Warning: fopen(/home/virtual/epih/journal/upload/ip_log/ip_log_2023-01.txt): failed to open stream: Permission denied in /home/virtual/lib/view_data.php on line 83 Warning: fwrite() expects parameter 1 to be resource, boolean given in /home/virtual/lib/view_data.php on line 84 The clinical meaning of the area under a Receiver Operating Characteristic curve for the evaluation of the performance of disease markers
Skip Navigation
Skip to contents

Epidemiol Health : Epidemiology and Health

OPEN ACCESS
SEARCH
Search

Articles

Page Path
HOME > Epidemiol Health > Accepted Articles > Article
Methods The clinical meaning of the area under a Receiver Operating Characteristic curve for the evaluation of the performance of disease markers
STEFANO PARODI1orcid , Damiano Verda2orcid , Francesca Bagnasco3orcid , Marco Muselli4orcid
Epidemiol Health 2022;e2022088
DOI: https://doi.org/10.4178/epih.e2022088 [Accepted]
Published online: October 17, 2022
  • 736 Views
  • 47 Download
  • 0 Crossref
  • 0 Scopus
1IRCCS Istituto Giannina Gaslini, GENOVA, Italy
2Rulex Innovation Labs, Via Felice Romani 9/2, 16122, Genova, Italy
3IRCCS Istituto Giannina Gaslini, Genova, Italy
4Rulex Innovation Labs, Via Felice Romani 9/2, 16122 Genoa, Italy, and Institute of Electronics, Computer and Telecommunication Engineering, National Research Council of Italy, Via De Marini, 6, 1614v9, Genova, Italy
Corresponding author:  STEFANO PARODI,
Email: stefanoparodi@gaslini.org
Received: 22 June 2022   • Revised: 5 October 2022   • Accepted: 17 October 2022

The area under a Receiver Operating Characteristic (ROC) curve (AUC) is a popular measure of pure diagnostic accuracy, which is independent from the proportion of diseased subjects in the analysed sample. However, its actual usefulness in clinical setting has been questioned, because it does not seem directly related to the actual performance of a diagnostic marker in identifying diseased and non-diseased subjects in real clinical settings. This study evaluates the relation between AUC and the proportion of correct classifications (global diagnostic accuracy, GDA) in relation to the shape of the corresponding ROC curves. We demonstrate that AUC represents an upward biased measure of GDA at an optimal accuracy cut-off for balanced groups. The size of bias depends on the shape of the ROC plot and on the proportion of diseased and non-diseased subjects. In proper curves the bias is independent from the diseased/non-diseased ratio and can be easily estimated and removed. Moreover, the comparison between two partial AUCs can be replaced by a more powerful test for the corresponding whole AUCs. Applications to three real data sets are provided, which include: a marker for a hormone deficit in children; two tumour markers for malignant mesothelioma; two gene expression profiles in ovarian cancer patients. AUC is a measure of accuracy with a potential clinical relevance for the evaluation of disease markers. Clinical meaning of ROC parameters should always be evaluated analysing the shape of the corresponding ROC curve.


Epidemiol Health : Epidemiology and Health