Research AppraisalRandomised Controlled Trial

OpenIO: An open framework for AI-native immunotherapy

Cancer cellWu, Yingcheng, Xiao, Hao, Jiang, Nan et al.13 July 2026DOI

Clinical Snapshot

25CEBM

Evidence: InsufficientRandomised Controlled Trial

PICO Framework

P — Population	Patients with neoplasms (cancer) considered for immunotherapy; also encompasses preclinical and computational model systems
I — Intervention	OpenIO framework — an integrated generative AI and multi-omics platform leveraging biological scaling laws and foundation models for rational engineering of immunotherapeutic interventions
C — Comparator	Conventional empirical immunotherapy screening and development approaches (implicit comparator; no explicit control arm described)
O — Outcomes	Transition from empirical to rational, AI-native design of immunotherapy; precision oncology outcomes (specific clinical endpoints not reported in the abstract)

Bottom Line

OpenIO is a conceptual framework paper proposing the integration of generative AI and multi-omics data to shift immunotherapy development from empirical screening toward rational, model-driven design. Published in Cancer Cell by a consortium of Chinese academic and clinical institutions, it articulates an ambitious vision grounded in biological scaling laws and foundation model architectures. However, as appraised against CEBM criteria, this paper provides no primary clinical data, no patient cohort, no quantified outcomes, and no statistical evidence to support its claims. It represents the lowest tier of clinical evidence — a framework proposal — and should be interpreted accordingly. The concept is scientifically credible and directionally important, but the gap between a proposed framework and a validated clinical tool is substantial. Clinicians should not modify immunotherapy practice based on this paper. Independent prospective validation, regulatory evaluation, and health technology assessment are prerequisites for any clinical translation. Australian oncologists and researchers may find value in engaging with the open framework for collaborative validation, but patient care decisions must continue to be guided by established evidence-based guidelines.

Evidence: Insufficient

Key Findings

P Value: Not reported
Effect Size: Not reported
Primary Outcome: No primary clinical or biological outcome reported; the paper proposes a conceptual framework for AI-native immunotherapy development
Nnt Or Sensitivity: Not applicable — no clinical trial, diagnostic accuracy study, or prognostic cohort data presented
Confidence Interval: Not reported

Clinical Application

Clinical implementation is not currently feasible based on available evidence. The framework requires prospective validation in independent cohorts, regulatory approval of any derived tools, and integration with existing clinical workflows. Infrastructure requirements for multi-omics data generation and computational processing are substantial and not universally available. No direct relevance to current Australian clinical practice can be established. The TGA has not approved any AI-native immunotherapy design tool derived from this framework. The PBS does not list any intervention arising from OpenIO. RACGP and COSA (Clinical Oncology Society of Australia) guidelines do not reference this framework. Australian oncologists should monitor this space but should not alter clinical practice based on this paper alone. The Australian Genomics Health Alliance and national tumour biobanks could theoretically contribute to future validation studies. Equity of access to multi-omics profiling in regional and remote Australia remains a significant barrier to any omics-dependent precision oncology framework. Theoretically applicable to patients with solid tumours being considered for immunotherapy, particularly those with available multi-omics profiling. Practical clinical applicability is not yet demonstrated for any specific cancer type, line of therapy, or patient subgroup.

Abstract

We propose Open Immune Oncology (OpenIO), a framework integrating generative AI and omics to advance precision oncology. By leveraging biological scaling laws and foundation models, we aim to transition immunotherapy from empirical screening to rational, AI-native engineering of therapeutic interventions.

References

1.Wu, Y., Xiao, H., Jiang, N., Hua, W., Ma, J., Ge, J., Liu, Y., Zhang, Z., Chen, J. X., Jin, R., Wang, Y., Zhou, J., Fan, J., Zheng, Z., Bai, L., Ye, H., Liu, Q., Guo, G., Zhang, Z., Sun, S., Guo, T., Zheng, S., & Gao, Q. (2026). OpenIO: An open framework for AI-native immunotherapy. Cancer Cell. https://doi.org/10.1016/j.ccell.2026.06.002

Share:X LinkedIn

Related Research

Systematic Review

15CEBM

Cancer medicine

Applications of Artificial Intelligence in Cancer Diagnosis and Treatment

1 Aug 2026

other

65CEBM

Acta oncologica (Stockholm, Sweden)

User perceptions of machine learning models as decision support for colorectal cancer multidisciplinary team conferences (AID-SIM-2): a qualitative simulation study.