Literature DB >> 10845401

A contaminated binormal model for ROC data: Part I. Some interesting examples of binormal degeneracy.

D D Dorfman1, K S Berbaum, E A Brandser.   

Abstract

RATIONALE AND
OBJECTIVES: Receiver operating characteristic (ROC) data with false-positive fractions of 0 are often difficult to fit with standard ROC methods and are sometimes discarded. Some extreme examples of such data were analyzed to evaluate the nature of these difficulties.
MATERIALS AND METHODS: Rating reports of fracture for single-view ankle radiographs were analyzed with the binormal ROC model and with two ROC models that keep the ROC curve from crossing the chance line. Because fractures were almost never reported that were not present, some views and locations yielded only ROC points with false-positive fractions of 0, while others yielded at least one ROC point with a non-0 false-positive fraction.
RESULTS: The models tended to yield ROC areas close to or equal to 1. ROC areas of 1 imply a true-positive fraction close to 1; yet the data contained no such fractions. When all false-positive fractions were 0, the true-positive fraction could be much higher for one view than another for all observers. ROC areas gave little or no hint of these unmistakable differences in performance.
CONCLUSION: These data challenge the validity and robustness of current ROC models. A key aspect of ankle fractures is that some may be visible on one view but not at all visible on another.

Entities:  

Mesh:

Year:  2000        PMID: 10845401     DOI: 10.1016/s1076-6332(00)80382-7

Source DB:  PubMed          Journal:  Acad Radiol        ISSN: 1076-6332            Impact factor:   3.173


  9 in total

1.  ROC curves predicted by a model of visual search.

Authors:  D P Chakraborty
Journal:  Phys Med Biol       Date:  2006-07-06       Impact factor: 3.609

2.  Comparing areas under receiver operating characteristic curves: potential impact of the "Last" experimentally measured operating point.

Authors:  David Gur; Andriy I Bandos; Howard E Rockette
Journal:  Radiology       Date:  2008-02-07       Impact factor: 11.105

3.  Experimental design and data analysis in receiver operating characteristic studies: lessons learned from reports in radiology from 1997 to 2006.

Authors:  Junji Shiraishi; Lorenzo L Pesce; Charles E Metz; Kunio Doi
Journal:  Radiology       Date:  2009-10-28       Impact factor: 11.105

4.  Using the mean-to-sigma ratio as a measure of the improperness of binormal ROC curves.

Authors:  Stephen L Hillis; Kevin S Berbaum
Journal:  Acad Radiol       Date:  2011-02       Impact factor: 3.173

5.  Operating characteristics predicted by models for diagnostic tasks involving lesion localization.

Authors:  D P Chakraborty; Hong-Jun Yoon
Journal:  Med Phys       Date:  2008-02       Impact factor: 4.071

6.  Comparative statistical properties of expected utility and area under the ROC curve for laboratory studies of observer performance in screening mammography.

Authors:  Craig K Abbey; Brandon D Gallas; John M Boone; Loren T Niklason; Lubomir M Hadjiiski; Berkman Sahiner; Frank W Samuelson
Journal:  Acad Radiol       Date:  2014-04       Impact factor: 3.173

7.  Arrow plot: a new graphical tool for selecting up and down regulated genes and genes differentially expressed on sample subgroups.

Authors:  Carina Silva-Fortes; Maria Antónia Amaral Turkman; Lisete Sousa
Journal:  BMC Bioinformatics       Date:  2012-06-26       Impact factor: 3.169

Review 8.  Multi-reader multi-case studies using the area under the receiver operator characteristic curve as a measure of diagnostic accuracy: systematic review with a focus on quality of data reporting.

Authors:  Thaworn Dendumrongsup; Andrew A Plumb; Steve Halligan; Thomas R Fanshawe; Douglas G Altman; Susan Mallett
Journal:  PLoS One       Date:  2014-12-26       Impact factor: 3.240

9.  Exploration of analysis methods for diagnostic imaging tests: problems with ROC AUC and confidence scores in CT colonography.

Authors:  Susan Mallett; Steve Halligan; Gary S Collins; Doug G Altman
Journal:  PLoS One       Date:  2014-10-29       Impact factor: 3.240

  9 in total

北京卡尤迪生物科技股份有限公司 © 2022-2023.