Compass Therapeutics, Inc.

2.1 History & Prior Art

Tovecimig (formerly CTX-009) is an investigational bispecific antibody that simultaneously blocks Delta-like ligand 4 (DLL4) and vascular endothelial growth factor A (VEGF-A) C014. The molecule originated from Korean biopharma development; an early Phase 1b study evaluated CTX-009 plus chemotherapy (irinotecan or paclitaxel) in advanced solid tumors and produced four BTC partial responses, with two patients on the 10 mg/kg + paclitaxel 80 mg/m2 regimen achieving tumor shrinkage of 41.4% and 61.6% with durable responses exceeding 9 months each C038. A follow-on Korean Phase 2 study enrolled 24 BTC patients in 2L/3L settings and reported 37.5% ORR with median PFS of 9.4 months and 53% one-year survival C039. Compass licensed and advanced the asset into the U.S. COMPANION-002 registrational program (NCT05506943) C009.

2.2 Novelty Assessment

Tovecimig is the first bispecific antibody to advance into a U.S. Phase 2/3 registrational trial in second-line BTC. The molecule is described as the only DLL4 x VEGF bispecific to demonstrate monotherapy activity in colorectal and gastric cancer C014. No DLL4 x VEGF-A competitor has reached Phase 2 in BTC. The modality (bispecific antibody) is well established as a regulatory and manufacturing pathway, removing a category of execution risk that typically drags novel platforms.

3.1 Target Biology

VEGF-A blockade is a validated oncology target with multibillion-dollar precedent in bevacizumab (Avastin) C014. DLL4-Notch1 signaling mediates resistance to anti-VEGF therapy and represents a novel angiogenesis target. The bispecific design anchors the antibody in the tumor microenvironment via DLL4 to localize the anti-angiogenic effect. By simultaneously blocking both pathways, tovecimig disrupts tumor vessel formation through dual mechanisms — VEGF-A blockade prevents new vessel formation while DLL4 blockade alters perfusion in existing tumor vessels C014.

3.2 Validation Layers

Three validation layers reinforce the mechanism. First, anti-VEGF biology is FDA-approved across colorectal, NSCLC, glioblastoma, ovarian, and other solid tumors via bevacizumab. Second, DLL4 signaling is a well-characterized resistance mechanism to anti-VEGF monotherapy with extensive preclinical literature. Third, competing PD-1 x VEGF-A bispecifics (ivonescimab) have shown that dual-pathway anti-angiogenic biology can produce clinically meaningful effects in solid tumors, providing class-level proof of concept even if the indications differ.

3.3 Target Validation Score

The mechanism scores high on validation. The drug works by blocking two well-characterized pathways with established oncology precedent. Investors should treat the mechanism as differentiated but not unprecedented — the COMPANION-002 results validate that dual pathway blockade produces a clinically and statistically meaningful effect in BTC.

4.1 Data Summary

The primary endpoint stands. Tovecimig plus paclitaxel achieved 17.1% ORR (19 of 111 patients) including one complete response, versus 5.3% (3 of 57) for paclitaxel alone, all confirmed by blinded independent central review C001. The 11.8-percentage-point absolute improvement was statistically significant at p=0.031. Baseline characteristics balanced across arms — median age 65 versus 63, ECOG 0 in 47.7% versus 47.4%, metastatic disease in 89.2% versus 91.2% C028 — so the response gap reflects drug effect rather than population skew.

The PFS readout vindicates the ORR signal. Tovecimig plus paclitaxel extended median PFS to 4.7 months versus 2.6 months for paclitaxel alone, a 56% reduction in the risk of progression (HR=0.44, p<0.0001) C002. Progression was confirmed by BICR. The PFS effect size is the kind of read that survives FDA scrutiny: small p-value, BICR-confirmed, hazard ratio that doesn't drift toward unity. Disease-control patterns reinforce the finding — progressive disease as best response occurred in 16.2% of the combination arm versus 42.1% of the control C029, a 25.9-point absolute reduction that matches the ORR and PFS direction.

