.5 mg/kg. We analyzed brain MRI data obtained 4.5 years before and after the initiation of cladribine treatment. The baseline (time = 0) was the first treatment week in Year 1. Data are presented as years relative to cladribine treatment.

Ethics

The Sheba Medical Center institutional review board committee approved the study (ethics approval number SMC-5596-08). Informed consent was waived as the study was non-interventional and retrospective. Data were collected, coded, and analyzed according to the ethical standards for human experimentation.

Patients

Inclusion criteria

Individuals aged 18 or older with highly active MS qualified for the study. Highly active MS referred to patients with either at least one relapse in the past year and at least one T1 gadolinium-enhancing lesion or nine T2 lesions while on another disease-modifying treatment (DMT), or two or more relapses in the previous year, regardless of DMT use; All participants completed oral cladribine treatment (2-year, two-dose cladribine treatment, receiving a cumulative dose of 3.5 mg/kg).

Exclusion criteria

Exclusion was based on manually visualizing MRI images exhibiting artifacts, and incorrect MRI acquisition parameters that precluded accurate software analysis.

Brain volume measurements

Brain MRI acquisition

T1-weighted MRI scans were performed using 1.5 or 3.0-Tesla MRI scanners with a high-resolution head coil. Imaging sequence employed a 3D protocol close to an almost isotropic voxel size of up to 2.2 × 2.2 × 2.2 mm³.

Post-acquisition processing. MRI scans were processed using the recon-all-clinical pipeline,¹⁷ available as a part of FreeSurfer 7.4.1 (https://surfer.nmr.mgh.harvard.edu/fswiki/recon-all-clinical), which enables fully automated cortical surface reconstruction and analysis of brain MRI scans of any modality, contrast, and resolution. The pipeline first performs volumetric segmentation and linear registration to Talairach space using SynthSeg,^18,19 followed by super-resolution of the input scan to 1-mm isotropic resolution using SynthSR²⁰ to facilitate visualization. Cortical surfaces are then reconstructed by predicting distance maps using SynthDist and fitting topologically accurate surfaces, which are parcellated according to three different atlases and used to estimate cortical thickness. All volumetric measurements were normalized as a percentage of baseline volume (the first treatment week in Year 1). Pseudo-atrophy, occurring within the first 6 months of treatment, was not addressed as our study focuses on later brain volume stabilization effects.¹⁶

The volume of the following brain regions was measured: Brain parenchyma (BP), Cerebral White Matter (WM), Total Gray Matter (GM), Cerebral GM, Subcortical GM, Deep GM (DGM), Thalamus, Cerebellum, and Brain Stem. BP was calculated as the summation of all segmented structures excluding the ventricles and CSF. DGM volume was calculated as the volume summation of the Thalamus, Caudate, Putamen, and Pallidum.

Statistical analyses

All statistical analyses and the creation of the plots were done using Python software version 3.10.12, Matplotlib (V3.7.1)²¹ and Seaborn (V0.12.2).²² The piecewise linear regression models were fitted using the piecewise-regression Python package (V1.5.0).²³ Descriptive statistics were used to summarize demographic, clinical, and radiologic variables, including minimum and maximum values, as means with standard deviation (SD) and medians with 25% to 75% interquartile range (IQR). Two-tailed statistics were used, and the significance level was p < 0.05.

As the exact time of a significant change in the relationship between the independent and dependent variables was a priori unknown, we could not incorporate a second linear growth factor by setting elements of the factor loading matrix to prespecified values. Therefore, we applied a piecewise linear regression that examined the relationship between brain-normalized volumes and the time relative to cladribine treatment. The model iteratively estimates the changes in gradient defined as a breakpoint that marks the transition point where the slope of the linear function shifts, indicating a change in the rate of brain volume decline.²⁴ The calculated α1 slope quantifies the rate of brain volume change before the breakpoint, while the α2 slope quantifies the rate of change after the breakpoint. A more negative α slope indicates a steeper decline in brain volume, signifying more rapid atrophy. The Davies statistical test was used to determine whether the breakpoint is statistically significant and to verify its existence independently of location.²⁵

