MNOV / mn-001-nafld-hypertriglyceridemia — MN-001 (tipelukast) for high blood triglycerides and fatty liver in people with type 2 diabetes

Program analysis · bpiq_drug_id 16136 · prepared 2026-08 · USD · framework v5.6.1 · Coverage: CLEARED

Company context, financials, ownership, options and the full pipeline: ../company.md.

How to read this. Written for a reader who has not studied pharmacology or finance. Terms specific to this program are in the Glossary below; general terms are defined once in framework/04-glossary.md. Tags: [VERIFIED — source] / [UNVERIFIED] / [WEB ESTIMATE — source, date].

Glossary

Only terms specific to this program. Generic terms are in framework/04-glossary.md.

Executive summary

• What it is (one sentence): An old oral anti-inflammatory molecule, repurposed to lower blood triglycerides and liver fat at the same time in people who have both problems alongside type 2 diabetes.
• The event and when (as disclosed): Topline results from the 40-patient placebo-controlled Phase 2 trial MN-001-NATG-202, guided by the company to “the third quarter of 2026” — a quarter, not a day. [VERIFIED — company release 2026-05-26] Last patient last visit was completed on 2026-05-26, so the clinical work is finished and only the analysis remains.
• The main reason it could work: In the company’s own earlier open-label study, the ten patients who had type 2 diabetes showed a 50.8% fall in triglycerides by Week 8, against 17.8% in the nine who did not. [VERIFIED — MediciNova poster, IDF World Diabetes Congress 2022, abstract LI2022-1192] If even part of that survives a placebo control, the triglyceride half of the readout should hit.
• The main risk: There are two co-primary endpoints, and the second one — liver fat — has no supporting human evidence at all. In that same earlier trial, liver fat measured by MRI went up by 0.64 percentage points over twelve weeks. [VERIFIED — ClinicalTrials.gov posted results, NCT02681055] And the trial’s other registered main endpoint, cholesterol efflux capacity, failed outright. [VERIFIED — same source, and J Atheroscler Thromb 2025, p = 0.6507]
• What it means for the stock: Less than the excitement would suggest. This is a 40-patient Phase 2 on a molecule whose composition-of-matter patent expired in 2009, and the company’s larger claim on the equity is the 234-patient ALS readout guided for year-end. The honest position is watch: the strict both-endpoints definition is unlikely to be met, but this ticker has historically risen on partial and even failed readouts, and the expected value sits within 3% of spot. There is no directional edge to take.

0. Program-tier coverage — CLEARED

One row per mandatory tool in the program-tier table in framework/02-connectors.md, in the order that file lists them. Company-tier coverage is in ../company.md C.0 — see there, not repeated.

Verdict: CLEARED. No mandatory row reads NOT CALLED. Two mandatory rows read BLOCKED, both after the full retry policy, both with the verbatim error recorded and the dependent claims tagged [UNVERIFIED]. Per 02-connectors.md § Three states, a BLOCKED tool does not make a program PROVISIONAL; it makes specific claims unverified, and those claims are named above.

A. Scientific & clinical assessment

A.1 Mechanism of action and scientific rationale

What the drug is. MN-001 (tipelukast) is a small molecule taken by mouth. It is not new. It was originated by Kyorin Pharmaceutical of Japan and licensed to MediciNova on 2002-03-14, and its first development programme was in asthma and interstitial cystitis, not in liver or lipid disease. [VERIFIED — MediciNova 10-K FY2025, filed 2026-03-10; and Drugs R&D 2007, PMID 17963431, [DOI 10.2165/00126839-200708060-00008](https://doi.org/10.2165/00126839-200708060-00008)] A Phase 3 programme in asthma was started in the United States in November 2006, in which 705 patients with mild to moderate asthma were to be randomised. [VERIFIED — same Drugs R&D record] No approved asthma product resulted, and no asthma indication appears anywhere in the company’s current pipeline. [VERIFIED — BPIQ fetch_company_drugs, read 2026-08-13, ten rows, none in asthma] That is a twenty-year record of the molecule not reaching a market, and it is context a reader is entitled to before anything below.

How it is said to work — and why “said to” is the right phrasing. The published pharmacology attributes an unusually long list of mechanisms to this one molecule: antagonism of leukotriene receptors, inhibition of phosphodiesterase 3 and phosphodiesterase 4, inhibition of 5-lipoxygenase, and inhibition of eosinophil migration. Older descriptions add phospholipase C and thromboxane A2. [VERIFIED — Drugs R&D 2007, PMID 17963431, [DOI 10.2165/00126839-200708060-00008](https://doi.org/10.2165/00126839-200708060-00008); and the MediciNova IDF 2022 poster, which states "leukotriene receptor antagonism, phosphodiesterase 3 and 4 inhibition, and 5 lipoxygenase inhibition"]

None of those is the mechanism the metabolic effect is attributed to. The metabolic story is a different one, and it belongs to the metabolite:

