Tests of Statistically Significant Differences between Two Imaging Systems in ROC Analysis: : Use of the Jackknife Method and its Application

Shiraishi, Junji; Utsunomiya, Akane

doi:10.6009/jjrt.kj00001355781

Cited by 12 publications

(3 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Finally, the confidence data obtained by the observers were analyzed by use of software for ROC analysis (DBM MRMC Version 2.1 Beta 3, C. E. Metz, The University of Chicago) [11][12][13][14][15]. Then, we used PROPROC as a curvefitting model, and we determined the statistically significant differences between the areas under the ROC curve (AUCs) for each maximum luminance setting of the LCD monitor with the jackknife method by use of DBM MRMC.…”

Section: Observer Studymentioning

confidence: 99%

Detectability of a lung nodule displayed on a liquid-crystal display monitor with different maximum luminance settings

Takahashi

Sueoka

Yoon

et al. 2009

Radiol Phys Technol

View full text Add to dashboard Cite

Our purpose in this study was to examine the detectability of a lung nodule at different maximum luminance settings of a liquid-crystal display (LCD) monitor by utilizing receiver operating characteristic (ROC) analysis. The LCD monitor used in this study was calibrated to the grayscale standard display function with different maximum luminance settings (670, 450, and 170 cd/m(2)). The average area under the ROC curve (AUC) and the standard deviation for all observers at 670, 450, and 170 cd/m(2) were 0.837 +/- 0.076, 0.832 +/- 0.051, and 0.830 +/- 0.078, respectively. There was no statistically significant difference in AUC as a function of the maximum luminance setting of the LCD monitor. Considering the results, setting the maximum luminance of an LCD monitor at a higher level may not provide a significant advantage in the detectability of a lung nodule.

show abstract

Section: Observer Studymentioning

confidence: 99%

Detectability of a lung nodule displayed on a liquid-crystal display monitor with different maximum luminance settings

Takahashi

Sueoka

Yoon

et al. 2009

Radiol Phys Technol

View full text Add to dashboard Cite

show abstract

“…ROC analysis [19] was performed for both standard scan conditions and CT-AEC with the DR-Wedge. The target objects were 60 images of simulated tumors in the right lung apex, tracheal bifurcation, and lung base and 60 images of simulated lung tissue around the simulated tumors.…”

Section: Image Analysis and Statistical Analysismentioning

confidence: 99%

Low-dose CT screening for lung cancer with automatic exposure control: phantom study

Gomi¹,

Muramatsu

Tsukagoshi

et al. 2008

Radiol Phys Technol

View full text Add to dashboard Cite

We conducted a study to determine optimal scan conditions for automatic exposure control (AEC) in computed tomography (CT) of low-dose chest screening in order to provide consistent image quality without increasing the collective dose. Using a chest CT phantom, we set CT-AEC scan conditions with a dose-reduction wedge (DR-Wedge) to the same radiation dose as those for low-tube current, fixed-scan conditions. Image quality was evaluated with the use of the standard deviation of the CT number, contrast-noise ratios (CNR), and receiver-operating characteristic (ROC) analysis. At the same radiation dose, in the scan conditions using CT-AEC with the DR-Wedge, the SD of the CT number of each slice position was stable. The CNR values were higher at the lung apex and lung base under CT-AEC with the DR-Wedge than under standard scan conditions (p < 0.0002). In addition, ROC analysis of blind evaluation by four radiologists and three technologists showed that the image quality was improved for the lung apex (p < 0.009), tracheal bifurcation (p < 0.038), and lung base (p < 0.022) in the scan conditions using CT-AEC with the DR-Wedge. We achieved improvement of image quality without increasing the collective dose by using CT-AEC with the DR-Wedge under low-dose scan conditions.

show abstract

“…For the case of a network with 35 hidden units, in which the performance was relatively good in the above experiment, the proposed system, the system using only the conventional Mahalanobis classifier, and the system using only the conventional MLP classifier were compared in terms of the ROC curve (receiver-operating characteristic curve) [12][13][14]. ROCKIT [14] produced by Metz's ROC Software Users Group was used for estimation and testing of the ROC curve.…”

Section: Test Of Rocmentioning

confidence: 99%

Enhancement of CAD system for breast cancers by improvement of classifiers

Hagihara¹,

Hagihara²,

Wei³

2005

Systems & Computers in Japan

View full text Add to dashboard Cite

SUMMARYIn this paper, we propose a method to improve the accuracy of classifiers by replacing the connection between the output layer and the immediately preceding hidden layer with an optimal linear transformer. This approach is intended to improve the performance of a breast cancer image diagnosis assistance system. The proposed classifier is composed of a three-layer MLP (multilayer perceptron) and a Mahalanobis classifier. The MLP has only one output unit, and produces output for two categories. If it is assumed that the value from the hidden layer immediately preceding the output layer forms a multivariable normal distribution for each class, that is, a Gaussian distribution, then the optimal linear transformer is a classifier based on the generalized Mahalanobis distance. Thus, the optimal classification is realized in the MLP after learning in which the generalized Mahalanobis distance with the hidden layer immediately preceding the output layer as the input is examined, and classification is performed on the basis of the likelihood. The proposed breast cancer image diagnosis assistance system, the system using only the conventional Mahalanobis classifier, and the system using only the conventional MLP classifier are compared. The best results are given by the proposed method, and it is shown that the performance of the breast image diagnosis assistance system can be improved.

show abstract

Tests of Statistically Significant Differences between Two Imaging Systems in ROC Analysis: : Use of the Jackknife Method and its Application

Cited by 12 publications

References 3 publications

Detectability of a lung nodule displayed on a liquid-crystal display monitor with different maximum luminance settings

Detectability of a lung nodule displayed on a liquid-crystal display monitor with different maximum luminance settings

Low-dose CT screening for lung cancer with automatic exposure control: phantom study

Enhancement of CAD system for breast cancers by improvement of classifiers

Contact Info

Product

Resources

About