Optimization of medical imaging display systems: using the channelized Hotelling observer for detecting lung nodules: experimental study

Ljiljana Platisa; Ewout Vansteenkiste; Bart Goossens; Cédric Marchessoux; Tom Kimpe; Wilfried Philips

doi:10.1117/12.812510

12 March 2009 Optimization of medical imaging display systems: using the channelized Hotelling observer for detecting lung nodules: experimental study

Ljiljana Platisa, Ewout Vansteenkiste, Bart Goossens, Cédric Marchessoux, Tom Kimpe, Wilfried Philips

Author Affiliations +

Proceedings Volume 7263, Medical Imaging 2009: Image Perception, Observer Performance, and Technology Assessment; 72630P (2009) https://doi.org/10.1117/12.812510
Event: SPIE Medical Imaging, 2009, Lake Buena Vista (Orlando Area), Florida, United States

Abstract

Medical-imaging systems are designed to aid medical specialists in a specific task. Therefore, the physical parameters of a system need to optimize the task performance of a human observer. This requires measurements of human performance in a given task during the system optimization. Typically, psychophysical studies are conducted for this purpose. Numerical observer models have been successfully used to predict human performance in several detection tasks. Especially, the task of signal detection using a channelized Hotelling observer (CHO) in simulated images has been widely explored. However, there are few studies done for clinically acquired images that also contain anatomic noise. In this paper, we investigate the performance of a CHO in the task of detecting lung nodules in real radiographic images of the chest. To evaluate variability introduced by the limited available data, we employ a commonly used study of a multi-reader multi-case (MRMC) scenario. It accounts for both case and reader variability. Finally, we use the "oneshot" methods to estimate the MRMC variance of the area under the ROC curve (AUC). The obtained AUC compares well to those reported for human observer study on a similar data set. Furthermore, the "one-shot" analysis implies a fairly consistent performance of the CHO with the variance of AUC below 0.002. This indicates promising potential for numerical observers in optimization of medical imaging displays and encourages further investigation on the subject.

Citation Download Citation

Ljiljana Platisa, Ewout Vansteenkiste, Bart Goossens, Cédric Marchessoux, Tom Kimpe, and Wilfried Philips "Optimization of medical imaging display systems: using the channelized Hotelling observer for detecting lung nodules: experimental study", Proc. SPIE 7263, Medical Imaging 2009: Image Perception, Observer Performance, and Technology Assessment, 72630P (12 March 2009); https://doi.org/10.1117/12.812510

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available