Figure - PMC

Skip to main content

An official website of the United States government

Here's how you know

Here's how you know

Official websites use .gov
A .gov website belongs to an official government organization in the United States.

Secure .gov websites use HTTPS
A lock ( ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites.

View full-text article in PMC

. 2015 Jan 20;25(4):932–939. doi: 10.1007/s00330-014-3487-0

Search in PMC
Search in PubMed
View in NLM Catalog
Add to search

© The Author(s) 2015

Open Access This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.

PMC Copyright notice

Fig. 2 — The black ROC curve with circles at diagnostic thresholds shows the data from Fig. 1, with AUC 0.83. The red curve with triangles describes the following data where test specificity is the same as in Fig. 1, but sensitivity has been increased. Here, assuming a rating of 3 or more conveys a diagnosis of cancer, then three positive cases are missed (sensitivity 94 %) and 17 negative cases are labeled positive (specificity 66 %); the AUC is 0.89: The blue ROC curve with squares describes the following data where test sensitivity is the same as Fig. 1, but specificity has been increased. Here, assuming a rating of 3 or more conveys a diagnosis of cancer, then seven positive cases are missed (sensitivity 86 %) and eight negative cases are labeled positive (specificity 84 %); the AUC is 0.89: The black ROC curve with solid circles describes the following data where test sensitivity is increased to the same level as the red curve but with specificity dropped by a corresponding amount. Assuming a rating of 3 or more conveys a diagnosis of cancer, then just three cases are missed (sensitivity 94 %) but 47 negative cases are labelled positive (specificity 6 %). The empirical ROC AUC is 0.50 and the “curve” is actually a straight line:

Inline graphic — The black ROC curve with circles at diagnostic thresholds shows the data from Fig. 1, with AUC 0.83. The red curve with triangles describes the following data where test specificity is the same as in Fig. 1, but sensitivity has been increased. Here, assuming a rating of 3 or more conveys a diagnosis of cancer, then three positive cases are missed (sensitivity 94 %) and 17 negative cases are labeled positive (specificity 66 %); the AUC is 0.89: The blue ROC curve with squares describes the following data where test sensitivity is the same as Fig. 1, but specificity has been increased. Here, assuming a rating of 3 or more conveys a diagnosis of cancer, then seven positive cases are missed (sensitivity 86 %) and eight negative cases are labeled positive (specificity 84 %); the AUC is 0.89: The black ROC curve with solid circles describes the following data where test sensitivity is increased to the same level as the red curve but with specificity dropped by a corresponding amount. Assuming a rating of 3 or more conveys a diagnosis of cancer, then just three cases are missed (sensitivity 94 %) but 47 negative cases are labelled positive (specificity 6 %). The empirical ROC AUC is 0.50 and the “curve” is actually a straight line: