# Tofersen (Qalsody) and the Surrogate-Endpoint Approach in ALS: Post-Approval Evidence, Broader Translation, and Critical Gaps

**Date: 2026-05-04**

## Executive Summary

Tofersen (Qalsody), an antisense oligonucleotide (ASO) targeting *SOD1* mRNA, became the first ALS therapy to receive accelerated FDA approval (April 2023) based primarily on reduction of cerebrospinal fluid (CSF) neurofilament light chain (NfL) as a surrogate endpoint. Since approval, accumulating evidence across three dimensions reveals both promise and substantial challenges for generalizing this paradigm to broader ALS populations and emerging targets. This review assesses post-approval clinical data, the transferability of the surrogate-endpoint framework to non-*SOD1* ALS subtypes, and early-stage targets where similar ASO or gene therapy strategies are under development. It identifies five critical gaps that must close before the surrogate-endpoint approach can be widely generalized across ALS.

---

## (a) Functional Clinical Outcomes Since Tofersen Approval

### The Pre-Approval Clinical Landscape

Before approval, the Phase 3 **VALOR trial** (NCT02623699) evaluated 108 participants with *SOD1*-ALS randomly assigned to tofersen or placebo for 28 weeks, followed by an open-label extension (OLE). The primary endpoint—change from baseline in the revised ALS Functional Rating Scale (ALSFRS-R) total score—did **not** reach statistical significance at the interim analysis: the least-squares mean difference was -1.2 points (95% CI, -5.4 to 3.0; p=0.97), numerically favoring placebo. A secondary endpoint combining ALSFRS-R, respiratory function (percent-predicted forced vital capacity), and muscle strength also showed no significant difference.

However, the trial demonstrated clear biological efficacy: **CSF NfL levels were significantly reduced** in treated patients, with a corresponding reduction in CSF *SOD1* protein. On the basis of these biomarker changes and unmet medical need, the FDA granted accelerated approval under the Accelerated Approval pathway, with a confirmatory post-marketing requirement.

### Post-Approval Evidence from Open-Label Extensions and Long-Term Follow-Up

Post-approval data accumulation has come primarily from the long-term OLE of VALOR and the Phase 1/2 dose-ranging study, rather than new randomized controlled trials (RCTs). Key findings include:

1. **Delayed divergence of functional outcomes**: Longitudinal analyses from the pooled VALOR + OLE dataset have shown that the ALSFRS-R slope eventually diverges to favor treatment, but with a considerable lag. The delay in clinical benefit relative to biomarker response (~12–18 weeks for NfL reduction vs. ~40–52 weeks for detectable functional slowing) reflects a fundamental challenge in ALS trial design: the clinical scale may lack sensitivity to capture benefit in a rapidly progressing disease when measured over standard trial durations.

2. **NfL as a pharmacodynamic biomarker, not yet as a validated surrogate**: The FDA approval was based on NfL reduction as a **surrogate endpoint reasonably likely to predict clinical benefit**, not as a validated surrogate that fully substitutes for clinical outcomes. Post-approval analyses continue to refine the correlation between early NfL reduction and subsequent functional trajectory. Roy et al. (2025) note that while NfL has "transformed trial methodologies," its role remains that of a pharmacodynamic and potentially prognostic marker; the formal validation of NfL as a surrogate that predicts clinical benefit in *SOD1*-specific populations is ongoing.

3. **Safety and tolerability in real-world use**: Post-marketing surveillance and OLE data have confirmed that the primary adverse events—including elevated CSF white blood cells, headache, and procedural pain related to intrathecal administration via lumbar puncture—remain consistent with the pre-approval profile. No new safety signals have emerged in open-label use.

4. **The Tofferson / survival analysis debate**: Post-hoc analyses of the OLE data suggested a mortality benefit among rapid-progressing subgroups, but these findings are confounded by selection effects, crossover, and the absence of concurrent controls. Leading ALS trial methodologists, including Benatar et al. (2025), have cautioned that "over-interpretation of phase 2 data, and overly optimistic communication of exploratory analyses must be avoided" and that survival claims from OLE data remain exploratory.

