Scenario wins: laertes (2) Mantic (1) Panshul42 (1) SynapseSeer (1)
Question: will Biogen publicly release CELIA (BIIB080) topline by 12 Aug 2026, and if so, will the first statement be a clear hit, a clear miss, or a miss paired with explicit positive clinical‑trend language? Forecasters converge that a pre‑deadline topline is plausible but not certain (roughly 60–68% chance) based on Biogen’s Q3‑2026 readout guidance, ClinicalTrials.gov timing, and natural disclosure windows (AAIC mid‑July and likely late‑July/early‑Aug earnings). A clean primary‑endpoint success is considered unlikely (~10–15% unconditional), while the most likely disclosed outcome if a topline appears is “missed primary but affirmative/favorable clinical trends” (largest conditional bucket; about 0.28 unconditional in one combined estimate), with a plain miss (~0.25 unconditional) next and truly ambiguous first wording rare (~5–10%). Key reasons: CELIA’s primary dose‑response design plus secondary arm tests create a realistic path for mixed messaging, Biogen already has public language templates emphasizing clinical trends and biomarkers, and historical precedent (including Biogen’s own tofersen disclosure) shows companies will often pair “did not meet primary” with trend language. Forecasters recommend monitoring Biogen investor slides, the Apr.‑29 Q1 materials, AAIC abstracts, and the exact first official document wording (press release vs. deck) because biomarker‑only claims won’t count as clinical positivity under the market rules.
Forecast:
Key driver is whether Biogen issues any official “top-line” CELIA disclosure by Aug 12, 2026 (London time). Multiple analysts cite (and at least one directly links) ClinicalTrials.gov listing an estimated primary completion date of May 14, 2026 for the Week-76 primary endpoint timepoint, and Biogen guidance has been framed as a Q2/Q3 2026 or “mid-2026” readout. That makes a pre–Aug 12 disclosure more likely than not, but the window is tight: Aug 12 is ~3 months after May 14, and multicenter AD trials often have nontrivial database lock/analysis lag, plus companies sometimes hold readouts for a later investor event/conference (especially if results are not cleanly positive). I therefore keep a material probability that no qualifying topline statement is made by the deadline.
On content: Alzheimer’s Phase 2 clinical endpoints have low-to-middling success base rates, and tau-targeting clinical success has historically been difficult. BIIB080’s earlier biomarker/pharmacology signals plausibly raise the chance of a “win” versus a generic AD Phase 2, but the stated primary endpoint is a dose–response effect on CDR-SB (a fairly high bar and potentially brittle to mixed dose patterns). If the primary endpoint is missed, Biogen may still highlight any favorable numerical directional effects and describe them as “encouraging/favorable trends,” but the resolution criteria require that Biogen explicitly frame these as clinical trends (not merely biomarker changes). I assign the largest mass to “missed endpoint but positive clinical trend” as the most common corporate framing when data are mixed, while still leaving substantial probability on a clear miss with no claimed positive clinical trend.
I keep the “ambiguous” communication bucket low but not negligible: large-cap companies usually state endpoint met/not met in the first topline disclosure, yet dose-ranging designs can yield messy narratives (e.g., disclosure of some results while deferring a binary primary-endpoint conclusion), which could produce an Option 4 resolution under the rules.
An amateur forecast might (a) anchor too hard on Biogen’s prior “mid-2026/Q2–Q3” guidance and set “no release” near zero, or (b) anchor too hard on Alzheimer’s failure history and over-allocate to a clear miss with no positive trend. My forecast instead (i) explicitly prices the tightness of the May→Aug operational window and the incentive to delay messy/negative readouts, while (ii) recognizing that if Biogen does disclose by the deadline, the disclosed outcome is somewhat selection-biased toward cleaner/more favorable narratives, and (iii) distinguishes “clinical trend” language from biomarker-only spin per the resolution criteria. Net effect: less extreme than the most optimistic/pessimistic analyst takes, closer to an error-minimizing ensemble with a modest adjustment toward delay risk.
Run 1: The main reasoning across the forecasts is that timing uncertainty and weak historical odds of success in Alzheimer’s Phase 2 trials dominate the outlook.
The collective view is that the most likely paths are either a delayed/no public top-line statement by August 12 or a result that does not clearly meet the primary endpoint but contains enough numerical signal for Biogen to present it favorably. A straightforward primary-endpoint win is treated as the least likely of the main scenarios.
Run 2: The forecasts converge on two main drivers: timing risk and likely interpretation of the clinical result.
Release timing is the biggest uncertainty.
The trial’s estimated primary completion date in mid-May 2026 leaves only about three months before the August 12 deadline. That is enough time for a top-line release in principle, but several forecasts stress that database lock, analysis, and corporate timing choices could easily delay disclosure past the cutoff.
If results are released, the most likely framing is a mixed/qualified one.
All forecasts treat a clean, unambiguous success as less likely than a result showing no primary endpoint win but some positive clinical trend. This reflects both the historically low success rate in Alzheimer’s Phase 2 studies and Biogen’s apparent willingness to emphasize favorable biomarker or trend signals even when the main endpoint is not clearly met.
Precedent from BIIB080 matters.
Prior reporting on BIIB080 included language about “favorable trends” alongside biomarker effects, which makes a similar communication style plausible for CELIA if the data are not definitively positive.
The collective view is that the dominant uncertainty is whether Biogen says anything by August 12, 2026 at all. If it does, the most probable first public statement is that CELIA did not meet its primary endpoint but showed a positive clinical trend.
Run 3: Overall, the forecasts converge on a few core drivers:
Timing is tight but not certain. The trial’s primary completion date in May 2026 leaves only a short window before August 12. Most reasoning assumes top-line data could arrive in time, but there is meaningful risk that database lock, cleaning, and internal review slip past the cutoff.
A clear statistical win is considered hard to achieve. As a Phase 2 Alzheimer’s trial using a clinical endpoint like CDR-SB, CELIA is viewed as a challenging setting, especially given the historically poor success rate of tau-targeting programs. That keeps the probability of Biogen announcing a clean primary-endpoint success relatively modest.
If the endpoint is missed, the likely framing is still constructive. Several rationales emphasize Biogen’s tendency to highlight “positive trends” or favorable numerical signals when results are mixed, especially if biomarker or subgroup data look encouraging. The prior BIIB080 signal package and novel mechanism support this possibility.
An entirely flat or negative readout is also plausible. The other side of the debate is that tau lowering may simply not translate into meaningful clinical benefit, in which case Biogen could say the trial missed the endpoint without pointing to a strong trend.
An ambiguous disclosure is less favored. Most reasoning expects that if top-line data are released, Biogen will state clearly whether the primary endpoint was met, even if the accompanying narrative is cautious or biomarker-heavy.
In short, the collective reasoning centers on a difficult Alzheimer’s clinical environment, a constrained disclosure timeline, and Biogen’s likely preference to emphasize any encouraging signal if the primary endpoint is not definitively achieved.
Question: What will Biogen’s first public top-line statement about the Phase 2 CELIA trial be by August 12, 2026? Final Prediction:
Total Cost: extra_metadata_in_explanation is disabled Time Spent: extra_metadata_in_explanation is disabled LLMs: extra_metadata_in_explanation is disabled Bot Name: extra_metadata_in_explanation is disabled
Forecaster 1:
Forecaster 2:
Forecaster 3:
Forecaster 4:
Forecaster 5:
The research provides an overview of Biogen’s Phase 2 CELIA trial for the drug BIIB080, an antisense oligonucleotide aimed at targeting tau protein in early Alzheimer’s disease. The trial involves 416 participants across 138 global sites, with a primary endpoint focused on assessing dose-response changes from baseline to Week 76 using the Clinical Dementia Rating-Sum of Boxes (CDR-SB). The estimated primary completion date is May 14, 2026, and Biogen has indicated that data is expected in mid-2026, with a plausible top-line readout occurring between July and August 2026. The research highlights both favorable signals, such as encouraging Phase 1b results and FDA Fast Track designation, as well as cautionary factors, including the historical low success rate of Alzheimer’s trials and the potential challenges of the trial’s dose-response endpoint.
The framing of the trial results by Biogen is critical, as the company has positioned CELIA as a “proof-of-concept study” rather than one strictly driven by statistical significance. This suggests that even if the primary endpoint is not met, Biogen may emphasize positive trends or biomarker effects in their communications. The research outlines various potential outcomes for the trial results, ranging from a clean positive result to ambiguous disclosures, while also noting the uncertainties surrounding data timing and the statistical framework used for analysis.
Sources used:
CELIA is Biogen’s global Phase 2 trial of BIIB080, an antisense oligonucleotide (ASO) targeting tau protein via MAPT mRNA, in early Alzheimer’s disease. The study enrolled 416 participants across 138 global sites, with a primary endpoint of dose-response in change from baseline to Week 76 on CDR-SB [10][11].
Assessment on Resolution Option 5 (No data by August 12, 2026): Given the May 14 primary completion date and mid-2026 guidance, it is likely but not certain that top-line results will be available before the August 12 deadline. There is some risk of delay (data quality issues, analysis complexity), but the timeline is well within range.
Favorable signals:
Cautionary signals:
The most critical insight for forecasting is Biogen’s own framing. Diana Gallagher at the March 2026 Biogen Conference explicitly stated CELIA is “designed as a proof-of-concept study rather than a p-value-driven decision” [3]. This strongly suggests:
No prediction markets were found on Metaculus, Polymarket, or other platforms specifically addressing CELIA trial outcomes or this particular question. This means there are no external crowd-sourced probability benchmarks available for calibration.
| Option | Description | Key Considerations |
|---|---|---|
| 1. Met primary endpoint | Clean positive | Low base rate for AD trials; dose-response endpoint is challenging; 416 participants may be underpowered. Phase 1b trends were encouraging but Phase 2 is a bigger test. |
| 2. Did not meet, no positive trend | Clean negative | Possible if results show no signal at all, but given Phase 1b biomarker and trend data, a complete absence of any positive signal would be somewhat surprising. |
| 3. Did not meet, but positive trend | Mixed result | Arguably the most consistent with Biogen’s own “proof-of-concept, not p-value driven” framing. If there’s a directional signal but not formal statistical significance, Biogen appears positioned to emphasize encouraging trends. |
| 4. Data released but ambiguous | Unclear framing | Possible given the proof-of-concept framing, but pharma companies usually make some directional statement. |
| 5. No data by August 12 | Timing miss | Less likely given May 14 primary completion and “mid-2026” guidance, but cannot be ruled out if analysis is delayed. |
Recommendation: The April 29 Biogen earnings call may provide critical updated guidance on CELIA timing. Monitor closely.
Let me check for any recent updates since the question opened:
The search confirms no new CELIA-specific updates since the background research. Let me now provide my final updated analysis.
The Round 2 critic makes an excellent point about resolution criteria specificity that I hadn’t fully appreciated:
Option 3 requires explicit “did not meet” language: The resolution states Biogen must “clearly state” the primary endpoint was not met while also describing positive clinical trends. Many companies avoid explicit negative language when highlighting positive elements.
Pharmaceutical communication patterns: For dose-response trials with mixed results, companies often:
This insight increases the probability of Option 4 (ambiguous disclosure) relative to Option 3 (explicitly stating endpoint failure with positive trend).
My assessment of timing (Option 5) and the overall probability ordering remains sound. The May 14, 2026 primary completion with consistent “mid-2026” guidance suggests ~79% probability of data release before August 12.
The search confirmed no material updates since April 22. The key framing consideration remains Biogen’s explicit “proof-of-concept” positioning, which now cuts both ways:
Taking the Round 2 critic’s refinements:
Option 1 (Met endpoint): Maintain at 14% — Phase 1b signals and FDA Fast Track justify this level
Option 2 (Failed, no trend): Reduce to 15% — with robust biomarker engagement, complete absence of any positive framing is unlikely
Option 3 (Failed, positive trend): Reduce to 33% — requires explicit “did not meet” language, which companies often avoid
Option 4 (Ambiguous): Increase to 17% — the resolution criteria specificity makes this more likely for nuanced dose-response results
Option 5 (No data): Maintain at 21% — timeline is tight but feasible
Updated prediction: Biogen states CELIA met its primary endpoint: 14% Biogen states CELIA did not meet its primary endpoint and no positive trend: 15% Biogen states CELIA did not meet its primary endpoint but showed a positive clinical trend: 33% Biogen releases top-line CELIA data by August 12, 2026 but does not clearly state whether the primary endpoint was met: 17% No top-line CELIA result is publicly released by August 12, 2026: 21%
1. Over-interpretation of “proof-of-concept” framing: The critic is right that this language is fairly standard pharmaceutical risk management. Many companies position Phase 2 trials this way for flexibility, and I may have over-weighted this as a signal of expected failure.
2. Crude base rate application: The ~99% AD Phase 2 failure rate includes many trials with weak preclinical rationale. BIIB080 has demonstrated 50-60% tau reduction and directional cognitive trends in Phase 1b—this does arguably place it in a more favorable reference class.
3. Logical inconsistency on failure distribution: The critic correctly identifies tension in my reasoning. If failures are the norm, why would “failure with positive trend” (Option 3) be 3.5x more likely than “failure without positive trend” (Option 2)? Not all failures produce positive trends to highlight.
4. Resolution criteria specificity for Option 3: The explicit requirement for “positive,” “favorable,” or “encouraging” language about clinical efficacy measures is a higher bar than I initially weighted. Biogen cannot claim such trends if the data doesn’t support them.
Timeline assessment remains sound: With May 14 primary completion and “mid-2026” guidance, the probability of meeting the August 12 deadline is high. 14-15% for Option 5 appropriately captures delay risk without over-weighting it.
Phase 1b biomarker signals suggest some biological effect: This still makes a complete absence of any signal (Option 2) somewhat less likely than a scenario where there’s at least directional movement worth discussing.
I’m adjusting my forecast to:
The critic’s suggested adjustments are directionally sound, though I’ll maintain slightly more weight on Option 3 than they suggest (44% vs 42%) because I still believe Biogen’s framing strategy indicates preparation for a nuanced “proof-of-concept success” narrative if any directional signal exists.
Updated prediction: Option_A: 12% Option_B: 15% Option_C: 44% Option_D: 16% Option_E: 13%
Looking at the critic’s Round 2 points, I need to carefully evaluate whether further adjustments are warranted.
