PolyPEPI1018

PolyPEPI1018 is an off-the-shelf, 12-peptide therapeutic cancer vaccine from Treos Bio aimed at microsatellite-stable (MSS) metastatic colorectal cancer (mCRC), the immunologically cold majority subtype (about 95% of mCRC) that does not respond to checkpoint inhibitors alone [1]. Unlike personalized neoantigen vaccines (autogene cevumeran, NOUS-209, GRT-C901/GRT-R902) that sequence each patient's tumor and manufacture a bespoke product, PolyPEPI1018 is a fixed cocktail of peptides from 7 shared tumor antigens given identically to every patient, betting that prevalent shared antigens can drive responses across a heterogeneous population [1]. The current Phase 2 trial (NCT05243862) pairs the vaccine with atezolizumab, an anti-PD-L1 antibody, in 18 patients [2]. The thesis: prime tumor-antigen-specific T-cells with the vaccine, then release the brake on those T-cells with checkpoint blockade.

Phase 2 single-arm combination study, active but not recruiting at n=18 [2]. No public FDA breakthrough, fast-track, or RMAT designations. Three Phase Ib datasets underpin the program. OBERTO (NCT03391232, n=11) tested PolyPEPI1018 with maintenance therapy and a PEPI companion diagnostic in mCRC; this trial is the basis of the Hubbard et al. Clin Cancer Res 2022 publication, which reported acceptable safety and detectable peripheral blood T-cell responses against vaccine peptides [1][3]. OBERTO-201 (NCT05130060, n=15), the Mayo Clinic-sponsored Phase Ib combining PolyPEPI1018 with trifluridine/tipiracil (TAS-102, standard third-line CRC chemotherapy), published in JCO Oncology Advances in November 2025, reported a 53.3% disease control rate (DCR), median progression-free survival (mPFS) of 4.0 months, and median overall survival (mOS) of 8.7 months, versus historical TAS-102 monotherapy benchmarks of roughly 44% DCR, 2.0 mo mPFS, and 7.1 mo mOS in the RECOURSE trial [4][5]. PolyPEPI1018-related adverse events were limited to Grade 1-2 injection-site reactions, with no systemic immune-related toxicities. Patients with longer PFS and OS mounted broader cellular and humoral responses to vaccine antigens, a suggestive but uncontrolled biomarker-outcome correlation in a 15-patient single-arm study [5]. In August 2025, Treos announced OBERTO-202, a separate potentially registration-enabling Phase 2 with toripalimab (Junshi Biosciences) at Charité-Berlin, the first major external partnership signal for the program [8]. Treos Bio is a small UK-headquartered private biotech with no SEC filings; PitchBook/Crunchbase show approximately $47M raised lifetime, including a roughly $14M Series B in March 2021 and a $2.1M bridge financing in 2025, indicating constrained runway [9]. Timeline to a NCT05243862 atezolizumab-combination readout is not publicly disclosed; given enrollment closed and standard 6-12 month follow-up windows for DCR endpoints, an H2 2026 to H1 2027 readout window is plausible. No FDA submission is on the horizon from this single 18-patient cohort.

