Combination disease‐modifying treatment in spinal muscular atrophy: A proposed classification

Abstract We sought to devise a rational, systematic approach for defining/grouping survival motor neuron‐targeted disease‐modifying treatment (DMT) scenarios. The proposed classification is primarily based on a two‐part differentiation: initial DMT, and persistence/discontinuation of subsequent DMT(s). Treatment categories were identified: monotherapy add‐on, transient add‐on, combination with onasemnogene abeparvovec, bridging to onasemnogene abeparvovec, and switching to onasemnogene abeparvovec. We validated this approach by applying the classification to the 443 patients currently in the RESTORE registry and explored the demographics of these different groups of patients. This work forms the basis to explore the safety and efficacy profile of the different combinations of DMT in SMA.


Introduction
[8][9][10][11][12][13] No systematic approach has been proposed to classify the various combination treatment scenarios for patients with SMA who receive DMTs and to facilitate analyses that will clarify potential benefit-risk differences between treatment regimens, including regimens incorporating one-time gene replacement therapy.The relative prevalence of different regimens is also poorly understood.We sought to devise a rational, systematic, broadly applicable approach for defining and grouping SMN-targeted DMT treatment scenarios that will facilitate future analyses aimed at exploring potential differences in clinical outcomes and address evidence gaps in the developing field of SMA DMTs.We also sought to describe the relative prevalence of various SMN-targeted DMT scenarios observed in the RESTORE registry (NCT04174157), a prospective, multicenter, multinational, noninterventional, treatment-neutral registry of patients with SMA, which was designed to integrate with existing real-world data. 14

Methods
To classify various treatment scenarios, definitions were developed by a group of neuromuscular experts representing a broad geographic range and varied clinical practice settings.Once consensus was reached on treatment definitions, these definitions were applied to patients enrolled in the RESTORE registry 14 (as of 20 December 2022) to examine the relative real-world prevalence of each treatment strategy.

Results
Our proposed treatment classification is primarily based on a two-part differentiation: initial therapy received (gene therapy or SMN2 splicing modifier) and persistence or discontinuation of subsequent DMT(s).This resulted in six polytherapy treatment categories: add-on, transient add-on, combination with onasemnogene abeparvovec, bridge to onasemnogene abeparvovec, switch to onasemnogene abeparvovec, and nusinersen/risdiplam combinations (Fig. 1).Because onasemnogene abeparvovec is a one-time therapy that provides ongoing SMN protein expression, any treatment administered after infusion with onasemnogene abeparvovec is termed add-on.Add-on treatment to onasemnogene abeparvovec that is discontinued is termed transient.Combination treatment includes initial treatment with nusinersen and/or risdiplam with ongoing or added treatment after infusion with onasemnogene abeparvovec.Bridge therapy is short-term treatment with nusinersen or risdiplam serving as a bridge to gene therapy with onasemnogene abeparvovec. 12Switching therapy is longer term treatment with nusinersen or risdiplam prior to receiving onasemnogene abeparvovec. 13idging and switching are distinguished based on duration of therapy with an initial SMN2 splicing modifier before the use of gene therapy, with bridging a shortduration (loading doses only for nusinersen or risdiplam treatment for ≤3 months) and switching a longer duration (one or more nusinersen maintenance dose or risdiplam treatment for >3 months).We defined discontinuation as two or more consecutive missed doses based on the expected dosing schedule (nusinersen) or no doses within the last 30 days (risdiplam).Durations for bridging and switching were adopted as reasonable surrogates for distinguishing these two treatment patterns in the absence of direct knowledge of caregiver/provider intent.Definitions for discontinuation were adopted based on previously published/presented analyses. 15,16mportantly, the actual time on nusinersen therapy before discontinuation is variable depending on phase of treatment (loading doses or maintenance doses).
We also assessed patient demographics and characteristics associated with monotherapy and the proposed addon and combination SMN-targeted DMT treatment scenarios (add-on, transient add-on, combination with onasemnogene abeparvovec, bridge to onasemnogene abeparvovec, and switch to onasemnogene abeparvovec) for patients from RESTORE who had received one-time gene replacement therapy with onasemnogene abeparvovec (Table S1).The greatest percentage of patients in each treatment group had two SMN2 copies (range: 44.4%-100%) and SMA type 1 (range: 33.3%-100%), except for patients receiving nusinersen monotherapy, with the greatest percentage of patients having three SMN2 copies (61.2%) and SMA type 3 (47.0%).Overall, patient demographics and clinical characteristics were similar between