1. Option 3 vs. Option 4 Balance: The critic’s persistent argument about this balance has merit. The resolution criteria for Option 3 require a very specific linguistic combination: Biogen must BOTH explicitly state “did not meet” AND claim positive clinical trends. Given the “proof-of-concept, not p-value driven” framing, language that sidesteps binary “met/not met” entirely becomes genuinely plausible.
2. Biomarker-to-Clinical Translation: The critic correctly notes that Phase 1b biomarker effects (50-60% tau reduction) don’t reliably predict Phase 2 clinical success in neurodegeneration. This is well-documented and supports keeping Option 2 in a meaningful range.
3. Dose-Response Complexity: The endpoint’s analytical complexity does create more interpretive flexibility, which could facilitate ambiguous communication.
SEC Disclosure Pressure: The critic’s own point about SEC requirements creates tension with the ambiguity thesis. Investor relations and regulatory norms typically force pharmaceutical companies to provide some statement about primary endpoints. This prevents Option 4 from overtaking Option 3 entirely.
Communication Incentives: If Biogen wants to continue development despite not meeting the formal endpoint, they have strong incentive to acknowledge the primary endpoint result while emphasizing positive signals - this is the Option 3 pathway. Complete ambiguity (Option 4) might invite more investor skepticism than a clear “didn’t technically meet, but encouraging signals” framing.
I’m accepting the critic’s argument to bring Options 3 and 4 closer to parity while maintaining a slight edge for Option 3 based on regulatory/investor disclosure norms. I’m also modestly increasing Option 2 to reflect the biomarker-to-clinical translation gap.
The “proof-of-concept” counterfactual point is well-taken - management’s framing does suggest they don’t expect statistical success but want optionality. This supports the heavier weighting toward mixed/ambiguous outcomes (Options 3+4) relative to clean success.
Updated prediction: Option_A: 9% Option_B: 15% Option_C: 31% Option_D: 29% Option_E: 16%
The critic raises several valid concerns that warrant adjustments to my forecast.
Points I Accept:
Over-reliance on “proof-of-concept” framing: The critic is right that this could be standard pharmaceutical hedging rather than insider signaling. Companies routinely manage expectations downward. I should weight this less heavily as a predictive signal.
Underweighting complete failure probability: The ~99% historical failure rate for AD Phase 2 trials is sobering. When AD trials fail, they often fail completely with no clinically meaningful signal. The J&J posdinemab failure (November 2025) is a relevant recent data point, even with the ASO mechanism difference.
Timing is probably tighter: With May 14 primary completion and consistent “mid-2026” guidance, the probability of missing the August 12 deadline is likely lower than my initial 18%. Biogen has strong investor expectations and limited wiggle room.
Points I Maintain:
Corporate incentives matter significantly: Biogen desperately needs pipeline wins after the Aduhelm controversy. If ANY positive signal exists, they will emphasize it. This should support Option 3 over Option 2, though perhaps not as strongly as I initially weighted.
Phase 1b data is meaningful: 50-60% CSF tau reductions with trends toward slower cognitive decline in higher-dose groups isn’t nothing—it suggests biological activity that MIGHT translate to clinical benefit.
New Considerations:
The critic’s point about dose-response statistical complexity is particularly insightful. If there’s significance at the highest dose but not across all doses, Biogen might face genuine ambiguity in characterizing results. This increases Option 4 probability.
The potential dissociation between biomarker and clinical outcomes is also worth more weight—strong tau reduction without clinical benefit would create an unusual disclosure pattern.
Revised Assessment:
I’m adjusting:
Updated prediction: Biogen states CELIA met its primary endpoint: 10% Biogen states CELIA did not meet its primary endpoint and no positive trend: 28% Biogen states CELIA did not meet its primary endpoint but showed a positive clinical trend: 32% Biogen releases top-line CELIA data by August 12, 2026 but does not clearly state whether the primary endpoint was met: 14% No top-line CELIA result is publicly released by August 12, 2026: 16%
The critic raises several compelling points that warrant adjustment to my forecast.
This genuine risk deserves more weight. I’m reducing Option 3 from 42% to 40%.
The “proof-of-concept” framing could cut both ways: The critic’s insight is valuable—if Biogen positions this as exploratory/learning, they may feel comfortable simply reporting “the dose-response endpoint was not met” without promotional spin. This could increase Option 2 probability. Adjusting from 18% to 19%.
Dose-response complexity favoring Option 4: Multiple scenarios don’t fit cleanly into resolution categories—monotonic but non-significant trends, only highest dose showing benefit, non-monotonic responses. This genuinely increases ambiguity risk. I’ll maintain 16% for Option 4 (already increased from 12% last round).
Option 3 still deserves to be modal: Despite language specificity concerns, Biogen has already used “favorable trends” language for clinical outcomes in peer-reviewed Phase 1b publications. The pharmaceutical industry strongly incentivizes positive spin, and Biogen’s proof-of-concept positioning suggests they’re preparing to emphasize any directional signal. If there’s ANY clinical signal, they will try to frame it positively.
Option 5 at 11% is well-calibrated: With 13 weeks between primary completion (May 14) and deadline (August 12), plus consistent “mid-2026” guidance and AAIC as a natural venue, the timeline is comfortable.
Option 1 at 14% balances base rates against differentiation: While historical AD Phase 2 failure rates are ~99%, BIIB080 has demonstrated biological proof-of-mechanism (tau PET reduction) that no previous tau therapy achieved. The Fast Track designation and strong Phase 1b trends justify modest optimism above strict base rates.
The critic’s point about J&J’s posdinemab complete failure (November 2025) is relevant for Option 2 probability. If the broader tau-targeting field is struggling, even Biogen’s mechanistically differentiated approach might show weak enough results that they don’t attempt to spin them as positive trends. This supports the slight increase to Option 2.
The critic’s overall direction of adjustment (more weight to Options 2 and 4, less to Option 3) is sound given the strict resolution criteria. My adjustments are slightly more conservative than the critic’s suggestions, as I still believe Biogen’s established “favorable trends” language patterns and pharmaceutical industry norms favor Option 3 as modal outcome.
Updated prediction: Option_A: 14% Option_B: 19% Option_C: 40% Option_D: 16% Option_E: 11%
Agent Forecast (multiple_choice, confidence: low-medium)
Base Rate: Recent anti-tau antibody Phase 2 trials (Roche semorinemab x2, UCB bepranemab, J&J posdinemab, 2020-2025) have achieved 0% primary endpoint success. BIIB080’s ASO mechanism is differentiated but unproven clinically. Broader AD Phase 2 primary success rate across mechanisms is roughly 10-15%. A 16% conditional probability for meeting primary sits slightly above that band to credit mechanistic differentiation and dose-response trial design.
Initial Reasoning:
Key Uncertainties:
Evidence Sources (7):
Grounding Facts (passed to reviewers):
Draft: Biogen states CELIA did not meet its primary endpoint but showed a positive clinical trend: 35.0%, No top-line CELIA result is publicly released by August 12, 2026: 30.0%, Biogen states CELIA met its primary endpoint: 14.0%, Biogen releases top-line CELIA data by August 12, 2026 but does not clearly state whether the primary endpoint was met: 13.0%, Biogen states CELIA did not meet its primary endpoint and no positive trend: 8.0%
Reviewer Panel:
Reviewer (reasoning)
The draft’s weakest link is its optimistic timeline and outcome assumptions: it underestimates the risk of delayed analysis past August 12 and overweights a positive clinical trend despite high Alzheimer’s Phase 2 failure rates and ambiguous dose–response designs. It also relies on limited analogies (e.g., UCB’s bepranemab) without properly calibrating to the broader base rate of nonmonotonic or null dose–response findings. Overall, the draft is too sanguine about both the “positive trend” outcome and the on-time data release.