Results

Demographic and clinical data of study patients

A total of 102 RRMS patients, 68 (66.7%) females, were included in the study. At the initiation of cladribine treatment, the mean ± SD age was 36.2 ± 10.5 years, median disease duration 6.7 years (25–75, IQR 3.5–13.5), and disability by the Expanded Disability Status Scale (EDSS) median 2.0 (1.1–4.0). Out of the 102 patients included in the study, 75.5% were previously treated with other medications for RRMS, while the remaining were newly diagnosed and started cladribine as their first treatment. Demographic and clinical variables are presented in Table 1.

Table 1.

Demographic and clinical variables at the initiation of cladribine treatment.

MS study population	RRMS patientsN = 102
Gender, n (%)
Male	34 (33.3)
Female	68 (66.7)
Age, Y
Mean ± SD	36.2 ± 10.5
Median (25–75 IQR)	33.4 (27.9–44.7)
Disease duration, Y
Mean ± SD	9.5 ± 8.3
Median (25–75 IQR)	6.7 (3.5–13.5)
Disability by EDSS
Mean ± SD	2.5 ± 1.8
Median (25–75 IQR)	2.0 (1.1–4.0)
IMD treated, n (%)	77 (75.5)
IMD treatment duration, Y
Mean ± SD	5.5 ± 3.4
Median (25–75 IQR)	4.8 (3.4–6.3)

Y: years; SD: standard deviation; IQR: interquartile range; IMD: immunomodulatory drugs.

MRI data

A total of 448 MRI exams were analyzed up to 4.5 years before and after the initiation of cladribine treatment. These exams were conducted at several time points, including 147 exams before the initiation of cladribine treatment, 102 at baseline, and 199 during follow-up (see Table 2 for yearly designations). The number of exams at each time point was relatively consistent, minimizing potential biases from uneven sample distribution.

Table 2.

MRI yearly designation.

Timepoint, years	Number of exams
−4	30
−3	30
−2	43
−1	44
0	102
1	75
2	57
3	39
4	28

Brain volume changes over time

Analyzing brain volume changes up to 4.5 years before treatment initiation is significant, as it enables the evaluation of brain atrophy in active RRMS patients under other immunomodulatory treatments. A clear progressive decline in the rate of atrophy was observed, as indicated by the α1 slope (Table 3). Significant volume reductions (α1 p < 0.05) were noted for all measured brain volumes except the brainstem. The most pronounced α1 slope values, indicating the fastest rate of brain atrophy before the breakpoint, were observed in the thalamus (−1.271), followed by the DGM (−1.036) and subcortical GM (−0.973). Following the breakpoint, the α2 slope became either positive or less negative (Table 3), and combined with the non-significant α2 p-values, this suggests a slowing or stabilization of brain atrophy following cladribine treatment.

Table 3.

Piecewise linear regression and Davies test analysis of brain volume changes related to the initiation of cladribine treatment.

	Davies p-value	Breakpoint	α1 slope	α1 p-value	α2 slope	α2 p-value
BP	0.036	2.522	−0.624	3.01 × 10⁻³⁴	0.124	0.724
Cerebral WM	0.023	2.235	−0.625	2.30 × 10⁻¹¹	0.302	0.507
Total GM	0.295	2.806	−0.65	2.02 × 10⁻³⁵	0.076	0.914
Cerebral GM	0.483	2.911	−0.649	1.93 × 10⁻³¹	−0.021	0.975
Subcortical GM	0.044	3.062	−0.973	1.85 × 10⁻⁵⁰	0.749	0.397
DGM	0.009	2.79	−1.036	3.06 × 10⁻⁴⁴	0.914	0.317
Thalamus	0.002	2.492	−1.271	2.74 × 10⁻²⁹	1.016	0.199
Cerebellum	9.25 × 10⁻⁴	0.048	−0.652	5.57 × 10⁻¹¹	−0.088	0.427
Brain Stem	1.258	−4.192	5.901	0.303	−0.547	2.85 × 10⁻²⁸

