Article

The Impact of Digital Histopathology Batch Effect on Deep Learning Model Accuracy and Bias

2020-12-04

Abstract excerpt

The Cancer Genome Atlas (TCGA) is one of the largest biorepositories of digital histology. Deep learning (DL) models have been trained on TCGA to predict numerous features directly from histology, including survival, gene expression patterns, and driver mutations. However, we demonstrate that these features vary substantially across tissue submitting sites in TCGA for over 3,000 patients with six cancer subtypes....

Topics

Open a Topic to create a Post that cites this publication.

Identifiers and source

Literature Corpus work: 4cc1f81a-f4e7-5c8f-8040-1b64802174ea
DOI: 10.1101/2020.12.03.410845

Open publication

Related research

Semantic proximity does not establish scientific evidence.

Click a neighbor to travelStep 1 · 12 closest

The Impact of Digital Histopathology Batch Effect on Deep Learning Model Accuracy and BiasDOI 10.1101/2020.12.03.410845Read

The impact of site-specific digital histology signatures on deep learning model accuracy and bias.Semantic proximity Read

Examining Batch Effect in Histopathology as a Distributionally Robust Optimization ProblemSemantic proximity Read

Evaluating reproducibility of AI algorithms in digital pathology with DAPPERSemantic proximity Read

CPU-READY DEEP LEARNING APPROACH FOR ROBUST TISSUE REGION SEGMENTATION ACROSS MULTI-COHORT H&E AND IHC-STAINED WHOLE SLIDE IMAGESSemantic proximity Read

Universal encoding of pan-cancer histology by deep texture representations.Semantic proximity Read

Deep Texture Representations as a Universal Encoder for Pan-cancer HistologySemantic proximity Read

Predicting cancer outcomes from histology and genomics using convolutional networksSemantic proximity Read

Built to last? Reproducibility and Reusability of Deep Learning Algorithms in Computational PathologySemantic proximity Read

Deceptive learning in histopathology.Semantic proximity Read

Weakly supervised contrastive learning infers molecular subtypes and recurrence of breast cancer from unannotated pathology imagesSemantic proximity Read

Cracks in the Foundation: How Data-Hungry and Sensitive to Domain Shift are Vision Foundation Models for Computational Pathology?Semantic proximity Read

The Overlooked Role of Specimen Preparation in Bolstering Deep Learning-Enhanced Spatial Transcriptomics WorkflowsSemantic proximity Read