# Post-Approval Evidence and the Surrogate-Endpoint Landscape for Tofersen in ALS **Date:** 2026-05-04 ## Executive Summary Tofersen (Qalsody) received FDA accelerated approval in April 2023 for SOD1-associated amyotrophic lateral sclerosis (SOD1-ALS) based on plasma neurofilament light chain (NfL) reduction as a surrogate endpoint. In the three years since approval, a substantial body of evidence has accumulated on long-term functional outcomes, while attempts to generalize the surrogate-endpoint approach to non-SOD1 subtypes have met significant setbacks. Several new targets for antisense oligonucleotide (ASO) and gene-therapy strategies have entered early development. Five critical gaps remain that must close before the NfL-based surrogate-endpoint model can be widely generalized across ALS etiologies. --- ## (a) Functional Clinical Outcomes Since Approval ### 1. Phase 3 VALOR + Open-Label Extension (148-Week Data) The most important post-approval evidence comes from the completed open-label extension (OLE) of the VALOR trial, published in *JAMA Neurology* in early 2026 (Smith et al.). Over 148 weeks of follow-up, investigators compared participants who initiated tofersen at randomization (early-start group) versus those who crossed over from placebo after 28 weeks (delayed-start group). | Outcome Measure | Early-Start Group (Week 148) | Delayed-Start Group (Week 148) | Interpretation | |---|---|---|---| | ALSFRS-R total score change | −9.9 points | −13.5 points | Numerically less functional decline with earlier treatment | | Slow vital capacity (SVC) | −6.9 | −10.4 | Preserved respiratory function in early starters | | Handheld dynamometry (HHD) | −18.6 | −35.1 | Markedly better muscle strength with early initiation | | Plasma NfL | −67% from baseline | −64% from baseline | Sustained, deep reduction in neurodegeneration marker | Key findings from this long-term analysis include: - **Disease stabilization or improvement:** Approximately one-quarter of participants experienced stabilization of symptoms and, in some cases, functional improvement in grip strength and respiratory function (Washington University, 2026). - **Survival benefit:** Survival in tofersen-treated patients exceeded that predicted from the expected natural history of SOD1-mutated ALS. - **Quality of life:** Consistent benefits were observed on patient-reported outcome measures including ALSAQ-5 and EQ-5D-5L. - **Responder analysis:** Early initiation of tofersen was associated with numerically less decline across all clinical domains and a lower risk of death-equivalent events compared to the delayed-start group. ### 2. Real-World Evidence Two independent real-world cohorts have confirmed the trial findings: **German Early Access Program (Meyer et al., 2024):** A 12-month, multicenter observational study of 24 SOD1-ALS patients confirmed sustained reductions in serum NfL and CSF phosphorylated neurofilament heavy chain (pNFH) consistent with VALOR/OLE. Therapy was reported as safe with no persistent serious adverse events. **Washington University Real-World Cohort (Smith et al., 2025):** Seven patients treated at Washington University from November 2021 to February 2024 showed sustained stabilization of disease in a real-world setting. The authors concluded that "our understanding of neurologic and functional recovery in SOD1 ALS on tofersen is just coming into existence," emphasizing the need to optimize rehabilitation strategies alongside ASO therapy. ### 3. Presymptomatic Prevention Trial (ATLAS) The ongoing Phase 3 ATLAS study (NCT04856982) is the confirmatory trial required by the FDA's accelerated approval. It randomizes ~150 clinically presymptomatic carriers of high-penetrance SOD1 variants with elevated plasma NfL to tofersen versus placebo. ATLAS will test whether initiating treatment before symptom onset can delay or prevent clinical manifestation. This trial is critical for validating both the clinical utility of NfL as a risk/susceptibility biomarker and the concept of presymptomatic intervention in ALS. --- ## (b) Translation of the Surrogate-Endpoint Approach to Non-SOD1 ALS Subtypes The tofersen precedent established that NfL reduction can serve as a surrogate endpoint "reasonably likely to predict clinical benefit" under accelerated approval. However, extending this framework to other ALS subtypes has