IMMP / efti-pembrolizumab-soft-tissue-sarcoma — Eftilagimod alfa with pembrolizumab and radiotherapy, given before surgery for soft tissue sarcoma

Program analysis · bpiq_drug_id 17452 · prepared 2026-08 · USD · framework v5.10.3 · 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

Executive summary

• What it is (one sentence): Eftilagimod alfa is a manufactured soluble immune protein that switches on antigen-presenting cells, tested here as a third agent added to radiotherapy and pembrolizumab before surgery in soft tissue sarcoma.
• The event and when (as disclosed): The full abstract publishes on the ESMO Congress 2026 website on Monday 19 October 2026, and an ePoster is presented at the congress in Madrid, 23–27 October 2026. [VERIFIED — Immutep press release 2026-07-24; ESMO congress dates confirmed in data/congresses.json against esmo.org]
• The main reason it could work: The trial already met its primary endpoint — a median 51.5% of the removed tumour turned to scar tissue, against roughly 15% historically with radiotherapy alone — with no grade 3 or higher toxicity attributed to either eftilagimod alfa or pembrolizumab. [VERIFIED — Immutep press release 2025-11-13, CTOS 2025 oral presentation]
• The main risk: This catalyst is not a test of whether the drug works. It is a quality-of-life analysis of a trial whose efficacy result was reported fifteen months ago, and the quality-of-life endpoint is not registered among the trial’s primary, secondary or other outcome measures on ClinicalTrials.gov. [VERIFIED — CT.gov get_trial_details NCT06128863, 2026-08-19]
• What it means for the stock: Very little on its own merits, and the window is not clean. The Nasdaq minimum-bid-price cure period expires on 26 October 2026, three days inside the congress, and a separate Phase 1 readout for a different drug is guided to the same half-year. This program’s attribution is CONTAMINATED.

0. Program-tier coverage — CLEARED

Company-tier coverage is in ../company.md C.0 — CLEARED, with the optional Yahoo options cross-check BLOCKED.

What the one block costs, stated rather than waved past. Open Targets is the only mandatory program-tier row that did not return. What it would have supplied is human genetic evidence that LAG-3 is causally involved in disease. Its absence matters less here than it would for most programs, for a specific reason: eftilagimod alfa is not a LAG-3 blocker, so the genetics of LAG-3 loss-of-function would not validate what this drug does anyway. The target-validation judgement in A.5 therefore rests on clinical precedent — including a 756-patient randomised Phase 3 of this exact hypothesis that was stopped for futility — which is a harder test than genetics would have been.

A. Scientific & clinical assessment

A.1 Mechanism of action and scientific rationale

What the drug is. Eftilagimod alfa is not an antibody. It is a manufactured, soluble, dimeric copy of the outer part of the human LAG-3 protein, fused to an antibody tail to keep it in circulation, and given as a 30 mg injection under the skin every two weeks. [VERIFIED — ChEMBL CHEMBL4594557, USAN stem "ef-; -imod" = Fc fusion protein, immunomodulator; dosing from Krebs et al., JTO Clin Res Rep 2024, [DOI](https://doi.org/10.1016/j.jtocrr.2024.100725)]

How it works. On the surface of a T cell, LAG-3 is a brake — which is why most drug companies build antibodies to block it, and why one such drug (relatlimab, with nivolumab) is approved in melanoma. Eftilagimod alfa does the opposite. Because the manufactured protein floats free rather than sitting on a T cell, its natural binding partner is what it goes looking for: MHC class II, a molecule on the surface of antigen-presenting cells. Binding there switches those cells on. The activated antigen-presenting cells mature, display tumour fragments to T cells, and recruit both CD4 and CD8 T cells into the fight, with measurable rises in interferon-gamma and CXCL10 in the blood. [VERIFIED — Atkinson et al., J Immunother Cancer 2020, [DOI](https://doi.org/10.1136/jitc-2020-001681); Forster et al., Clin Cancer Res 2024, [DOI](https://doi.org/10.1158/1078-0432.CCR-24-0473). According to PubMed.]

