Joeky Senders

155 Deep Learning and NLP learning curves Conclusion We developed an open-source natural language processing pipeline that can determine the histopathological diagnosis of brain tumor patients based on free-text pathology reports. A deep learning architecture was identified that learns two to eight times more efficiently than competing deep learning and regressions-based approaches. This framework could facilitate clinical research by providing an automatic and deterministic method for medical text analysis. Supplementary material Supplementary Table S1 and Figure S1 are available online at: https://innovations.bmj.com/content/6/4/192