Development of a Novel Composite Measure (SCACOMS) to Assess Disease Progression in Spinocerebellar Ataxia
Gilbert L'Italien1, Evan Popoff2, Basia Rogula2, Michele Potashman1, Lauren Powell2, Melissa Beiner1, Vladimir Coric1, Susan Perlman3, Jeremy Schmahmann4, Suzanne Hendrix5
1Biohaven Pharmaceuticals, Inc., 2Broadstreet HEOR, 3UCLA School of Medicine, 4Massachusettes General Hospital, 5Pentara Corporation
Objective:
Develop a composite scale derived from natural history datasets to assess disease progression and treatment effects more accurately in patients with spinocerebellar ataxia (SCA).
Background:

SCAs are rare inherited neurodegenerative disorders characterized by progressive ataxia affecting limb coordination, balance, and speech. Due to symptom heterogeneity, existing scales may not capture changes in response to disease-modifying treatments with sufficient sensitivity over one year (typical timeframe for interventional studies).

Design/Methods:

Using two natural history datasets, Spinocerebellar Ataxia Composite Scales (SCACOMS) were derived from Clinical Global Impression of change [CGI], Friedreich’s ataxia rating scale – functional stage (FARS-FUNC), and the Modified-functional Scale for the Assessment and Rating of Ataxia [f-SARA]).  A partial least squares (PLS) regression model was used to objectively select the optimal combination and weighting of items for measuring progression over time. The resulting sum of weighted item scores is designated SCACOMS.  Scales were assessed for ability to detect disease progression using mean-to-standard deviation ratios (MSDR).  SCACOMS were applied to BHV4157-206, a 48-week study evaluating troriluzole in SCA subjects (NCT03701399) to measure treatment effects (between group differences in least-squared mean [LSM] change from baseline).

Results:

Two composite scales were developed and cross-validated in each natural history dataset (SCA and SCA3 subset). Using PLS regression, the CGI, FARS-FUNC, and f-SARA gait, speech, and stance items were considered for inclusion. Improved sensitivity of SCACOMS to progression was observed (e.g., increased MSDR compared to the original scales).  Examination of SCACOMS treatment effects in BHV4157-206 showed a statistically significant treatment effect in SCA3 subjects favoring troriluzole (LSM-difference=3.60, p=0.0030) and Cohen’s d of 0.59.

Conclusions:

Using natural history datasets, SCACOMS was successfully developed, validated, and used to assess clinical decline in trial patients with SCA. The use of SCACOMS can allow for development of future studies with increased power to detect small but meaningful clinical impact of disease modifying treatments.

10.1212/WNL.0000000000205169