Why combine it with radiotherapy and pembrolizumab in this disease. Soft tissue sarcomas are “immune-cold”: they carry few mutations and attract few T cells, which is why checkpoint drugs alone work poorly in them. The combination attacks that from three directions at once. Radiotherapy kills tumour cells and spills their contents, supplying the antigen. Eftilagimod alfa switches on the cells whose job is to pick that antigen up and show it to T cells. Pembrolizumab then removes the PD-1 brake so the T cells that arrive can act. [VERIFIED — trial design, CT.gov NCT06128863; mechanism rationale, Wiśniewska et al., Autoimmun Rev 2026, [DOI](https://doi.org/10.1016/j.autrev.2026.103992)]

How well the target is validated — and the honest answer is: contested by the sponsor’s own data. The pharmacology is reproducible. Across five trials, patients given eftilagimod alfa showed statistically significant rises in absolute lymphocyte count, and the patients whose counts rose lived longer. [VERIFIED — Immutep press release 2026-05-28] But that is a retrospective association across single-arm and open-label studies, not proof that the drug caused the survival. The one properly controlled test of the hypothesis — TACTI-004, 756 patients planned, randomised, double-blind, placebo-controlled, in first-line lung cancer — was stopped for futility on 2026-03-13. [VERIFIED — Immutep press release 2026-03-13; CT.gov NCT06726265] The company’s own explanation, four months later, was that preliminary immune monitoring in TACTI-004 “showed a markedly different immune activation profile versus prior studies”. [VERIFIED — Immutep press release 2026-07-13] That is the sponsor saying its drug did not do in the Phase 3 what it had done everywhere else. It is the single most important scientific fact about this molecule and it is not explained.

The exact scientific step the October 2026 readout must prove. None, in efficacy terms — and this is the point a reader most needs to be given plainly. The primary endpoint of EFTISARC-NEO was met and announced on 2025-05-27, presented in full at ESMO Congress 2025 on 2025-10-20 and again as an oral presentation at CTOS 2025 on 2025-11-13. The October 2026 catalyst is a health-related quality-of-life analysis of the same 40 patients. What it must show is that patients receiving nine weeks of injections on top of five weeks of radiotherapy, before major surgery, did not feel materially worse for it.

The honest scientific risk, in three parts.

1. A single-arm trial cannot separate the drug from the backbone. Every one of the 40 patients received radiotherapy and pembrolizumab as well as eftilagimod alfa. The 51.5% median hyalinization is compared against a historical figure of roughly 15% for radiotherapy alone — but pembrolizumab plus radiotherapy is itself an active regimen with a randomised Phase 3 behind it. Nothing in EFTISARC-NEO isolates eftilagimod alfa’s contribution.
2. The endpoint is a surrogate. Percent tumour hyalinization is a pathologist’s measure of dead tissue, not a measure of whether patients live longer or relapse less. In sarcoma it is not an established regulatory surrogate.
3. The quality-of-life endpoint being presented is unregistered. It appears nowhere among the one primary and seven secondary outcome measures on the trial’s ClinicalTrials.gov record, and the record lists no “other” outcomes at all. [VERIFIED — CT.gov get_trial_details NCT06128863, 2026-08-19] There is therefore no pre-specified analysis plan a reader can hold the result against.

What ChEMBL does and does not establish. compound_search returned exactly one record — CHEMBL4594557, EFTILAGIMOD ALFA, synonyms IMP321 and Immufact, maximum clinical phase 2 — which confirms the molecule’s identity and that ChEMBL knows of no approval. It supplies no selectivity or off-target data: structure_type is NONE, molecule_properties is null and there is no SMILES string, because ChEMBL’s bioactivity data are built around small molecules and this is a recombinant fusion protein. That is the same limitation 02-connectors.md records for an antibody, and the cost is small for the same reason. [VERIFIED — ChEMBL compound_search, 2026-08-19]

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

Milestone dates as far as they are known. First patient in 2023-07-17. Last patient out for the primary endpoint is not separately disclosed; the registry’s primary completion date is 2025-04-30 and topline was announced 27 days later on 2025-05-27. Database lock is not disclosed. Overall study completion, which covers the 24-month post-surgical follow-up for disease-free and overall survival, is 2027-04-30. [VERIFIED — CT.gov get_trial_details NCT06128863, 2026-08-19; announcement date from the Immutep press feed]

