SNM Annual Meeting Abstracts
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
 QUICK SEARCH:   [advanced]


     




J Nucl Med. 2013; 54 (Supplement 2):313
This Article
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Google Scholar
Right arrow Articles by Foster, B.
Right arrow Articles by Mollura, D.
PubMed
Right arrow Articles by Foster, B.
Right arrow Articles by Mollura, D.

Instrumentation & Data Analysis

Data Analysis & Management IV: Organ Imaging

Affinity propagation clustering determines distributed uptake regions in PET images: A computer-aided approach for quantification of pulmonary infections in small animals

Brent Foster1, Ulas Bagci1, Ziyue Xu1, Bappaditya Dey2, Brian Luna2, William Bishai3, Sanjay Jain2 and Daniel Mollura1

1 Radiology and Imaging Sciences, National Institutes of Health, Bethesda, MD 2 Department of Medicine, Johns Hopkins University, Baltimore, MD 3 KwaZulu-Natal Research Institute, Durban, South Africa

Abstract No. 313

Objectives: To design a robust and efficient computational platform to accurately quantify pulmonary infections in rabbits through precise segmentation of high-distributed uptakes from 18F-FDG-PET images.

Methods: PET-CTs were analyzed from 10 rabbits infected with aerosolized Mycobacterium tuberculosis (H37Rv strain) over various time points (0 to 38 weeks). An overview of the proposed framework is as follows: (i) the histogram of the PET image(s) were estimated and smoothed via diffusion based kernel density estimation; (ii) the similarity of all points on the smoothed histogram was calculated using a novel similarity function, with the assumption that closer points on the histogram are more similar and more likely to belong to the same class; and (iii) the Affinity Propagation (AP) method was used to cluster the data points in order to find optimal thresholding levels to separate the high-uptake regions into several classes, which may further aid in the definition of target boundaries.

Results: The Dice Similarity Coefficient (DSC), sensitivity, and specificity were calculated between the segmentation region that was found by the proposed method and compared to two expert delineations. The average DSC was 89.06 ±9.82% with a sensitivity of 97.87±7.09% and a specificity of 83.70±15.32%. The correlation coefficient between the delineation performances of the two expert observers was R2=0.85 (p<0.01), while the correlation coefficient between the proposed method and the average of the observers segmentations was R2=0.91 (p<0.01).

Conclusions: Our proposed segmentation method quantified the distributed uptake regions with high accuracy and within seconds; hence, it outperformed the state-of-the-art methods.

Research Support: This research is supported by CIDI, the intramural research program of the National Institute of Allergy and Infectious Diseases (NIAID) and the National Institute of Biomedical Imaging and Bioengineering (NIBIB). Dr. Jain acknowledges the NIH Director’s New Innovator Award (OD006492). The rabbit infection study is funded by HHMI, NIAD R01AI079590, and R01A1035272.





This Article
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Google Scholar
Right arrow Articles by Foster, B.
Right arrow Articles by Mollura, D.
PubMed
Right arrow Articles by Foster, B.
Right arrow Articles by Mollura, D.