• MN-002, the active metabolite of MN-001, increases how much ABCA1 and ABCG1 protein cells make, and increases ApoA-I-mediated cholesterol efflux, independently of protein kinase A. Molecular docking simulations in the same paper suggest MN-002 may bind PPAR-alpha, a nuclear receptor central to fat metabolism. [VERIFIED — J Atheroscler Thromb 2025, PMID 40850750, [DOI 10.5551/jat.65669](https://doi.org/10.5551/jat.65669)]
• In vitro work reported by the sponsor states that MN-001 downregulates CD36 and upregulates ABCG1 messenger RNA, both described as highly associated with type 2 diabetes. [VERIFIED — MediciNova IDF 2022 poster, abstract LI2022-1192, citing Ogura, The Liver Meeting 2021 and the 19th International Symposium on Atherosclerosis 2021]

A drug with five described mechanisms and no established primary one is not mechanistically clear. That is a statement about the evidence, not a criticism of the chemistry. It matters here because the two co-primary endpoints of the pending trial rest on two different proposed mechanisms — cholesterol/lipid handling for triglycerides, fatty-acid uptake for liver fat — and neither has been isolated as the operative one in a human being.

How well the target is validated. Weakly, and this document cannot do better than say so. Open Targets was BLOCKED on three attempts (section 0), so no independent human-genetic evidence for any of these targets against either endpoint could be obtained. ChEMBL was BLOCKED on four attempts, so the molecule’s off-target selectivity profile is not independently confirmed either. What remains is: an approved drug class exists for one named mechanism (leukotriene receptor antagonists such as montelukast are marketed asthma drugs, which validates the target as druggable but says nothing about liver fat or triglycerides); and the metabolic mechanism rests on the sponsor’s own in vitro experiments plus a docking simulation. [UNVERIFIED — no independent target-validation source was obtainable this session]

The exact scientific step the next readout must prove. That 500 mg/day of MN-001 for 24 weeks, against placebo, in adults who have all three of NAFLD, type 2 diabetes and raised triglycerides, produces both (1) a fall in liver fat as measured by CAP score at Week 24 and (2) a fall in fasting serum triglycerides at Week 24. [VERIFIED — ClinicalTrials.gov NCT05464784, primary outcome measures, read 2026-08-13]

The honest scientific risk, stated plainly. The sponsor has already run this drug in this disease area once, and the results of that trial are public in three places that do not fully agree with each other. Taking the registry — the source the sponsor filed with the government — at face value:

• Liver fat went up. Liver fat by MRI, baseline to Week 12, changed by +0.64 percentage points (standard deviation 5.54) in 19 patients. [VERIFIED — ClinicalTrials.gov posted results, NCT02681055, secondary outcome "Measure the Effect of MN-001/002 on Percentage of Fat in the Liver"] That is the direct conceptual predecessor of the CAP-score co-primary now pending. It is not a signal; it is slightly the wrong direction.
• The other main endpoint failed. Cholesterol efflux capacity, baseline to Week 12, changed by −0.013 (standard deviation 0.1022) — essentially nothing. [VERIFIED — same posted results, primary outcome] The 2025 paper puts a p-value on the same question in diabetic patients: p = 0.6507, described in the abstract as “no significant improvement”. [VERIFIED — J Atheroscler Thromb 2025, PMID 40850750, [DOI 10.5551/jat.65669](https://doi.org/10.5551/jat.65669)]

The comparator for both statements is the drug’s own earlier data, which is the strongest available comparator and the one rule 17 asks for. The scientific risk is therefore specific rather than generic: the triglyceride half of this readout has a real prior signal and the liver-fat half has a prior result pointing the wrong way, and the trial is formally judged on both.

A.2 Clinical development plan, timeline, feasibility, resourcing

Trials

First and last patient in, last patient out, database lock and topline. First patient in 2022-08-22 [VERIFIED — NCT05464784 start date]. Last patient in on or before 2025-11-04, the date the company announced enrolment complete [VERIFIED — company release 2025-11-04]. Last patient last visit 2026-05-26 [VERIFIED — company release 2026-05-26]. Database lock is not disclosed [UNVERIFIED]. Topline is guided to Q3 2026 and is modelled in the Readout section below.

A source conflict that is not smoothed over. The predecessor trial’s triglyceride result is reported two different ways by the same sponsor, in the same units, for the same 19 patients, at the same Week-8 timepoint:

The two differ by a factor of 6.4. Both are sponsor-supplied. The registry entry names the same denominator (19 participants), the same unit (mg/dL) and the same timepoint, so the obvious reconciliations — a different population, a different visit, a percentage misread as an absolute — do not survive inspection. This conflict is reported, not resolved and not averaged (01-rules.md rule 24). It matters because the poster figure is the number the entire investment case rests on, and the registry figure — the one filed with the government — is six times smaller. Worth noting in fairness: the registry figure is still statistically meaningful. A mean of −21.67 with a standard deviation of 27.89 on n = 19 gives a standard error of 6.40 and t ≈ −3.4. [UNVERIFIED — arithmetic on the two verified registry figures] So the smaller number is a real effect; it is just a far more modest one.