proven far more challenging. ### 1. C9orf72-Associated ALS The hexanucleotide repeat expansion in *C9orf72* is the most common genetic cause of ALS (~10% of familial, ~1–2% of all ALS). Three ASO programs have been tested: | Program | Approach | Status | Key Finding | |---|---|---|---| | **WVE-004** (Wave Life Sciences) | Variant-selective ASO targeting G4C2 repeat-containing transcripts | **Discontinued (May 2023)** | Achieved robust, sustained reductions in poly(GP) (up to 50%) but showed no clinical benefit in Phase 1b/2a FOCUS-C9 trial | | **BIIB078 / tadnersen** (Biogen/Ionis) | ASO targeting C9orf72 sense strand | **Discontinued** | CNS distribution was demonstrated, but post-mortem analysis (McEachin et al., *Cell*, 2025) found **no reduction** in C9ORF72 transcripts, dipeptide repeat proteins, or TDP-43 pathology despite ASO penetration | | **afinersen** (undisclosed sponsor) | ASO targeting repeat expansion | Preclinical/early clinical | Limited public data available | These failures are instructive: WVE-004 showed target engagement (poly(GP) reduction) but no clinical benefit; BIIB078 showed CNS delivery but failed target engagement at the molecular level. **Neither program produced a meaningful NfL reduction that tracked with clinical outcomes**, underscoring that NfL's surrogacy does not automatically generalize across genetic etiologies. ### 2. ATXN2-Associated ALS Intermediate-length polyglutamine expansions in *ATXN2* are a known genetic risk factor for ALS. Biogen and Ionis developed BIIB105 (aka IONIS-ATXN2Rx), an ASO designed to lower ATXN2 protein. - **May 2024 discontinuation:** Despite significant reduction of ATXN2 protein in CSF, BIIB105 showed **no reduction in NfL and no clinical benefit** over the 6-month placebo-controlled period. - Biogen's stated rationale: "While BIIB105 lowered ATXN2 protein, it did not reduce neurofilament, which gives us confidence that BIIB105 did not slow the disease process." This failure is particularly salient for surrogate-endpoint generalization because it directly contradicts the SOD1-ALS logic: target engagement (ATXN2 lowering) occurred, but the expected NfL response—and clinical benefit—did not materialize. ### 3. Sporadic (Non-Genetic) ALS Approximately 90–97% of ALS cases are sporadic, with TDP-43 proteinopathy present in the majority. Currently, there is **no FDA-recognized biomarker that substitutes for NfL in sporadic ALS**. Trials such as the HEALEY ALS Platform Trial (NCT04297683) have incorporated NfL as a secondary/Pharmacodynamic biomarker alongside traditional clinical endpoints (ALSFRS-R, SVC, survival), but NfL has not been accepted as a standalone surrogate endpoint for sporadic ALS because: - NfL is nonspecific to ALS pathophysiology (elevated in other neurodegenerative diseases, trauma, and inflammatory conditions). - The causal link between an intervention in sporadic ALS and NfL change is far less direct than in SOD1, where reducing the mutant protein is the mechanism. - No sporadic ALS drug has yet demonstrated that NfL reduction predicts clinical benefit. --- ## (c) Emerging Targets in Early Development Despite setbacks with C9orf72 and ATXN2, the ASO/gene-therapy pipeline in ALS continues to expand, with several targets in early clinical or late preclinical stages: ### 1. FUS-ALS Fused in sarcoma (FUS) mutations cause a rare but aggressive form of ALS. Two programs are advancing: - **Jacifusen:** An investigator-initiated, open-label case series at five international sites (four US, one Switzerland) was published in *The Lancet* in May 2025. The ASO targets FUS pre-mRNA and showed potential to slow functional decline in a small first-in-human cohort. - **Ulefnersen (Ionis/Biogen):** The FUSION Phase 1–3 clinical trial is actively recruiting individuals with FUS-ALS internationally under the expanded access and clinical trial framework. This represents the first sponsor-initiated registrational program in FUS-ALS. ### 2. STMN2 (Stathmin-2) STMN2 is a TDP-43-regulated gene critical for neurite extension and neuromuscular junction maintenance. TDP-43 depletion causes the inclusion of a cryptic exon in STMN2 mRNA, leading to loss-of-function. Because TDP-43 pathology is present in ~97% of ALS, restoring STMN2 expression is viewed as a **broadly applicable strategy**. - A **first-in-human Phase 1/2 study** of an