.png){kind=icon}
.png){kind=icon}
Return
Artificial intelligence and carcinogenesis
.png){kind=icon}
PRE
AI
DOI：10.1007/s00761-025-01888-w.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
BackgroundIn recent years, artificial intelligence (AI) has emerged as a new tool in oncology that can support in the detection of (pre)neoplastic lesions and deciphering carcinogenesis. In particular, modern AI models open up new perspectives for the identification and characterization of early cancer, for molecular subtyping, and for automated hypothesis generation from large multimodal datasets.Materials and methodsThis narrative review is based on a selective literature search in PubMed on current developments in the use of AI for early cancer detection, carcinogenesis, and research. For this purpose, the terms AI early cancer detection/screening, AI tumor biomarker, AI carcinogenesis, and AI hypothesis generation oncology were used, with a particular focus on recent studies published between 2022 and 2025.ResultsThe scoping assessment reveals a wide range of current and emerging AI approaches, for example for use in early cancer detection. In parallel, multimodal AI methods, serum marker-based tumor screening, and foundation models for histopathology and multi-omics datasets are being intensively studied. Advances in AI agents and hypothesis-generating AI are also being addressed.ConclusionArtificial intelligence is already improving routine cancer diagnostics and can broaden the range of methods available for prevention and early detection. AI systems can furthermore drive precision medicine when combined with clinical, pathological, and molecular genetic data. At the same time, limitations such as model generalizability and safety must be continuously addressed before widespread clinical implementation can be achieved.
Keywords:
Maschinelles Lernen
Intelligente Systeme
Biomarker
Algorithmen
Onkologie
Machine learning
Intelligent systems
Biomarkers
Algorithms
Oncology