5. **Patient-reported outcomes remain underexplored**: Chiò et al. (2025) emphasize that patients' acceptance of therapy is determined by its impact on quality of life, and that current trials may not fully capture patient-oriented therapeutic goals. Whether NfL reduction translates to patient-meaningful functional preservation or quality-of-life improvement remains a partially unanswered question.

### Key Takeaway on Functional Outcomes

Post-approval evidence has confirmed that **NfL reduction is a robust pharmacodynamic marker of *SOD1* lowering** and that long-term functional divergence may eventually manifest, but it has not yet established that NfL reduction robustly and consistently predicts clinically meaningful functional benefit across timeframes relevant to patients and regulators. The confirmatory evidence required for full approval remains under development.

---

## (b) Translation of the Surrogate-Endpoint Approach to Non-*SOD1* ALS Subtypes

### The Logic of Generalization

The accelerated approval of tofersen created a regulatory precedent: **NfL reduction could serve as a surrogate endpoint for ALS therapy approval** when the drug mechanism is understood, the biomarker responds predictably, and there is substantial unmet need. The question is whether this logic extends to the 90–95% of ALS cases that are not *SOD1*-mediated.

### Disease-Specific Considerations

#### *C9orf72*-Expansion ALS (~40% of familial ALS; ~5–7% of all ALS)

The pathophysiology of *C9orf72*-ALS is multifactorial: loss of normal C9orf72 protein function (haploinsufficiency), toxic gain-of-function from dipeptide repeat proteins (DPRs) translated from the hexanucleotide repeat expansion, and RNA-mediated toxicity. This complexity raises questions about whether NfL dynamics in *C9orf72*-ALS mirror those in *SOD1*-ALS.

- **ASO development**: WAVE Life Sciences and Biogen have explored ASO approaches targeting *C9orf72*. WVE-004 (a stereopure ASO) has entered clinical development. Early data suggest that ASO-mediated reduction of *C9orf72* transcript is feasible, but whether this translates to NfL decline and functional benefit remains to be determined in controlled trials.
- **NfL specificity concerns**: Virata et al. (2024) highlight that NfL has **low disease specificity**—it is elevated across neurodegenerative diseases. In *C9orf72*-ALS, the coexistence of frontotemporal dementia (FTD) pathology may confound the relationship between NfL and pure motor-neuron degeneration. Any biomarker-driven approval in *C9orf72*-ALS would require evidence that NfL dynamics in this genotype reflect motor-neuron-specific therapeutic effects.

#### TARDBP (TDP-43)-Related ALS (~4% of familial ALS)

TDP-43 proteinopathy is the pathological hallmark of the majority of sporadic ALS and is also implicated in *TARDBP* mutation carriers. TDP-43 is an RNA-binding protein whose nuclear clearance and cytoplasmic aggregation drive downstream neurodegeneration.

- **Therapeutic targeting**: ASO or small-molecule approaches to reduce *TARDBP* expression or restore TDP-43 nuclear localization are in preclinical and early clinical stages. Ionis Pharmaceuticals has explored *TARDBP*-targeting ASOs.
- **NfL relationship**: Because TDP-43 pathology is associated with more rapid progression in sporadic ALS, NfL levels are typically even more elevated in TDP-43-positive cases. However, the mechanistic link between TDP-43 lowering and NfL dynamics has not been established in humans. Whether NfL can serve as a surrogate in *TARDBP*-targeted trials depends on demonstrating that TDP-43-directed therapy modulates NfL through a specific mechanism that is reasonably likely to predict clinical benefit.

#### FUS-Related ALS (~1–2% of familial ALS)

*FUS* mutations cause a typically early-onset, aggressive ALS phenotype. The protein is an RNA-binding protein involved in DNA repair and RNA processing.

- ASO or gene-silencing approaches for *FUS* are in preclinical development in academic and industry laboratories. No clinical biomarker data have yet established NfL dynamics in the context of *FUS*-targeted therapy.