Trials

Feasibility matrix

Resourcing sufficiency. For this catalyst, entirely sufficient and essentially free: the trial is run and funded by an academic institution in Warsaw, enrolment is closed, and what remains is an ePoster and a 24-month follow-up the sponsor institution carries. Immutep does not need to spend to reach 19 October 2026. For anything beyond this catalyst the answer inverts. A randomised registrational trial in this indication would cost a multiple of the company’s entire cash balance (A$68.87M at 2026-06-30 — see ../company.md C.3), no such trial has been announced, and the company has just cut headcount and wound down its only Phase 3. There is no funded path from this result to an approval.

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

A.3a Target label. Eftilagimod alfa, added to neoadjuvant radiotherapy and pembrolizumab, for adults with resectable, non-metastatic, FNCLCC grade 2–3 soft tissue sarcoma of the extremities, girdles or superficial trunk, with a primary tumour larger than 5 cm or any locally recurrent tumour.

A.3b Target product profile

A.3c Strategic Go/No-Go questions. The set that matches this asset’s next decision is Pre-Phase-III: EFTISARC-NEO is a completed Phase 2 whose result either does or does not justify a registrational trial.

A.3d Regulatory designations.

• FDA Orphan Drug Designation, granted 2026-04-15 for eftilagimod alfa in soft tissue sarcoma. What it grants: seven years of US market exclusivity for the designated indication if the drug is approved, exemption from the FDA application fee, and tax credits on qualifying US clinical trial costs. It grants nothing about approvability and involves no efficacy judgement. [VERIFIED — Immutep press release 2026-04-15]
• No Breakthrough Therapy, Fast Track, Priority Review or Special Protocol Assessment was found for this indication in the press feed or the EDGAR filings feed. [UNVERIFIED — derived from absence in two searches, not from a confirmed empty register]

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.

A.5 Evidence quality and endpoints

Endpoints

A.5b Key opinion leaders.

Panel as of. 2026-08-19.

Investigators

The registry record for NCT06128863 lists no overall officials, no principal investigator and no site contacts — its single location entry carries contacts: null. A search_investigators call against the sponsor institution returned 51 investigators across 18 other sarcoma trials and none on this one. The two people named below are therefore recorded on inference from published authorship at the sponsoring department rather than from the trial record itself, and are tagged accordingly. This is thinner than a normal investigator table and the thinness is the finding.

Independent voices

One voice was found, and it comments on the drug class rather than on this trial’s endpoint. That distinction is stated rather than blurred: nobody outside the sponsoring institution was found commenting on neoadjuvant pathologic response or on quality of life in soft tissue sarcoma with this combination.

Judgement

Dissent

Empty. Endpoint supported is UNKNOWN rather than SUPPORTIVE_CONTESTED, so there is no claimed support for anyone to contest. Inventing a disagreement nobody holds would be worse than an empty table.

B. Commercial assessment

B.0 Current treatment algorithm

What a patient with a large, high-grade, resectable soft tissue sarcoma of the arm, leg or trunk receives today, in order:

1. Diagnosis and staging. Imaging, biopsy, and grading on the FNCLCC 1–3 scale. Patients whose tumour is small, superficial or low-grade go straight to surgery and never reach this trial’s population.
2. Neoadjuvant radiotherapy. For large, deep, grade 2–3 tumours, roughly five weeks of radiotherapy before the operation. This is standard, long-established, and it is what the historical 15% hyalinization figure comes from.
3. Neoadjuvant chemotherapy, for selected histologies. An anthracycline with ifosfamide, in some centres and some tumour types. Contested, and not universal.
4. Pembrolizumab added to the radiotherapy — the new step, and the one that matters here. Following SARC032, adding pembrolizumab to neoadjuvant radiotherapy improved disease-free survival and is now recognised in NCCN guidance for the relevant histologies. [WEB ESTIMATE — SARC032 Lancet publication and NCCN recognition via web search, 2026-08-19]
5. Wide surgical excision, five to six weeks after radiotherapy ends.
6. Relapse, if it comes, is treated systemically — doxorubicin-based regimens, then later lines. Roughly half of high-risk patients relapse, and most relapses are distant rather than local.

