Annotation-free glioma grading from pathological images using ensemble deep learning Running title: Glioma grading using ensemble deep learning

Background: Glioma grading is critical for treatment selection and the fine classification between glioma grades II and III is still a challenging problem. Traditional systems based on single deep learning (DL) model can only achieve relatively low accuracy in distinguishing glioma grades II and III. 
Methods: We developed ensemble DL models by combining DL and ensemble learning techniques to achieve annotation-free glioma grading (grade II or III) from pathological images. We established multiple tile-level DL models using residual network ResNet-18 architecture, then used DL models as component classifiers to develop ensemble DL models to achieve patient-level glioma grading.