Mu He1, Wei Zha2, Fei Tan3, Leith Rankine4, Sean Fain5, Bastiaan Driehuys6. 1. Department of Electrical and Computer Engineering, Duke University, Durham, North Carolina. 2. Department of Medical Physics, University of Wisconsin-Madison, Madison, Wisconsin. 3. Department of Biomedical Engineering, Duke University, Durham, North Carolina. 4. Medical Physics Graduate Program, Duke University, Durham, North Carolina. 5. Department of Medical Physics, University of Wisconsin-Madison, Madison, Wisconsin; Department of Biomedical Engineering, University of Wisconsin-Madison, Madison, Wisconsin; Department of Radiology, University of Wisconsin-Madison, Madison, Wisconsin. 6. Department of Biomedical Engineering, Duke University, Durham, North Carolina; Medical Physics Graduate Program, Duke University, Durham, North Carolina; Department of Radiology, Duke University, Durham, North Carolina.
Abstract
RATIONALE: Hyperpolarized 129Xe MRI enables quantitative evaluation of regional ventilation. To this end, multiple classifiers have been proposed to determine ventilation defect percentage (VDP) as well as other cluster populations. However, consensus has not yet been reached regarding which of these methods to deploy for multicenter clinical trials. Here, we compare two published classification techniques-linear-binning and adaptive K-means-to establish their limits of agreement and their robustness against reduced signal-to-noise ratio (SNR). METHODS: A total of 29 subjects (age: 38.4 ± 19.0 years) were retrospectively identified for inter-method comparison. For each 129Xe ventilation image, 7 images with reduced SNR were generated with equal decrements relative to the native SNR. All 8 sets of images were then analyzed using both methods independently to classify all lung voxels into four clusters: VDP, low-, medium-, and high-ventilation-percentage (LVP, MVP and HVP). For each cluster, the percentage of the lung it comprised was compared between the two methods, as well as how these values persisted as SNR was degraded. RESULTS: The limits of agreement for calculating VDP were [+0.2%, +4.0%] with a +1.5% bias for binning relative to K-means. However, the inter-method agreement for the other clusters was moderate, with biases of -5.7%, 8.1%, and -4.0% for LVP, MVP, and HVP, respectively. As SNR decreased below ∼4, both methods began reporting values that deviated substantially from the native image. By requiring VDP to remain within ≤1.8% of that calculated from the native image, the minimum tolerable SNR values were 2.4 ± 1.0 for the linear-binning, and 3.5 ± 1.5 for the K-means. CONCLUSIONS: Both methods agree well in quantifying VDP, but agreement for LVP and MVP remains variable. We suggest a required SNR threshold be two standard deviations above the minimum value of 3.5 ± 1.5 for robust determination of VDP, suggesting a minimum SNR of 6.6. However, robust quantification of the ventilated clusters required an SNR of 13.4.
RATIONALE: Hyperpolarized 129Xe MRI enables quantitative evaluation of regional ventilation. To this end, multiple classifiers have been proposed to determine ventilation defect percentage (VDP) as well as other cluster populations. However, consensus has not yet been reached regarding which of these methods to deploy for multicenter clinical trials. Here, we compare two published classification techniques-linear-binning and adaptive K-means-to establish their limits of agreement and their robustness against reduced signal-to-noise ratio (SNR). METHODS: A total of 29 subjects (age: 38.4 ± 19.0 years) were retrospectively identified for inter-method comparison. For each 129Xe ventilation image, 7 images with reduced SNR were generated with equal decrements relative to the native SNR. All 8 sets of images were then analyzed using both methods independently to classify all lung voxels into four clusters: VDP, low-, medium-, and high-ventilation-percentage (LVP, MVP and HVP). For each cluster, the percentage of the lung it comprised was compared between the two methods, as well as how these values persisted as SNR was degraded. RESULTS: The limits of agreement for calculating VDP were [+0.2%, +4.0%] with a +1.5% bias for binning relative to K-means. However, the inter-method agreement for the other clusters was moderate, with biases of -5.7%, 8.1%, and -4.0% for LVP, MVP, and HVP, respectively. As SNR decreased below ∼4, both methods began reporting values that deviated substantially from the native image. By requiring VDP to remain within ≤1.8% of that calculated from the native image, the minimum tolerable SNR values were 2.4 ± 1.0 for the linear-binning, and 3.5 ± 1.5 for the K-means. CONCLUSIONS: Both methods agree well in quantifying VDP, but agreement for LVP and MVP remains variable. We suggest a required SNR threshold be two standard deviations above the minimum value of 3.5 ± 1.5 for robust determination of VDP, suggesting a minimum SNR of 6.6. However, robust quantification of the ventilated clusters required an SNR of 13.4.
Authors: Saba Samee; Talissa Altes; Patrick Powers; Eduard E de Lange; Jack Knight-Scott; Gary Rakes; John P Mugler; Jonathan M Ciambotti; Bennet A Alford; James R Brookeman; Thomas A E Platts-Mills Journal: J Allergy Clin Immunol Date: 2003-06 Impact factor: 10.793
Authors: Pierrick Coupé; José V Manjón; Elias Gedamu; Douglas Arnold; Montserrat Robles; D Louis Collins Journal: Med Image Anal Date: 2010-03-20 Impact factor: 8.545
Authors: Mu He; S Sivaram Kaushik; Scott H Robertson; Matthew S Freeman; Rohan S Virgincar; H Page McAdams; Bastiaan Driehuys Journal: Acad Radiol Date: 2014-09-26 Impact factor: 3.173
Authors: Nicholas J Tustison; Brian B Avants; Philip A Cook; Yuanjie Zheng; Alexander Egan; Paul A Yushkevich; James C Gee Journal: IEEE Trans Med Imaging Date: 2010-04-08 Impact factor: 10.048
Authors: Ziyi Wang; Scott Haile Robertson; Jennifer Wang; Mu He; Rohan S Virgincar; Geoffry M Schrank; Elianna A Bier; Sudarshan Rajagopal; Yuh Chin Huang; Thomas G O'Riordan; Craig R Rackley; H Page McAdams; Bastiaan Driehuys Journal: Med Phys Date: 2017-05-18 Impact factor: 4.071
Authors: Rohan S Virgincar; Zackary I Cleveland; S Sivaram Kaushik; Matthew S Freeman; John Nouls; Gary P Cofer; Santiago Martinez-Jimenez; Mu He; Monica Kraft; Jan Wolber; H Page McAdams; Bastiaan Driehuys Journal: NMR Biomed Date: 2012-10-13 Impact factor: 4.044
Authors: Meilan K Han; Ella A Kazerooni; David A Lynch; Lyrica X Liu; Susan Murray; Jeffrey L Curtis; Gerard J Criner; Victor Kim; Russell P Bowler; Nicola A Hanania; Antonio R Anzueto; Barry J Make; John E Hokanson; James D Crapo; Edwin K Silverman; Fernando J Martinez; George R Washko Journal: Radiology Date: 2011-07-25 Impact factor: 11.105
Authors: David J Roach; Matthew M Willmering; Joseph W Plummer; Laura L Walkup; Yin Zhang; Md Monir Hossain; Zackary I Cleveland; Jason C Woods Journal: Acad Radiol Date: 2021-08-12 Impact factor: 3.173
Authors: Matthew M Willmering; Peter J Niedbalski; Hui Wang; Laura L Walkup; Ryan K Robison; James G Pipe; Zackary I Cleveland; Jason C Woods Journal: Magn Reson Med Date: 2019-12-01 Impact factor: 4.668