Where eftilagimod alfa would fit: step 4, as a third drug. It displaces nothing. It is added on top of a two-drug regimen that itself only recently became standard — and it is added on top of the specific regimen whose benefit was demonstrated without it. That is the commercial position in one sentence, and it is a harder position than “first in a setting where nothing works”: the incumbent here is not an absence, it is a randomised Phase 3 with a met endpoint.

B.1 Competitive landscape and positioning

Where this asset wins, and the single fact the thesis rests on. It wins on being the only drug of its kind, in a disease where the immune approach has just been shown to work and where the obvious next question is how to make it work better. It also wins, uniquely among efti’s indications, on having a regulatory designation. The thesis rests on one fact: that a 51.5% median hyalinization in 38 uncontrolled patients reflects something eftilagimod alfa added, rather than what pembrolizumab plus radiotherapy does on its own. Nothing in the evidence base separates those two readings, and no trial that could separate them is funded.

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.

B.2 Addressable market

Launch markets would be the United States and the EU5, in that order — the markets Immutep retained when it licensed the rest of the world to Dr. Reddy’s and Greater China to EOC Pharma. [VERIFIED — Immutep / Dr. Reddy's release 2025-12-08]

B.3 Value and feasibility

B.3a Expected peak sales. Built bottom-up for the soft tissue sarcoma indication only — the indication this catalyst is about — and conditional on an approval that is not currently funded to be pursued. It excludes every other efti indication, all Dr. Reddy’s and EOC Pharma territories, and any value from IMP761.

Two things must be said about that table at once, because either alone misleads. First, these are small numbers by oncology standards — an orphan neoadjuvant indication is a fraction of what first-line lung cancer would have been, and the loss of that indication is not made up here. Second, they are enormous relative to what this equity currently costs: the enterprise value in ../company.md C.4 is about US$9.6M, so even the low scenario is roughly twice the entire enterprise value in annual peak revenue. Both facts are true. Which one dominates depends entirely on whether a registrational trial is ever funded, and today it is not.

No third-party peak-sales figure is passed through. The dedicated peak-sales web search returned none for eftilagimod alfa in any indication. It did return a “consensus price target is AU$1.30” figure — roughly 23× the current ordinary-share price — which is rejected under 01-rules.md rule 6 as a stale pre-discontinuation number no source dated or stood behind. The current, dated analyst figures are recorded in the Market and timing section below instead.

B.3b Feasibility

B.4 Product-development risk

Framing questions.

1. Does the development plan support the target product profile claims? No, because there is no development plan. The target product profile in A.3b describes a third agent added to the SARC032 backbone, and supporting that claim requires a randomised trial against that backbone. No such trial exists, is registered, or is funded. What exists is a completed single-arm study, a 24-month follow-up that runs to 2027-04-30, and an ePoster.
2. Will the identified risks affect the target product profile? The dominant risk — that the pathological response reflects the backbone rather than the added drug — attacks the target product profile’s efficacy row directly and cannot be mitigated by anything other than the trial that is not funded. The financing and listing risks in the grid below do not touch the profile; they touch whether there is a company left to pursue it.
3. If a risk cannot be mitigated, is the asset still differentiated from competitors? Yes, on mechanism — it remains the only MHC class II agonist in the clinic, and that is not contingent on this trial. But differentiation on mechanism without a controlled efficacy result is a scientific distinction, not a commercial one.

Readout

Sources

The modelled estimate. The arithmetic, stated so it can be argued with: the registry’s primary_completion_date for NCT06128863 is 2025-04-30, plus the stated default lag of two to four months to database lock and analysis (01-rules.md rule 34; data/benchmarks/readout-lag.json holds no observations, so the default applies and the tag is [UNVERIFIED — modelled, default lag]), gives 2025-06-30 to 2025-08-31. That window closed a year ago. The modelled method dates the trial’s primary readout, and the primary readout already happened — 27 days after primary completion, on 2025-05-27, which is faster than the default lag predicts. The method has nothing to say about when a congress will publish a secondary analysis of a trial that finished last year, and pretending otherwise would be worse than saying so.