Feasibility matrix

Resourcing sufficiency. Not a constraint for this readout, and a severe one immediately after it. The company holds $27.3 million with a runway to roughly 2027-12 on its own guided spend (see ../company.md C.3), and the remaining work on NATG-202 is analysis of a completed dataset, which costs very little. So the readout itself is fully funded.

What is not funded is anything that follows it. A registration-quality trial in this disease area means biopsy histology in hundreds of patients per arm, which is what MAESTRO-NASH and ESSENCE both did. MediciNova’s entire guided operating spend for 2026 is $16.2 million across ten pipeline rows (../company.md C.3). It cannot run that trial. The company has said as much about its other Phase 3-ready asset, in as many words: the BPIQ note on the progressive multiple sclerosis row reads “plan to find partner to help fund Ph3 trial”. [VERIFIED — BPIQ fetch_company_drugs, row 15141, read 2026-08-13] A positive readout here is therefore a partnering event, not a development event, and that is the correct frame for everything in section B.

A.3 Target label, target product profile, strategic questions, designations

A.3a Target label. Adults with non-alcoholic fatty liver disease and raised fasting triglycerides in the setting of type 2 diabetes mellitus, already established on stable oral antidiabetic therapy. [VERIFIED — NCT05464784 title and eligibility criteria]

A.3b Target product profile

A.3c Strategic Go/No-Go questions. The asset’s next decision is Go-to-Phase-III, so the second table is the live one. The first is answered retrospectively, because the Phase 2 decision was taken in 2022 and how it was taken is informative.

Pre-Phase-II (Go-to-Phase-II) — answered retrospectively, as of the 2022 decision:

Pre-Phase-III (Go-to-Phase-III / registration) — the live decision, which the pending readout is the input to:

A.3d Regulatory designations. None, for this program. No Fast Track, Breakthrough Therapy, Orphan Drug or other designation for MN-001 in NAFLD or hypertriglyceridemia was found in the company’s 168-item press feed, in the FY2025 10-K, or in the ClinicalTrials.gov records for either trial. [VERIFIED — absence checked across BPIQ fetch_company_press_releases, the 10-K FY2025 and both registry records, 2026-08-13]

For contrast, and to show the absence is meaningful rather than an artefact of not looking: the company’s other molecule has such designations, and announces them. MN-166 (ibudilast) holds Orphan Drug designation in glioblastoma (October 2018) and received FDA Fast Track designation for methamphetamine dependence. [VERIFIED — BPIQ fetch_company_press_releases, items dated 2018-10-06 and one undated ADVFN archive item, read 2026-08-13] MN-001 in this indication has nothing comparable, which is consistent with a Phase 2 asset in a non-orphan metabolic indication.

A.4 Target product profile valuation matrix

Calibration examples, not asset claims: in oncology each incremental month of overall survival is worth roughly $150–300M in peak-sales potential; oral formulations command roughly 15–25% price premiums over injectables. The second of those is the only one that bears on this asset at all, and it is a reason MN-001 has some positioning against weekly injectable semaglutide.

A.5 Evidence quality and endpoints

Endpoints

A.5b Key opinion leaders.

Panel as of. 2026-08-13 — the date the investigator and independent-voice searches below were run.

Investigators

No investigators were returned for this program’s pivotal trial, and that is a finding rather than an omission. get_trial_details on NCT05464784 returns two facilities with no contacts, no overall officials and no responsible-party investigator named — an unusually bare registry record for a trial that has completed its clinical phase. search_investigators across the condition returned exactly one person, Junbo Ge of Zhongshan Hospital, Shanghai, and he is the contact on NCT07693777, an unrelated Phase 3 of a different company’s drug (DR10624) in severe hypertriglyceridemia — not this program, and not recorded here.

Independent voices

None found, and the search that failed to find one is worth describing. Every named person attached to this program’s data is a MediciNova employee. The IDF 2022 poster carrying the pivotal prior result names three authors — Kazuko Matsuda, Malath Makhay and Yuichi Iwaki — and its own disclosure section states “Matsuda K, Makhay M, and Iwaki Y are employees of MediciNova, Inc.” [VERIFIED — the poster itself, read 2026-08-13] Matsuda is the Chief Medical Officer and Iwaki the President and Chief Executive Officer. [VERIFIED — BPIQ fetch_company_insider_transactions, executive titles, read 2026-08-13]

Two academic candidates were considered and excluded: Masatsune Ogura and Takashi Miida of the Department of Clinical Laboratory Technology, Juntendo University, co-authors on the 2025 cholesterol-efflux paper. [VERIFIED — PubMed PMID 40850750, author affiliations] They are excluded because MediciNova announced a formal MN-001 research collaboration with Juntendo University School of Medicine on 2022-06-22 [VERIFIED — company release 2022-06-22], which is a disclosed relationship with the sponsor. 03b-program-spec.md is explicit that a person with a disclosed relationship belongs in the Investigators table rather than among voices held out as independent; they are not investigators on this trial either, so they appear in neither table and are named here instead.

Judgement

Dissent

Empty. Endpoint supported is UNKNOWN rather than SUPPORTIVE_CONTESTED, so no dissent is owed, and inventing a disagreement nobody has expressed would be worse than an empty table.

B. Commercial assessment

B.0 Current treatment algorithm

What a patient with type 2 diabetes, fatty liver and raised triglycerides actually receives today, in order:

1. Lifestyle first, for both problems. Weight loss, diet and exercise are guideline-recommended first-line therapy for NAFLD and for raised triglycerides alike. Adherence is poor, which is why the drug market exists at all. [WEB ESTIMATE — MASLD treatment reviews, read 2026-08-13]
2. Glycaemic control, which now doubles as liver therapy. Metformin first, then — and this is the commercially decisive fact — GLP-1 receptor agonists (semaglutide, now approved for MASH), dual GIP/GLP-1 agonists (tirzepatide), SGLT2 inhibitors and pioglitazone. All of these reduce liver fat, and several improve steatohepatitis. [WEB ESTIMATE — AASLD practice guidance November 2025 update, PMID 41201884, and Current Diabetes Reports 2026, read 2026-08-13]
3. Lipid management. A statin for cardiovascular risk regardless of triglycerides. For triglycerides in the 150–499 mg/dL band, the primary lever is treating the diabetes itself; for ≥500 mg/dL, a fibrate, icosapent ethyl or prescription omega-3. [WEB ESTIMATE — hypertriglyceridemia treatment reviews, read 2026-08-13]
4. Liver-specific drug therapy. Resmetirom for non-cirrhotic MASH with F2–F3 fibrosis; semaglutide 2.4 mg for MASH. Neither is labelled for plain NAFLD without steatohepatitis, which is the population NATG-202 enrols. [WEB ESTIMATE — MASLD approval reviews, read 2026-08-13]

Exactly where MN-001 would fit, and the problem with that spot. It would be an add-on oral to a patient already on stable oral antidiabetic therapy — which is not an inference, it is the trial’s own entry criterion. So the comparator is not “nothing”; the comparator is escalating the existing antidiabetic to a GLP-1 receptor agonist, which addresses blood sugar, weight, liver fat and cardiovascular risk simultaneously, is already approved for MASH, and is already in the prescriber’s hand. MN-001 would add a pill and address two laboratory numbers. That is the commercial problem in one sentence, and no readout can fix it — only a partner with a strategic reason could.

B.1 Competitive landscape and positioning

Where this asset wins, and the single fact the thesis rests on. The genuine differentiation is positional rather than mechanistic: MN-001 is aimed squarely at the intersection of raised triglycerides and fatty liver in a diabetic patient, as a single cheap oral add-on, and no approved drug is labelled for that intersection. If the poster’s numbers are the right ones, the triglyceride effect in diabetics (−50.8% at Week 8) is far larger than icosapent ethyl’s, which the literature describes as modest.

The single fact the thesis rests on is that the 50.8% Week-8 triglyceride reduction seen open-label in ten diabetic patients survives a placebo control. That fact carries three specific problems, all of them checkable. First, the sponsor’s own registry filing reports the all-patient version of the same endpoint as −21.67 mg/dL rather than −138.8 mg/dL (A.2). Second, the subgroup’s baseline was 444.7 mg/dL — near-severe hypertriglyceridemia — and an uncontrolled fall from a high baseline in ten people is the textbook setting for regression to the mean, with no placebo arm to separate the two. Third, the between-group difference the poster reports for triglycerides was p = 0.098, which is not statistically significant; the only comparison in the entire prior dataset that reached significance was HDL, at p<0.0002. [VERIFIED — MediciNova IDF 2022 poster, read 2026-08-13]

Calibration example, not an asset claim: oncology first-movers in novel mechanisms have shown roughly 3.2× higher peak-sales potential but about 1.8× higher development risk. Not applicable here — this asset is neither first-mover nor novel-mechanism.

B.2 Addressable market

Launch markets would be the United States first, then Europe. Japan, China, South Korea and Taiwan are outside MediciNova’s territory entirely under the Kyorin licence, which removes the largest non-US diabetes market from the opportunity before any commercial modelling starts. [VERIFIED — 10-K FY2025, MN-001 licence terms]

B.3 Value and feasibility

B.3a Expected peak sales. Built bottom-up as eligible patients × annual net price × peak penetration, with every input tagged and the undefensible ones named. All figures are conditional on approval, which requires a Phase 3 that has not been designed or funded. All figures exclude Japan, China, South Korea and Taiwan (outside MediciNova’s territory) and exclude the royalty owed to Kyorin, whose rate is not disclosed anywhere (see B.3b).

One third-party figure is rejected rather than passed through, under 01-rules.md rule 6. A market-research source gives the hypertriglyceridemia therapeutics market as $12.82 billion in 2026 rising to $18.29 billion by 2033. [WEB ESTIMATE — Coherent Market Insights via web search, read 2026-08-13] That aggregate necessarily includes generic fibrates, statins and over-the-counter omega-3s, none of which is an addressable branded opportunity for this asset, and quoting it beside a Phase 2 asset would imply a scale that does not exist. The narrower figure — a severe hypertriglyceridemia market of roughly $1.4 billion across the seven major markets in 2023 [WEB ESTIMATE — DelveInsight via web search, read 2026-08-13] — is more usable but still does not fit: NATG-202 enrols patients from 151 mg/dL, and severe means ≥500 mg/dL.

