CrisPRO for the Next STC-1010 Program: What We Do, Where We Fit, What We Cannot Do

A CrisPRO Capability Brief for Prospective STC-1010 Program Partners

Prepared by: CrisPRO Platform Intelligence
Indication: MSS/pMMR Metastatic Colorectal Cancer — First-Line Haptenated Whole-Cell Vaccine
Purpose: Honest capability positioning for prospective program partners

What this document is. CrisPRO does not oversell. This brief tells you exactly where our computational platform adds genuine, defensible value to an STC-1010-type program, where it cannot help, and what the critical gaps are that no computational platform can close. If you are evaluating CrisPRO as a partner for a next-generation STC-1010 program, this is the honest version of that conversation.

THE PROGRAM CONTEXT

An STC-1010-type program sits at the intersection of three hard problems:

1. The MSS CRC biology problem: ~95% of metastatic CRC is structurally resistant to checkpoint immunotherapy. Every broad IO approach has failed. The responding subgroup exists but has never been prospectively enrolled for.
2. The attribution problem: mFOLFOX6 ± bevacizumab achieves ~80–90% disease control rate in first-line mCRC. Any single-arm trial will be statistically indistinguishable from SOC effect without a pharmacodynamic endpoint mechanistically linked to the vaccine — not the chemotherapy.
3. The biomarker problem: No validated predictive biomarker for haptenated whole-cell vaccine response exists in MSS CRC. Without one, the program cannot design a Phase IIa that can detect a real signal, cannot build a regulatory strategy, and cannot differentiate itself from the long list of IO agents that have failed in unselected populations.

CrisPRO's value is in addressing problems 1 and 3. Problem 2 is a clinical trial design problem that requires a randomized control arm — no computational platform can substitute for that.

WHERE CRISPRO ADDS GENUINE VALUE

1. Biomarker Stratification Engine

Fit level: High

The single most urgent unmet need in any first-line MSS CRC IO program is the absence of a validated predictive biomarker for response. CrisPRO's biomarker stratification engine integrates multi-omic data streams — cytokine profiles, immunophenotyping, PBMC markers, ctDNA, HLA expression, tumor necrosis, immune infiltration, tertiary lymphoid structure (TLS) evolution — with the published MSS CRC biomarker literature to identify which combination of baseline features predicts pharmacodynamic response and ultimately 12-month non-progression.

What this looks like in practice:

• A ranked biomarker panel with clinical accessibility classification (standard-of-care vs. research-grade)
• A composite enrichment score for Phase IIa patient selection
• A pre-specified biomarker analysis plan for the Phase IIa protocol — credible to regulators because it is pre-specified, not post-hoc

When this is needed: Now, during Phase I. The exploratory biomarker data from the first cohort of patients is being collected now. Analysis must begin immediately to inform Phase IIa design before enrollment opens. Post-hoc biomarker analysis of Phase I data is hypothesis-generating but not credible for Phase IIa enrichment without pre-specification.

What CrisPRO needs from you: Access to the Phase I exploratory biomarker dataset under NDA/CDA. Standardized assay platforms (if not yet specified, CrisPRO can recommend them). The antigen characterization dataset for the vaccine (mass spectrometry or equivalent) to enable HLA-peptide binding modeling.

2. TME Profiling

Fit level: High

An STC-1010-type program collects mandatory tumor biopsies at baseline and on-treatment. These biopsies are the most valuable scientific asset in the program — but only if the assay platforms are standardized, the response thresholds are pre-specified, and the analysis plan is in place before samples are processed.

CrisPRO's TME profiling capability provides multi-omic characterization of the immune contexture from these biopsies, mapping the six immunosuppressive layers that determine whether vaccine-primed T cells can reach and kill tumor cells:

What this produces: The first human TME dataset for an STC-1010-type program — currently entirely absent from the public record. This dataset is the scientific foundation for every subsequent biomarker, regulatory, and combination strategy decision.

When this is needed: Now, during Phase I. Biopsy collection is ongoing; assay platform decisions must be made before samples are processed. Retrospective assay platform changes invalidate cross-patient comparisons.

