Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Search research articles

Related Experiment Video

Updated: Nov 19, 2025

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches

Published on: December 15, 2023

Hidden Variables in Deep Learning Digital Pathology and Their Potential to Cause Batch Effects: Prediction Model

Max Schmitt¹, Roman Christoph Maron¹, Achim Hekler¹

¹Digital Biomarkers for Oncology Group, National Center for Tumor Diseases, German Cancer Research Center (DKFZ), Heidelberg, Germany.

Journal of Medical Internet Research

|February 2, 2021

Related Concept Videos

Survival Tree

Survival Tree

Survival trees are a non-parametric method used in survival analysis to model the relationship between a set of covariates and the time until an event of interest occurs, often referred to as the "time-to-event" or "survival time." This method is particularly useful when dealing with censored data, where the event has not occurred for some individuals by the end of the study period, or when the exact time of the event is unknown.
Building a Survival Tree
Constructing a...

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

Smartphone Applications in Dermatology With Special Focus on AI: A European-Continent-Based Review.

International journal of dermatology·2026

Same author

How to benchmark medical AI agents.

PLoS medicine·2026

Same author

AI-based selection of tumor regions for genomic profiling in neuropathology.

Neuro-oncology advances·2026

Same author

Clinical decision support in hematological malignancies using a case-grounded AI agent.

Nature medicine·2026

Same author

Infusion-site reactions to foslevodopa/foscarbidopa in Parkinson's disease.

Journal der Deutschen Dermatologischen Gesellschaft = Journal of the German Society of Dermatology : JDDG·2026

Same author

A deep learning framework for efficient pathology image analysis.

Nature communications·2026

Summary

Artificial intelligence (AI) in digital pathology can be compromised by hidden variables in whole slide images. These variables, such as patient age and scanner type, are learnable by neural networks and can introduce batch effects, impacting cancer diagnosis accuracy.

Area of Science:

Digital pathology
Artificial intelligence in medical imaging
Machine learning for cancer diagnosis

Background:

AI shows promise for cancer diagnosis from whole slide images, necessitating large, diverse datasets.
Data diversification can introduce hidden variables, potentially leading to batch effects and reduced AI accuracy.
Understanding these hidden variables is crucial for developing robust AI systems in pathology.

Purpose of the Study:

To analyze the learnability of common hidden variables in digital pathology datasets.
To investigate potential batch effects caused by patient age, slide preparation date, slide origin, and scanner type.
To assess the impact of these variables on AI-based classification systems.

Main Methods:

Trained four separate convolutional neural networks (CNNs) to learn specific hidden variables.

Keywords:

artifacts artificial intelligence clinical pathology convolutional neural networks deep learning digital pathology machine learning neural networks pathology pitfalls

Related Experiment Videos

Last Updated: Nov 19, 2025

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches

Published on: December 15, 2023

Utilized a dataset of digitized whole slide melanoma images from five institutes.

Ensured robustness through multiple repetitions of CNN training and evaluation, with a balanced accuracy threshold of 50.0%.

Main Results:

All four analyzed hidden variables (patient age, slide preparation date, slide origin, scanner type) were learnable by CNNs.
Mean balanced accuracy exceeded 50.0% for all tasks, with significant variation across variables.
Performance ranged from 56.1% for slide preparation date to 100% for slide origin.

Conclusions:

Learnable hidden variables in dermatopathology datasets can create detrimental batch effects for AI classification.
Awareness of these pitfalls is essential for developing and evaluating AI systems in digital pathology.
Data set stratification is recommended to mitigate batch effect issues in AI development.