Our publications
hema.to speeds up diagnostic workflow and improves quality
We're incredibly proud to share that hema.to has been clinically validated in an international, four-center clinical trial! Blood cancer diagnoses weren't just twice as fast, they were made with 8% more sensitivity and 10% more specificity when made with the help of our beautiful, easy-to-use decision-support software. You can download the white paper here.
Knowledge transfer enhances performance
... as shown by Nanditha Mallesh and co-authors in this Cell Patterns paper. In this collaboration between the hema.to team, the Munich Leukemia Lab, Charité and additional renowned university clinics, we showed that data from multiple labs can be combined to increase the clinical performance of AI models for each lab.
Deep learning can classify blood cancer on an expert level
... as shown by the hema.to team in collaboration with several experienced hematologists and the university of Bonn in this Cytometry A paper. With this paper, Zhao et al. have shown deep learning can be used to aid the routine clinical workflow using cytometry data.
Knowledge transfer enhances performance
... as shown by Nanditha Mallesh and co-authors in this 2019 Blood paper. With this paper, we showed that improvement of diagnostic performance is generated by combining multiple datasets, an approach we call "knowledge transfer".
Neural nets provide highly reliable diagnostic support in a routine setting
... as shown by a collaboration between hema.to, the Munich Leukemia Lab and the University of Bonn in this 2019 Blood paper. This paper shows that AI can support hematologists in their routine workflows for blood cancer diagnosis by providing highly reliable recommendations.