Objections:
Reviewer (calibration)
The draft forecast’s estimate for the data release timeline appears overly optimistic, underestimating the probability of no top-line results by the deadline given the ‘Q2/Q3 2026’ guidance. While it correctly identifies key considerations like Biogen’s ‘signal-finding’ framing and incentives to highlight positive trends, the conditional probability for meeting the primary endpoint seems slightly too high considering the CEO’s expressed uncertainty and the historical difficulty of AD trials. The distribution of conditional outcomes could be refined to better reflect these challenges.
Objections:
Reviewer (fact-checker)
3 unsupported claim(s). Verdict: minor_issues
Unsupported claims:
Adjustment Rationale:
Adjusted from draft in response to reviewer convergence on three issues. (1) Lowered P(release by Aug 12) from 70% to 62% — reviewers correctly noted that Biogen’s Q2/Q3 2026 guidance combined with the CEO’s ‘nobody has moved the needle’ tone and typical Alzheimer’s Phase 2 timeline slippage skews the readout later in the window, and Aug 12 falls only ~47% into Q3. (2) Lowered conditional P(met primary) from 20% to 16% — the 0% Phase 2 primary endpoint success rate for anti-tau antibodies (Roche, UCB, J&J) is a strong base rate anchor that should pull harder against optimism, even given BIIB080’s ASO mechanistic differentiation and dose-response primary design. (3) Raised conditional P(missed, no trend) from 12% to 16% and P(ambiguous) from 18% to 20% — truly null readouts do occur, and nonmonotonic dose-response patterns often produce ambiguous disclosure language. Final P(missed, positive trend) of 30% aligns with the 29.8% ensemble signal. Also softened confidence from medium to low-medium per calibration reviewer’s note about compound uncertainty across timing and outcome.
forecast_sdk | 396.1s | research $0.7597 + reviewers $0.0142 + reconcile $0.1622 = $0.9361
All Predictions: [{‘Biogen states CELIA met its primary endpoint’: 0.1, ‘Biogen states CELIA did not meet its primary endpoint and no positive trend’: 0.22, ‘Biogen states CELIA did not meet its primary endpoint but showed a positive clinical trend’: 0.26, ‘Biogen releases top-line CELIA data by August 12, 2026 but does not clearly state whether the primary endpoint was met’: 0.12, ‘No top-line CELIA result is publicly released by August 12, 2026’: 0.3}, {‘Biogen states CELIA met its primary endpoint’: 0.11, ‘Biogen states CELIA did not meet its primary endpoint and no positive trend’: 0.21, ‘Biogen states CELIA did not meet its primary endpoint but showed a positive clinical trend’: 0.3, ‘Biogen releases top-line CELIA data by August 12, 2026 but does not clearly state whether the primary endpoint was met’: 0.13, ‘No top-line CELIA result is publicly released by August 12, 2026’: 0.25}, {‘Biogen states CELIA met its primary endpoint’: 0.14, ‘Biogen states CELIA did not meet its primary endpoint and no positive trend’: 0.3, ‘Biogen states CELIA did not meet its primary endpoint but showed a positive clinical trend’: 0.21, ‘Biogen releases top-line CELIA data by August 12, 2026 but does not clearly state whether the primary endpoint was met’: 0.05, ‘No top-line CELIA result is publicly released by August 12, 2026’: 0.3}, {‘Biogen states CELIA met its primary endpoint’: 0.14, ‘Biogen states CELIA did not meet its primary endpoint and no positive trend’: 0.35, ‘Biogen states CELIA did not meet its primary endpoint but showed a positive clinical trend’: 0.22, ‘Biogen releases top-line CELIA data by August 12, 2026 but does not clearly state whether the primary endpoint was met’: 0.06, ‘No top-line CELIA result is publicly released by August 12, 2026’: 0.23}, {‘Biogen states CELIA met its primary endpoint’: 0.14, ‘Biogen states CELIA did not meet its primary endpoint and no positive trend’: 0.15, ‘Biogen states CELIA did not meet its primary endpoint but showed a positive clinical trend’: 0.4, ‘Biogen releases top-line CELIA data by August 12, 2026 but does not clearly state whether the primary endpoint was met’: 0.01, ‘No top-line CELIA result is publicly released by August 12, 2026’: 0.3}] Median Reasoning: EXTRACTED_PROBABILITIES: [14, 30, 21, 5, 30]
Model’s Answer: The forecasting problem has two fairly separate layers: first, whether Biogen will publicly release CELIA top-line results by the deadline, and second, if it does, what the company’s first characterization of those results will be. On timing, the best hard datapoint in the research is the ClinicalTrials.gov estimate of primary completion around mid-May 2026, together with Biogen/partner commentary pointing to a 2026 or mid-2026 readout. That makes a release by August 12 more likely than not, especially because Biogen has obvious venues to disclose it: a press release, AAIC timing, or Q2 earnings materials. But it is not close to certain. Registry dates can slip, last-patient-last-visit timing can drift, and database lock plus analysis for a 416-patient global AD study can push a “mid-2026” expectation into late Q3 or later. So I do not make “release by deadline” overwhelming.
On outcome substance, I start from base rates: Alzheimer’s disease-modifying Phase 2 trials on clinical endpoints are hard, and first-in-class tau programs should be viewed skeptically. The primary endpoint here is not a biomarker but a clinical dose-response on CDR-SB at Week 76. That is a meaningful hurdle. The Phase 1b story for BIIB080 does provide some reason not to anchor too low: there was biomarker evidence of tau lowering, the mechanism is more biologically direct than failed extracellular anti-tau antibody strategies, and the Phase 2 design is explicitly dose-response based, which can sometimes rescue signal detection relative to a single-dose comparison. Still, the central forecast should remain that outright primary-endpoint success is less likely than failure.
Among failures, I split meaningfully between “no positive trend” and “miss but positive clinical trend.” This is important because the resolution criteria care about Biogen’s wording, not just the underlying statistics. For a company in a strategically important CNS program, if there is any directional clinical encouragement across secondary measures or dose levels, management often says so. On the other hand, the criteria require affirmative language about clinical efficacy, not just biomarker or safety positives, so not every spun release qualifies for the “positive trend” bucket. That keeps “miss and no positive trend” as the single most likely released-result category, but only modestly ahead of “miss with positive trend.”
I assign only a small probability to the ambiguous-disclosure option. Biogen is a large public company, and for a material Phase 2 readout I would usually expect fairly explicit language such as “met primary endpoint” or “did not meet primary endpoint,” even if surrounded by nuance. Ambiguity is possible if the release emphasizes data tables, dose-response complexity, or secondary findings without a blunt headline, but that is a minority case.
The assumptions doing the most work are:
Putting those together, my median view is: a topline readout before the deadline is somewhat more likely than not, but not enough to crush the no-release bucket; if data are released, a miss is more likely than a hit; and conditional on a miss, Biogen has a meaningful but not dominant chance of using “encouraging clinical trend” language.
Option_A: 14 Option_B: 30 Option_C: 21 Option_D: 5 Option_E: 30
All Predictions: [{‘Biogen states CELIA met its primary endpoint’: 0.18, ‘Biogen states CELIA did not meet its primary endpoint and no positive trend’: 0.27, ‘Biogen states CELIA did not meet its primary endpoint but showed a positive clinical trend’: 0.3, ‘Biogen releases top-line CELIA data by August 12, 2026 but does not clearly state whether the primary endpoint was met’: 0.06, ‘No top-line CELIA result is publicly released by August 12, 2026’: 0.19}, {‘Biogen states CELIA met its primary endpoint’: 0.21, ‘Biogen states CELIA did not meet its primary endpoint and no positive trend’: 0.24, ‘Biogen states CELIA did not meet its primary endpoint but showed a positive clinical trend’: 0.31, ‘Biogen releases top-line CELIA data by August 12, 2026 but does not clearly state whether the primary endpoint was met’: 0.14, ‘No top-line CELIA result is publicly released by August 12, 2026’: 0.1}, {‘Biogen states CELIA met its primary endpoint’: 0.18, ‘Biogen states CELIA did not meet its primary endpoint and no positive trend’: 0.26, ‘Biogen states CELIA did not meet its primary endpoint but showed a positive clinical trend’: 0.37, ‘Biogen releases top-line CELIA data by August 12, 2026 but does not clearly state whether the primary endpoint was met’: 0.04, ‘No top-line CELIA result is publicly released by August 12, 2026’: 0.15}, {‘Biogen states CELIA met its primary endpoint’: 0.38, ‘Biogen states CELIA did not meet its primary endpoint and no positive trend’: 0.18, ‘Biogen states CELIA did not meet its primary endpoint but showed a positive clinical trend’: 0.36, ‘Biogen releases top-line CELIA data by August 12, 2026 but does not clearly state whether the primary endpoint was met’: 0.03, ‘No top-line CELIA result is publicly released by August 12, 2026’: 0.05}] Median Reasoning: EXTRACTED_PROBABILITIES: [18, 26, 37, 4, 15]
Model’s Answer:
I need to evaluate: (1) probability of data release by August 12, 2026, (2) probability of meeting the primary endpoint, and (3) how Biogen would characterize results if the endpoint is missed.
The ClinicalTrials.gov record shows primary completion estimated May 14, 2026, with the trial “Active, not recruiting” and 416 enrolled. After last patient’s Week 76 visit, database lock typically takes 4-8 weeks (June-July 2026), followed by 4-8 weeks of statistical analysis (July-September 2026). The August 12 deadline falls squarely in the middle of this analysis completion window.
However, Biogen has consistently guided to “mid-2026” across multiple investor presentations (Q4 2025 earnings, Stifel CNS Day March 2026). Companies typically wouldn’t commit to “mid-2026” without confidence in their timeline. Also, AAIC 2026 (typically late July) presents a natural venue for presentation before the deadline. Still, there’s meaningful probability (~15%) that analysis isn’t complete or ready for release by August 12.
Factors favoring success (~20% primary endpoint met):
Factors against success:
Conditional on the primary endpoint NOT being met, Biogen’s characterization matters. Given BIIB080’s strong biomarker profile, even if the dose-response test fails, individual dose groups may show numerical improvements. Companies routinely highlight any favorable signal. I estimate ~55-60% probability of “positive trend” language if primary is missed, because: (a) Phase 1b signals were real; (b) some doses may show numerical benefit; (c) Biogen would want to preserve option value for the program.
The most likely scenario is that Biogen releases data by August 12 (85% probability). Among release scenarios, the highest probability outcome is that the primary endpoint is not met but exploratory signals or numerical trends lead Biogen to characterize results positively (37%). An outright failure with no favorable language is next most likely (26%), followed by meeting the primary endpoint (18%). Ambiguous communication is unlikely from a sophisticated company (4%).
Option_A: 18 Option_B: 26 Option_C: 37 Option_D: 4 Option_E: 15
I’ll work through this systematically.
High-quality factual sources:
Moderate-quality sources:
Key tension: Phase 1b clinical signal disputed — [S20] says no group differences; [S22]/[S3] say “favorable trends.” These may refer to different phases (MAD vs. LTE) but introduce genuine ambiguity about the underlying biology.
Strong evidence:
Primary completion date May 14, 2026 + multiple convergent “summer 2026” signals ([S14], [S15], [S39], [S38]): Multiple independent sources confirm data available ~summer 2026. Strong support for pre-August 12 disclosure. This is the most important timing factor.
Zero tau-targeting therapies have ever met a clinical endpoint in AD Phase 2 ([S27], [S26], [S4]): 30+ agents, all failures. Historical pattern with strong predictive power. This is the dominant factor against “met primary endpoint.”
Enrollment shortfall (416 vs. 735 planned, 44% below target) ([S14], [S5]): Structural factor reducing statistical power for the dose-response primary endpoint. Directly lowers probability of meeting primary endpoint even if true effect exists.
Moderate evidence:
AAIC 2026 (July 12–15, London) within resolution window ([S11]): Natural venue for AD trial disclosures; Biogen has engaged AAIC for this program ([S22], [S23]). Supports pre-August 12 disclosure.
Biogen’s prior “favorable trends” language for Phase 1b ([S3], [S22], [S25]): Company has already used this framing for weaker data. Moderate evidence they’ll use similar language for CELIA results, supporting “positive clinical trend” outcome.
Biogen management “cautious” tone at JPM 2026 ([S4]): Suggests management doesn’t expect a clean positive result. Moderate evidence against “met primary endpoint.”
Weak evidence:
Phase 1b biomarker results (30–49% CSF tau reduction) ([S18]): Proof-of-mechanism but not proof-of-clinical-efficacy. Weak positive signal for the trial outcome.
BIIB080 ASO mechanism distinct from all prior failures ([S4], [S31]): Intracellular tau reduction vs. antibody clearance. Theoretically superior but unproven. Weak upward adjustment for “met primary endpoint.”
Timeframe: April 22 to August 12, 2026 = ~112 days.
If timeframe halved (~56 days, to ~June 17): Probability of disclosure drops substantially — data lock alone would barely be complete by then. “No disclosure” probability rises to ~60–70%.
If timeframe doubled (~224 days, to ~December 2026): “No disclosure” probability drops to ~5%. CTAD (typically November) would be within scope. Nearly all uncertainty shifts to outcome.
The current 112-day window is well-calibrated for a summer readout — it captures AAIC and allows 4 weeks post-AAIC for a standalone press release. The main risk is data lock taking longer than 12 weeks (pushing past August 12) or a strategic decision to wait for CTAD.
Timing (P(disclosure by Aug 12)): The convergence of [S14] (primary completion May 14), [S15] (UBS “this summer”), [S39] (“H1 2026”), and [S38] (Ionis 2026 highlight) strongly supports a pre-August 12 disclosure. AAIC 2026 (July 12–15) is the natural venue and falls 4 months after primary completion — sufficient time for data lock and analysis. The enrollment shortfall doesn’t delay the primary completion date, only reduces power. I maintain ~70% probability of disclosure by August 12, with the main risk being a longer-than-expected analysis period or strategic delay to CTAD.
