AI in Oncology
Research and news where artificial intelligence is the method — diagnosis, drug discovery, pathology, clinical decision support.
Development and validation of a machine learning model based on routine laboratory tests for predicting FLT3-ITD mutations in acute myeloid leukaemia.
Song J, et al
Researchers developed and validated a gradient boosting machine (GBM) model using routine laboratory tests to predict FLT3-ITD mutations in newly diagnosed acute myeloid leukaemia (AML), achieving area-under-the-curve (AUC) scores of 0.847 in training, 0.772 in internal temporal validation, and 0.76…
Source →Assessing Foundation Models for Computational Pathology in Endometrial Cancer.
et al
Benchmark of seven state-of-the-art computational-pathology foundation models on 3,293 endometrial cancer patients from randomized trials and clinical cohorts, across morphological, molecular and prognostic tasks. H-OPTIMUS-0, CONCH and VIRCHOW2 achieved the highest mean performance, but the best mo…
Source →Dynamic multimodal survival prediction in multiple myeloma integrating gene expression, longitudinal laboratory measurements, and treatment history.
Jia S, et al
A dynamic multimodal deep learning framework for predicting survival in multiple myeloma achieved a concordance index (C-index) of 0.773 ± 0.024 and a 1-year time-dependent AUC of 0.789 ± 0.021 on a held-out validation cohort of 752 patients, outperforming established baselines including DeepSurv, C…
Source →Artificial intelligence-based tumour infiltrating lymphocyte quantification in patients with triple-negative breast cancer: an independent validation study
Dixon-Douglas JR, et al
This independent external validation (the CATALINA study) tested whether artificial intelligence algorithms that quantify tumour-infiltrating lymphocytes (TILs) in triple-negative breast cancer hold up on data they were not trained on. The material was pooled from seven randomised clinical trials wi…
Source →Interpretable graph-conditioned CNNs for dose-volume histogram prediction in radiotherapy.
Zhu X, et al
A novel deep learning framework combining a three-dimensional convolutional neural network (CNN) with a graph neural network (GNN) directly predicts dose-volume histograms (DVHs) for multiple organs-at-risk in radiotherapy planning, cutting the mean dose prediction error for the brain stem from 1.24…
Source →Spinal meningiomas: histopathological grading using a benchmark radiomics model with notes on disease control.
Palla A, et al
A benchmark radiomics ensemble model combining k-nearest neighbors, random forest, and support vector machine classifiers achieved an area under the receiver operating characteristic curve of 0.879 and a mean F1 score of 0.748 in preoperatively predicting the histological grade of spinal meningiomas…
Source →Deep learning predicts overall survival in high-grade meningioma using MRI and clinical data.
Zaidan SK, et al
A deep learning model integrating T1-weighted MRI with clinical covariates (age, sex, and tumor grade) achieved a mean cross-validated concordance index (c-index) of 0.731 ± 0.112 — ranging from 0.571 to 0.895 across five folds — in predicting overall survival for patients with high-grade meningioma…
Source →Integrative Multi-Omics Analysis Identifies Thrombosis-Associated Molecular Features Linked to Germline Susceptibility and Immune Cell Communication in Gastric Cancer.
Lu X, et al
Researchers identified 22 thrombosis-associated genes differentially expressed in gastric cancer and constructed a machine-learning-based 14-gene prognostic signature that reliably stratified patients into high- and low-risk groups in both training and validation cohorts. Integrative genomic analyse…
Source →Prediction of Ki-67 Status in Breast Cancer Using a Habitat-Guided 2.5D Multiparametric MRI Deep Learning Model.
Miao Z, et al
A habitat-guided 2.5D deep learning model based on multiparametric MRI achieved an area under the curve (AUC) of 0.821 and a sensitivity of 0.804 in predicting Ki-67 proliferation status in breast cancer across an independent cross-vendor test set of 100 patients, significantly outperforming both a …
Source →Integrated multi-omics analyses identify an RAS-SLC11A2-associated molecular framework linking iron metabolism with PCOS-related cardiometabolic risk.
Zhang S, et al
Researchers have identified an RAS-SLC11A2 molecular framework that mechanistically links iron metabolism dysregulation to elevated cardiometabolic risk in polycystic ovary syndrome (PCOS), clarifying how ovarian dysfunction, chronic inflammation, oxidative stress, and hypertension are molecularly i…
Source →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.