AI for prostate cancer analysis

A new study led by Joakim Lundeberg (KTH/SciLifeLab) demonstrates that tumor heterogeneity in prostate cancers can be characterized using an AI approach. The results are published in Nature Communications.  

Each cancer is unique and also alters its features as it progresses. The underlying genetic changes of this heterogeneity can be of clinical relevance to cancer patients. The current study presents an unsupervised AI-based approach that enables mapping of gene activity patterns that are not detectable using conventional histology.

By applying the unsupervised analysis tool to prostate cancer spatial transcriptomics data, where gene expression have been quantified in distinct segments of a tumor, the researchers present an novel way to gain further insight into the mechanisms behind tumor progression and patient treatment response.

Read the full paper in Nature Communications

Read the press release by KTH Royal Institute of Technology (in Swedish)