Outcome (conditional on disclosure):
Met primary endpoint: The historical base rate is effectively 0% for tau-targeting therapies (30+ failures). The ASO mechanism is genuinely novel, but Phase 1b showed no clinical signal [S20], and the enrollment shortfall (44% below plan) severely reduces power. Even if there’s a true biological effect, statistical significance on CDR-SB dose-response is unlikely. I adjust slightly upward from 0% to ~5% for the novel mechanism, but this remains very low.
Did not meet, no positive trend: Companies rarely characterize failures without positive spin, especially for proof-of-concept trials where program continuation is at stake. Biogen has already used “favorable trends” language for weaker Phase 1b data [S3], [S22]. However, if the data is clearly negative (e.g., no directional signal at all), they might not claim a trend. The enrollment shortfall increases the chance of a null result with no directional signal. I estimate ~25% conditional probability.
Did not meet but positive clinical trend: This is the modal outcome for failed AD trials. Companies routinely claim “positive trends” or “favorable signals” when the primary endpoint is missed. Biogen’s proof-of-concept framing gives them flexibility to highlight any directional signal. Their prior language (“favorable trends in exploratory clinical outcomes” from Phase 1b) establishes this as their preferred framing. I estimate ~55% conditional probability.
Ambiguous: Possible if Biogen releases data at a conference without a clear press release statement on primary endpoint status, or if the primary endpoint framing is unclear. Less likely given explicit resolution criteria and Biogen’s typical disclosure practices. ~10% conditional.
Scenario decomposition:
Scenario 1: Disclosure by August 12 (P = 0.70)
Scenario 2: No disclosure by August 12 (P = 0.30)
Conditional on Scenario 1 (disclosure):
Final probabilities:
Sum: 3.5 + 17.5 + 40.6 + 8.4 + 30 = 100 ✓
Comparing to outside view: Met (4→3.5), Not met no trend (21→17.5), Not met positive trend (38→40.6), Ambiguous (7→8.4), No disclosure (30→30).
The inside view adjustments are modest. The main shift is slightly reducing “not met, no trend” and slightly increasing “not met, positive trend” based on the strong evidence of Biogen’s prior framing language. The timing probability remains at 70%.
The “positive trend” outcome at ~41% is the modal outcome. This reflects: (1) the near-certainty that BIIB080 will not meet its primary endpoint given historical base rates and enrollment shortfall, and (2) the high probability that Biogen will characterize any directional signal as a “positive trend” given their proof-of-concept framing and prior language.
A ±10% shift in “no disclosure” (from 30% to 20% or 40%) would proportionally shift all other categories. At 20% no-disclosure, “positive trend” rises to ~46%, which seems too confident given genuine timing uncertainty. At 40% no-disclosure, “positive trend” drops to ~35%, which seems too pessimistic given the strong convergent timing signals.
Options paraphrase: By Aug 12, 2026: (A) CELIA met CDR-SB primary endpoint; (B) didn’t meet, no positive trend; (C) didn’t meet but positive clinical trend; (D) data released but ambiguous; (E) no top-line data released. ✓
Outside view base rates: Met: 4%, Not met no trend: 21%, Not met positive trend: 38%, Ambiguous: 7%, No disclosure: 30%. Inside view: 3.5%, 17.5%, 40.6%, 8.4%, 30%. Rooted in base rates with modest adjustments. ✓
Consistency check: Most likely = C (not met, positive trend); Least likely = A (met primary endpoint). Makes sense given tau-targeting failure history and Biogen’s framing incentives. ✓
Key evidence:
Blind spot: BIIB080’s ASO mechanism is genuinely novel — if it works differently from antibodies and the enrollment shortfall doesn’t prevent detection of a large effect, the trial could meet its primary endpoint. This would make the ~3.5% assignment look too low. However, given zero successes in the reference class and Phase 1b clinical null, I maintain this low probability.
Status quo: The status quo is “no disclosure yet” (April 22). The question is whether disclosure happens in the next 112 days. Given primary completion in 22 days and strong convergent timing signals, the status quo will change — disclosure is likely but not certain.
Technicalities: 3.5 + 17.5 + 40.6 + 8.4 + 30 = 100. ✓ Rounding to integers: 4 + 18 + 40 + 8 + 30 = 100. ✓
Probabilities: [4, 18, 40, 8, 30]
I’ll work through this systematically.
By Aug 12, 2026: (A) CELIA met primary endpoint; (B) missed, no positive trend; (C) missed but positive clinical trend; (D) data released but ambiguous on endpoint; (E) no top-line data released. Resolution based on Biogen’s first public statement.
High quality (primary/official sources):
Moderate quality (reliable secondary):
Lower quality (opinion/derivative):
Strong evidence:
Primary completion ~May 14, 2026 + AAIC July 12–15 within window [S14, S11]: The last patient’s Week 76 visit completes mid-May, with AAIC falling 8-9 weeks later. This is within the typical 5-15 week lag for comparable trials. Multiple independent sources (ClinicalTrials.gov, Alzforum, UBS, BrainTrials) converge on summer 2026 timing. This is the strongest factor supporting a pre-August 12 disclosure.
Zero tau-targeting therapies have met clinical endpoints in AD Phase 2 [S27, S26, S4]: 30+ agents failed. UCB bepranemab (Nov 2024) and J&J posdinemab (Nov 2025) both missed CDR-SB endpoints recently. This is the most powerful base rate for outcome prediction.
Enrollment shortfall: 416 vs. 735 planned (44% fewer) [S14, S5]: Dramatically reduces statistical power for the dose-response primary endpoint. Even if there is a true biological effect, the trial may lack power to detect it statistically.
Moderate evidence:
Phase 1b showed no clinical signal [S20]: Independent ADDF review states “no group differences” in clinical measures. Biogen’s own framing of “trends toward clinical improvement” [S22] appears to be spin on exploratory data. This moderately lowers probability of meeting primary endpoint.
Biogen frames CELIA as “proof-of-concept, not registrational” [S31]: This framing suggests Biogen may be managing expectations and could be more willing to present mixed results. It also means they may use “positive trend” language even for marginal results.
AAIC 2026 Developing Topics deadline passed April 11 [S12]: If Biogen submitted efficacy data, notification comes “early May 2026.” No confirmation of submission. This creates uncertainty about whether AAIC is the venue.
Weak evidence:
UBS “this summer” upgrade [S15]: Analyst opinion, not primary source. Consistent with other timing signals but adds no independent information.
Biogen “cautious” tone at JPM 2026 [S4]: Journalistic characterization; suggests management is hedging but not conclusive.
Current timeframe: 111 days (≈3.6 months) from April 22 to August 12, 2026.
Key dates within window:
If timeframe halved (to ~55 days, ~June 15): Probability of release drops sharply—data lock wouldn’t even be complete. Option E would dominate (~70-75%).
If timeframe doubled (to ~7 months, ~November 2026): Release probability approaches ~90%+. Options A-D would collectively dominate, with E dropping to ~5-10%.
The current window is tight but plausible—AAIC is the natural venue and falls within the window. The key uncertainty is whether data lock completes in time for AAIC or whether disclosure slips to Q3 earnings (late October) or CTAD (November/December).
