A Fully Automated Multi-Scale Pipeline for Oral Epithelial Dysplasia Grading and Outcome Prediction

Adam Shephard, Neda Azarmehr, Raja Muhammad Saad Bashir, SHAN E AHMED RAZA, Hanya Mahmood, Syed Ali Khurram, Nasir Rajpoot

Keywords: Oral Epithelial Dysplasia, Histopathology, HoVer-Net+, Multiple Instance Learning, Deep Learning, Instance Segmentation, Segmentation

TL;DR: Oral epithelial dysplasia grading and outcome prediction using whole slide images.

Abstract: Oral epithelial dysplasia (OED) is a premalignant histopathological diagnosis given to lesions of the oral cavity, characterised by changes to the nuclear morphometry and the epithelial layers. In this work, we have finetuned HoVer-Net+ for the simultaneous segmentation of nuclei and the epithelial layers in heamatoxylin and eosin (H&E) stained whole slide images (WSIs). We then employed a multi-scale attention-based multiple instance learning architecture for the prediction of OED status, grade, recurrence and malignant transformation. The impressive results have demonstrated the potential of such methods.

Registration: I acknowledge that acceptance of this work at MIDL requires at least one of the authors to register and present the work during the conference.

Authorship: I confirm that I am the author of this work and that it has not been submitted to another publication before.

Paper Type: novel methodological ideas without extensive validation

Primary Subject Area: Application: Histopathology

Secondary Subject Area: Segmentation

Confidentiality And Author Instructions: I read the call for papers and author instructions. I acknowledge that exceeding the page limit and/or altering the latex template can result in desk rejection.