Multi-peptide cancer vaccines work by training the immune system to recognize specific protein fragments (peptides) that tumors display on their surface. PolyPEPI1018 is a cocktail of 12 peptides derived from 7 tumor-associated antigens commonly overexpressed in colorectal cancer, including cancer-testis antigens (proteins normally made only in testis tissue but aberrantly switched on in many tumors) [1]. Peptides are selected via Treos's PEPI platform, which screens for sequences that can bind multiple HLA types. HLA molecules are the 'presentation trays' every person's cells use to show what is inside them to T-cells, and the platform aims to pick peptides where most patients, regardless of HLA genotype, can mount a T-cell response [3]. Combining with atezolizumab is biologically reasonable: the vaccine should induce tumor-antigen-specific T-cells, but those T-cells get switched off when they encounter PD-L1 on the tumor surface, and the checkpoint inhibitor releases that brake. The deeper problem with MSS colorectal cancer is that even successfully primed peripheral T-cells often cannot reach the tumor. Cold MSS tumors secrete a chemokine profile low in CXCL9/10/11 (the signals that normally pull T-cells into a tissue) and recruit suppressive stromal cells that physically exclude T-cells from the tumor interior. A strong T-cell response measured in peripheral blood can therefore coexist with zero intratumoral infiltration and zero antitumor effect. This is why peripheral immunogenicity (the strength of T-cell or antibody responses measured in blood, as opposed to inside the tumor) has historically been a weak proxy for clinical benefit in cancer vaccines. The broader history is unforgiving: GVAX, PROSTVAC, IMA901, and tergenpumatucel-L all generated peripheral immune responses without delivering randomized survival benefit, and PROSTVAC failed its Phase 3 PROSPECT readout in prostate cancer in 2019 [6].

NCT05243862 is an open-label, single-arm Phase 2 study of PolyPEPI1018 plus atezolizumab in metastatic colorectal cancer, enrolling 18 patients at Treos Bio sites [2]. The primary endpoint is treatment-related adverse event incidence and severity, not efficacy, which signals a safety and signal-finding study rather than a registrational design. Secondary endpoints likely cover objective response rate (the fraction of patients whose tumors shrink by at least 30% in target lesion diameter per RECIST criteria) and immune correlates, though the public registry entry foregrounds safety. There is no comparator arm. An n=18 cohort is too small to establish efficacy versus historical controls in MSS CRC, where response rates to checkpoint monotherapy run in the low single digits. A reasonable read-through: if the vaccine plus atezolizumab produces meaningful objective responses or durable disease control in MSS patients, that is a signal worth advancing. If responses look like background atezolizumab monotherapy, the program stalls. Patient selection criteria are not visible in the public summary, but OBERTO used a PEPI companion diagnostic to enrich for patients predicted to respond to multiple vaccine peptides [3], and the Mayo Phase Ib retrospectively associated broader antigen-specific responses with longer survival [5], suggesting Treos has been running PEPI-based selection. Whether the Phase 2 carries that selection forward will determine how interpretable the data are. Strategically, PolyPEPI1018 sits in a different camp from the leading cancer vaccine competitors. Personalized neoantigen vaccines (autogene cevumeran from Genentech/BioNTech, NOUS-209 from Nouscom, GRT-C901/GRT-R902 from Gritstone bio) sequence each patient's tumor, identify private mutations unique to that patient, and manufacture a bespoke mRNA or viral-vector product per case. PolyPEPI1018 instead uses one off-the-shelf cocktail of shared antigens deployed identically to every patient. Trade-offs: shared-antigen vaccines are cheaper, manufacturable at scale, and immediately available to any newly diagnosed patient, but they depend on target antigens being widely expressed across the patient population and being immunogenic on common HLA backgrounds (which is exactly what the PEPI platform tries to make sure). Personalized vaccines theoretically target every patient's truly tumor-specific neoantigens but cost six figures per patient to manufacture, require weeks of bespoke production, and have so far also failed to produce randomized survival benefit in solid tumors outside niches like pancreatic adjuvant. The off-the-shelf approach has a clearer commercial path if it works; the open question is whether shared tumor-associated antigens can drive enough antitumor immunity in MSS CRC.

Our model estimates a 6% chance this drug is eventually approved. It starts from the historical base rate for Phase 2 drugs in this area (about 13%), then adjusts using ten facts about the trial and sponsor. What moves the number most: it is helped by a non-randomized design; it is held back by the sponsor's thin or weak approval record, smaller-than-typical enrollment for this phase, and weak or limited earlier-phase results. The other facts land near average for this stage, so they leave the estimate roughly where the base rate put it.