intrathecally delivered STMN2-targeting ASO commenced in January 2026 (64 patients, multiple-ascending dose design). - Preclinical data show that ASO correction of STMN2 misplicing restores protein levels independently of TDP-43 binding in humanized mouse models. STMN2 is particularly notable because it offers a potential bridge from monogenic (SOD1) to sporadic ALS therapeutics, though no validated surrogate biomarker for STMN2 rescue yet exists. ### 3. UNC13A Like STMN2, UNC13A contains cryptic exons regulated by TDP-43. Misplicing truncates the protein, impairing synaptic vesicle release. UNC13A variants are also genetic risk factors for ALS. ASO-mediated correction of UNC13A splicing is in preclinical development and could theoretically complement STMN2 strategies. ### 4. AAV-Based ATXN2 Approaches While BIIB105 (ASO) was discontinued, AAV-mediated RNAi targeting ataxin-2 has shown efficacy in TAR4/4 sporadic ALS mouse models (*Nature Communications*, October 2025). Preclinical work suggests that lower ATXN2 expression via AAV delivery may be more effective than ASO-based approaches, providing a rationale for continued development despite the ASO setback. ### Summary of Pipeline Landscape | Target | Therapy Type | Stage | ALS Subtype Applicability | |---|---|---|---| | SOD1 | ASO (tofersen) | Approved | SOD1-ALS only | | FUS | ASO (jacifusen, ulefnersen) | Phase 1–3 | FUS-ALS (~1% familial) | | STMN2 | ASO | Phase 1/2 (started Jan 2026) | Potentially ~97% (TDP-43+) | | UNC13A | ASO (preclinical) | Preclinical | Sporadic/genetic risk | | ATXN2 | AAV-RNAi | Preclinical | Risk-modifier, C9orf72 synergy | | C9orf72 | ASO | Discontinued ×2 | — | | ATXN2 | ASO (BIIB105) | Discontinued | — | --- ## Five Gaps That Must Close for the Surrogate-Endpoint Approach to Be Widely Generalized ### Gap 1: Mechanism-Specific Validation of the Target–Biomarker–Outcome Chain The tofersen precedent established a causal chain for SOD1-ALS: lower SOD1 protein → lower NfL → slower functional decline. This chain **does not transfer automatically** to other genetic targets. The failures of BIIB105 (ATXN2 lowering without NfL or clinical improvement) and WVE-004 (poly(GP) lowering without clinical benefit) demonstrate that target engagement alone is insufficient. For each new target, sponsors must independently validate that pharmacodynamic change (target modulation) correlates with NfL reduction, and that NfL reduction correlates with clinical benefit in that specific disease context. This requires more than biomarker correlation; it requires mechanistically grounded evidence. ### Gap 2: The TDP-43 Majority Requires Different Biomarkers or Targets Since the vast majority of ALS is TDP-43 proteinopathy, and no ASO can safely silence TDP-43 itself (loss-of-function is embryonic lethal in mouse models), emerging therapies target downstream consequences (STMN2, UNC13A). However, **there is no validated circulating biomarker for TDP-43 pathology**. NfL reflects generic neuronal injury, not TDP-43–specific pathophysiology. Whether restoring STMN2 expression will produce a predictable, NfL-measurable signal—and whether that signal correlates with clinical benefit—remains entirely unknown. The field urgently needs TDP-43–specific biomarkers (e.g., CSF TDP-43, TDP-43 seeding assays) or validated multibiomarker panels (NfL + MAP2 + NPTX2) to bridge this gap. ### Gap 3: Formal Biomarker Qualification Beyond Orphan Indications NfL has not been formally qualified by the FDA or EMA for any context-of-use in ALS. Tofersen's accelerated approval was possible in part because SOD1-ALS is a rare orphan disease, and the FDA's threshold for "reasonably likely to predict clinical benefit" is lower under accelerated approval. **Generalizing NfL as a surrogate endpoint to larger populations (sporadic ALS, platform trials) would require formal qualification studies**, including: - Large natural history databases linking baseline NfL to disease trajectory across genotypes. - Interventional data showing that NfL change predicts clinical outcome within a prespecified margin. - Harmonization of assay platforms (Quanterix Simoa, etc.) and reference ranges. - Demonstration that NfL reduction is not merely a pharmacodynamic readout but a validated surrogate for how patients feel, function, or survive. ### Gap 4: Trial Duration and Design Mismatch