One observation this analysis produced but may not write down, recorded here instead. EFTISARC-NEO gives a confirmed lag: registry primary completion 2025-04-30 against a topline announcement of 2025-05-27, 27 days, both dates from primary sources. data/benchmarks/readout-lag.json is empty and sits outside this analysis’s lane, so nothing was added to it; the observation is left here for whoever maintains that file.

Window

Basis. All three edges are the same date, because only one moment discloses the numbers. Immutep stated on 2026-07-24 that the full abstract publishes on the ESMO Congress 2026 website on Monday 19 October 2026 — “full” is the company’s own word, so the data is readable that day rather than held back for the podium. The ePoster displayed during the congress on 23–27 October shows that same content, and an ePoster is not a second readout: nothing is disclosed by hanging the same numbers on a board four days later. The congress tail is a market-attention window rather than a readout window, and it already has its own home in this document — the stock-direction window below runs 2026-10-19 to 2026-11-03 precisely to cover the abstract drop, the poster and any release accompanying it. Precision is DAY because a source names that specific day. Confidence is HIGH for the unusual reason that there is almost nothing left to go wrong: the trial is complete, the data exist, the abstract is accepted, and the date is a publisher’s calendar rather than a sponsor’s forecast.

Corrected after the first lock. This block was written with latest at 2026-10-27, which double-counted the congress tail into a second place and claimed DAY precision over an eight-day spread. The window was wrong, not the label: a named calendar Monday from the company is a disclosed day, not a half-year, so the fix narrows the window rather than relabelling precision to PERIOD. Nothing downstream moves — date_confidence still reads DAY plus HIGH, and the priority score is unchanged at 55.

Disagreement. CONSISTENT. BPIQ’s 2026-10-19 and the congress’s 23–27 October look like a conflict and are not one: the company’s own release of 2026-07-24 states both, as the abstract publication date and the congress dates respectively. Both are carried, neither is averaged, and the window spans them.

Date slippage. Two dated statements, one transition, zero slips. The catalyst has never moved.

Attribution

Status.

Conflicts

Computed with lib/clustering.mjs’s attributionFor over this ticker’s pipeline at CATALYST_CLUSTER_MIN_MONTHS = 6, for bpiq_drug_id 17452, and transcribed rather than estimated.

Note. The conflicting row is IMP761, Immutep’s Phase 1 LAG-3 agonist antibody in autoimmune disease, guided to “H2 2026”. Because that text names no day, the clustering computation widens it to the whole enclosing half-year, 2026-07-01 to 2026-12-31 — which contains this program’s entire readout window, so the gap is zero. The conflict is confirmed because this side rests on a disclosed date: readout.precision is DAY.

The contamination is not merely arithmetic, and this is the part that matters for the stock call. IMP761 is the only asset at this company that is not eftilagimod alfa; it is the only programme whose Phase 1 has met a primary endpoint since the Phase 3 failure; and it is the only pipeline row BPIQ flags is_high_mgmt_interest. ../company.md C.2 records it as dominant, against meaningful for this program. If both land inside the same weeks, the larger of the two moves will not be this one — and on ../company.md C.7’s evidence, IMP761 news has moved this stock more than efti news has, in both directions (−4.64% on positive EULAR data, +0.73% on a positive December update).

A third event sits in the same window that the clustering computation cannot see at all, because it is not a drug catalyst: the Nasdaq minimum-bid-price cure period expires on 2026-10-26, three days inside the ESMO congress (../company.md C.4). At US$0.40 the shares would have to more than double and hold for ten consecutive sessions to cure it, which this catalyst cannot deliver. A change to the ADS ratio — the ordinary remedy for a dual-listed Australian issuer — would multiply the quoted US dollar price without changing what anyone owns, and would land squarely on top of the readout. None has been announced. The scenario prices below are written in today’s ADS terms and would need re-basing if one is.