Overall survival is the only soft spot, and it has a clean explanation. ITT median OS was 8.9 months for combination versus 9.4 months for control (HR=1.05, p=0.78) C003. The pre-specified RPSFT crossover-adjusted analysis returned similar numbers (HR=1.13, p=0.65), but Compass disclosed that RPSFT statistical assumptions were not met and called the result "largely uninterpretable" C004. Crossover drove the pattern: 31 of 57 control patients (54%) crossed over to receive tovecimig after centrally confirmed progression C005, so 142 of 168 patients in the trial (85%) ultimately received tovecimig.

Three pre-specified and post-hoc analyses test whether tovecimig drove OS when patients actually received it. In the pre-specified PFS2 analysis, the same 31 crossover patients had median PFS of 1.9 months on paclitaxel pre-crossover versus 3.5 months on tovecimig post-crossover (HR=0.36, p=0.0016) C007. In a post-hoc OS comparison within the control arm, crossover patients reached median OS of 12.8 months versus 6.1 months for non-crossover patients (HR=0.54, p=0.04) C006. Notably, these crossover patients were not the healthy half — they had progressed faster on paclitaxel monotherapy than non-crossover patients (1.9 versus 3.6 months, HR=2.31, p=0.007) C035. Sicker patients lived longer once they got tovecimig.

4.2 Anti-Anchor Check

The Korean Phase 2 study reported a 37.5% ORR in 24 patients C039, materially higher than the U.S. COMPANION-002 17.1% ORR C001. Treat the Korean data as hypothesis-generating rather than confirmatory — the small sample, mixed 2L/3L population, and Korean-only enrollment limit translation. The U.S. COMPANION-002 read is the relevant base case for BLA discussions. Investors who anchor to the Korean numbers will be disappointed by the registrational study; those who anchor to the U.S. ITT result and the FDA's acceptance of ORR-based endpoints in unmet-need oncology will be appropriately calibrated.

The data tells a coherent story. ORR rose. Progression slowed. Patients who received tovecimig at any point in the trial pulled the survival curves up. The ITT OS miss reflects trial-design philosophy (enabling ethical crossover) more than drug failure.

5.1 FAERS Disproportionality

Tovecimig is investigational — the FDA Adverse Event Reporting System (FAERS) contains no real-world post-marketing data because the drug is not yet approved or widely used outside clinical trials. FAERS disproportionality analysis is therefore not applicable at this time and will become relevant only after BLA approval and commercial launch. Class context (anti-VEGF agents) shows established disproportionality signals for hypertension, proteinuria, and hemorrhagic events; investors should expect tovecimig's post-marketing FAERS profile to align with bevacizumab once data accumulates.

5.2 Class-Wide Signals

Tovecimig's safety profile reflects its anti-VEGF mechanism. Any-grade hypertension occurred in 69% of combination patients, with related Grade 3+ hypertension at 44% C010. Rate compared with 4% related Grade 3+ in the paclitaxel arm — a 22-fold elevation C044. Proteinuria followed the same pattern: 34% versus 9% any-grade C045. Both are known class effects of VEGF-A blockade documented across bevacizumab, ramucirumab, and aflibercept programs, and clinically manageable in oncology practice with standard antihypertensives and dose modifications. The class signals are present but not exaggerated relative to the parent class.

Hematologic toxicity was driven by paclitaxel rather than tovecimig. Neutropenia of Grade 3+ severity occurred in 36% of combination patients C010, comparable to the 26% Grade 3+ rate seen in the paclitaxel control arm. This pattern indicates the bispecific does not materially worsen the chemotherapy-driven cytopenias.