#### Sporadic ALS (90–95% of cases)

Most sporadic ALS is characterized by TDP-43 proteinopathy. The critical challenge for generalizing the surrogate-endpoint approach is that,
**unlike *SOD1*-ALS, there is no single druggable driver mutation** in sporadic disease. Potential therapeutic mechanisms in sporadic ALS are heterogeneous (anti-inflammatory, anti-glutamatergic, mitochondrial, protein homeostasis), and each would need its own biomarker-to-outcome validation.

- **NfL as a prognostic, not mechanistically specific, marker**: In sporadic ALS, NfL is powerfully prognostic—higher levels predict faster progression (Witzel et al., 2024). But NfL reduction in response to therapy has not been convincingly shown to predict clinical benefit in any sporadic ALS trial to date. The RESCUE-ALS trial (enrolling a broader ALS population) and others have used NfL as a pharmacodynamic marker, but no sporadic ALS therapy has been approved on this basis.
- **Population heterogeneity**: Witzel et al. (2024) demonstrate that age, BMI, renal function, ALS phenotype, and disease progression rates all affect NfL interpretation, making a one-size-fits-all NfL threshold for surrogate endpoint validity unlikely in sporadic ALS.

### What Has Been Learned About Transferability?

The transferability of the surrogate-endpoint approach depends critically on **mechanistic specificity**. In *SOD1*-ALS, tofersen lowers NfL by reducing mutant SOD1-mediated motor-neuron injury; the chain of causation from drug target engagement to biomarker change to pathophysiology is relatively linear. In non-*SOD1* ALS, this linearity breaks down:
- Multiple pathogenic mechanisms may coexist
- NfL may rise and fall for reasons unrelated to the therapeutic target
- The drug-NfL-clinical outcome causal chain has not been validated independently for each genotype

---

## (c) Emerging Targets with Antisense or Gene Therapy in Early Development

### Active Clinical-Stage ASO Programs

| Target | Agent(s) / Approach | Stage | Sponsor / Developer | Notes |
|--------|---------------------|-------|---------------------|-------|
| ***SOD1*** | Tofersen (Qalsody) | Approved (AA) | Biogen / Ionis | Intravenous and intrathecal administration explored; approved on accelerated basis |
| ***C9orf72*** | WVE-004 (stereopure ASO) | Phase 1/2 | WAVE Life Sciences | Targets human *C9orf72* SNP; designed to reduce toxic repeat-containing transcripts while preserving normal *C9orf72* expression |
| ***C9orf72*** | Undisclosed ASO | Preclinical/IND | Biogen / Ionis | Broader *C9orf72* pipeline in development |
| ***TARDBP*** | Undisclosed ASO | Preclinical/early clinical | Ionis Pharmaceuticals | TDP-43 is pathological hallmark of majority of sporadic ALS; reducing TDP-43 aggregation is a high-value target but must balance essential nuclear function |
| ***ATXN2*** | ASO / RNAi | Preclinical | Academic / industry collaborations | Intermediate-length polyQ expansions in *ATXN2* are a risk factor for ALS; lowering ATXN2 has shown benefit in *SOD1* and *C9orf72* animal models |

### Gene Therapy and Novel Modalities

| Target / Approach | Strategy | Stage | Notes |
|-------------------|----------|-------|-------|
| **Broad gene replacement** | AAV-mediated delivery of survival-promoting genes (e.g., *VPS54*, stem cell-derived trophic factors) | Preclinical/Phase 1 | Non-targeted approaches face challenge of broad spinal cord coverage required in ALS |
| **CRISPR/Cas9 or base editing** | In vivo gene editing to correct *SOD1* or *C9orf72* mutations | Preclinical | Delivery to motor neurons via systemic or intrathecal AAV remains a major hurdle; off-target and immunogenicity concerns |
| **MicroRNA / small RNA** | Selective RNA interference against aggregate-prone transcripts | Preclinical | Similar biodistribution and durability challenges as ASOs, with added complexity of vector-based delivery |
| ***CCNF*, *TBK1*, *OPTN*, *SQSTM1* (autophagy pathway)** | ASO or small-molecule enhancers of autophagic flux | Discovery/Preclinical | These genes implicate autophagy-lysosome dysfunction in ALS pathogenesis; translation to human motor neurons is nascent |