A range this wide is the honest output. Rule 10 forbids a point estimate on unverified inputs, and the two inputs that would narrow it — a verified patient count and a defensible net price for a method-of-use-only oral — are both missing. What the range does establish is a useful comparison: MediciNova’s entire enterprise value is $38.1 million (../company.md C.4), which is below the low end of even the pessimistic scenario. The market is pricing close to zero probability on all of this, across all ten pipeline rows.

B.3b Feasibility

B.4 Product-development risk

Framing questions.

1. Does the development plan support the target product profile claims? Partly, and it stops short. NATG-202 is a well-shaped test of the two laboratory claims — randomised, blinded, placebo-controlled — so it can support or refute “lowers triglycerides and liver fat”. It cannot support the claims that matter commercially, which are in the payer row of A.3b: nothing in the plan compares MN-001 against a GLP-1 agonist, measures a patient-reported outcome, or generates the histological evidence every approved agent in this area was approved on.
2. Will the identified risks affect the target product profile? Yes, and one of them is decisive. The endpoint risk is not a delay risk, it is a profile risk: if CAP score does not move, the target profile collapses from “one pill for both problems” to “another triglyceride drug”, and the triglyceride market is served by generics. The IP risk compounds it — a method-of-use-only position cannot defend a branded price for a plain lipid indication.
3. If a risk cannot be mitigated, is the asset still differentiated from competitors? On the triglyceride claim alone, barely. A large triglyceride effect in diabetics would still be interesting and would still be positionally distinct from the approved MASH drugs. But without a liver-fat effect it is competing with generic fibrates and icosapent ethyl on price, with no composition-of-matter protection. That is differentiation without defensibility.

Readout

Sources

The modelled estimate. 01-rules.md rule 34 asks for the registry’s primary_completion_date plus the lag to database lock and analysis. That input cannot be used here and the substitution is stated rather than hidden: the registry’s primary_completion_date of 2026-12-31 is contradicted by the sponsor’s own announcement that the last patient’s last visit occurred on 2026-05-26, so adding a lag to it would model a readout after an event that has already happened. The arithmetic used instead takes the same conceptual input from the primary source that supersedes the registry:

Last patient last visit 2026-05-26 + 2 to 4 months to database lock, unblinding and topline analysis = 2026-07-26 to 2026-09-26. data/benchmarks/readout-lag.json holds no observation for a comparable trial (its observations array is empty by design), so the stated default lag of two to four months applies and the tag is [UNVERIFIED — modelled, default lag] rather than benchmarked n=N.

Note what this arithmetic already tells us: the front edge of the modelled range, 2026-07-26, has already passed without an announcement. That is why earliest below sits at the three-month mark rather than the two-month mark.

Window

Basis. earliest is last patient last visit plus three months, because the two-month edge of the modelled range (2026-07-26) has already elapsed with no announcement, so the earliest still-live date is the next round point in the same arithmetic. likeliest sits in the last week of September: inside the company’s guided quarter, consistent with the modelled range’s back half, and consistent with this sponsor’s habit of announcing at the end of a guided period rather than the start. latest extends five weeks past the guided quarter end to allow one further slip of the size already on record — the one documented slip moved guidance by about a month — while stopping short of AASLD (2026-11-05), which the abstract-deadline argument above makes an unlikely venue. Precision is PERIOD because no source names a day or a month for the topline: the company names a quarter and BPIQ synthesizes that quarter’s last day. Confidence is MEDIUM because three of the four obtainable sources agree on the quarter and the clinical phase is verifiably finished, but the window rests on an undisclosed database-lock date and one source contradicts the rest outright.

Disagreement. URESOLVED. bpiq, company and modelled all point to the third quarter of 2026; ctgov points to 2026-12-31 and, with the default lag applied, to 2027-02 through 2027-04. The disagreement is not averaged and is not resolved by picking one (01-rules.md rule 24): the registry is reported as it reads, and the window above is built from the three sources that agree while the fourth is recorded as contradicting them. The reason to believe the three over the one is stated rather than assumed — a company announcement that the last patient completed their last visit is a statement about an event that has occurred, whereas a registry completion date is a sponsor-maintained projection that this sponsor has not updated.

Date slippage. One slip.

[VERIFIED — BPIQ fetch_company_drugs note field, row 16136, read 2026-08-13] Two dated statements, one transition, and that transition is a genuine slip rather than a reiteration: “summer 2026” ends in August in the northern hemisphere and “Q3 2026” ends on 2026-09-30, so guidance moved back by roughly a month. One slip is a good record by the standards of this corpus, and it is worth saying so plainly rather than only reporting the count.

Attribution

Status.

Conflicts

Transcribed from lib/clustering.mjs’s attributionFor("MNOV", <MNOV.json>, [this program], CATALYST_CLUSTER_MIN_MONTHS, 16136), run 2026-08-13. Every field below is computed, none typed from a guess.

