AI in Oncology
Research and news where artificial intelligence is the method — diagnosis, drug discovery, pathology, clinical decision support.
First successful phase 3 trial of a personalised mRNA cancer vaccine in melanoma (INTerpath-001)
The phase 3 INTerpath-001 trial met its primary endpoint of recurrence-free survival (RFS) and the key secondary endpoint of distant metastasis-free survival (DMFS) in 1,137 patients with completely resected stage IIB-IV cutaneous melanoma, randomised 2:1 to the personalised mRNA vaccine intismeran …
Read the full story →Computable longitudinal patient journeys from structured and unstructured EHR data
A suite of large pre-trained language models demonstrated the ability to accurately extract computable clinical data from unstructured electronic health records and integrate findings into knowledge graphs representing longitudinal patient journeys. The system facilitates understanding of real-world…
Read the full story →A clinically-oriented foundation model for intraoperative pathology
CRISP, a vision-based foundation model trained exclusively on frozen section slides, outperformed existing foundation models in supporting intraoperative treatment decisions, according to its authors. Validation included a prospective patient cohort. Specialising in frozen sections is intended to ad…
Read the full story →AI model predicts which breast-cancer drugs work best
An AI model trained on millions of protein measurements can predict which drugs will be most effective in tissue samples from people with triple-negative breast cancer. It gauges drug effectiveness directly from proteomic profiling of biopsy material, in a subtype with historically limited targeted …
Read the full story →Proximity-guided graph learning reveals tumour-associated proximity antigens
A proximity-mapping atlas of tumour cell surfaces has defined a new class of targets called tumour-associated proximity antigens — membrane proteins whose spatial co-localisation on cancer cells distinguishes them from healthy tissue. The study identified disease-associated membrane spatial communit…
Read the full story →An operational perturbation proteomics-based virtual cell model
Researchers developed ProteinTalks, a virtual cell model built from temporal protein-abundance measurements in systematically perturbed breast cancer cell lines, enabling computational prediction across diverse drug discovery tasks. The model leverages large-scale perturbation proteomics to capture …
Read the full story →DeepMind’s new genome ‘atlas’ charts effects of all 9 billion human gene mutations
DeepMind's AlphaGenome model computationally predicts the functional consequences of all roughly 9 billion possible single-nucleotide variants in the human genome, estimating for each DNA letter change its effect on gene expression, splicing and regulatory elements. In oncology it is intended to sup…
Read the full story →From algorithms to patient outcomes — lessons from one of the first randomized trials of AI in medicine
A Nature Medicine commentary argues that the next generation of medical AI should be judged not by whether algorithms match clinicians but by whether integrated human–AI systems measurably improve patient outcomes. The authors draw lessons from the first completed randomized controlled trials of AI …
Read the full story →AI-supported decision-making after endoscopic resection of early gastric cancer: what the evidence shows
A review in Clinical Endoscopy examines how machine learning could support the decision made after endoscopic resection of early gastric cancer: proceed to additional gastrectomy with lymph node dissection, or observe. The clinical problem is real — lymph node metastasis occurs in roughly 5-10% of e…
Read the full story →Large language models agreed with tumour board decisions in 71% of thyroid cancer cases, with discordance concentrated in complex ones
In a prospective single-centre study, anonymised clinical data of 59 consecutive thyroid cancer patients discussed by a multidisciplinary tumour board were submitted to two general-purpose language models. ChatGPT 5.2 matched the board's decision in 71.2% of cases (42/59, kappa 0.623) and Gemini 3.0…
Read the full story →Generative AI in neoantigen cancer vaccine design — a review of promise and limits
A review published in Biotechnology Advances (1 September 2026) sets out where generative models can assist in designing therapeutic neoantigen cancer vaccines: tumour antigen discovery, ranking candidates by predicted immunogenicity, and assembling multi-epitope constructs. The authors also state w…
Read the full story →SlideChat is a multimodal generative artificial intelligence assistant for whole-slide computational pathology across cancer types
Chen et al. developed SlideChat, a multimodal generative AI assistant for whole-slide computational pathology, and benchmarked it across 27 pathology tasks spanning 33 cancer types. Expert pathologists rated the system as clinically relevant and accurate, suggesting its potential utility as a diagno…
Read the full story →Artificial intelligence-based tumour infiltrating lymphocyte quantification in patients with triple-negative breast cancer: an independent validation study
An independent external validation (the CATALINA study) tested whether computational tumour-infiltrating lymphocyte scoring (cTIL) matches pathologist assessment in triple-negative breast cancer. Individual data were collated from 1759 patients pooled from seven randomised adjuvant trials, of whom 1…
Read the full story →Manual, digital, and AI tumour-infiltrating lymphocyte scoring: a secondary analysis of the APHINITY randomised trial
In a secondary analysis of the APHINITY randomised trial in HER2-positive early breast cancer, standardised manual scoring of stromal tumour-infiltrating lymphocytes (sTILs) proved reproducible, while digital and AI-based methods delivered consistent prognostic and treatment-benefit stratification d…
Read the full story →First successful AI assisted brain tumour surgery saves UK patient’s sight
Neurosurgeons at London's National Hospital for Neurology and Neurosurgery carried out what is described as the world's first brain tumour operation assisted by artificial intelligence analysing the live surgical video feed, rather than pre-operative scans alone. The tumour lay near the pituitary gl…
Read the full story →A knowledge-driven framework for predicting single-cell responses for unprofiled drugs
Feng et al. present MAP, an AI framework that integrates biological mechanism knowledge to predict how individual cells respond to chemical perturbations, including drugs never previously profiled. The system demonstrated improved generalization to untested compounds compared to existing methods and…
Read the full story →Moderna's personalised mRNA vaccine and melanoma recurrence: what the phase 2b trial actually showed, and what comes next
The personalised mRNA-4157 (V940) vaccine combined with pembrolizumab after complete resection of melanoma prolonged recurrence-free survival compared with pembrolizumab alone in the randomised phase 2b KEYNOTE-942 trial (Lancet 2024, n=157: 107 vs 50 patients). At 18 months, 79% of patients in the …
Read the full story →Transfer learning with deployment-covariate recalibration for survival prediction under covariate shift
CoxRTL is a new recalibrated transfer-learning framework that borrows information from larger external cohorts to improve Cox-model survival predictions when the target site has limited training data and no deployment-phase outcome labels. The method explicitly addresses covariate shift — the common…
Read the full story →Local AI agent brings tumor board expertise to hematological malignancies
Nature Cancer reports a locally deployed AI agent, running on a hospital's own infrastructure, intended to reproduce the work of a multidisciplinary tumour board in haematological malignancies. Local deployment means patient data never leave the institution, which is the main advantage claimed over …
Read the full story →AI-based augmentation of oncology clinical trials
A comprehensive review in Nature Reviews Clinical Oncology examines how artificial intelligence can be applied across the entire lifecycle of oncology clinical trials to reduce their historically high failure rates and improve generalizability of results. The authors identify AI opportunities spanni…
Read the full story →This page is educational — it is not medical advice and does not replace consultation with an oncologist. Diagnostic and treatment decisions are made solely by specialist physicians.