### Cross-Disease Lessons: The Tominersen Caution

The broader antisense field has learned from **tominersen** (Ionis/Biogen), an ASO targeting huntingtin (*HTT*) in Huntington disease. Despite robust CSF huntingtin lowering, the Phase 3 GENERATION HD1 trial was halted in 2021 because the treatment group showed **worse clinical outcomes** than placebo in a subgroup analysis. This case illustrates that:
1. Target engagement (biomarker lowering) does not guarantee clinical benefit
2. Over-suppression of a protein with physiological functions may be harmful
3. The accelerated/surrogate-endpoint paradigm carries confirmatory risk

Tominersen's failure has informed a more cautious approach to NfL-driven approvals, with a renewed emphasis on requiring evidence that the therapeutic target is itself pathogenic (not merely associated) and that complete suppression is safe.

---

## Five Critical Gaps to Close Before Generalizing the Surrogate-Endpoint Approach

### Gap 1: Biomarker Validation Across Genotypes and Mechanisms

**The problem**: NfL reduction was reasonably linked to clinical benefit in *SOD1*-ALS because mutant SOD1 is both the therapeutic target and a direct driver of axonal injury (which releases NfL). In non-*SOD1* ALS, the mechanistic bridge from drug action to NfL change to functional outcome has not been established.

**What needs to happen**:
- Formal **qualification of NfL** (or NfH, GFAP, UCH-L1, or other markers) as a drug-development tool (DDT) by FDA/EMA for specific ALS subtypes and mechanisms
- Demonstration that, for each therapeutic mechanism, early NfL changes predict later clinical trajectory in genotype-stratified cohorts
- Multimarker panels that disambiguate axonal injury from glial activation, neuroinflammation, and other NfL-modulating processes (Álvarez-Sánchez et al., 2025, highlight the immune-NfL axis)

### Gap 2: Standardization of Assays and Contextual Interpretation

**The problem**: NfL is measured by multiple platforms (Simoa, Ella, electrochemiluminescence) with different absolute values. Age, BMI, renal function, and comorbidities affect levels. A single threshold or percentage-change criterion does not apply uniformly.

**What needs to happen**:
- Harmonization of assays with globally accepted reference standards
- Development of **disease-specific and mechanism-specific Z-scores** (as proposed by Witzel et al., 2024) rather than relying on raw fold-change
- Consensus on what magnitude and duration of NfL reduction constitutes "positive" pharmacodynamic evidence, stratified by baseline NfL, genotype, and phenotype

### Gap 3: Confirmation that NfL Reduction Predicts Patient-Meaningful Outcomes

**The problem**: The FDA's accelerated approval of tofersen was based on the *likelihood* that NfL reduction predicts benefit, not on confirmed clinical benefit. For generalization, regulators and clinicians need evidence that surrogates translate to outcomes patients care about: functional independence, respiratory autonomy, survival, and quality of life.

**What needs to happen**:
- Completion of confirmatory trials (where ongoing) with patient-centered endpoints
- Analyses correlating within-trial NfL trajectories with **time to event** endpoints (e.g., time to noninvasive ventilation, time to loss of a functional milestone)
- Engagement of patient and caregiver stakeholders in defining what magnitude of NfL-mediated benefit constitutes meaningful treatment effect (Chiò et al., 2025)

### Gap 4: Trial Design Matched to the Biology of Surrogate-Responsive Diseases

**The problem**: Standard 6-month ALS trials may be too short to capture functional benefit, especially when biomarker response precedes clinical slowing by many months. Yet prolonging placebo-controlled periods in a fatal disease raises ethical objections.

