Radiation Oncology-Biology Integration Network (ROBIN): Bridging the gap between biological research and clinical practice

Authors

Fabiana Gregucci

Christine Lee-Poturalski

Madelyn Dirrim

Jadyn Stewart

Douglas Caruthers

Charles D. Stiles

Kathy Hentrich

Jessica Cardenas

Timothy A. Chan

David S. Yu

Phuoc Tran

Nicole Simone

Amit Sawant

Julie Schwarz

Clifford Robinson

Daphne Haas-Kogan

Franziska Michor

Claire Vanpouille-Box

John Ng

Yasmin Hasan

Jennifer Yu

Lei Ren

Luigi Marchionni

Geoffrey Hugo

Ross Berbeco

Jean L. Nakamura

James Lindsay

Xiaofeng Yang

Joseph O. Deasy

Ralph R. Weichselbaum

Silvia C. Formenti

Doi

10.1158/1078-0432.CCR-25-1216

PMID: 41941260 · DOI: 10.1158/1078-0432.CCR-25-1216 · Journal: Clin Cancer Res (2026)

TL;DR

This is a white-paper / program review of the NIH/NCI-funded Radiation Oncology–Biology Integration Network (ROBIN), a U54 consortium that embeds mechanistic biology inside prospective radiation-therapy (RT) clinical trials PMID:41941260. ROBIN comprises five disease-focused centers — OligoMET (oligometastatic prostate cancer), ImmunoRad (rectal cancer), GenRad (bladder cancer and HNSCC), METEOR (cervical and pancreatic cancer), and KIDSROBIN (pediatric diffuse midline glioma and high-risk neuroblastoma) — each running a Molecular Characterization Trial (MCT) with longitudinal biospecimen collection, single-cell and spatial multi-omics, advanced imaging (CBCT, MRI), and integrative data science. The paper summarizes the consortium’s mission, governance (Admin, CTC, DSIA, and CBCT working groups), and preliminary findings across centers; no new primary data are analyzed PMID:41941260.

Cohort & data

  • Design: consortium white paper / review, not a primary-data study. Data availability statement explicitly notes: “No new data were generated or analyzed for this review” PMID:41941260.
  • Funded centers (U54 grants):
    • OligoMET — U54 CA273956 — University of Maryland Baltimore, Weill Cornell Medicine, Thomas Jefferson University PMID:41941260.
    • ImmunoRad — U54 CA274291 — Weill Cornell + 9 collaborating centers in the U.S. and Europe (MSKCC, University of Chicago, Rutgers, Cedars-Sinai, ICR/Royal Marsden, Glasgow, Manchester, Leeds, Gustave Roussy) PMID:41941260.
    • GenRad — U54 CA274513 — Cleveland Clinic and Emory University PMID:41941260.
    • METEOR — U54 CA274318 — Washington University, St. Louis PMID:41941260.
    • KIDSROBIN — U54 CA274516 — Harvard University and University of California, San Francisco PMID:41941260.
  • Active trials referenced:
    • TERPS (NCT05223803) — Phase II RCT of metastasis-directed therapy in oligometastatic PRAD; 80+ screened, 47 randomized (23 UMB, 9 UPMC, 5 each at UCSD, Virginia Bon Secours, Thomas Jefferson) PMID:41941260.
    • KNIGHTS (NCT06212583) — first integral-biomarker-designed randomized SABR trial in oligometastatic prostate cancer, catalyzed by OligoMET metabolic signatures PMID:41941260.
    • ImmunoRad MCT1 (NCT05943210) — preoperative short-course radiotherapy (SCRT) in READ; 39/50 enrolled, 140+ tissue samples banked and processed for spatial transcriptomics PMID:41941260.
    • ImmunoRad MCT2 (NCT05024097) — Phase I/II SCRT (25 Gy/5 fx) + etrumadenant (A2AR/A2BR antagonist) + zimberelimab (anti–PD-1) + FOLFOX in rectal cancer; 21 enrolled (Part I 5 evaluable; Part II Stage 1 15 accrued, 11 completed) PMID:41941260.
    • RAD-SG (NCT05833867, GenRad MCT part A) — RT + sacituzumab-govitecan for organ-preservation in muscle-invasive BLCA PMID:41941260.
    • GenRad HNSCC MCT (NCT03521570, part 2) — IMRT re-irradiation ± PD-1 inhibition vs. chemoradiation in recurrent/second-primary HNSC; trial completed and results reported in PMID:38780927.
    • METEOR-CRATR (NCT05975593) — longitudinal sampling in locally advanced CESC and PAAD; 26 cervical + 12 pancreatic patients enrolled, 75 tumor samples (65 cervical, 10 pancreatic), 22 cervical and 3 pancreatic with serial samples, 126 DICOM-RT plans archived, 11 cervical patients underwent experimental DBSI MRI PMID:41941260.
    • PNOC023 and COG ANBL1531 — pediatric cooperative-group trials in DMG and high-risk NBL; 33 patients accrued to PNOC023 arms A/B; ANBL1531 accrual complete PMID:41941260.
  • Additional KIDSROBIN data assets: 2,039 serial MRI scans from 253 DMG patients; 8,057 clinical notes from 99 DMG patients and 7,597 clinical notes from 182 NBL patients secured for radiomics and NLP studies PMID:41941260.
  • Assays referenced: 10x Genomics Visium HD spatial transcriptomics (visium-hd), 10x Xenium single-cell spatial profiling (xenium), single-cell RNA-seq (scrna-seq), bulk RNA-seq, 16S/ITS microbiome sequencing (16s-rrna-seq), ctDNA nucleosome positioning (ctdna-dynamics), serum miRNA profiling, flow-cytometry immune profiling (multiparameter-flow-cytometry), plasma proteomics, T-cell receptor repertoire, digital pathology / multimodal AI, CBCT delta radiomics (cbct-delta-radiomics), MRI (including DBSI), 123I-/131I-MIBG SPECT/CT dosimetry, and DICOM-RT plan archiving (dicom-rt-planning) PMID:41941260.