3. Trial Design Intelligence

Fit level: High

The structural weaknesses in a first-line single-arm MSS CRC IO trial are well-characterized from the trial history:

• Single-arm design: The 12-month PFS rate primary endpoint will be compared against historical benchmarks subject to selection bias and temporal confounding
• No pre-specified historical control benchmark: Without pre-specification, any historical control chosen post-hoc will be challenged by regulators
• No adaptive enrichment mechanism: Unselected MSS CRC enrollment will dilute any signal from the responding subgroup

CrisPRO's trial design intelligence addresses all three:

Adaptive enrichment design: A Phase IIa design that concentrates enrollment in the biomarker-defined responding subgroup as Phase I data accumulates. This is not a standard statistical design — it requires integration of the biomarker stratification engine with the trial enrollment algorithm.

Historical control benchmarking: A pre-specified historical control analysis using individual patient data from published first-line MSS mCRC trials, with explicit adjustment for the key confounders (liver metastasis status, ECOG PS, prior adjuvant therapy, bevacizumab use). This is the difference between a historical control that regulators accept and one they challenge.

ctDNA co-primary endpoint: The GRANITE trial experience shows that low-ctDNA patients at maintenance initiation have a meaningfully different outcome than high-ctDNA patients. ctDNA clearance rate is a pharmacodynamic endpoint mechanistically linked to the vaccine (not the chemotherapy backbone) and can serve as an early stopping rule for go/no-go decisions.

When this is needed: Soon — Phase IIa design decisions must be made before Phase I completes. The window for protocol amendment is now.

4. Combination Timing Optimization

Fit level: Medium

A four-component regimen (vaccine + cyclophosphamide + GM-CSF + mFOLFOX6 ± bevacizumab) has multiple pharmacodynamic interaction points that affect vaccine efficacy:

• mFOLFOX6 lymphopenic nadir: Occurs at days 7–10 post-infusion. Vaccine administration during this window severely impairs T cell priming. Timing matters.
• Bevacizumab vascular normalization window: Peak CD8+ T cell infiltration occurs at cycle 2 (~days 14–28). Vaccine timing should coincide with this window to maximize T cell trafficking to tumor.
• Cyclophosphamide MDSC depletion timing: Low-dose CY depletes MDSCs and Tregs with a nadir at approximately 3–5 days post-administration. Vaccine administration at this nadir maximizes the immunological window.

CrisPRO's combination timing optimization models these interactions and generates a recommended vaccination timing window relative to each backbone component.

Limitation: Modeling precision is constrained by the absence of published clinical dose data for the vaccine component. Models rely on preclinical data and published analogues from other vaccine trials. Recommendations are testable hypotheses for protocol amendment, not validated clinical guidance.

5. Resistance Mechanism Mapping

Fit level: Medium

Paired baseline/progression biopsies from an STC-1010-type program create an opportunity to map acquired resistance mechanisms as patients progress. CrisPRO's resistance mechanism mapping systematically characterizes which of the six suppression layers is operative at progression in each patient, generating a resistance taxonomy that informs next-line combination strategies.

What this produces: A patient-level resistance profile that answers: did this patient fail because of TGF-β exclusion, MDSC suppression, MHC-I loss, liver immune tolerance, T cell exhaustion, or antigen escape? Each failure mode points to a different next-line strategy.

Limitation: This capability requires paired baseline/progression biopsies. If on-treatment biopsies are conditional ("only if not representing unacceptable clinical risk"), the dataset will be limited. CrisPRO cannot generate resistance data that the trial does not collect.

WHERE CRISPRO CANNOT HELP — AND WHY THAT MATTERS

Honest capability positioning requires naming the gaps. These are not CrisPRO limitations — they are fundamental problems that no computational platform can solve.

Gap 1: The Attribution Problem

The problem: With no control arm and n=6, a 100% disease control rate cannot be attributed to the vaccine. mFOLFOX6 ± bevacizumab alone achieves ~80–90% DCR in first-line mCRC. The vaccine contribution is statistically indistinguishable from SOC effect.

Why CrisPRO cannot close it: This is a clinical trial design problem. No amount of modeling can separate vaccine contribution from chemotherapy contribution in a single-arm trial. The only solution is a randomized control arm or a pharmacodynamic endpoint mechanistically linked to the vaccine (DTH response magnitude, neoantigen-specific T cell expansion, ctDNA clearance rate) that is not driven by chemotherapy.

