Our Science
Collaborative science. Published evidence.
Papers, abstracts and posters from Sonrai scientists and the clinical and academic teams we work with. Each entry says what Sonrai contributed, from multiomic and digital pathology analysis in clinical trials to diagnostics and the translation of AI into practice.
01
Scientific questions
Work starts with a question from the clinical or translational team.
02
Collaborative outputs
Papers and abstracts are written with our partners and published in their own right.
03
Our contribution
Where we have confirmed it, each entry states what Sonrai did.
Clinical trial biomarker analysis
Multiomic and digital pathology analysis of trial data for aglatimagene besadenovec (CAN-2409). Sonrai's part is the analysis.
Prolonged survival after systemic immune reactivation by aglatimagene besadenovec plus valacyclovir in patients with unresectable stage III/IV NSCLC with an inadequate response to ICI
Sonrai's contribution
Sonrai applied multiomics factor analysis (MOFA) to integrate the trial's longitudinal proteomic, flow cytometry and CyTOF data, identifying coordinated immune programmes and examining how they relate to survival and tumour histology. These analyses were exploratory.
J Immunother Cancer 2026;14(9):e015821
Read the paperConference presentations
Advanced analytics identify a differential immune response to CAN-2409+valacyclovir in non-squamous vs squamous NSCLC, linked to improved survival in patients with progressive ICI-refractory NSCLC
Sonrai's contribution
Sonrai applied multiomics factor analysis (MOFA) to integrate flow cytometry, CyTOF and proteomic data from the trial, examining how immune profiles differ between non-squamous and squamous disease.
Abstract 513. J Immunother Cancer 2025;13(Suppl 2):A582
Read the abstractDigital Pathology in Relation to Immune Activation Following Aglatimagene Besadenovec Immunotherapy in Localized Prostate Cancer
Machine learning analysis of digitised H&E prostate biopsies from patients treated with aglatimagene besadenovec plus radiotherapy, quantifying lymphocyte infiltration and the spatial relationship between lymphocytes and tumour.
Late-breaking abstract LBA 30, poster presentation, 2026
Read the abstractClinical and immunologic effects of aglatimagene besadenovec combined with radiotherapy in intermediate-risk localized prostate cancer: integrated results from the PRTK03 and PRTK05 clinical trials
Integrated results from two trials of aglatimagene besadenovec with radiotherapy in intermediate-risk localised prostate cancer, including AI-enabled H&E analysis of lymphocytic infiltration and spatial lymphocyte and tumour interactions.
Prostate Cancer Foundation, 2026
Read the abstractProgrammes and consortia
Collaborative programmes Sonrai has taken part in, and our role in each.
AI tissue diagnostics for colorectal and lung cancer
A programme with Queen's University Belfast's Precision Medicine Centre, funded through the UK AI in Health and Care Award, to develop AI-based tissue tests for cancer diagnosis.
Sonrai's role
Sonrai leads the award, developing and training the AI algorithms on a cloud platform where pathologists can upload biopsy images. Applications named in Sonrai's October 2023 announcement include MSI testing for Lynch syndrome in colorectal cancer and PD-L1 tumour proportion scoring in non-small cell lung cancer.
Future Medicines Institute
A £55m industry-led hub for research, innovation and skills, hosted by Queen's University Belfast with Ulster University. Backed by the Northern Ireland Executive and a seven-member industry consortium, with delivery managed by UKRI.
Sonrai's role
Sonrai is one of the seven industry consortium partners, specialising in cloud-enabled data assimilation software and platforms.
PathLAKE
A £15m UK centre of excellence in AI for pathology, funded through the Industrial Strategy Challenge Fund and led by University Hospitals Coventry and Warwickshire NHS Trust with Warwick, Oxford, Nottingham and Queen's. It set out to build a pathology image data lake and AI techniques for early cancer diagnosis.
Sonrai's role
Sonrai is named as an industry partner. It contributed to the programme's knowledgebase and helped assess around 6,000 annotated digital images from Queen's University Belfast. It also explored how the data could be warehoused, accessed and visualised through its software platform.
ACTIONED
A consortium of more than £7m led by Queen's University Belfast with Roche Diagnostics, supported by Innovate UK. It aimed to combine tissue pathology, genomics and AI in a colorectal cancer diagnostic workflow.
Sonrai's role
Sonrai joined the consortium as the AI and deep learning specialist, focused on big data and bespoke algorithms to improve personalised medicine.
All publications and presentations
Filter by topic or type, or browse everything from our team.
-
2026
Prolonged survival after systemic immune reactivation by aglatimagene besadenovec plus valacyclovir in patients with unresectable stage III/IV NSCLC with an inadequate response to ICI
Sonrai's contribution
Sonrai applied multiomics factor analysis (MOFA) to integrate the trial's longitudinal proteomic, flow cytometry and CyTOF data, identifying coordinated immune programmes and examining how they relate to survival and tumour histology. These analyses were exploratory.