5.3 Kill Shots

The data presents no kill-shot safety signal. The DMC's recommendation to continue without modification at every pre-specified review C011 anchors the safety story. No new signal emerged across four DMC meetings spanning the trial's full enrollment. Tovecimig was generally well tolerated with no new safety signals C010. The hypertension and proteinuria profile will shape the FDA's labeled dose modifications and contraindications and may produce a black-box warning, but neither is a program-killer. Investors should monitor the BLA review for any new safety findings from the integrated safety database that emerge during agency review.

6.1 Chain of Title

Tovecimig was originally developed in South Korea and licensed in by Compass Therapeutics. The molecule has been advanced through a U.S. Phase 2/3 registrational trial under Compass's IND. Specific patent assignments and license terms are not disclosed in the source materials reviewed for this DD pass and would require a deeper pull from USPTO assignment records and Compass's 10-K disclosure. The Form 8-K signed by Chief Accounting Officer Neil Lerner C037 is the controlling document for today's disclosure but does not address IP. A buy-side analyst should request the Compass 10-K for fiscal year 2025 (filed March 5, 2026) for full IP disclosure including patent terms, license obligations, and any royalty arrangements with the original Korean inventor.

6.2 Maintenance & Lapse Risk

Orphan Drug Designation in BTC granted April 2026 C013 provides seven-year market exclusivity at U.S. approval, layered on top of any underlying composition-of-matter or method-of-use patent protection. Fast Track Designation granted April 2024 C013 does not provide IP protection but enables priority regulatory review. The orphan exclusivity window is the largest single contributor to commercial protection in the BTC indication.

6.3 FTO Gaps

FTO (freedom to operate) analysis would require examination of competing DLL4 and VEGF-A bispecific patent estates. Compass faces no known DLL4 x VEGF bispecific competitor in BTC at Phase 2 or beyond. Adjacent class patents (ivonescimab, faricimab, etc.) target different bispecific combinations and should not block tovecimig's commercial path. A full FTO opinion is recommended before BLA submission.

7.1 Verified Credentials

The leadership team has the relevant track record C023. Thomas J. Schuetz, MD, PhD, is President, CEO, and Vice Chairman; Cynthia Sirard, MD is CMO; Bing Gong, PhD is CSO; Barry Shin, JD, MBA is CFO; Arjun Prasad, MBA, MPH is CCO; Jon Anderman, JD is SVP General Counsel. Neil Lerner, CPA, MIM serves as Chief Accounting Officer and signed today's 8-K C037. The bench includes Karin Herrera (SVP Clinical Operations) and Kris Sachsenmeier, PhD (VP Translational Science) C023.

7.2 Publication Record

Compass scientists have authored or co-authored multiple peer-reviewed publications relevant to the program. Minori Rosales, MD, PhD serves as Study Director for COMPANION-002 C031 and is a co-author on the trial-design publication (Azad et al., Future Oncology 2024, PMID 38861293) C030. The Cholangiocarcinoma Foundation 2024 conference posters C038C039 include multiple Compass authors and collaborate with academic investigators including Drs. Katie Kelley (UCSF), Scott Paulson (Texas Oncology Baylor Sammons), Mitesh Borad (Mayo Arizona), Lipika Goyal (Stanford), Nilofer Azad (Johns Hopkins), and Milind Javle (MD Anderson).

7.3 Flags

No team flags surface in the disclosure record reviewed. Standard buy-side practice should include LinkedIn / news verification of executive bios, patent inventor lookups, and review of any prior litigation involvement. The Korean origin of tovecimig (originally developed by Korean biopharma) suggests an in-license relationship that should be confirmed via the 10-K filings.

8.1 Capital Raised vs. Claimed

As of December 31, 2025, Compass held $30.6 million in cash and equivalents and $178.3 million in current marketable securities — $208.9 million total C018. The company guides this to runway into 2028 C022. Total liabilities of only $22.8 million against stockholders' equity of $196.8 million C021 means the balance sheet is essentially debt-free. Common shares outstanding rose from 137.8 million at year-end 2024 to 178.3 million at year-end 2025 C020, a 29% increase, with the share count jumping from 138.3 million at June 30 to 177.6 million at September 30, 2025 — indicating a large Q3 2025 follow-on offering.