Market and timing for this event

• Plain takeaway. A two-month wait for a secondary quality-of-life poster on a trial whose real result came out fifteen months ago, in a stock trading at 1.20× its own net cash with an options chain that cannot price anything, no short interest to squeeze, no insider buying, and a listing deadline landing inside the event window. The date is one of the best-known in this corpus and the event behind it is one of the least informative.
• Months to this catalyst. 2.0 months — 61 days from 2026-08-19 to readout.window.earliest of 2026-10-19, measured from the earliest edge and never from catalyst_date (rule 23; here the two happen to coincide, because BPIQ’s date is a real disclosed day rather than a placeholder). readout.precision is DAY and all three window edges are that same day, so this is a distance to a known date, not to the opening of a window, and there is no later edge to quote. The congress runs 23–27 October and the ePoster shows the data the abstract already carries, which is why the congress sits in the stock-direction window below rather than in the readout window.
• Expected move around this event. A bracket, not a point estimate, because ../company.md C.6 records the chain as unusable — there is no strike at or below spot on any expiry, and the nearest expiry to the catalyst (2026-10-16) expires three days before the abstract publishes. Bracket: roughly ±5% to ±15%. That sits well below the +12% / −16% Phase 2 class averages in framework/07-benchmarks.md [WEB ESTIMATE — IQVIA 2024], deliberately: those averages are measured on primary-endpoint reported dates, and this is not one. Where the class prior and a program-specific fact disagree the specific fact wins, and the specific facts here are three clean, positive, non-registrational data events on this ticker that produced +0.73%, −0.65% and −4.64% (../company.md C.7).
• Nearest comparable past reaction. ../company.md C.7’s 2026-05-28, −0.65% — the pooled five-trial lymphocyte-and-survival analysis on a clean press feed. It is the closest analogue because it has the same information structure: a secondary analysis of data already reported, presented favourably by the sponsor, with nothing new about whether the drug works. The 2026-06-04, −4.64% IMP761 EULAR row is the second-closest and carries the sell-the-news warning. The row a reader would want — 2025-05-26, EFTISARC-NEO meeting its primary endpoint — returns null for both move fields in BPIQ, and with no price cache for this ticker its size cannot be established at all. That is a real gap in the evidence behind the stock call and it is stated rather than filled.
• Materiality. ../company.md C.2 records this program as meaningful at the programme level, immaterial at the event level. The stock-direction call below is made consistent with the event, not the programme (rule 27): a no-edge call on a company whose strongest remaining asset this is, precisely because this catalyst does not test it.
• Date slippage. Zero slips (see Readout, above). The catalyst has never moved.

Spot. US$0.40 per ADS, read 2026-08-19, cited from ../company.md C.1 and C.4.

Scenario prices

Expected value. Method: the 85% probability below applied to the midpoint of each range — 0.85 × $0.41 + 0.15 × $0.32 = $0.3965, which is −0.88% against spot $0.40. This is arithmetic, not advice, and it is not a price target. It agrees with the no-edge stock-direction call rather than quietly contradicting it, which is the point of computing it.

Run-up

Entry and exit

Predicted move and peak

Priority score drivers

Priority score. Priority score 55 · formula_version 1.0.0

Transcribed from lib/runup.mjs’s priorityScore over the seven drivers above, never hand-computed. Read it against its own composition: a perfect date-confidence gate and a near-perfect priced-in score are doing the lifting, and a maximal clustering risk is doing the sinking. It is not a score that says the trade is good; it is a score that says the date is knowable and the price is low.

Settlement. 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, so a settlement will require scripts/fetch-prices.mjs to be run for IMMP first.

Verdict

What I would do. Watch.

Why. The date is excellent and the event is not. This is one of the best-specified catalysts in the corpus — a publisher’s calendar date, two independent sources, zero slips, DAY precision — and what it delivers is a quality-of-life ePoster on a 40-patient single-arm trial whose efficacy result was reported fifteen months ago, on an endpoint that is not even registered among the trial’s outcome measures. Meanwhile the interesting question about this company is not sarcoma at all: it is whether IMP761, the only asset here that is not eftilagimod alfa, is worth more than the US$9.6M enterprise value the market is assigning to everything. That programme’s own readout is guided to the same half-year, which is what makes this one CONTAMINATED, and the Nasdaq cure period expires three days into the congress. Buying a $0.40 stock at 1.20× net cash is not an absurd thing to do — but if you do it, you are buying IMP761 and an option on a restructuring, and you should say so rather than telling yourself you bought a sarcoma catalyst.

