ALX Oncology Holdings (ALXO) reported Q1 2026 on May 8 with two material disclosures: a $140.4M February equity raise that takes pro-forma cash to ≈$188M against ≈$300M market cap, and the full ESMO Breast Cancer 2026 biomarker dataset for evorpacept + zanidatamab in HER2+ metastatic breast cancer. The cross-tumor pattern — CD47 expression as a predictive biomarker for response — now spans two independent trials in two cancers. At spot $2.08, the market prices the cash floor cleanly and the bull case skeptically. The disagreement is concentrated and quantifiable.

What the filing shows

Clinical (ESMO Breast 2026, May 7): Among centrally confirmed HER2+ patients in the Phase 1b/2 evorpacept + zanidatamab study (24 ITT; 10 centrally confirmed HER2+; all post-T-DXd; median 5 prior HER2-directed therapies), the CD47-high subgroup (≥20% IHC total membrane staining) showed 100% ORR (5/5), median DOR 20.2 months, median PFS 22.1 months. CD47-low subgroup: 1/4 responded, mPFS 3.4 months. The 6.5x PFS delta replicates the ASPEN-06 randomized gastric pattern (HR 0.39 mPFS, 65% ORR CD47-high). Two trials, two HER2+ cancers, consistent biomarker effect.

Capital structure: February 2026 equity raise of $140.4M net brings pro-forma cash to ≈$188M against ≈$300M market cap. Quarterly burn ≈$5-6M implies 8+ years of runway. Going-concern risk eliminated.

Trial design and partners: ASPEN-09 (post-T-DXd HER2+ mBC, evorpacept + trastuzumab, single-arm) expanded from 80 to 120 patients; primary endpoint changed to ORR in CD47-high subgroup; Ventana (Roche subsidiary) named publicly as companion diagnostic partner. CMO explicitly ruled out accelerated approval — Phase III required. Approval timeline approximately 2030-2031.

What the market is pricing

Spot $2.08 decomposes to ≈$1.31/share cash + ≈$0.77/share pipeline option value (post Feb raise, ≈144M shares outstanding). The Jan 2027 ATM call ($2 strike, 247 days) trades at $0.90 with IV 145%. Calls outweigh puts 15.7:1 in OI; positioning is concentrated at the Jan 2027 expiry (1,452 of 2,089 total OI), consistent with LEAP positioning ahead of the ASPEN-09 mid-2027 readout.

Reverse-engineering state probabilities that produce spot $2.08:

The cash floor agreement is tight: 89% market-implied survival above $1.00 vs our 92%. Disagreement is concentrated in the bull state — modeled at 45%, market-implied at ≈10%. Probability-weighted target $4.15 vs spot $2.08; PV at 18% biotech required return = $3.24, ≈56% upside discounted to today.

Why the gap exists

Three reasons.

CD47 graveyard pricing. Magrolimab (Gilead, $4.9B Forty Seven acquisition) failed for IgG4 Fc binding to red blood cells causing hemolytic anemia and infection deaths — ENHANCE-3 AML discontinued 2024. Maplirpacept (Pfizer's $2.3B Trillium asset) terminated Q1 2024 for "business reasons" (lack of differentiation). ImmuneOnco IMM2520 axed in pipeline prioritization. None of these were biomarker-selection failures. The market reads "CD47 = bad drug class" and prices accordingly. Evorpacept's three differentiators (inactive Fc by design, biomarker enrichment via CD47 expression cut, combination architecture rather than monotherapy) address all three failure modes specifically.

Post-T-DXd HER2+ mBC competitive set is thinner than consensus believes. SATEEN (sacituzumab + trastuzumab) failed for futility (1/27 PR). ZW49 discontinued 2024. MCLA-128 (Merus zenocutuzumab) approved December 2024 only for NRG1-fusion, not HER2+ mBC. Daiichi HER3-DXd Phase 3 (HERTHENA-Breast04) is HR+/HER2-, not post-T-DXd HER2+. The only credible direct Phase 3 competitor remaining is Jazz EmpowHER-303 (zanidatamab monotherapy). The TAM (≈$2-4B at 50-70% CD47-high prevalence across ≈20K addressable patients in US/EU5/Japan) holds up better than the implicit consensus pricing suggests.

T-DXd CD47 upregulation is published biology, not management spin. AACR 2023 Abstract 2944, Cancer Research 2025 Abstract 2128, and Nature Communications 2017 super-enhancer paper independently confirm that T-DXd raises surface CD47 expression and that CD47 blockade synergizes with T-DXd in HER2-transgenic mouse models. The structural tailwind — standard of care creates the post-treatment CD47-high population that evorpacept targets — is published infrastructure across multiple labs.

The market prices class risk that doesn't apply to evorpacept's specific design.

Risks

1. ASPEN-09 fails to replicate. The 100% ORR finding is n=5; confidence intervals enormous. Trastuzumab combination differs from prior zanidatamab combination. CMO explicitly stated 100% ORR is "not expected to replicate at scale." If the 80-patient interim shows ORR <25% in CD47-high or no separation between CD47-high and CD47-low, the thesis breaks. Stock approaches cash + ALX2004 option value (≈$1.40 floor). 20% modeled probability with -64% downside to current.

2. Phase III timeline slippage. Management guides "ready for registrational studies by end of 2027." Even conditional on positive ASPEN-09 interim mid-2027, dosing patient #1 in Phase III typically slips 6-12 months. Approval realistically 2030-2031. Multiple future capital raises required for full Phase III funding (current $188M is sufficient through Phase III initiation, not through completion).

3. CD47 cut point not yet established. Management's 50-70% CD47-high prevalence estimate assumes a specific IHC threshold. If the optimal cut is set higher to enrich for response (improving PFS but reducing eligible population), addressable TAM compresses materially. The interim readout will define this jointly — ORR magnitude AND cut-defined prevalence trade off.

4. Liquidity constraint on deployment. ADV $1.9M. A 5% position on a $10M book is 26% of ADV — ruinous to entry. Deployable alpha is roughly 5-10% of theoretical alpha. This is a sleeve position by liquidity construction, not a conviction-position by thesis size.

5. HER2 central vs local discordance. Only 10 of 24 patients in the exploratory study were centrally confirmed HER2+ (42% discordance, archival vs fresh biopsy and local vs central testing). Phase III will require central confirmation as gating criterion. Manageable but adds enrollment friction.

Catalysts

The next 60-90 days are catalyst-empty. Insider buying is 8 months stale: CEO Lettmann and CFO Shantharam bought $158K combined at ≈$1.00 in Aug-Sep 2025 — now sitting at ≈$2.08 with no fresh additions disclosed.

What would change the thesis

• Pullback to $1.40-$1.60: improves prob-weighted asymmetry from 8.4x to 12-15x. Adds become aggressive.
• Insider P-code at $2+: confirmation at higher price (LR 2.0+); P_bull updates toward 0.55, target ≈$4.95.
• 8-K Item 1.01 partnership announcement: any strategic deal beyond Ventana CDx (development pact, license, equity investment) re-rates the strategic-optionality leg.
• ASPEN-09 80-patient interim ORR <25% in CD47-high: thesis breaks. Stock approaches cash floor.
• Adverse safety signal from evorpacept in any indication: pause and re-evaluate.
• Jazz EmpowHER-303 hits before ASPEN-09: competitive set re-rates around zanidatamab; ALXO's CD47-stratification distinctness becomes more important, not less.
• ALXO surprise raise >$10M before mid-2027: signals lower confidence in current runway through Phase III initiation. Distinguish opportunistic-into-strength from defensive by issue price relative to spot.

Evidence