Equity Research

Immunic — two trials, one bar

Immunic (IMUX) has one asset that matters: vidofludimus calcium, a once-daily oral DHODH inhibitor for relapsing multiple sclerosis. Its two Phase 3 trials, ENSURE-1 and ENSURE-2, read out together: topline data are guided by the end of 2026, with registry primary completion estimated for November 2026. Each trial needs an observed hazard ratio of about 0.74 or lower on time to first relapse — roughly what teriflunomide did against placebo — and the statistical design behind that bar is undisclosed, so the briefing computes it under stated assumptions. From the MRI data, the class precedent and the 2024 futility pass, the model gives both trials winning 62% and at least one 87%, and values the share in each outcome: $14.34 if both are positive, $3.96 if they split, $1.00 if both fail. That weights to $9.99 a share against a spot of $7.63 (7 October 2026 close); on base economics the market implies about a 44% chance of a clean win. Disclosure (updated 2026-10-09): the author holds a long position in IMUX, entered after the briefing was written.

Annexon — nineteen events, one verdict

Annexon (ANNX) is a complement company with two products near a decision. Vonaprument (ANX007) is an eye injection for geographic atrophy; its Phase 2 missed its primary endpoint, and the vision-loss signal behind the stock — 4 of 89 patients against 15 of 89 on sham — came from a secondary analysis on 27 events. In the fourth quarter of 2026 the independent data monitoring committee for the Phase 3 ARCHER II gives one verdict on the Month 15 primary: met, continue to Month 24, or futility, with no patient-level data either way. Tanruprubart (ANX005) for Guillain-Barré syndrome has a US filing targeted for the same quarter and an EU application under review. The briefing computes the trial odds live and values the share in each verdict state: $5.69 if the DMC says met, $1.15 on continue, $0.52 on futility. At the model's 42% chance of "met" that weights to $2.96 a share against a spot of $4.19 (2 October 2026 close); at spot the market implies about 68%. Disclosure: the author holds no position in ANNX.

Monte Rosa — two degraders, no disease efficacy yet

Monte Rosa Therapeutics (GLUE) makes molecular-glue degraders. MRT-8102 degrades NEK7 and cuts hsCRP by the same ~80–85% the oral NLRP3 class reports, but after ZEUS that carries no cardiovascular value on its own; its first disease-level test is a ~60-patient gout-prevention study reading out in the second half of 2027. MRT-6160, a VAV1 degrader, has seven days of healthy-volunteer dosing behind it, and Novartis runs, funds and can end its two Phase 2s, which read out around December 2028 and February 2029; about 85% of its base value to Monte Rosa is milestones under a licence whose amounts are redacted. The briefing is a live sum of parts: base $8.42 a share against a spot of $11.26 (1 October 2026 close), bear $4.63, bull $35.79. At spot the market prices about $0.37B for MRT-8102, above the model's $0.02B base. Disclosure: the author holds no position in GLUE.

VK2735 — an early lead on small numbers

Viking Therapeutics' VK2735 is a GLP-1/GIP dual agonist, the mechanism of Lilly's tirzepatide, in Phase 3 for obesity with an oral tablet behind it. After the September 2026 maintenance study, bulls called it better than tirzepatide. The briefing tests that claim at matched weeks: VK2735 leads by three to five points of absolute weight loss even after both drugs have spent three months at top dose, but the week-33 figure rests on thirteen patients and an exploratory endpoint, and Lilly's triple agonist retatrutide files first. What is the lead worth in a market where net prices fall every year, and what is monthly maintenance dosing worth on top of it?

The model is built for a chronic, price-eroding market rather than a line of therapy: treated patients × share × months on therapy × net price, by formulation and geography, through a risk-adjusted net present value to a per-share figure that carries September's convertible notes one way only. Disclosure: the author holds no position in VKTX.

Varseta-M — the same bet, twice

CytomX's varsetatug masetecan (CX-2051) is a masked EpCAM antibody-drug conjugate with a topoisomerase-1 payload, headed for a registrational trial in late-line colorectal cancer in the first half of 2027. Merck KGaA's precemtabart tocentecan is a CEACAM5 antibody-drug conjugate with the same payload class, in the same line, and already dosing its Phase 3. AbbVie's telisotuzumab adizutecan started a third. The briefing is built around that fact: what is differentiation on confirmed response and on the absence of myelosuppression worth, against twelve to twenty-four months of lead time?

It runs a bottom-up model — indication × line × geography, bear, base and bull — through a risk-adjusted net present value to a per-share figure, then ranks every lever by how far it moves that figure. The competitive landscape table is monotherapy like-for-like, with a provenance tier on every cell; the tolerability radar carries its sources and scale in a legend; and the loss-of-exclusivity date is derived on the page from the patent families and regulatory exclusivity rather than assumed.

What this is, and what it is not

It is investment research, not investment advice, and not a recommendation to buy or sell any security. Disclosures: the author is long CTMX and IMUX (the IMUX position was entered after its briefing was written) and holds no position in VKTX, GLUE or ANNX. Every Varseta-M figure is Phase 1, single-arm and sponsor-reported, at a data cutoff of January 2026; the registrational design has not been agreed with the FDA; and the largest structural assumption on the page — whether the label reads third-line or fourth-line — is a slider precisely because nobody yet knows. Parameters the author set rather than sourced are flagged [A], and the page asks you to stress them before quoting anything.

The engine is a JavaScript model with a Python twin; the two are held to agree within one part in a million on every scenario before a page is published, and the Monte Carlo runs in your browser. Nothing is fetched at runtime.

Read it

Vidofludimus calcium (Immunic) — the interactive briefing →
The hazard ratio each ENSURE trial must clear by placebo relapse rate, the readout odds computed live from the Phase 2 and the passed interim, a value per share if both trials win, if they split and if both fail, the competitive landscape against teriflunomide and the BTK inhibitors, the cap table, catalysts, a Monte Carlo and sliders over the full parameter set. Opens in a new tab.

Vonaprument and tanruprubart (Annexon) — the interactive briefing →
The ARCHER Phase 2 audited by definition, the X debate adjudicated claim by claim, the DMC odds computed live with a slider for your own, a value per share for met, continue and futility, the GBS launch, a sum of parts, catalysts, a Monte Carlo and sliders over the full parameter set. Opens in a new tab.

MRT-8102 and MRT-6160 (Monte Rosa) — the interactive briefing →
The GFORCE-1 data audited, NEK7 degradation against the NLRP3 class, the Novartis VAV1 licence with known and redacted terms side by side, the Phase 2 timeline, a sum-of-parts bridge, a tornado of the levers, and sliders over the full parameter set. Opens in a new tab.

VK2735 (Viking Therapeutics) — the interactive briefing →
The superiority claim adjudicated at matched weeks, a like-for-like landscape with an estimand toggle, tolerability, maintenance economics, a failure-point monitor, revenue fan, stress grids, an rNPV waterfall, a tornado of the levers, and sliders over the full parameter set. Opens in a new tab.

Varseta-M (CX-2051) — the interactive briefing →
Thesis chart, competitive landscape, tolerability profile, headline outcomes, failure-point monitor, revenue fan and stress grids, an rNPV waterfall, a tornado of the levers, and sliders over the full parameter set. Opens in a new tab.