What would change this. Two observables, in opposite directions. To the upside: any disclosure of a funded, randomised registrational trial in soft tissue sarcoma — a partner, a cooperative-group collaboration, a registered protocol. That is the missing piece between a real Phase 2 result and any commercial value at all, and its absence is what holds this at watch rather than buy. To the downside: an announced change to the ADS ratio, or any other corporate action taken to cure the Nasdaq deficiency before 2026-10-26. It would not destroy value by itself, but it would re-base every dollar figure in this document and would mean the market’s attention in the catalyst window is on the listing, not on the data.

What to watch.

• 2026-10-12 (approximately) — the T-5 exit rule resolves here, one week ahead of the abstract.
• 2026-10-19 — the full ESMO abstract publishes on esmo.org. This is the catalyst; read the actual quality-of-life instrument and the questionnaire completion rate, not the headline.
• 2026-10-23 to 2026-10-27 — ESMO Congress 2026, Madrid; the ePoster is displayed and Immutep’s practice is to issue a release on the day.
• 2026-10-26 — the Nasdaq 180-day cure period expires. Watch for an ADS-ratio announcement at any point before this.
• Between now and then — any IMP761 Phase 1 update, guided to H2 2026 and the reason this program’s attribution is CONTAMINATED.
• By roughly 2026-10-24 — the FY2026 Form 20-F is due; last year’s was filed 2025-10-24 and it is the first place a current share count and a debt position will appear.
• Open-ended — an EFTISARC-NEO publication. Fifteen months after the primary endpoint there is still none in PubMed, and its appearance would be a stronger signal about this programme’s future than the poster is.

Locked prediction

• Outcome-direction (will the readout succeed on its primary endpoint?): positive — probability 85%, band 72–92% [UNVERIFIED — modelled]. The probability is that the pre-registered settlement definition below is met — that the ESMO 2026 abstract and ePoster report quality of life maintained or improved rather than clinically meaningfully worse. It is high for four reasons and capped below 90 for three. High because: the trial’s safety record shows no grade ≥3 toxicity attributed to either study drug; efti has a direct favourable precedent on the same class of endpoint, with AIPAC reporting EORTC QLQ-C30 global health status sustained under efti and deteriorated under placebo in a far sicker population; the treatment adds five subcutaneous injections to visits the patient already attends; and there is a selection effect — the sponsor chose to submit this abstract and to announce its acceptance in July, which is not what a company does with a deteriorating result. Capped below 90 because: the endpoint is not a registered outcome measure on NCT06128863, so there is no pre-specified analysis plan to hold it to; the instrument is not disclosed; and nine weeks of therapy around five weeks of radiotherapy and major surgery is genuinely burdensome, so an honest report of a transient dip at one timepoint is a real possibility. No third party has published a probability on this event.
• Stock-direction (which way do the shares move?): no-edge — confidence medium — window 2026-10-19 to 2026-11-03, basis: opens on the day the abstract publishes and closes one week after the congress ends, covering both the abstract drop and the ePoster presentation with any accompanying release. Materiality, per rule 27: ../company.md C.2 records this program as meaningful at the programme level and immaterial at the event level, and the call is made consistent with the event. Attribution, per rule 36: this program is CONTAMINATED — the IMP761 Phase 1 readout is guided to a half-year window that contains this one entirely, and the Nasdaq cure period expires on 2026-10-26, three days inside the congress. No move in this window can be presented as attributable to the sarcoma data alone.
• Scenario prices: positive $0.36–$0.46 · miss $0.28–$0.36
• Expected value: $0.3965, −0.88% against spot $0.40 (arithmetic, not advice)
• Run-up: entry $0.40 on 2026-08-19, exit rule T-5 trading days before readout.window.earliest — predicted move −10% to +15%, predicted peak 0% to +25% around 2026-10 — priority score 55, formula_version 1.0.0
• Settles on the publication of the EFTISARC-NEO health-related quality-of-life abstract on the ESMO Congress 2026 website on 2026-10-19 and the accompanying ePoster presentation at the congress. Positive is defined as: the abstract, ePoster or accompanying Immutep release reports health-related quality of life maintained or improved from baseline through the neoadjuvant period — that is, no statistically significant or clinically meaningful deterioration (≥10 points on a 0–100 EORTC QLQ-C30-type global health status scale, or the stated equivalent on whichever instrument is used) in the primary quality-of-life measure reported. A result presented as mixed, or one reporting a clinically meaningful deterioration on the headline measure, settles as a miss; so does a withdrawal of the presentation.
• Locked: yes · Settled: no