Key findings

This is a program description; the items below are preliminary results the authors highlight from each center.

  • OligoMET (oligometastatic PRAD): digital-pathology multimodal AI, plasma proteomics, and T-cell receptor repertoire analyses yielded signatures associated with stereotactic ablative radiation (SABR) benefit. Preclinical dietary interventions modulated radiosensitivity — low-fat and calorie-restricted regimens enhanced radiation response, whereas ketogenic diets unexpectedly promoted radio-resistance. Systemic metabolic signatures linked to SABR-MDT response informed the biomarker-designed KNIGHTS (NCT06212583) trial PMID:41941260.
  • ImmunoRad (READ): visium-hd spatial transcriptomics on 140+ biopsies identified RT-associated spatial reorganization of lymphocytes and altered macrophage wound-healing signals. scrna-seq of pre/post-RT PBMCs showed stress responses in B cells, ECM-remodeling/immune-activation in monocytes, and MYC-driven activation in CD8+ T cells — interpreted as RT-induced trained immunity. 16S/ITS sequencing showed post-RT increases in Clostridiales and Trichocomaceae. In MCT2, preliminary CR rate was 1/5 (20%) in Part I and 11/15 stage-1 patients had 82% CR in Part II (SCRT + etrumadenant + zimberelimab + FOLFOX) PMID:41941260.
  • GenRad (BLCA and HNSC): RT + sacituzumab-govitecan (RAD-SG, NCT05833867) for definitive bladder cancer was safe and feasible with only grade-1 toxicities observed at preliminary read-out, with comprehensive genomics, serum proteomics, imaging, and ctDNA analysis underway. ADC+RT increased myeloid-lineage immune cells during treatment — distinct from the molecular drivers of efficacy in patients treated with ICI alone. In recurrent/second-primary HNSCC (NCT03521570), IMRT re-irradiation + PD-1 inhibition was tolerable and effective; shorter PFS correlated with a 1.5-fold increase in PD1+Ki67+CD4+ T cells from baseline to week 2–4 PMID:38780927. Analysis of the Javelin HN100 chemoradiation ± avelumab expansion cohort identified intratumoral bacteria in HNSCC as a major source of therapy resistance PMID:41941260.
  • METEOR (CESC and PAAD): among 26 cervical and 12 pancreatic patients (75 tumor samples), chemoradiation (CRT) remodeled the TME with consistent myeloid-derived-cell infiltration and upregulation of TP53/MDM2 in cervix tumor cells, without significant upregulation of CD274/PDCD1 axes. SBRT altered cDC1 infiltration in human and murine pancreatic tumors with impaired antigen-presenting-cell phenotype (altered cDC1/cDC2 ratios) in draining lymph nodes. HPV genotype and gene expression, rather than histology, drove cervical tumor biology PMID:41941260.
  • KIDSROBIN (pediatric DMG and NBL): Xenium (xenium) single-cell spatial profiling on baseline and post-131I-MIBG second-look neuroblastoma samples identified differentially depleted/enriched cell populations. ctDNA nucleosome positioning identified differential transcription-factor profiles in samples with subsequent ctDNA rise; serum miRNA profiling identified differentially modulated miRNAs post-131I-MIBG. Paired 123I-/131I-MIBG SPECT/CT enabled tumor-absorbed dose estimation per unit administered activity. The team frames pediatric tumors as a tractable model for intra-tumoral heterogeneity given low mutational burden PMID:41941260.
  • Infrastructure: the DSIA WG targets FAIR data via the NCI Cancer Research Data Commons, adopting XNAT for imaging informatics and pycerr (Python Computational Environment for Radiotherapy Research) with AI auto-segmentation and radiomics. The CBCT WG standardizes Cone-Beam CT acquisition protocols across institutions to enable cbct-delta-radiomics (high-temporal-resolution treatment-response imaging), exploring HyperSight CBCT, AI-driven image enhancement, and integration with MR/CT/PET/genomic datasets PMID:41941260.