Starting from outside view base rates:
Adjustments:
For E (no release by Aug 12):
For A (met primary):
For C (missed, positive trend):
For B (missed, no positive trend):
For D (ambiguous):
Recalibration after adjustments:
Scenario decomposition:
Scenario 1: Data released before August 12 (P = 0.61)
Scenario 2: No data released before August 12 (P = 0.39)
Conditional on release (Scenario 1):
| Outcome | Base rate (8-trial ref class) | Adjusted |
|---|---|---|
| A (met primary) | 12.5% | 11% (enrollment shortfall, no prior tau successes, no Ph1b signal) |
| B (missed, no trend) | ~40% | 38% (Biogen likely to spin, but some trials just fail cleanly) |
| C (missed, positive trend) | ~38% | 42% (Biogen’s “proof-of-concept” framing, prior “favorable trends” language) |
| D (ambiguous) | ~10% | 9% |
Check: 11+38+42+9 = 100% ✓
Unconditional probabilities:
Sum: 7+23+26+5+39 = 100% ✓
Status quo check: The status quo is “no data yet released.” The world changes slowly, and 111 days is not a long time. However, the primary completion date is imminent (May 14), making a release within the window genuinely plausible rather than speculative.
Consistency check: Most likely = C (missed, positive trend) or E (no release); Least likely = A (met primary) or D (ambiguous). This makes sense given the historical failure rate of tau therapies and the tight timing window.
Three most significant evidence pieces:
Blind spot: The scenario most likely to make this forecast look silly is if Biogen presents at AAIC with a clean positive result (met primary endpoint). The ASO mechanism genuinely is different from antibodies, and if tau production reduction translates to clinical benefit, A could be much higher. However, the historical base rate and Phase 1b clinical data argue strongly against this.
A ±10% shift: Moving E from 39% to 29% would require strong evidence of AAIC submission or Q2 earnings disclosure. Moving C from 26% to 36% would require stronger evidence of Biogen’s tendency to spin mixed results. Neither shift seems warranted without new information.
Probabilities: [7, 23, 26, 5, 39]
Analysis:
Paraphrase / criteria check (<30 words)
Resolves to Biogen’s first official topline CELIA disclosure by Aug 12, 2026: met primary, failed/no trend, failed/positive trend, ambiguous topline, or no topline released.
Outside view base rates previously established
Starting point from outside view:
Facts
Opinions / interpretations
Timing anchor: primary completion May 14, 2026 + repeated 2026/summer guidance
Tau clinical failure base rate is extremely poor
Enrollment shortfall: 416 actual vs 735 planned
BIIB080 mechanism is meaningfully distinct from prior failed tau antibodies
Biogen has already used favorable-trend language in earlier BIIB080 communications
Management reportedly sounding cautious in early 2026
AAIC as likely venue
Proof-of-concept framing means more willingness to discuss mixed signals
Forecast horizon is about 112 days from Apr 22, 2026 to Aug 12, 2026.
This is short enough that:
If the timeframe were halved (deadline around mid-June 2026), “No topline by deadline” would jump substantially, likely becoming the plurality outcome, because data lock after May 14 would be very tight.
If the timeframe were doubled (deadline around late November 2026), “No topline” would fall sharply, and probabilities would redistribute mainly into the explicit outcome buckets, especially “did not meet but positive trend” and “did not meet/no trend.”
So the biggest driver versus the outside view is not efficacy; it is whether Biogen gets the data out by this particular date.
The outside view already captured the broad shape well: low primary-success odds, meaningful but not overwhelming chance of no disclosure, and the modal disclosed outcome being a miss with positive spin.
My inside-view changes are modest.
There is fairly convergent timing evidence:
A large pharma with a high-profile Alzheimer’s asset usually does not sit on a clean topline for long, especially if summer guidance has been given. That argues for trimming “No topline” somewhat from 30.
The historical base rate for tau programs is terrible, and CELIA’s reduced enrollment is a substantial headwind. Even if BIIB080 has real biology, the question is not “does it affect biomarkers?” but whether Biogen’s first topline will clearly say the primary endpoint was met. That is a high bar. A novel ASO mechanism is enough to keep success above zero, but not enough for a big upward move.
This combines:
Biogen has incentives to preserve option value for the asset if the primary analysis misses but secondaries or subgroup/dose-response patterns are directionally supportive. This is a classic setup for “did not meet primary endpoint, but showed encouraging clinical trends.”
I do not want to over-assume PR spin. If the data are broadly flat or inconsistent, and especially if clinical measures fail to line up, Biogen may simply say it did not meet the endpoint and avoid overstating. Also, if safety/tolerability or delivery burden weakens enthusiasm, a positive-trend framing becomes harder.
Biogen usually has the sophistication to state a topline outcome clearly if it chooses to disclose topline at all. Ambiguity is possible if conference materials are sparse or emphasize biomarker findings, but this still seems less likely than an explicit success/failure statement.
I’ll use a two-stage decomposition:
Let:
This is a moderate shift from outside view 0.70 / 0.30, based on convergent timing evidence and the natural AAIC/press-release window.
So immediately:
I decompose released-readout states into 4 mutually exclusive scenarios:
Conditions:
Conditions:
Conditions:
Conditions:
These conditional probabilities sum to 1.00.
Now multiply by P(Release)=0.74:
Rounded and normalized:
Compared with outside view:
Why the shifts?
Probability calibration
If the current status quo continues, the most likely path is:
That points to Option 3 as the plurality outcome.
A 10-point move from Option 5 to the disclosed buckets would require stronger evidence that AAIC submission or a fixed disclosure date exists; I do not have that.
A 10-point move from Option 3 to Option 2 would require stronger evidence that BIIB080’s earlier “favorable trends” were weak PR gloss and that CELIA is likely cleanly negative. We have some hints, but not enough for that large a shift.
So I keep the distribution fairly close to the outside view, with only moderate adjustment.
Consistency check
Most likely category/categories = did not meet primary endpoint but showed a positive clinical trend, then did not meet and no positive trend / no topline; Least likely = met primary endpoint; yes, aligned.
Most significant evidence (<20 words each)
Blind-spot statement Biggest risk: Biogen submitted/secured AAIC disclosure and CELIA unexpectedly shows a clearly positive dose-response, making Option 1 too low.
Status quo outcome Status quo modestly favors a disclosed but negative/mixed result over either a breakthrough success or prolonged silence.
Technicalities Probabilities are integers, 0–100, sum to 100, no percent signs.
Checklist:
Probabilities: [5, 26, 36, 7, 26]
=== FORECASTING CHECKLIST ===
Options in plain words (<30):
A = Meets primary endpoint; B = Fails, no positive trend; C = Fails, but Biogen claims positive clinical trend; D = Top-line out but unclear on endpoint; E = No top-line by 12 Aug 2026.
Outside-view distribution (22 Apr 2026): [5, 22, 27, 8, 38].
Consistency line: Most likely = E (no release); Least likely = A (meets); this matches tau failure base-rate and timing risk.
Key evidence lines (<20 words each):
• Primary completion ≈ 14 May 2026 – gives ≤ 90 days to deadline.
• AAIC late-breaking abstracts closed 11 Apr – lowers July conference likelihood.
• All 30 + tau trials have failed on cognition.
• BIIB080 shows strong tau lowering but no Phase 1b clinical signal.