Breakpoint

Applying the piecewise linear regression model, we captured the distinct phases of brain volume change: a steeper decline before the breakpoint and a slower, more stabilized trend after. The significance of the breakpoint, as confirmed by the Davies test,²⁵ further validates this shift in the trajectory of brain atrophy following treatment (Table 3). It is important to note that the piecewise model independently identified the breakpoint in most brain volumes approximately 2 years after cladribine treatment initiation. This finding is crucial because the breakpoint represents the moment when the rate of brain volume changes and the process of brain atrophy undergoes a significant shift. This timing suggests that the therapeutic effects of cladribine may take some time to manifest fully in terms of stabilizing or slowing down brain atrophy. Following the breakpoint, a nonsignificant α2 slope for all brain structures analyzed is shown in Table 3 from fitting the piecewise linear model. The estimated intercept of the linear function before breakpoint, α1, is between −1.271 and −0.624 for various brain structures. The estimated slope of the linear function after the breakpoint, α2, is slowing and not statistically significant, suggesting that after initiating treatment, brain volume is more stabilizing without further decrease (Figure 1).

Discussion

To the best of our knowledge, the current study is the first to compare long-term brain atrophy rates before and after cladribine treatment, where highly active RRMS patients were on other immunomodulatory drugs (IMDs) before switching to cladribine. Prior to the breakpoint, patients were receiving other IMDs that have been reported to have modest effects on brain atrophy.^26,27 The identification of a consistent breakpoint around 2 years after cladribine was initiated (at time 0) suggests that the full therapeutic effects of cladribine begin to manifest during this period. This finding highlights cladribine’s significant ability to slow brain atrophy, contrasting with the less effective treatments used before the breakpoint.

The innovative approach of using the piecewise linear regression model applied in this study enabled us to evaluate brain atrophy changes over time. Moreover, the model independently identified the breakpoint in brain volume dynamics across various brain structures. The model consistently identified this breakpoint around 2 years after cladribine treatment initiation, marking the time when brain atrophy underwent a significant shift. To our understanding our study is the first to report this approach that allows for a clear comparison of brain atrophy rates before and after treatment, providing more precise insights into the temporal effects of cladribine on brain volume. The use of this model underscores the importance of statistical methods in pinpointing critical moments where IMD treatment impacts become more pronounced.

Before the breakpoint, the α1 slopes indicated a steeper decline in brain volume, reflecting ongoing atrophy in untreated active RRMS patients and in patients who received other IMDs despite treatment. Our findings align closely with those of Eshaghi et al.,²⁸ showing similar patterns of decline in brain volume across various structures. Specifically, we observed comparable rates of atrophy in the whole brain (our α1 = −0.624 vs. −0.34), cortical GM (our α1 −0.649 vs. −0.67), DGM (our α1 −1.036 vs. −1.34), and cerebellum (our α1 −0.652 for combined WM + GM vs. −0.58 for cerebellar GM). The consistency in these measurements across studies supports the robustness of our findings. Following the breakpoint, the α2 slopes became either positive or less negative, indicating stabilization or a reduction in the rate of brain atrophy. This suggests that cladribine exerts a potential neuroprotective effect, particularly after the 2-year mark, when the cumulative impact of the treatment is fully realized.

The consistent identification of this breakpoint across various brain regions, including the BP, cerebral WM, subcortical GM, DGM, and thalamus, reinforces the validity of the findings. It suggests that cladribine’s effect on brain volume is not isolated to a single brain region but has a broad, systemic impact across the brain. The statistical significance of the breakpoints, supported by low p-values from the Davies test, confirms that these shifts in brain volume dynamics are not random but represent meaningful changes directly attributable to cladribine treatment. The similarity in group sizes at key points mitigates concerns about bias introduced by uneven sample distribution.