8.2 Cash Runway

FY2025 operating expenses totaled $72.8 million, split between R&D of $56.0 million and G&A of $16.9 million; net loss reached $66.5 million C019. At a roughly $5.5M monthly burn, $208.9 million of liquidity buys roughly 38 months at current run-rate — long enough to file the BLA, build the commercial organization, and reach approval if the path goes clean. Burn will accelerate during commercial pre-launch scale-up; a 12-month launch slip plus pre-commercial expansion could compress runway to 18-24 months and force a 2027-2028 raise.

8.3 Market Reaction and Enterprise Value

CMPX closed at $1.79 on April 27, 2026, the day of the data release, down $3.24 or 64.4% from the prior trading day's close of $5.03 C051. At the closing price and 178.3M shares outstanding C020, market capitalization is approximately $319M; subtracting $208.9M of cash and marketable securities yields enterprise value of approximately $110M C052. The market is pricing tovecimig and the broader pipeline at roughly $110M total — a level historically associated with failed Phase 2 oncology programs rather than BLA-ready assets in unmet-need orphan indications. The price action reflects buy-side fear that the FDA will demand a confirmatory trial; the regulatory toolkit and the magnitude of the ORR plus PFS effect C001C002 argue against this concern. Cash per share at year-end 2025 was approximately $1.17 (208.9M / 178.3M), so the closing price of $1.79 prices tovecimig and pipeline at roughly $0.62 per share above cash.

8.4 Insider Transactions

Insider transaction detail (Form 4 filings) was not pulled in this DD pass. A buy-side analyst should review Form 4 filings since the August 2025 follow-on offering for any signal of insider buying or selling near the catalyst date. The Q3 2025 share-count jump indicates institutional participation in the financing; specific buyers (mutual funds, healthcare hedge funds) can be identified through 13F filings.

9.1 FDA Precedent

The FDA-meeting language in today's 8-K is deliberately spare: "In the coming months, Compass intends to meet with the U.S. Food and Drug Administration (FDA) to discuss these data in advance of a planned BLA submission" C012. Read this as a Type B end-of-Phase 2 / pre-BLA meeting in Q3-Q4 2026, BLA filing in late 2026 or H1 2027, and potential accelerated-approval acceptance under FDA's Subpart H framework if the agency agrees that ORR is reasonably likely to predict clinical benefit in this setting. The precedent base for ORR-based oncology accelerated approvals is substantial — multiple targeted therapies and bispecifics have received accelerated approval on ORR effect sizes smaller than the 11.8-point absolute and 3.2-fold relative effect Compass demonstrated.

9.2 Designation Status

Tovecimig's regulatory toolkit is built for accelerated review. Fast Track Designation in BTC was granted in April 2024 C013, opening rolling review and priority review eligibility. Orphan Drug Designation followed in April 2026 C013, delivering seven-year market exclusivity at approval plus tax credits. The trial design publication (Azad et al., Future Oncology 2024) C030C046 anchored the ORR primary endpoint in the original IND submission — the FDA would not have allowed a Phase 2/3 registrational design with ORR as primary if the agency objected philosophically.

9.3 Pathway Feasibility

The accelerated-approval case rests on three facts. The unmet need is severe — there is no FDA-approved second-line treatment for the ~85% of BTC patients without an actionable mutation C015C026, and off-label chemo regimens deliver ORR ~5% and OS ~6 months C015. The ORR effect size is large by oncology accelerated-approval standards. The PFS confirmation strengthens the read — ORR alone often stretches the FDA, but ORR plus a 56% reduction in progression risk plus a coherent explanation for OS confounding makes a regulatory denial of the path difficult to write.

The risk is the agency demanding a second confirmatory trial pre-approval rather than accepting confirmatory commitment post-approval. We assign ~25% probability to that outcome — material, but the unmet-need argument and the BICR-confirmed PFS effect tilt the decision toward acceptance.