Note. Not required for INDETERMINATE, and deliberately not written. What the reader needs instead is in the stock-direction call below, where rule 36 requires it.

What is actually going on here, for a reader who wants to know. Row 14393 is COMBAT-ALS, the 234-patient Phase 2b/3 in amyotrophic lateral sclerosis guided to year-end 2026, and row 14184 is the OXTOX Phase 2b in chemotherapy-induced peripheral neuropathy guided to H2 2026. Neither has been analysed, so the clustering pass widens each to its whole enclosing half-year, 2026-07-01 to 2026-12-31, which swallows this program’s window entirely — hence a gap of zero days on both. Neither side of either pair rests on a disclosed date: all three catalysts are period placeholders, so confirmed is false on both and the status is INDETERMINATE rather than CONTAMINATED. That is the correct reading and it is not reassuring: COMBAT-ALS is by far the larger claim on the equity, and the reason its collision with this program cannot be confirmed is simply that nobody has disclosed a date for either.

Market and timing for this event

Company-level figures live in ../company.md and are cited by section here rather than restated (01-rules.md rule 26).

• Plain takeaway. The market is not positioned for this readout in either direction. The shares sit at 20.3% of their 52-week range (../company.md C.4), no hedge fund holds a position at all and institutions hold 11.2% (C.5), the options chain cannot price the event because its nearest strike is 88% above spot (C.6), and average dollar volume is roughly $52,650 a day (C.4). The one thing that has moved is short interest, which tripled in six weeks to 8.75 days of average volume (C.5). Nothing here says the readout is priced in; nothing says it is priced out either. It says almost nobody is watching.
• Months to this catalyst. 0.4 months — 13 days — from 2026-08-13 to readout.window.earliest of 2026-08-26. Say plainly: readout.precision is PERIOD, so this is 13 days to the front edge of a nine-week window, not 13 days to a known event. The likeliest date is 1.4 months out and the latest is 2.6 months out.
• Expected move around this event. The chain cannot price it, so a bracket is given instead of a number. ../company.md C.6 records the whole chain as unusable: the nearest strike to a $1.33 share price is $2.50, no at-the-money straddle can be constructed, and all ten put contracts sit on the documented 0.01488 implied-volatility floor artifact. The expiries nearest this catalyst (2026-09-18 and 2026-10-16) carry 0 and 14 contracts of open interest at the $2.50 strike. The bracket, taken from this ticker’s own eleven-row catalyst history in C.7, is −5% to +10% intraday for an ordinary readout, with an open gap of up to +25% for a headline the company can present as a win. [UNVERIFIED — bracket from C.7's own range, not an options-implied figure]
• Nearest comparable past reaction. 2026-05-26 is the closest in subject — the same program’s own last-patient-last-visit announcement — and it traded −2.14% with a zero open gap (../company.md C.7). But the closest in shape is 2018-04-02, and it is the more instructive of the two: the same molecule, the same trial (NATG-201), the same two-co-primary structure, announced as “achieving one of the two main endpoints (data from the second endpoint has not been disclosed)” — and the stock opened +25.24% and closed +8.98%. It is comparable because the structural set-up is identical and the sponsor’s disclosure behaviour is the thing being predicted. It is not comparable in one important way, which has to be said: the withheld endpoint in 2018 is now public and it failed, so the market paid +25% for a result it could not fully see. Also, the 2022-12-07 announcement of the lipid-panel findings from the same trial — the closest analogue for a triglyceride result specifically — paid a much more modest +5.18%.
• Materiality. Meaningful, not dominant, as recorded in ../company.md C.2: a 40-patient Phase 2 in a crowded metabolic indication on a molecule with no composition-of-matter protection, whose news value is its timing rather than its size, against a 234-patient ALS readout guided for year-end that is the larger claim on the equity. The stock-direction call below is consistent with this and is deliberately not a large one.
• Date slippage. One slip, from “top-line data by summer 2026” (2025-11-04) to “Q3 2026” (2026-05-26). See readout.slips above. One slip on a program whose clinical phase is complete is a good record.

Spot. $1.33, read 2026-08-13, cited from ../company.md C.1 and C.4. Previous close $1.32. Every figure below is measured against $1.33.

Scenario prices

Expected value. $1.37, +2.6% against spot of $1.33. Method: probability_pct of 15% applied to the midpoint of each scenario range — 0.15 × $1.875 + 0.85 × $1.275 = $1.365. This is arithmetic, not advice, and it is not a price target. Read it with the caveat in the Miss row: the miss range is internally bimodal, so its arithmetic midpoint is a cruder summary than the positive range’s is.

Run-up

A run-up call is written, and its own priority score of 9 out of 100 is the finding. 01-rules.md rule 39 withholds a run-up call only when date_confidence floors at zero, which happens exactly when readout.precision is UNKNOWN. Here precision is PERIOD, so the driver scores 20 rather than 0 and the call is permitted. What the arithmetic then says is that there is almost nothing left to trade: the exit rule resolves to roughly four trading days after the entry.