Genes & alterations

The paper does not make gene-level variant claims; genes are discussed as functional markers or pathway readouts in preliminary results.

  • MYC — MYC-driven transcriptional activation observed in CD8+ T cells post-RT in rectal-cancer PBMC scRNA-seq (ImmunoRad), interpreted as part of an RT-induced trained-immunity signal PMID:41941260.
  • TP53 / MDM2 — p53/MDM2 axis upregulated in human cervix tumor cells after chemoradiation in METEOR Project 1 PMID:41941260.
  • CD274 (PD-L1) / PDCD1 (PD-1) — the PD-1/PD-L1 axis was not significantly upregulated after CRT in the METEOR cervical cohort PMID:41941260. In GenRad HNSCC, PD1+Ki67+CD4+ T-cell expansion associated with shorter PFS under re-irradiation + PD-1 blockade PMID:38780927.

No somatic mutation, copy-number, fusion, or gene-panel profiling is presented as primary data in this review.

Clinical implications

  • Biomarker-driven radiation oncology: the MCT design — longitudinal tissue, blood, stool, and imaging across pre-, on-, and post-RT timepoints — is framed as a scalable template for generating actionable radiation-response biomarkers. OligoMET’s metabolic and imaging signatures already catalyzed the KNIGHTS integral-biomarker RCT (NCT06212583) PMID:41941260.
  • Radiation + systemic therapy combinations: consortium data support feasibility and biological activity of RT combined with antibody-drug conjugates (sacituzumab-govitecan in MIBC), adenosine-pathway + PD-1 inhibitors (etrumadenant + zimberelimab with FOLFOX in rectal cancer), and PD-1 blockade (nivolumab) in re-irradiated HNSCC PMID:41941260.
  • Intratumoral microbiome as a resistance axis: HNSCC intratumoral bacteria identified in the Javelin HN100 expansion cohort are highlighted as a newly druggable resistance mechanism for chemoradiation-immunotherapy combinations PMID:41941260.
  • Disparities research: OligoMET explicitly investigates molecular drivers of poorer outcomes in African-American prostate-cancer patients with the goal of informing trials beyond prostate cancer PMID:41941260.
  • Pediatric translatability: KIDSROBIN positions single-cell and spatial profiling of genetically “simple” pediatric tumors (low TMB) as a tractable model for intra-tumoral heterogeneity whose mechanistic findings are expected to generalize to adult cancers PMID:41941260.
  • Imaging-omics integration: CBCT delta radiomics is positioned as a research axis for predicting RT response and normal-tissue toxicity once protocols are harmonized across institutions PMID:41941260.