Efficacy risk dominates. The dominant failure mode is the one that has killed nearly every therapeutic cancer vaccine since 2000: peripheral T-cell responses get measured in blood, the team publishes immunogenicity data, and the patients still progress. MSS colorectal cancer is the toughest immuno-oncology setting in solid tumors because the tumors do not recruit T-cells into the tumor microenvironment in the first place and they downregulate antigen presentation. Vaccinating against tumor antigens does not by itself fix the cold tumor problem. Without biomarker-selected responders, the n=18 atezolizumab cohort is unlikely to produce interpretable efficacy data, though OBERTO-201 suggests baseline activity above TAS-102 monotherapy benchmarks when paired with chemotherapy [5]. Safety risk is the most benign part of this profile: peptide vaccines paired with PD-L1 inhibitors generally show manageable toxicity, and OBERTO-201 reported only Grade 1-2 injection-site reactions related to PolyPEPI1018, with immune-related adverse events from the checkpoint partner being the main concern in combination [1][5]. Execution risk is real because Treos Bio is small and privately held with approximately $47M raised lifetime and a $2.1M bridge in 2025 [9]; runway, cash to fund OBERTO-202 with Junshi [8], and trial completion all depend on inputs not visible from outside. The Junshi/Charité collaboration is an external validation signal but does not appear to include upfront capital sufficient for a US registrational program. Commercial risk: even with a positive Phase 2 signal, a small biotech with a single asset in MSS CRC faces a long, capital-intensive Phase 3 path against an indication where Merck, Roche, and Bristol have all tried and largely failed in the MSS setting. Payer scrutiny of cancer vaccines is severe given the field's track record, and Treos would almost certainly need a US-based pharma partner beyond the Junshi regional deal to fund a registrational trial.

Watch but discount heavily. The signal worth checking for is a published or presented disease control rate clearly above 50% in PEPI-selected MSS CRC patients on the atezolizumab combination, or a credible prospectively-defined biomarker correlation between vaccine-induced T-cell breadth and clinical benefit. OBERTO-201 (JCO Oncology Advances 2025) is the strongest existing data point, reporting 53.3% DCR, 4.0 mo mPFS, and 8.7 mo mOS in 15 patients on PolyPEPI1018 + TAS-102; these numbers exceed historical TAS-102 monotherapy benchmarks but only in single-digit-percent terms in a small uncontrolled single-arm trial [5]. The Hubbard 2022 OBERTO publication established peripheral immunogenicity but not durable benefit [1]. Treos has been at this since at least 2018 (OBERTO), which is either patient pursuit of a hard problem or a sign the program has not generated breakthrough data that triggers a large pharma deal; the Junshi/Charité collaboration in August 2025 [8] is the most credible external validation to date but is regional (China + Germany) rather than a US-pharma transaction. Commercial context: MSS represents approximately 95% of metastatic CRC; US mCRC incidence is roughly 53,000 newly diagnosed metastatic patients per year, with the 2L+ treatable population for a late-line asset like PolyPEPI1018 in the rough range of 30,000-40,000 patients annually in the US. This is a meaningful commercial opportunity if the drug works, but it is also the proving ground where every major IO program has failed. Catalysts to track: an ASCO 2026 or ESMO 2026 abstract or oral presentation of NCT05243862; as of mid-2026, Treos has not publicly announced abstract acceptance at either meeting based on available sources, so the most actionable trigger is the next ASCO/ESMO abstract embargo window. Also watch for updated OBERTO-201 follow-up data, any new US-based or large-pharma partnership announcement, and OBERTO-202 enrollment kickoff metrics. Until those land, treat this as a low-probability lottery ticket with high downside correlation to the broader history of therapeutic cancer vaccines; we land at the 4-7% PoS range rather than the model's 9.4%, putting weight on the historical cancer-vaccine LOA being below Wong's 6.7% all-oncology Phase 2 LOA. If a major US pharma partnership announcement materializes or NCT05243862 reports a DCR materially above 50% in PEPI-selected patients, the read changes.