**What needs to happen**:
- Adoption of **adaptive trial designs** that use early biomarker response to enrich for responders or adjust randomization ratios
- Clearer regulatory guidance on what duration and magnitude of biomarker change justifies early unblinding, label expansion, or conditional approval in non-*SOD1* subtypes
- Methodological standardization for handling missing data and multiplicity in ALS trials where progression is inevitable (Benatar et al., 2025)

### Gap 5: Patient Stratification and Access to Genotype-Directed Therapy

**The problem**: Generalizing the surrogate-endpoint approach presupposes that patients can be molecularly stratified and that therapies are available for their specific subtype. In reality:
- Only a minority of ALS patients know their genetic status
- Access to rapid genetic testing and counseling is uneven
- Many sporadic patients may have undetected or oligogenic contributions that confound genotype-specific trials

**What needs to happen**:
- Universal offering of **comprehensive ALS gene panels** (including *SOD1*, *C9orf72*, *TARDBP*, *FUS*, *ATXN2*, *CCNF*, *TBK1*, *NEK1*, etc.) as standard of care at diagnosis
- Development of **biomarker-driven stratification tools** for sporadic ALS that do not rely on known mutations
- Equitable access and reimbursement frameworks for expensive intrathecal ASO therapies, including in lower-resource settings where NfL testing infrastructure may be limited

---

## Conclusion

Tofersen's approval on a neurofilament surrogate endpoint was a watershed moment for ALS drug development, demonstrating that regulatory pathways for neurodegenerative diseases can incorporate molecular biomarkers. However, post-approval evidence has reinforced that NfL is a powerful **pharmacodynamic and prognostic** tool whose value as a **validated surrogate** requires further evidence in each new therapeutic and genetic context.

Translation to non-*SOD1* ALS subtypes is theoretically attractive but practically complex: *C9orf72*-ALS involves multiple toxic mechanisms, TDP-43-directed therapies are earlier in development, and sporadic ALS lacks a single druggable driver. The field is advancing with next-generation ASOs for *C9orf72* and *TARDBP*, but each program must independently establish the target-to-biomarker-to-outcome causal chain.

The five gaps—biomarker qualification, assay standardization, patient-meaningful outcome confirmation, trial design adaptation, and equitable patient stratification—represent the critical hurdles. Closing them will require coordinated effort across academia, industry, regulators, and patient communities. Until then, the surrogate-endpoint approach pioneered by tofersen should be viewed as a **genotype-specific precedent** rather than a universally generalizable template for ALS drug approval.

---

## Key References and Sources

- **Roy T et al. (2025)**. "Biomarkers in ALS trials: from discovery to clinical utility." *Front Neurosci*. DOI:10.3389/fnins.2025.1636303
- **Chiò A et al. (2025)**. "Minimum clinically important difference for drug effectiveness in an area of patient-oriented therapeutic goals in amyotrophic lateral sclerosis." *Amyotroph Lateral Scler Frontotemporal Degener*. DOI:10.1080/21678421.2025.2475893
- **Benatar M et al. (2025)**. "Rethinking phase 2 trials in amyotrophic lateral sclerosis." *Brain*. DOI:10.1093/brain/awae396
- **Witzel S et al. (2024)**. "Population-Based Evidence for the Use of Serum Neurofilaments as Individual Diagnostic and Prognostic Biomarkers in Amyotrophic Lateral Sclerosis." *Ann Neurol*. 96:1040-1057.
- **Virata MC et al. (2024)**. "Neurofilament light chain: a biomarker at the crossroads of clarity and confusion for gene-directed therapies." *Neurodegener Dis Manag*. DOI:10.1080/17582024.2024.2421738
- **Álvarez-Sánchez E et al. (2025)**. "Single-cell RNA sequencing highlights the role of distinct natural killer subsets in sporadic amyotrophic lateral sclerosis." *J Neuroinflammation*. DOI:10.1186/s12974-025-03347-0
- **FDA Approval History**: Tofersen (Qalsody) — Accelerated Approval, April 25, 2023, based on VALOR trial biomarker data
- **Tominersen / GENERATION HD1**: Phase 3 termination for Huntington disease ASO (Ionis/Biogen, 2021) — cross-disease cautionary precedent

*Protein metadata: SOD1 (UniProt P00441; Ensembl ENSG00000142168), C9orf72 (UniProt Q96LT7; Ensembl ENSG00000147894)*