Cantor Daily Research Highlights

Research evidence excerpt

Cantor Daily Research Highlights

r out in the tail (~0.1% of random 4-patient draws) and the dose response is clean and consistent across

cohorts. Rummey’s placebo-response work shows mFARS can be inflated by frequent testing and older baselines – and

RESTORE-FA’s cohorts are on the older side with mFARS measures at baseline and 1-month, this tempers our mFARS

read while strengthening our view of the USS data.

■ We place more limited value on FXN biomarkers here: the acute functional changes reduce the need for a biomarker

surrogate, protein is hard to measure cleanly in blood and muscle, carrier and late-onset levels overlap (no clear

threshold), and LXEO recently removed frataxin as a co-primary (discussed below).

■ We see limited interpretability in the natural history (NH) model we had referenced: it enters frataxin on a log scale

(only apparent after author correspondence), so the modeled effect is small and tied to percent change – even a doubling

barely moves the modeled trajectory. We read this as an artifact of the regression structure rather than a confirmed

biological ceiling.

■ To be clear, this is not a dismissal of frataxin: we will keep a close eye on biomarker updates and, given our track record

predicting mRNA and protein, continue to incorporate them – but approval will depend on functional improvement,

not biomarkers.

USS versus mFARS: USS appears less subject to placebo effect

A 2026 placebo-response analysis makes the case for USS as the cleaner FA endpoint – placebo gains concentrate in the

limb domains of mFARS, not in the USS

■ Placebo response is domain-specific: limb (appendicular) items show trends in early placebo improvement while axial

gait and balance stays flat or declines, which strengthens USS as the more independent measure.