The α2 slopes following the breakpoint were not statistically significant, suggesting that brain volume stabilized under cladribine treatment. This stabilization suggests a beneficial therapeutic success, as cladribine appears to halt further brain atrophy, even though it does not necessarily reverse the damage that occurred before the breakpoint. This stable phase post-breakpoint indicates that the treatment effectively prevents continued neurodegeneration, providing patients with a more stable disease course after the initial 2 years of treatment.

It is important to note that cladribine is administered once yearly over 2 years, and the consistent identification of the breakpoint around this 2-year mark aligns with the treatment schedule. The delayed yet significant therapeutic effect of cladribine suggests that the cumulative impact of the full 2-year treatment regimen is necessary to achieve brain atrophy stabilization. This reinforces the importance of completing the entire treatment course to harness the full neuroprotective potential. Furthermore, clinically, these findings suggest that skipping the second year of cladribine treatment may limit cladribine therapeutic benefit in delaying or reducing the stabilization of brain atrophy. By adhering to the full treatment regimen, patients are more likely to experience durable stabilization of brain volume and a reduction in the progression of neurodegeneration. This observation is supported by previous studies, such as the CLARITY study, which demonstrated that cladribine significantly reduced whole-brain atrophy compared with placebo over a period of 2 years.¹⁵ Our study further contributes by extending the data beyond the CLARITY study, providing insight into the longer-term effects of cladribine on brain atrophy in RRMS patients. The findings suggest a slowing or stabilization of brain atrophy following cladribine treatment.

In conclusion, this study provides new insights into the timing and cumulative effects of cladribine treatment on brain atrophy in RRMS patients. The innovative use of the piecewise linear regression model allowed the identification of a consistent breakpoint, demonstrating that the full therapeutic benefits of cladribine take time to manifest, becoming most evident after 2 years of treatment. Completing the full 2-year regimen is essential for achieving robust and durable stabilization of brain volume, preventing further atrophy, and offering long-term neuroprotective effects for patients with MS.

Footnotes

Data Availability Statement

The data supporting the findings of this study are available from the corresponding author upon reasonable request.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The study was partially supported by an investigator-initiated grant from Merck Serono Ltd (AN: MS700568_0218).

ORCID iD

Yehuda Warszawer

References

Dastidar

Heinonen

Lehtimäki

, et al. Volumes of brain atrophy and plaques correlated with neurological disability in secondary progressive multiple sclerosis. J Neurol Sci 1999; 165: 36–42.

Grossman

Udupa

, et al. Brain atrophy in relapsing-remitting multiple sclerosis and secondary progressive multiple sclerosis: Longitudinal quantitative analysis. Radiology 2000; 214: 665–670.

Chylińska

Komendziński

Wyszomirski

, et al. Brain atrophy as an outcome of disease-modifying therapy for remitting-relapsing multiple sclerosis. Mult Scler Int 2023; 2023: 4130557.

De Stefano

Airas

Grigoriadis

, et al. Clinical relevance of brain volume measures in multiple sclerosis. CNS Drugs 2014; 28: 147–156.

Calabrese

Rinaldi

Grossi

, et al. Basal ganglia and frontal/parietal cortical atrophy is associated with fatigue in relapsing-remitting multiple sclerosis. Mult Scler 2010; 16: 1220–1228.

Pellicano

Gallo

, et al. Relationship of cortical atrophy to fatigue in patients with multiple sclerosis. Arch Neurol 2010; 67: 447–453.

Nygaard

de Rodez Benavent

Harbo

, et al. Eye and hand motor interactions with the Symbol Digit Modalities Test in early multiple sclerosis. Mult Scler Relat Disord 2015; 4: 585–589.

Charil

Dagher

Lerch

, et al. Focal cortical atrophy in multiple sclerosis: Relation to lesion load and disability. NeuroImage 2007; 34: 509–517.