Entry and exit

exit on the prediction record is null, and for a stated reason rather than an omission: lib/runup.mjs’s resolveExit requires a committed price series, and data/prices/MNOV.json does not exist (../company.md C.8). No close can be read for the resolved date, so the plan is recorded and what it resolved to is left null until the cache covers this ticker.

Predicted move and peak

Priority score drivers

Five of the seven are transcribed from lib/runup.mjs’s scoreDriver, run 2026-08-13; two are the analyst’s own judgement read from this document.

Priority score. Transcribed from lib/runup.mjs’s priorityScore, run 2026-08-13 — not hand-computed.

Priority score 9 · formula_version 1.0.0

How 9 comes about, since a single digit deserves an explanation. The six weighted drivers combine to a base of 46.05, which is unremarkable rather than bad. Then date_confidence’s structural gate multiplies it by 0.20, giving 9.21. The additional hard ceiling of 15 does not bind, because 9.21 is already below it. A quarter-precision readout date is what sinks this, not the science and not the balance sheet — and it is the correct answer: you cannot time an entry and an exit around a nine-week window, and the exit rule resolving to four trading days after the entry is the same fact stated a second way.

Settlement. Left null at lock time. Settled only on an explicit user request, from a confirmed primary source, against the committed price cache — which for this ticker does not yet exist (../company.md C.8), so a run-up settlement here is blocked until scripts/fetch-prices.mjs covers MNOV.

Verdict

What I would do. Watch.

Why. The strict test this readout will be scored on is that both co-primary endpoints hit, and the sponsor’s own prior data in this exact population says one of them is unlikely to: liver fat measured by MRI went up 0.64 percentage points over twelve weeks, and the same trial’s other registered main endpoint failed outright. Against that, the triglyceride endpoint has a genuine prior signal — though the sponsor reports its magnitude two ways that differ by a factor of 6.4, which is a serious problem with the one number the whole case rests on. The asset is also structurally hard to own even if it works: composition-of-matter protection expired in 2009, leaving method-of-use patents only; the company cannot fund the histology-based Phase 3 the regulatory precedent requires; and the clinical comparator is escalating a drug the patient is already taking to an approved GLP-1 agonist that does more. Finally, the trade itself has almost no room left — the exit rule resolves four trading days after the entry, and the priority score is 9 out of 100.

What would change this. A CAP-score reduction significant against placebo at Week 24 would change the whole thesis, because it is the endpoint with no supporting evidence and therefore the one carrying all the information. The second thing that would change it is a named partner: this program’s value can only be realised through one, so a licence or co-development deal would matter more than the readout it followed. A third, narrower trigger: the sponsor reconciling its own two triglyceride figures, or publishing NATG-201 in a peer-reviewed journal, would remove the largest single evidence problem in this document.

What to watch.

• Now to 2026-08-26 — the Q2 2026 report, previewed in the press feed on 2026-08-08 and not yet filed at the time of this analysis. Watch for a restatement of the Q3 topline guidance, the cash balance (the newest figure is 2026-03-31), and any drawdown on the $50 million Lucid at-the-market facility, which has never been used.
• 2026-08-26 to 2026-10-31 — the topline window. Read the release for which co-primary is reported first and whether both are reported at all: the 2018-04-02 precedent on this exact trial announced one endpoint and withheld the other for seven years, and that is the specific disclosure behaviour to check against.
• Next FINRA settlements (mid- and end-August, published roughly two weeks late) — short interest tripled to 346,440 shares over the six weeks to 2026-07-31. Whether that continues, stalls or reverses before the window opens is the only positioning signal this illiquid name offers.
• Whether ClinicalTrials.gov NCT05464784 is updated — the record still carries a primary_completion_date of 2026-12-31 eleven weeks after last patient last visit. An update to that field, or the posting of results, would be an independent confirmation of the timeline.
• 2026-11-05 to 2026-11-09, AASLD The Liver Meeting, Denver — not currently a source for this readout, because no company statement names it. If the company announces a presentation there, the window above needs re-cutting past 2026-10-31.
• Year-end 2026 — COMBAT-ALS topline. It is not this program, but it is the larger claim on the equity and its window overlaps this one, which is exactly why attribution.status is INDETERMINATE.

Locked prediction

Three separate calls, per framework/05-prediction-protocol.md.

