Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Comparing three-class diagnostic tests by three-way ROC analysis.

S Dreiseitl1, L Ohno-Machado, M Binder

  • 1Brigham and Women's Hospital, Division of Health Sciences and Technology, Harvard Medical School, Massachusetts Institute of Technology, Boston, USA. sdreisei@dsg.harvard.edu

Medical Decision Making : an International Journal of the Society for Medical Decision Making
|August 10, 2000
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

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

Sort by
Same author

Adoptive cell therapy with tumor-infiltrating lymphocytes in combination with nivolumab in patients with advanced melanoma.

Immuno-oncology technology·2025
Same author

Adverse impact of a high CD4/CD8 ratio in the allograft may be overcome by methotrexate- but not mycophenolate- or post-transplant cyclophosphamide-based graft versus host disease prophylaxis.

European journal of haematology·2023
Same author

Pre-activated antiviral innate immunity in the upper airways controls early SARS-CoV-2 infection in children.

Nature biotechnology·2021
Same author

101 <i>Dothideomycetes</i> genomes: A test case for predicting lifestyles and emergence of pathogens.

Studies in mycology·2020
Same author

Correction: Role of growth arrest-specific gene 6-mer axis in multiple myeloma.

Leukemia·2020
Same author

Prosthetics in antiquity-An early medieval wearer of a foot prosthesis (6th century AD) from Hemmaberg/Austria.

International journal of paleopathology·2018
Same journal

Flexible Survival Extrapolation with Blended Hazards: Accounting for Treatment Effect Waning in Health Technology Assessment.

Medical decision making : an international journal of the Society for Medical Decision Making·2026
Same journal

A Microsimulation Model for Chronic Kidney Disease Progression in Type 2 Diabetes Patients in the United States: Michigan Model for Diabetes-Chronic Kidney Disease Model.

Medical decision making : an international journal of the Society for Medical Decision Making·2026
Same journal

Cardiovascular Risk Estimation and Statin Adherence: A Historical Cohort Study.

Medical decision making : an international journal of the Society for Medical Decision Making·2026
Same journal

Taste or Scale? Methodological Approach to Health Preferences Comparison across Groups.

Medical decision making : an international journal of the Society for Medical Decision Making·2026
Same journal

Mind the Gap: Impact of New Labels on Public Perceptions and Calculated Risk of Adverse Outcomes after a Melanoma In Situ Diagnosis-A Secondary Analysis of an Online Randomized Experiment.

Medical decision making : an international journal of the Society for Medical Decision Making·2026
Same journal

A Metamodel-Based General-Purpose Autocalibration Tool for Simulation Models.

Medical decision making : an international journal of the Society for Medical Decision Making·2026
See all related articles

This study introduces a method to estimate variance for three-way ROC surfaces, enabling statistical comparisons of diagnostic tests. Neural networks trained on clinical data improved performance for pigmented skin lesion classification.

Area of Science:

  • Biostatistics
  • Machine Learning
  • Medical Diagnostics

Background:

  • Receiver Operating Characteristic (ROC) analysis is standard for binary diagnostic tests.
  • Generalizing ROC analysis to three-class diagnostic tests requires new statistical measures.
  • The volume under the ROC surface quantifies discriminatory power for three-class tests.

Purpose of the Study:

  • To develop a method for calculating nonparametric variance estimates of the volume under the ROC surface.
  • To enable statistical comparison of two diagnostic tests with three possible outcomes.
  • To validate the method and apply it to a real-world dataset.

Main Methods:

  • Utilized Mann-Whitney U statistics for nonparametric variance estimation.
  • Extended the method to calculate covariance estimates for the volume under the ROC surface.

Related Experiment Videos

  • Validated derived formulas using synthetic data and applied to a dataset of pigmented skin lesions.

    • Main Results:

    • A method for estimating variance and covariance of the volume under the ROC surface was successfully developed.
    • The method allows for statistical comparison of three-class diagnostic tests.
    • A neural network trained on both clinical data and lesion features outperformed one trained solely on lesion features for classifying pigmented skin lesions.

    Conclusions:

    • The proposed method provides a robust statistical framework for evaluating three-class diagnostic tests.
    • The findings support the use of combined clinical and feature data for improved diagnostic accuracy in machine learning models.
    • This work advances the statistical methodology for analyzing complex diagnostic scenarios.