10.1 Verified TAM

Compass estimates the U.S. 2L BTC opportunity at $1B+ C027. The estimate assumes a price point in the $150-200K annual treatment range typical of orphan oncology biologics and capture of a meaningful share of the ~15,000 non-targetable 2L patients per year C026. Penetration will be gated by BLA timing, commercial launch execution, and any later-arriving competitive entrants.

10.2 Prevalence Data

BTC affects roughly 26,500 U.S. patients annually C015, projected to grow to ~34,000 by 2037 C047. Approximately 90% receive first-line treatment; ~70% of first-line patients progress to second-line therapy. The second-line addressable population is approximately 15,000 patients per year. Approved targeted therapies address only the ~15-20% of BTC patients with FGFR2, IDH1, or HER2 alterations C042. The remaining ~85% have no FDA-approved second-line option C026.

10.3 SOC & Pricing

First-line standard of care is gemcitabine + cisplatin + an immune checkpoint inhibitor (durvalumab in TOPAZ-1, pembrolizumab in KEYNOTE-966), but two-year OS in TOPAZ-1 is 24.9% C040, so most patients reach a second-line decision. Off-label use of FOLFIRI delivered ORR 4.0%, PFS 2.1 months, and OS 5.7 months in 59 patients (Choi 2021, PMID 34303267) C016. Off-label FOLFOX in the same comparison delivered ORR 5.9%, PFS 2.8 months, and OS 6.2 months in 59 patients C017. Approved targeted therapies (pemigatinib, futibatinib, ivosidenib, zanidatamab) carry their own toxicity profiles including ocular toxicity and hyperphosphatemia C042.

11.1 Competitor Stage Map

Tovecimig's COMPANION-002 results lift the bar on every metric versus existing 2L BTC chemotherapy. ORR rises from ~5% to 17.1%. PFS rises from ~2.5 to 4.7 months. OS in patients who actually received tovecimig (the 85% pooled cohort and the 12.8-month crossover subset) C005C006 tracks materially above the 6.1-month non-crossover control and the published off-label benchmarks. If approved, tovecimig becomes the first targeted-pathway therapy for the non-genomically-defined majority of 2L BTC.

Competitor exposure is limited but worth tracking. Ivonescimab (Akeso / Summit) is the most-watched dual-pathway angiogenic-immunology bispecific (PD-1 x VEGF-A) but is positioned in NSCLC rather than BTC, and Compass demonstrated preclinical superiority versus ivonescimab in a head-to-head mouse PD-1/PD-L1 humanized model C024. No DLL4 x VEGF bispecific competitor has reached Phase 2 in BTC.

11.2 Differentiator Durability

The differentiator (DLL4 x VEGF-A in BTC) is durable on three vectors. First, orphan exclusivity at approval delivers a seven-year U.S. market window. Second, no DLL4 x VEGF-A bispecific competitor has even reached Phase 2 in BTC, so any competitor would face a 4-6 year clinical development gap. Third, the 26,500-patient BTC market C015 is small enough that it does not attract the same competitive density as larger oncology indications, while still being large enough to support a $1B+ commercial opportunity C027.

12.1 Contradicted Findings

The Korean Phase 2 ORR was ~37.5% C039; the U.S. COMPANION-002 ORR was 17.1% C001. The 20-point gap suggests the larger U.S. registrational study captured a more representative patient population and a more accurate effect size. Investors who anchored to the Korean numbers will see today's read as a disappointment; the more disciplined view is that the U.S. data is the relevant base case. No findings are directly contradicted between the 8-K disclosure and the underlying SEC/EDGAR/CT.gov primary records.