• Outcome-direction (will the readout succeed on its primary endpoint?): miss — probability 15%, band 8–28% [UNVERIFIED — modelled]. The probability that the settlement definition below is met, and nothing vaguer. Reasoning: the triglyceride co-primary carries a real prior signal and is plausible at roughly 55–65% on its own; the CAP-score co-primary has a prior result pointing the wrong way (+0.64 percentage points liver fat by MRI at Week 12), no established MCID, meaningful measurement variability, and 20 patients per arm, and sits at roughly 20–25%; both together, allowing for modest correlation, land at 15%. No third party has published a probability on this event; the four published analyst targets ($5.00–$10.00) are price views, not probabilities, and are not treated as one.
• Stock-direction (which way do the shares move?): no-edge — confidence low — window 2026-08-13 to 2026-11-14, basis: from the lock date to two weeks past readout.window.latest of 2026-10-31, so the window covers the whole plausible readout period plus the settling of the reaction. Attribution could not be determined for this program (01-rules.md rule 36): the COMBAT-ALS and OXTOX catalysts both sit in overlapping period placeholders that swallow this window, so a price move inside it cannot be attributed to this program alone — and COMBAT-ALS is the larger claim on the equity. Materiality, per ../company.md C.2, is meaningful but not dominant, and this call is deliberately small to match. no-edge agrees with the expected value below, which sits +2.6% from spot — a rounding error against a bracket of −5% to +25% for the event itself. The direction is declined because the two live paths point opposite ways and nothing in the evidence separates them: a partial hit spun as a win, which this ticker’s own history says trades up, against a clean double miss, which the sponsor’s own prior data says is a real possibility.
• Scenario prices: positive $1.60–$2.15 · miss $1.05–$1.50.
• Expected value: $1.37, +2.6% against spot $1.33. Arithmetic, not advice.
• Run-up: entry $1.33 on 2026-08-13, exit rule T-5 trading days before readout.window.earliest — predicted move −4% to +7%, predicted peak 0% to +12% around 2026-08 — priority score 9, formula_version 1.0.0. The exit rule resolves to roughly 2026-08-19, four trading days after entry; exit is null because no committed price cache exists for MNOV.
• Settles on — Positive definition: MediciNova reports, in a dated public release or an SEC filing, that MN-001-NATG-202 (NCT05464784) met both co-primary endpoints at Week 24 with statistical significance versus placebo at the conventional two-sided p<0.05 — that is, (1) mean change from baseline in liver fat by controlled attenuation parameter score and (2) mean change from baseline in fasting serum triglycerides. A result on only one co-primary is a miss, however it is presented, and a release that reports one endpoint while withholding the other is a miss until the withheld endpoint is disclosed and also met. Source: MediciNova’s own press release or SEC filing, cross-checked against ClinicalTrials.gov posted results for NCT05464784. Settles on: the topline announcement of MN-001-NATG-202.
• Locked: yes · Settled: no

Program data-quality flags

• The sponsor reports its own prior primary endpoint two ways, differing by 6.4×. NATG-201’s Week-8 triglyceride change is −21.67 mg/dL (SD 27.89, n=19) in ClinicalTrials.gov’s posted results and −138.8 mg/dL / −40.2% (345.7 → 206.9, n=19) on the sponsor’s own IDF 2022 poster. Same endpoint, same timepoint, same denominator, same units. Reported as an unresolved conflict per 01-rules.md rule 24; not averaged and not resolved by picking one. This is the most consequential flag in the document, because the poster figure is the number the investment case rests on.
• ClinicalTrials.gov NCT05464784 is stale. It still carries primary_completion_date and completion_date of 2026-12-31 eleven weeks after the sponsor announced last patient last visit on 2026-05-26. The registry field was therefore not used as the input to the modelled readout estimate, and that substitution is stated in the Readout section rather than made silently.
• Open Targets BLOCKED — Rate limit exceeded for client: global, three attempts. No independent human-genetic target validation appears anywhere in this document; every target-validation claim in A.1 and A.5 is tagged [UNVERIFIED] as a result.
• ChEMBL BLOCKED — Internal error: Error calling tool 'compound_search': HTTP error in tool compound_search: 500 from upstream API, four attempts. The molecule’s formal identity and off-target selectivity profile are not independently confirmed; A.1’s mechanism description rests on published literature and sponsor statements instead.
• PubMed’s index conflates MN-001 with the manganese crystal surface Mn(001). Three of the seven hits (PMIDs 22989203, 12633385 and, in part, the term’s presence elsewhere) are condensed-matter physics papers. A naive hit count of seven would overstate this molecule’s evidence base by more than double; the real figure is three non-bibliography papers in twenty years.
• There is no peer-reviewed publication of NATG-201 at all. The trial that constitutes the entire prior evidence base for the pending readout exists publicly only as ClinicalTrials.gov registry entries and a conference poster whose three authors are all declared employees of the sponsor. This is recorded as a data-quality flag and not only as an evidence score, because it means no independent peer reviewer has ever examined the numbers the investment case rests on.
• No investigator is named on the pivotal trial’s registry record. get_trial_details on NCT05464784 returns two facilities with no contacts, no overall officials and no responsible-party investigator. The kol.investigators array is empty as a consequence, and that emptiness is a finding about disclosure rather than about the search.
• No exposure–response analysis and no statistical analysis plan are public. The trial carries two co-primary endpoints at 20 patients per arm with no disclosed multiplicity or alpha-allocation approach, which is material to how a partial result should be read.
• The Kyorin royalty rate is not disclosed anywhere. The 10-K states only that MediciNova is “obligated to pay a royalty on net sales of the licensed products”. This is why B.3b writes no eNPV rather than an eNPV built on a guessed rate.
• No price cache exists for this ticker. data/prices/MNOV.json is absent, so the run-up exit is null, scorePricedIn was fed BPIQ’s verified 52-week range rather than a committed series, and a run-up settlement on this program cannot be performed until the cache is populated. Company-tier consequences of the same gap are in ../company.md C.8.