Nat commun: artificial intelligence assisted diagnosis and treatment of prostate cancer

July 24, 2018 / BIOON / - the latest research results of scientists from Sweden show that data-driven AI can help us understand more about how prostate cancer develops, and ultimately help improve the clinical diagnosis and treatment results of this disease Each kind of cancer is unique, and its characteristics change with time The so-called tumor heterogeneity is due to the competitive cloning of cells in the tumor and acquired gene mutations that may increase the probability of metastasis Photo source: kth the Royal Institute of technology researchers from the Swedish life sciences laboratory have developed an AI method for data training to help us better understand the main events related to prostate cancer and its surrounding microenvironment heterogeneity Under the leadership of Joakim Lundeberg, Professor of molecular biology at kth, the research team from the Royal Institute of Technology (KTH) and Karolinska Research Institute in Sweden obtained data from 6750 tumor tissue samples for spatial transcriptomics (a histological method for quantitative analysis of genes, developed jointly by researchers at kth and Karolinska Research Institute) Relevant research results were recently published in nature communications Lundeberg said the spatial information played a crucial role Compared with the traditional molecular methods of tumor samples, this method has a great advantage in prostate cancer gene activity "We have found significant differences in gene activity in cancer cells and non cancer cells from different parts of the same prostate cancer tissue, for example, cells related to inflammation may be related to patient prognosis." These abundant information about gene expression patterns obtained by AI methods can not be found by naked eyes Therefore, the analysis results of these large amount of tissue samples can be used as the basis of clinical diagnosis of tumor tissue based on AI, and can be used to explore gene expression patterns in tumor microenvironment "Ai simply helped us create a digital tissue anatomy." He said He said further exploration of the underlying mechanisms behind cancer is crucial to studying tumor progression and how patients respond to treatment Molecular data have been successfully used to guide the treatment of other epithelial cancers, such as breast cancer Emilie Berglund, co-author of the study, said the new study showed that the data could also help prostate cancer "Early treatment of prostate cancer is effective, but it's still difficult to distinguish between possible progression to malignancy and which patients will benefit from it." Berglund said "We hope this research can help in these areas." Reference materials: emelie Berglund et al Spatial maps of prostate cancer scripts review an unexplored landscape of ergonomics, nature communications (2018) Doi: 10.1038/s41467-018-04724-5