12.2 Unsupported Claims

The "$1B+" 2L BTC opportunity C027 rests on company-commissioned market research that is not independently verifiable; treat it as a company estimate. The 4-PR Phase 1b detail C038 and the 37.5% ORR Korean Phase 2 C039 come from CCF 2024 conference posters, which are not peer-reviewed in the traditional sense (T3 evidence). The SCLC complete-response narrative on slide 30 conflates the CTX-471 program with CTX-8371 C025 — the SCLC CR is from CTX-471, not CTX-8371; a careful reader should not confuse the two programs.

12.3 Existential vs. Manageable

The buy-side thesis breaks if any of these scenarios plays out:

• Existential — FDA demands a confirmatory trial pre-approval. The OS miss C003 creates agency room to argue ORR alone is insufficient. We model 25% probability; outcome would defer approval by 24-36 months and force a dilutive raise.
• Material — The full DoR data underwhelms. Complete dataset including Duration of Response will be presented at a medical conference later in 2026 C041. Short DoR in a small ORR population could weaken the BLA filing.
• Manageable — Hypertension labeling restricts utilization. 44% related Grade 3+ hypertension C010 will likely produce a black-box warning or prominent safety section.
• Material — Single-trial registration risk. COMPANION-002 is a 168-patient single pivotal study C008. The FDA can decline single-trial packages, though the strength of unmet need and the PFS effect mitigate.
• Manageable — Capital structure dilution if launch slips. $208.9M cash funds 38 months at current burn C018C019, but commercial pre-launch spend will accelerate burn meaningfully.
• Manageable — Open-label trial design. Only the radiology assessor was blinded C033. Patient and investigator awareness could bias subjective endpoints; ORR by BICR reduces this concern materially.

13.1 Platform Beyond Lead

Compass's pipeline provides cheap optionality beyond tovecimig BTC. The corporate pipeline includes three other clinical-stage assets C024: CTX-471 (CD137/4-1BB agonist) advancing to Phase 2 NCAM (CD56)+ basket trial in mid-2026 with prior Phase 1b post-checkpoint cohort showing four PRs and one complete response in small cell lung cancer; CTX-8371 (PD-1 x PD-L1 bispecific) with Phase 1 producing three responses in the first 15 patients including confirmed PRs in NSCLC, TNBC, and Hodgkin lymphoma C025; and CTX-10726 (PD-1 x VEGF-A bispecific) with IND cleared by FDA in early 2026 and Phase 1 initiated Q1 2026 C024.

13.2 Repurposing Potential

Tovecimig itself has expansion paths beyond BTC. A Phase 2 trial in DLL4+ solid tumors (CRC, gastric, ovarian, renal, HCC) is planned for mid-2026 initiation C024. An MD Anderson-led investigator-sponsored trial is enrolling patients to study tovecimig added to first-line gemcitabine / cisplatin / durvalumab in BTC C050 — if positive, this opens the much larger 1L BTC market without requiring another Compass-funded registrational trial.

13.3 Strategic Acquirer Logic

Compass's profile — BLA-stage oncology asset in a high-unmet-need orphan indication, small but growing market, pipeline of three additional clinical bispecifics, $208.9M cash, $196.8M stockholders' equity C018C021 — fits the acquirer logic for mid-cap oncology specialists or pharma companies seeking to expand their gastrointestinal oncology footprint. Companies such as AstraZeneca (durvalumab in TOPAZ-1), Merck (pembrolizumab in KEYNOTE-966), Servier, Helsinn, and Incyte all maintain BTC presence and could find tovecimig complementary. The 26,500-patient U.S. BTC market C015 fits an orphan-style commercial scale that mid-cap oncology specialists can execute. M&A interest will likely intensify post-BLA acceptance.

The trade structures cleanly. Position-sizing should reflect single-trial single-indication binary risk; weighted around 60-75% probability of FDA acceptance of the BLA path over a 6-9 month window suggests holding a weight calibrated to the position's binary-gap valuation. The capital structure (178M shares, $208.9M cash, no debt) C020C018 frames a clean enterprise-value calculation that incorporates expected probability-weighted launch.