Limitations & open questions

  • No primary data. This is a consortium white paper; every quantitative result is labeled preliminary and none has undergone formal peer-reviewed primary publication within this article PMID:41941260.
  • Accrual is ongoing for most MCTs (e.g. ImmunoRad MCT1 at 39/50; OligoMET TERPS at 47 randomized; PNOC023 at 33) — statistical power for the stated biomarker claims is not yet established PMID:41941260.
  • Small sample sizes in several preliminary readouts (e.g. 11/15 patients in MCT2 Part II Stage 1; 10 patients contributing spatial profiling from Glasgow; Part I CR 1/5) mean effect sizes may be unstable PMID:41941260.
  • Assay heterogeneity across sites — the DSIA and CBCT working groups explicitly flag the need for harmonized ontologies, acquisition protocols, and artifact correction before cross-center integrative analysis is robust PMID:41941260.
  • Translation gap for radiation-immunotherapy — the authors note that despite strong preclinical signal, successful translation of RT-induced immune effects to the clinic is still lacking PMID:41941260.
  • CBCT quantitative limits: scatter, noise, and metal artifacts restrict most current CBCT applications to anatomy-based setup, motivating the WG’s correction and harmonization work PMID:41941260.
  • Open: how RT-induced microbial shifts (Clostridiales, Trichocomaceae; intratumoral bacteria in HNSCC) causally shape response vs. resistance — the consortium flags this as a major forward-looking mechanistic question PMID:41941260.
  • Open: whether the “trained immunity” / MYC-driven CD8+ activation signal in PBMCs post-RT is a reproducible systemic-response biomarker across disease sites PMID:41941260.

Citations from this paper used in the wiki

  • “The Radiation Oncology-Biology Integration Network (ROBIN) initiative addresses critical gaps in radiation oncology by integrating advanced biological research, technological innovation, and clinical practice.” — Abstract.
  • “Currently, the ROBIN program comprises five centers, including OligoMET, ImmunoRad, GenRad, METEOR, and KIDSROBIN (Table 1).” — Introduction.
  • “To date, over 80 patients have been screened, with 47 enrolled and randomized” — OligoMET TERPS (NCT05223803) enrollment.
  • “Project 2 preclinical studies demonstrated that dietary interventions can modulate radiosensitivity; low-fat and calorie-restricted regimens enhanced radiation responses, whereas ketogenic diets unexpectedly promoted radio-resistance.” — OligoMET.
  • “Over 140 tissue samples have been banked, extensively quality-controlled, and processed for spatial transcriptomic profiling using the 10x Genomics Visium HD platform.” — ImmunoRad MCT1.
  • “MYC-driven activation in CD8+ T cells, suggesting that RT induces systemic immune reprogramming, aligning with the concept of trained immunity.” — ImmunoRad scRNA-seq.
  • “16S/ITS sequencing of stool samples revealed shifts in bacterial and fungal populations – notably, increased Clostridiales and Trichocomaceae post-RT.” — ImmunoRad microbiome.
  • “In part II, 11/15 accrued patients to stage 1 completed treatment and are assessed for response, reporting a CR rate of 82%.” — ImmunoRad MCT2 (NCT05024097).
  • “The preliminary results show that RT+SG for definitive bladder cancer treatment is safe and feasible, with only grade 1 toxicities observed.” — GenRad RAD-SG (NCT05833867).
  • “A shorter progression-free survival was associated with a 1.5-fold increase in PD1+Ki67+CD4+ T cells from baseline to week 2 or 4 of treatment” — GenRad HNSCC MCT (NCT03521570).
  • “One notable finding is that intratumoral bacteria in HNSCC is a major source of resistance to therapy.” — GenRad Javelin HN100 analysis.
  • “CRT remodels the TME in consistent ways: inducing infiltration by novel populations of myeloid-derived cells and upregulating p53/MDM2 in human cervix tumor cells, while not significantly upregulating PDL-1 and PD-1 axes.” — METEOR Project 1.
  • “KIDSROBIN has also secured 2,039 serial MRI scans from 253 patients for DMG radiomics studies, and 8,057 clinical notes from 99 patients with DMG and 7,597 clinical notes from 182 patients with NBL for natural language processing (NLP) studies.” — KIDSROBIN.
  • “XNAT, an open-source imaging informatics platform [62], has been adopted for secure, multi-user, scriptable data management.” — DSIA WG.
  • “Available tools include Wikis, Jupyter notebooks, R Studio environments, and integration with pyCERR – a Python-based evolution of the Computational Environment for Radiotherapy Research with AI auto-segmentation tools and radiomics capabilities.” — DSIA WG.

This page was processed by crosslinker on 2026-05-04.