Predicting Endometrial Cancer Subtypes and Molecular Features from Histopathology Images Using Multi-resolution Deep Learning Models
AbstractThe determination of endometrial carcinoma histological subtypes is a critical diagnostic process that directly affects patients’ prognosis and treatment options. Recently, molecular subtyping and mutation status are increasingly utilized in clinical practice as they offer better inform prognosis and offer the possibility of individualized therapies. Compared to the histopathological approach, however, the availability of molecular subtyping is limited as it can only be obtained by genomic sequencing, which may be cost prohibitive. Here, we implemented deep convolutional neural network models that predict not only the histological subtypes, but also molecular subtypes and 18 common gene mutations based on digitized H&E stained pathological images. Taking advantage of the multi-resolution nature of the whole slide images, we introduced a customized architecture, Panoptes, to integrate features of different magnification. The model was trained and evaluated with images from The Cancer Genome Atlas and Clinical Proteomic Tumor Analysis Consortium. Our models achieved an area under the receiver operating characteristic curve (AUROC) of 0.969 in predicting histological subtype and 0.934 to 0.958 in predicting the copy number high (CNV-H) molecular subtype. The prediction tasks of 4 mutations and microsatellite high (MSI-H) molecular subtype also achieved a high performance with AUROC ranging from 0.781 to 0.873. Panoptes showed a significantly better performance than InceptionResnet in most of these top predicted tasks by up to 18%. Feature extraction and visualization revealed that the model relied on human-interpretable patterns. Our results suggest that Panoptes can help pathologists determine molecular subtypes and mutations without sequencing, and our models are generalizable to independent datasets.