Program data-quality flags

• Open Targets search_entities BLOCKED — verbatim Rate limit exceeded for client: global, on all three attempts. The standing platform throttle 02-connectors.md documents, reproduced. Cost: no independent human-genetics validation of LAG-3, so the target-validation row in A.5 rests on clinical precedent alone. Smaller than usual here because eftilagimod alfa is an MHC class II agonist rather than a LAG-3 blocker, so LAG-3 genetics would not have validated what this drug does.
• PubMed’s null-authors bug did NOT fire on this ticker. All ten articles returned complete author lists with affiliations, which is what made the A.5b investigator and independent-voice tables possible at all. Recorded because 02-connectors.md describes that bug as the most expensive one on its page and a reader should know when it did not occur.
• No EFTISARC-NEO publication exists in PubMed fifteen months after the primary endpoint was announced. search_articles on “eftilagimod” returned 10 of 10 hits and none is this trial. Not a connector failure — a real absence, and a finding about the programme.
• The catalyst’s own endpoint is not registered. Health-related quality of life appears among neither the primary, the seven secondary, nor any “other” outcome measures on NCT06128863, and the registry lists other_outcomes: null. The instrument is not disclosed in the abstract-acceptance release either. The settlement definition above therefore has to name a threshold generically rather than against a pre-specified analysis.
• CT.gov lists no investigators for NCT06128863. No overall officials, no principal investigator, and the single Warsaw location carries contacts: null. search_investigators against the sponsor institution returned 51 investigators across 18 other sarcoma trials and none on this one. The two names in A.5b’s investigator table are inferred from institutional authorship and tagged [UNVERIFIED] for that reason.
• No conflict-of-interest statements were retrievable. PubMed get_article_metadata returns affiliations but no COI field on either article used in A.5b, so every conflicts_checked entry records where the search was made and what it could not establish. Per 02-connectors.md § KOL sources, the absence of a disclosure is not evidence of no conflict.
• Source conflict recorded rather than resolved, then reconciled by a third source. BPIQ’s catalyst_date of 2026-10-19 does not fall inside the ESMO Congress 2026 dates in data/congresses.json (2026-10-23 to 2026-10-27). Neither was picked and neither was averaged; the company’s own release of 2026-07-24 states both, as the abstract-publication date and the congress dates, and the readout window spans them. Recorded because a reader comparing the two files without that release will think one of them is wrong.
• The modelled readout estimate is uninformative and is recorded anyway. Primary completion 2025-04-30 plus the default two-to-four-month lag gives a window that closed in August 2025. The method dates the trial’s primary readout, which already happened; it cannot date a congress presentation of a secondary analysis. data/benchmarks/readout-lag.json is empty, so the tag is [UNVERIFIED — modelled, default lag].
• A confirmed readout-lag observation was produced and not written down. EFTISARC-NEO’s registry primary completion of 2025-04-30 against its topline announcement of 2025-05-27 is a confirmed 27-day lag from two primary sources — exactly what data/benchmarks/readout-lag.json exists to collect. That file sits outside this analysis’s lane and was not touched.
• No price cache for this ticker. data/prices/IMMP.json does not exist and a read-only Yahoo chart cross-check returned the body Too Many Requests. Consequences, each stated at its point of use: the 52-week range in ../company.md C.4 comes from BPIQ fields rather than a derived series; the priced_in run-up driver was scored on a synthetic two-bar stand-in carrying those same high and low values; two rows of ../company.md C.7 cannot be sized at all; and a future settlement of this prediction will require the cache to be fetched first.