What this means for a partner: If Phase IIa is also single-arm and the 12-month PFS rate is consistent with historical FOLFOX benchmarks, the program will face a "no added value" regulatory conclusion. The GRANITE Phase II trial — neoantigen vaccine + checkpoint blockade + maintenance chemo — failed its primary endpoint in unselected MSS CRC. That is the precedent.

Gap 2: Antigen Characterization

The problem: CrisPRO's biomarker stratification and HLA-peptide binding modeling require knowledge of which antigens the vaccine presents to the immune system. Without the antigen characterization dataset (mass spectrometry or equivalent), CrisPRO cannot model antigen coverage, predict HLA-peptide binding, or assess whether the vaccine's antigen repertoire matches the patient's tumor antigen landscape.

Why CrisPRO cannot close it: This requires disclosure of the antigen dataset under NDA/CDA, or direct provision of the mass spectrometry data. It is a data access problem, not a computational problem.

What this means for a partner: Antigen characterization is the scientific foundation for a companion diagnostic strategy. Without it, the program cannot develop a mechanistic rationale for patient selection based on tumor antigen expression, cannot design a CDx strategy, and cannot respond to regulatory questions about the biological basis for efficacy in specific patient subgroups.

Gap 3: Manufacturing and CMC

The problem: Batch-to-batch consistency, GMP manufacturing validation, and CMC scalability are wet-lab, regulatory, and engineering problems.

Why CrisPRO cannot close it: This is entirely outside the scope of computational precision oncology. CrisPRO has no capability in this domain.

Gap 4: Peer-Reviewed Clinical Validation

The problem: Computational biomarker hypotheses require prospective clinical validation. CrisPRO can generate hypotheses and analysis plans, but cannot validate them without clinical data from a biomarker-enriched cohort.

Why CrisPRO cannot close it: Validation requires Brenus to collect the right data, in the right patients, with the right assays, in a prospectively designed biomarker-enriched cohort. CrisPRO is a hypothesis-generation and analysis platform, not a substitute for clinical validation.

THE ENGAGEMENT PRIORITY MATRIX

For a next-generation STC-1010 program partner, CrisPRO recommends the following engagement sequence:

WHAT MAKES THIS PROGRAM SCIENTIFICALLY INTERESTING TO CRISPRO

An STC-1010-type program is scientifically interesting precisely because it is hard. The haptenation mechanism — generating a broad polyclonal immune response against >200 tumor antigens via DNFB haptenation and immunogenic cell death — is mechanistically distinct from every prior MSS CRC IO approach. It does not rely on a single neoantigen, a single checkpoint, or a single immune pathway. That breadth is both the scientific rationale and the analytical challenge.

The analytical challenge is this: with >200 antigens, which ones are driving the immune response? Which HLA alleles are presenting them? Which patients have tumors that express the relevant antigens? Which patients have the antigen presentation machinery (TAP, MHC-I) to present them? These are questions CrisPRO is built to answer — but only with access to the antigen characterization dataset and the Phase I biomarker data.

The program is at an inflection point. Phase I safety data is in hand. Phase IIa design decisions are imminent. The biomarker data being collected now will either inform a Phase IIa that can detect a real signal — or it will be analyzed post-hoc and generate hypotheses that are too late to act on.

CrisPRO's value is in making the Phase IIa design decision the right one, before the window closes.

THE HONEST BOTTOM LINE

CrisPRO can help you find the responding subgroup, design a trial that can detect the signal in that subgroup, and build the biomarker strategy that makes the program credible to regulators and partners.

CrisPRO cannot give you a control arm, cannot validate biomarkers without clinical data, and cannot substitute for the peer-reviewed clinical publication that sophisticated partners and regulators will require.

The program has the right mechanism. The question is whether the trial design is built to prove it.

This document represents CrisPRO's honest capability assessment for an STC-1010-type program. Full capability specifications, engagement terms, and data requirements are available under NDA/CDA. Contact CrisPRO to initiate the engagement conversation.