The VALOR trial did not meet its primary endpoint at 28 weeks on ALSFRS-R; clinical benefit emerged only in the OLE over years. Current ALS trial paradigms often rely on 6-month placebo-controlled phases (e.g., HEALEY platform trial regimens), which are likely too short for NfL-driven clinical benefit to manifest. If surrogate endpoints are to be used regulatorily, sponsors and regulators must agree on: - **Minimum duration** for surrogate-validity assessment. - **Adaptive/platform trial designs** that allow biomarker-driven dose escalation and early futility without losing statistical power. - **Natural history controls** as valid comparators when placebo-controlled arms become ethically or practically infeasible in ultra-rare subtypes. The regulatory precedent for SOD1 does not establish a design template that scales to sporadic ALS, where heterogeneity and placebo effects dilute endpoint signals further. ### Gap 5: Confirmatory Clinical Data Still Pending Tofersen itself remains under accelerated approval contingent on confirmatory evidence. The ATLAS presymptomatic trial is ongoing. The 148-week OLE data, while encouraging, **relied on natural history comparisons and lacked a concurrent placebo control**. Before the surrogate-endpoint model can be generalized, the field needs: - **Completed confirmation of tofersen's clinical benefit** in a randomized, placebo-controlled setting (ATLAS). - **At least one successful non-SOD1 ASO** that demonstrates both target engagement, NfL reduction, *and* clinical benefit, proving the model is not SOD1-specific. - **Post-marketing real-world outcome data** showing that tofersen's observed effects translate to routine clinical practice outside specialized trial centers. Without this confirmatory scaffolding, regulators are unlikely to extend accelerated approval based on NfL to other ALS targets, particularly in more common sporadic forms where a single surrogate biomarker would need to substitute for larger, longer efficacy trials. --- ## Conclusion Tofersen's post-approval data provide strong evidence that early, sustained SOD1 suppression slows functional decline and may extend survival in SOD1-ALS. However, the leap from this single-gene success to a generalizable surrogate-endpoint framework for ALS as a whole faces formidable obstacles. NfL is a valuable pharmacodynamic biomarker, but its surrogacy is mechanism- and genotype-dependent, not universal. The recent failures of C9orf72 and ATXN2 ASOs—despite target engagement—demonstrate that the SOD1-ALS template cannot be naively extrapolated. Closing the five gaps identified here will require deeper mechanistic biomarker research, formal regulatory qualification, validated TDP-43–specific biomarkers, redesigned trial architectures, and, most critically, at least one successful confirmatory validation outside the SOD1 context. --- ## References (Selected) - Smith JT, et al. Long-Term Tofersen in SOD1 Amyotrophic Lateral Sclerosis. *JAMA Neurology.* Published online December 2025 / February 2026. - Meyer T, et al. Clinical and patient-reported outcomes and neurofilament response during tofersen treatment in SOD1-related ALS—A multicenter observational study over 18 months. *Muscle & Nerve.* 2024. - Smith JT, et al. Tofersen treatment leads to sustained stabilization of disease in SOD1 ALS in a "real-world" setting. *Annals of Clinical and Translational Neurology.* 2025. - McEachin ZT, et al. Molecular impact of antisense oligonucleotide therapy in C9orf72-associated ALS. *Cell.* 2025. - Wave Life Sciences. Topline Results from Phase 1b/2a FOCUS-C9 Study of WVE-004. May 2023. - Ito T. Rethinking antisense oligonucleotide therapeutics for amyotrophic lateral sclerosis. *Annals of Clinical and Translational Neurology.* 2024. - Benatar M, et al. Design of a randomized, placebo-controlled, phase 3 trial of tofersen initiated in clinically presymptomatic SOD1 variant carriers: the ATLAS study. *Amyotrophic Lateral Sclerosis and Frontotemporal Degeneration.* 2023. - Lancet. Antisense oligonucleotide jacifusen for FUS-ALS: an investigator-initiated, multicentre, open-label case series. May 2025. - NeurologyLive. First-Ever Study of STATHMIN-2 Therapy in ALS Commences. January 2026. - AAV-based delivery of RNAi targeting ataxin-2 improves survival and pathology in TDP-43 mice. *Nature Communications.* 2025.