首页|Washington University School of Medicine Reports Findings in Endometrial Hyperpl asia (End-to-end deep learning method for predicting hormonal treatment response in women with atypical endometrial hyperplasia or endometrial cancer)
Washington University School of Medicine Reports Findings in Endometrial Hyperpl asia (End-to-end deep learning method for predicting hormonal treatment response in women with atypical endometrial hyperplasia or endometrial cancer)
扫码查看
点击上方二维码区域,可以放大扫码查看
原文链接
NETL
NSTL
By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News Daily News-New research on Uterine Diseases and C onditions - Endometrial Hyperplasia is the subject of a report. According to new s originating from St. Louis, Missouri, by NewsRx correspondents, research state d, "Endometrial cancer (EC) is the most common gynecologic malignancy in the Uni ted States, and atypical endometrial hyperplasia (AEH) is considered a high-risk precursor to EC. Hormone therapies and hysterectomy are practical treatment opt ions for AEH and early-stage EC." Our news journalists obtained a quote from the research from the Washington Univ ersity School of Medicine, "Some patients prefer hormone therapies for reasons s uch as fertility preservation or being poor surgical candidates. However, accura te prediction of an individual patient's response to hormonal treatment would al low for personalized and potentially improved recommendations for these conditio ns. This study aims to explore the feasibility of using deep learning models on whole slide images (WSI) of endometrial tissue samples to predict the patient's response to hormonal treatment. We curated a clinical WSI dataset of 112 patient s from two clinical sites. An expert pathologist annotated these images by outli ning AEH/EC regions. We developed an end-to-end machine learning model with mixe d supervision. The model is based on image patches extracted from pathologist-an notated AEH/EC regions. Either an unsupervised deep learning architecture (Autoe ncoder or ResNet50), or non-deep learning (radiomics feature extraction) is used to embed the images into a low-dimensional space, followed by fully connected l ayers for binary prediction, which was trained with binary responder/non-respond er labels established by pathologists. We used stratified sampling to partition the dataset into a development set and a test set for internal validation of the performance of our models. The autoencoder model yielded an AUROC of 0.80 with 95% CI [0.63, 0.95] on the i ndependent test set for the task of predicting a patient with AEH/EC as a respon der vs non-responder to hormonal treatment."
St. LouisMissouriUnited StatesNort h and Central AmericaCancerCyborgsEmerging TechnologiesEndometrial Cance rEndometrial HyperplasiaGynecologyHealth and MedicineHormonesMachine L earningOncologyUterine Diseases and ConditionsWomen's Health