Discussion
Our description of the relative prevalence of the various SMN-targeted DMT treatment scenarios observed in the RESTORE registry has some limitations.Although a global disease registry, RESTORE has primarily enrolled patients treated in the United States, and treatment patterns in other parts of the world may differ.In addition, treatment regimens that include only the two SMN2 splicing modifiers (nusinersen and risdiplam) may be underrepresented, and our proposed classification does not differentiate between various potential combination scenarios involving only the two SMN2 splicing modifiers (nusinersen and risdiplam).However, our classification system reflects the authors' collective expert opinion that the primary interest in differentiating treatment sequences or combinations for SMA lies with regimens that include onasemnogene abeparvovec because, unlike nusinersen or risdiplam, the gene therapy will theoretically continue to produce sufficient functional SMN protein over the lifetime of the patient.Several potential treatment scenarios are encompassed within the "Combination with onasemnogene abeparvovec" group.Although it is possible that these various scenarios could be associated with important differences in clinical outcomes, multiple variables (e.g., age at diagnosis, duration of symptoms, timing of treatments, disease severity at treatment initiation) confound answering this question.Small patient numbers representing each specific (often unique) scenario preclude meaningful analysis of outcomes within this group.All treatment groups include infants identified at risk for SMA by newborn screening as well as clinically diagnosed patients, indicating a potentially significant degree of heterogeneity within each group that may complicate analysis of differential outcomes by treatment regimen.Alternative means of classifying and differentiating SMA combination treatment groups are possible.We believe, however, that our proposed system is broadly applicable and does not preclude stratification according to other criteria (e.g., according to patient age during combination therapy, or analyses that consider ongoing exposure to previously administered SMN2 splicing modifiers [i.e., analyses that consider the half-life of these treatments]) that may be more suitable to approach specific clinical questions.Our classification exercise was limited to SMN-targeting treatments for SMA.Although clinical trials investigating new classes of SMA treatments (e.g., apitegromab, a selective myostatin inhibitor) are currently underway, 17 present experience with these new treatments in real-world practice is insufficient to inform rationale for integration into a classification schema such as this.Clinical experience with investigational myostatin inhibitors has thus far been limited to older patients with SMA (≥2 years of age), and the majority of the incident SMA population now receiving initial (and potentially additional) treatment are well below this age.Our classification schema could be expanded in the future to include non-SMN-directed treatments, as well as other treatments currently under investigation for older SMA patients, such as intrathecal onasemnogene abeparvovec.
Although add-on treatment was ongoing as of the data cutoff in the majority of cases, this percentage is certain to fluctuate over time.We expect that patient "migration" between treatment groups will be common as treatments are initiated and discontinued.In this cohort from RESTORE, the majority of patients who initiated treatment with either nusinersen or risdiplam continued on to receive onasemnogene abeparvovec.Future analyses will explore the rationales for add-on and combination treatments, and alignment of stated rationale with patient status at the time of add-on initiation.We will also explore clinical outcomes (including motor function assessments, motor milestone achievements, adverse events, and duration of additional treatment) for patients in the RESTORE registry according to these treatment definitions.

2156 ª 2023
The Authors.Annals of Clinical and Translational Neurology published by Wiley Periodicals LLC on behalf of American Neurological Association.