Moser

Hoepner

Schwenker

, et al. Cladribine alters immune cell surface molecules for adhesion and costimulation: Further insights to the mode of action in multiple sclerosis. Cells 2021; 10: 3116.

10.

Fissolo

Calvo-Barreiro

Eixarch

, et al. Immunomodulatory effects associated with cladribine treatment. Cells 2021; 10: 3488.

11.

Rammohan

Coyle

Sylvester

, et al. The development of cladribine tablets for the treatment of multiple sclerosis: A comprehensive review. Drugs 2020; 80: 1901–1928.

12.

Giovannoni

Comi

Cook

, et al. A placebo-controlled trial of oral cladribine for relapsing multiple sclerosis. N Engl J Med 2010; 362: 416–426.

13.

Giovannoni

Cook

Rammohan

, et al. Sustained disease-activity-free status in patients with relapsing-remitting multiple sclerosis treated with cladribine tablets in the CLARITY study: A post-hoc and subgroup analysis. Lancet Neurol 2011; 10: 329–337.

14.

Leist

Comi

Cree

BAC

, et al. Effect of oral cladribine on time to conversion to clinically definite multiple sclerosis in patients with a first demyelinating event (ORACLE MS): A phase 3 randomised trial. Lancet Neurol 2014; 13: 257–267.

15.

De Stefano

Giorgio

Battaglini

, et al. Reduced brain atrophy rates are associated with lower risk of disability progression in patients with relapsing multiple sclerosis treated with cladribine tablets. Mult Scler 2018; 24: 222–226.

16.

Cortese

Battaglini

Sormani

, et al. Reduction in grey matter atrophy in patients with relapsing multiple sclerosis following treatment with cladribine tablets. Eur J Neurol 2023; 30: 179–186.

17.

Gopinath

Greve

Das

, et al. Cortical analysis of heterogeneous clinical brain MRI scans for large-scale neuroimaging studies. In: Greenspan

Madabhushi

Mousavi

, et al. Lecture notes in computer science. Cham: Springer Nature Switzerland, 2023, pp. 35–45.

18.

Billot

Greve

Puonti

, et al. SynthSeg: Segmentation of brain MRI scans of any contrast and resolution without retraining. Med Image Anal 2023; 86: 102789.

19.

Billot

Magdamo

Cheng

, et al. Robust machine learning segmentation for large-scale analysis of heterogeneous clinical brain MRI datasets. Proc Natl Acad Sci USA 2023; 120: e2216399120.

20.

Iglesias

Billot

Balbastre

, et al. SynthSR: A public AI tool to turn heterogeneous clinical brain scans into high-resolution T1-weighted images for 3D morphometry. Sci Adv 2023; 9: eadd3607.

21.

Hunter

. Matplotlib: A 2D graphics environment. Comput Sci Eng 2007; 9: 90–95.

22.

Waskom

. Seaborn: Statistical data visualization. J Open Source Softw 2021; 6: 3021.

23.

Pilgrim

. Piecewise-regression (aka segmented regression) in Python. J Open Source Softw 2021; 6: 3859.

24.

Muggeo

VMR

. Estimating regression models with unknown break-points. Stat Med 2003; 22: 3055–3071.

25.

Davies

. Hypothesis testing when a nuisance parameter is present only under the alternative. Biometrika 1987; 74: 33–43, https://academic.oup.com/biomet/article-lookup/doi/10.1093/biomet/74.1.33

26.

Borges

Shea

Ohayon

, et al. The effect of daclizumab on brain atrophy in relapsing-remitting multiple sclerosis. Mult Scler Relat Disord 2013; 2: 133–140.

27.

Sotirchos

Gonzalez-Caldito

Dewey

, et al. Effect of disease-modifying therapies on subcortical gray matter atrophy in multiple sclerosis. Mult Scler 2020; 26: 312–321.

28.

Eshaghi

Prados

Brownlee

, et al. Deep gray matter volume loss drives disability worsening in multiple sclerosis. Ann Neurol 2018; 83: 210–222.