• Biogen guidance “mid/summer-26” repeated Q4-25, UBS 22 Apr 26.
Blind-spot: Biogen secretly accelerated LPLV to March 2026 enabling April data-lock and AAIC abstract; if true, probability of A or C rises sharply.
Status-quo tilt: Drug development delays common; a small upward nudge to “no release” (E) is warranted.
Technicalities: Probabilities below are 0–100, sum = 100, no “%”.
——————————————————————–
Analysis
(a) Source analysis (quality & fact/opinion split)
• [S14] ClinicalTrials.gov (May 2025 update) – highest-quality factual: enrollment = 416; primary completion est. 14 May 26.
• [S3] Biogen PR (Apr 2025) – factual on enrollment & 2026 readout; opinionated “favorable trends”.
• [S11/12/30] AAIC 2026 websites – factual: conf. 12-15 Jul; Developing Topics deadline 11 Apr 26.
• [S18/19/25] Phase 1b data – factual biomarker drop; opinion “encouraging”.
• [S26/27] Tau field review & UCB failure – factual negative precedent.
• [S4/15/31] Analyst & forum notes – single-source, partially opinion, moderate quality on timing & company tone.
(b) Evidence weighting relative to resolution criteria
Strong:
• Tau agents’ 0/30 success record [S26/27].
• Primary completion 14 May 26 [S14] – dictates earliest possible read-out.
Moderate:
• Repeated Biogen/analyst “mid-26” guidance [S3][S4][S15][S38].
• Enrollment shortfall 416 vs 735 – lowers power (single factual source).
Weak:
• Sponsor spin historically yields “positive trend” language – anecdotal but plausible.
• AAIC presence pattern – prior baseline/poster only; not proof of 2026 slot.
(c) Timeframe analysis
Window to deadline = 112 days. Halving (56 d) would raise “no release” (E) by ~+20 pts; doubling (224 d) would cut E roughly in half.
(d) Justification of inside-view shifts
(e) Quantitative model
Define 3 mutually-exclusive scenarios (sum = 1):
S-timely (0.60): Data lock & PR before 12 Aug.
S-delay (0.30): Analysis/strategy slippage → no topline by deadline.
S-ambiguous (0.10): Only biomarker/safety headline; endpoint status unclear.
Conditional outcome probabilities
A B C D E
S-timely: 0.07 0.37 0.43 0.13 0
S-delay : 0 0 0 0 1
S-ambig.: 0 0.20 0.20 0.60 0
Combined (weighted):
A = 0.600.07 ≈ 0.042
B = 0.600.37+0.100.20 ≈ 0.198
C = 0.600.43+0.100.20 ≈ 0.258
D = 0.600.13+0.100.60 ≈ 0.089
E = 0.301 ≈ 0.300 (rounded upward to 0.40 after status-quo check)
Adjust for checklist step-6 (status-quo tilt) by shifting 0.10 from B/C evenly into E.
Final calibrated probabilities:
A 4 | B 22 | C 26 | D 9 | E 40
——————————————————————–
Probabilities: [4, 22, 26, 9, 40]
=== FORECASTING CHECKLIST STEP-BY-STEP ===
Options in plain English (<30 words).
A = Biogen says primary endpoint met.
B = Says endpoint NOT met and no positive trend.
C = Says endpoint NOT met but positive clinical trend.
D = Releases topline but unclear on endpoint result.
E = No topline result released by 12 Aug 2026 (23:59 BST).
Outside-view base rates (established earlier):
[A 9, B 24, C 21, D 6, E 40] (sum = 100).
Single-line consistency check.
Most likely = E (no data) and B (clean miss); least likely = A (success). Matches argument that tau field has 0/30 wins and timing is tight.
Top five evidence items (≤20 words each).
• Primary completion estimated 14 May 2026 (S14) → ≤12 weeks to deadline.
• AAIC 2026 (12-15 Jul) sits inside window; Biogen routine AAIC participation (S22,S23).
• Developing-Topics abstract deadline (11 Apr) passed before completion → conference disclosure less certain (S12).
• Trial under-enrolled 416/735 → ~30 % power loss.
• Tau DMTs 0/30 clinical wins; two more failures 2024-25 (S26,S4).
Blind-spot scenario.
Rapid database lock (≤4 weeks) plus late-breaker slot secured quietly at AAIC → early July press release claiming success. Would boost A to >30 %.
Status-quo outcome.
Most AD mid-stage tau trials either slip or fail; status-quo therefore favours “no data” or “negative data” framing.
Technicalities check.
Probabilities below are 0-100 integers, sum to 100, no “%”.
ANALYSIS (succinct version)
(a) Source quality assessment (illustrative selection)
S14 ClinicalTrials.gov (May-2025) – high-quality factual timing & enrolment.
S22 NeurologyLive (Aug-2025) – moderate; quotes Biogen scientist, useful tone.
S26 FierceBiotech (Nov-2024) – high reliability for peer-company failure news.
S15 UBS note (22-Apr-2026) – opinion; only guides likely timing.
S31 Stifel CNS Forum transcript (Mar-2026) – primary source for “proof-of-concept” framing.
(b) Evidence weighting
Strong: • 0/30 tau successes (structural).
• May-14 primary completion gives ≤12 wk runway (historical median 9 wk). Moderate:• AAIC venue inside window but abstract deadline passed.
• Biogen history of issuing press releases quickly (but also of slipping AD timelines). Weak: • Analyst “this summer” language (sentiment, not binding).
(c) Timeframe sensitivity
Window = 111 days. If window halved (≈55 days, ends mid-June), release probability drops to ~25 %, pushing E to ~75 %. If doubled (≈7 months), release probability rises to ~85 %, redistributing mainly to B and C.
(d) Inside-view justification of shifts
(e) Quantitative model (scenario approach)
Scenario 1 – “Data released” (P = 0.63)
• Endpoint met P = 0.11 → 0.07 overall
• Miss, pos. trend P = 0.40 → 0.25
• Miss, no trend P = 0.45 → 0.28
• Ambiguous P = 0.13 → 0.08
Scenario 2 – “No data” (P = 0.37) → maps entirely to option E.
CALIBRATED INSIDE-VIEW FORECAST
Probabilities: [7, 28, 25, 8, 32]
In synthesizing our forecast for the CELIA trial results, our team focused on three primary factors: historical success rates in Alzheimer’s research, the trial’s specific design, and Biogen’s own communications.
Historically, Phase 2 Alzheimer’s trials have an extremely high failure rate, particularly when attempting to move cognitive scales like the primary endpoint used here. We noted that Biogen reduced the trial’s enrollment and distributed patients across multiple dosing groups, which significantly limits the statistical power needed to “meet” the primary endpoint. Consequently, we viewed a clean success as less likely than a miss.
A key nuance in our thinking was management’s language during earnings calls. They explicitly stated they would be looking for “trends and clinical data” to guide internal decisions. This suggests the company is prepared for a scenario where the main goal is missed, but they can still highlight a “positive clinical trend” to justify moving into Phase 3. We weighed this heavily against the possibility of a total failure with no positive trends.
Finally, we analyzed the timing. Biogen’s “mid-2026” guidance aligns with their typical late-July or early-August earnings schedule, making a public statement before the August 12 deadline highly probable, though we allowed for minor clinical delays.