Open AccessEvaluation of automatic contour propagation in T2‐weighted 4 DMRI for normal‐tissue motion assessment using internal organ‐at‐risk volume ( IRV )
Author(s) -
Zhang Jingjing,
Markova Svetlana,
Garcia Alejandro,
Huang Kirk,
Nie Xingyu,
Choi Wookjin,
Lu Wei,
Wu Abraham,
Rimner Andreas,
Li Guang
Publication year - 2018
Publication title -
journal of applied clinical medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.83
H-Index - 48
ISSN - 1526-9914
DOI - 10.1002/acm2.12431
Subject(s) - jaccard index , ground truth , magnetic resonance imaging , similarity (geometry) , nuclear medicine , volume (thermodynamics) , cardiac cycle , bellows , image registration , breathing , computer science , artificial intelligence , medicine , pattern recognition (psychology) , radiology , anatomy , image (mathematics) , physics , mechanical engineering , quantum mechanics , engineering
Purpose The purpose of this study was to evaluate the quality of automatically propagated contours of organs at risk ( OAR s) based on respiratory‐correlated navigator‐triggered four‐dimensional magnetic resonance imaging ( RC ‐4 DMRI ) for calculation of internal organ‐at‐risk volume ( IRV ) to account for intra‐fractional OAR motion. Methods and Materials T2‐weighted RC ‐4 DMRI images were of 10 volunteers acquired and reconstructed using an internal navigator‐echo surrogate and concurrent external bellows under an IRB ‐approved protocol. Four major OAR s (lungs, heart, liver, and stomach) were delineated in the 10‐phase 4 DMRI . Two manual‐contour sets were delineated by two clinical personnel and two automatic‐contour sets were propagated using free‐form deformable image registration. The OAR volume variation within the 10‐phase cycle was assessed and the IRV was calculated as the union of all OAR contours. The OAR contour similarity between the navigator‐triggered and bellows‐rebinned 4 DMRI was compared. A total of 2400 contours were compared to the most probable ground truth with a 95% confidence level (S95) in similarity, sensitivity, and specificity using the simultaneous truth and performance level estimation ( STAPLE ) algorithm. Results Visual inspection of automatically propagated contours finds that approximately 5–10% require manual correction. The similarity, sensitivity, and specificity between manual and automatic contours are indistinguishable ( P > 0.05). The Jaccard similarity indexes are 0.92 ± 0.02 (lungs), 0.89 ± 0.03 (heart), 0.92 ± 0.02 (liver), and 0.83 ± 0.04 (stomach). Volume variations within the breathing cycle are small for the heart (2.6 ± 1.5%), liver (1.2 ± 0.6%), and stomach (2.6 ± 0.8%), whereas the IRV is much larger than the OAR volume by: 20.3 ± 8.6% (heart), 24.0 ± 8.6% (liver), and 47.6 ± 20.2% (stomach). The Jaccard index is higher in navigator‐triggered than bellows‐rebinned 4 DMRI by 4% ( P < 0.05), due to the higher image quality of navigator‐based 4 DMRI . Conclusion Automatic and manual OAR contours from Navigator‐triggered 4 DMRI are not statistically distinguishable. The navigator‐triggered 4 DMRI image provides higher contour quality than bellows‐rebinned 4 DMRI . The IRV s are 20–50% larger than OAR volumes and should be considered in dose estimation.