This paper proposes an explainable AI framework for oncology decision support, integrating multimodal cancer biomarkers. The system improves prediction accuracy and physician trust, offering transparent reasoning behind treatment recommendations using fused clinical, radiomic, and molecular profiles within smart hospital infrastructures.

Jeremy Owen Radcliffe, Harriet Louise Grafton, Louis Anthony Woodbridge, Charles Peter Ellingham, Beatrice Frances Doyle

This study explores edge-based AI and federated learning to support privacy-preserving, real-time monitoring of vital signs in wearable biomedical devices. The system enables continuous health tracking while ensuring patient confidentiality across distributed networks in digital health environments.

Julian Thomas Whitaker, Annabelle Rose Hensley, Douglas Charles Fenton, Chloe Eleanor Parkhurst, Samuel Victor Brigham

This research introduces a temporal deep neural network using autoencoders to identify anomalies in electronic health records. By capturing sequential dependencies in medical histories, the model supports early detection of outliers in diagnosis, treatment paths, and hospitalization patterns within clinical data repositories.

Frederick Lawrence Shields, Victoria Anne Cartwright, Maxwell Gordon Langston, Georgia Elise Banbridge, Oliver Zachary Whitmore

This study proposes a microservice-oriented architecture for scalable genomic data pipelines in cloud environments. Through Kubernetes-based orchestration, it enhances efficiency, fault tolerance, and resource allocation, making high-throughput genomic analysis workflows more cost-effective and reproducible across biomedical informatics platforms.

Benjamin Isaac Alderidge, Lydia Caroline Mowbray, Nicholas Rupert Haslam, Francesca Violet Harkness, William Thomas O’Connell

This paper presents a knowledge graph embedding approach for personalized drug recommendations. Using heterogeneous biomedical datasets, the system models patient-specific interactions and predicts effective therapies. The model supports informed decision-making in precision medicine and addresses multi-drug compatibility across diverse health profiles.

Henry Rupert Chadwick, Clara Penelope Sykes, Dominic Russell Haversham, Eloise Matilda Griffin, Alfred George Beckett

This study introduces a deep transfer learning model that analyzes longitudinal MRI data to detect early Alzheimer’s indicators. The model improves diagnostic precision, leveraging spatial-temporal imaging features to distinguish progressive cognitive decline patterns in patients with mild cognitive impairment.

Isabelle Margaret Thornton, Edward James Whitmore, Sophia Beatrice Callahan, Thomas Alaric Bremmer, Harriet Joanne Sanderson

This research introduces a predictive model using temporal attention networks to forecast hospital readmission risk. By aggregating data from EHRs, lab reports, and prescription records, the system provides actionable insights to reduce preventable readmissions and support patient-centric clinical workflows.

Archibald Felix Dunmore, Matilda Eleanor Rosebridge, Jonathan Miles Hargrave, Clara Felicity Statham, Oliver Hugh Redgrave