J Immunother Cancer 2026;14(9):e015821
Read the paper -
2026
Digital Pathology in Relation to Immune Activation Following Aglatimagene Besadenovec Immunotherapy in Localized Prostate Cancer
Machine learning analysis of digitised H&E prostate biopsies from patients treated with aglatimagene besadenovec plus radiotherapy, quantifying lymphocyte infiltration and the spatial relationship between lymphocytes and tumour.
Late-breaking abstract LBA 30, poster presentation, 2026
Read the abstract -
2026
Clinical and immunologic effects of aglatimagene besadenovec combined with radiotherapy in intermediate-risk localized prostate cancer: integrated results from the PRTK03 and PRTK05 clinical trials
Integrated results from two trials of aglatimagene besadenovec with radiotherapy in intermediate-risk localised prostate cancer, including AI-enabled H&E analysis of lymphocytic infiltration and spatial lymphocyte and tumour interactions.
Prostate Cancer Foundation, 2026
Read the abstract -
2025
Advanced analytics identify a differential immune response to CAN-2409+valacyclovir in non-squamous vs squamous NSCLC, linked to improved survival in patients with progressive ICI-refractory NSCLC
Sonrai's contribution
Sonrai applied multiomics factor analysis (MOFA) to integrate flow cytometry, CyTOF and proteomic data from the trial, examining how immune profiles differ between non-squamous and squamous disease.
Abstract 513. J Immunother Cancer 2025;13(Suppl 2):A582
Read the abstract -
2025
Integrative discovery of a multimodal cancer immunotherapy using machine learning and viral vector engineering
Abstract 893. J Immunother Cancer 2025;13(Suppl 2):A1011
Read the abstract -
2025
AI driven pre-regulatory validation of PD-L1 analysis in lung cancer
Describes a deep learning tool that assists pathologists in scoring PD-L1 tumour proportion in non-small cell lung cancer, with human review of cases near the clinical thresholds.
Sci Rep 2025;15:45728
Read the paper -
2025
The impact of artificial intelligence on biomarker discovery
A review of the AI toolkit available for biomarker discovery, examples of AI-powered biomarkers across therapeutic areas, and the challenges in bringing them to the clinic.
Emerg Top Life Sci 2025;8(2):89-105
Read the paper -
2024
A comparative analysis of model and age independence of a plasma-based real-time qPCR clinical assay for the non-invasive detection of stage I breast cancer
Evaluates a plasma-based qPCR assay across age groups and classification methods, including XGBoost and logistic regression, using an eight gene mRNA signature.
Clin Cancer Res 2025;31(12 Suppl):P5-10-09. Presented December 2024
Read the abstract -
2024
KRASFormer: a fully vision transformer-based framework for predicting KRAS gene mutations in histopathological images of colorectal cancer
A vision transformer framework for predicting KRAS gene mutations from histopathology images of colorectal cancer.
Biomed Phys Eng Express 2024;10(5)
Read the paper -
2023
Development of a plasma-based real-time qPCR gene expression assay for targeted screening and diagnosis of early-stage breast cancer
Investigates a plasma-based qPCR gene expression assay for detecting early-stage breast cancer, using RNA biomarkers for targeted screening and diagnosis.
Cancer Res 2024;84(9 Suppl):PO1-28-06. Presented December 2023
Read the abstract -
2023
True-T: Improving T-cell response quantification with holistic artificial intelligence based prediction in immunohistochemistry images
A holistic AI-based approach to improving the quantification of T-cell response in immunohistochemistry images.
Comput Struct Biotechnol J 2023;23:174-185
Read the paper -
2023
Translation of tissue-based artificial intelligence into clinical practice: from discovery to adoption
A review of the steps needed to take tissue-based artificial intelligence from discovery through to clinical adoption.
Oncogene 2023;42(48):3545-3555
Read the paper -
2023
Development of a novel blood-based RNA gene expression platform for early-stage lung cancer diagnosis
Applies machine learning to blood-based microRNA expression to explore early detection of lung cancer.
Abstract 8557. J Clin Oncol 2023;41(16_suppl):8557
Read the abstract -
2022
Particle swarm optimization artificial intelligence technique for gene signature discovery in transcriptomic cohorts
A particle swarm optimisation AI technique for discovering gene signatures in transcriptomic cohorts.
Comput Struct Biotechnol J 2022;20:5547-5563
Read the paper -
2022
ICOSeg: Real-Time ICOS Protein Expression Segmentation from Immunohistochemistry Slides Using a Lightweight Conv-Transformer Network
A lightweight conv-transformer network for real-time segmentation of ICOS protein expression in immunohistochemistry slides.
Cancers (Basel) 2022;14(16):3910
Read the paper -
2022
General Roadmap and Core Steps for the Development of AI Tools in Digital Pathology
A roadmap and the core steps for developing AI tools in digital pathology.
Diagnostics (Basel) 2022;12(5):1272
Read the paper