Vinicius de Padua V Alves1, Jonathan R Dillman1,2, Andrew T Trout3,4,5. 1. Department of Radiology, Cincinnati Children's Hospital Medical Center, Cincinnati, OH, USA. 2. Department of Radiology, University of Cincinnati College of Medicine, 3333 Burnet Ave, Kasota Building MLC 5031, Cincinnati, OH, 45226, USA. 3. Department of Radiology, Cincinnati Children's Hospital Medical Center, Cincinnati, OH, USA. andrew.trout@cchmc.org. 4. Department of Radiology, University of Cincinnati College of Medicine, 3333 Burnet Ave, Kasota Building MLC 5031, Cincinnati, OH, 45226, USA. andrew.trout@cchmc.org. 5. Department of Pediatrics, University of Cincinnati College of Medicine, Cincinnati, OH, USA. andrew.trout@cchmc.org.
Abstract
INTRODUCTION: Multiple region-of-interest (ROI) sampling strategies have been described for liver fat quantification by MRI PDFF. While adult studies have shown that sampling strategies including as few as four ROIs provide a reasonable tradeoff between laboriousness and quantitative performance, there is a paucity of similar data for pediatric patients. PURPOSE: To assess agreement between different ROI sampling strategies for liver MRI PDFF analysis in children and adolescents. MATERIALS AND METHODS: This retrospective, internal review board-approved study included clinical MRI PDFF acquisitions for 50 children and adolescents. Four different ROI sampling paradigms reported in the literature were reproduced to measure mean liver PDFF. An 18-ROI (2 in each Couinaud segment) paradigm was considered the reference standard. Spearman correlation, intraclass correlation coefficients (ICCs), and Bland-Altman analyses were used to quantify agreement. RESULTS: Mean age for the 50 participants was 14 ± 2.5 years (range 8-17 years). Based on the 18-ROI paradigm, mean PDFF was significantly higher for the right lobe (24.0 ± 13.7% right, 22.0 ± 13.1% left; p = 0.001). PDFF values for each individual Couinaud segment were highly correlated with the reference standard (ρ = 0.977 to 0.993, p < 0.0001). PDFF values derived from all sampling paradigms, including strategies using large free-hand ROIs, were strongly correlated with the reference standard (ρ = 0.995 to 0.998, p < 0.0001) with excellent agreement (ICC range 0.995 to 0.998). CONCLUSION: Liver PDFF sampling paradigms using large ROIs showed strong correlation, excellent agreement, and nonsignificant mean differences from a reference standard paradigm sampling every Couinaud segment in children. Paradigms that exclusively sample the right lobe may overestimate liver PDFF.
INTRODUCTION: Multiple region-of-interest (ROI) sampling strategies have been described for liver fat quantification by MRI PDFF. While adult studies have shown that sampling strategies including as few as four ROIs provide a reasonable tradeoff between laboriousness and quantitative performance, there is a paucity of similar data for pediatric patients. PURPOSE: To assess agreement between different ROI sampling strategies for liver MRI PDFF analysis in children and adolescents. MATERIALS AND METHODS: This retrospective, internal review board-approved study included clinical MRI PDFF acquisitions for 50 children and adolescents. Four different ROI sampling paradigms reported in the literature were reproduced to measure mean liver PDFF. An 18-ROI (2 in each Couinaud segment) paradigm was considered the reference standard. Spearman correlation, intraclass correlation coefficients (ICCs), and Bland-Altman analyses were used to quantify agreement. RESULTS: Mean age for the 50 participants was 14 ± 2.5 years (range 8-17 years). Based on the 18-ROI paradigm, mean PDFF was significantly higher for the right lobe (24.0 ± 13.7% right, 22.0 ± 13.1% left; p = 0.001). PDFF values for each individual Couinaud segment were highly correlated with the reference standard (ρ = 0.977 to 0.993, p < 0.0001). PDFF values derived from all sampling paradigms, including strategies using large free-hand ROIs, were strongly correlated with the reference standard (ρ = 0.995 to 0.998, p < 0.0001) with excellent agreement (ICC range 0.995 to 0.998). CONCLUSION: Liver PDFF sampling paradigms using large ROIs showed strong correlation, excellent agreement, and nonsignificant mean differences from a reference standard paradigm sampling every Couinaud segment in children. Paradigms that exclusively sample the right lobe may overestimate liver PDFF.
Authors: Susanne Bonekamp; An Tang; Arian Mashhood; Tanya Wolfson; Christopher Changchien; Michael S Middleton; Lisa Clark; Anthony Gamst; Rohit Loomba; Claude B Sirlin Journal: J Magn Reson Imaging Date: 2014-06 Impact factor: 4.813
Authors: Miriam B Vos; Stephanie H Abrams; Sarah E Barlow; Sonia Caprio; Stephen R Daniels; Rohit Kohli; Marialena Mouzaki; Pushpa Sathya; Jeffrey B Schwimmer; Shikha S Sundaram; Stavra A Xanthakos Journal: J Pediatr Gastroenterol Nutr Date: 2017-02 Impact factor: 2.839
Authors: Michael S Middleton; Elhamy R Heba; Catherine A Hooker; Mustafa R Bashir; Kathryn J Fowler; Kumar Sandrasegaran; Elizabeth M Brunt; David E Kleiner; Edward Doo; Mark L Van Natta; Joel E Lavine; Brent A Neuschwander-Tetri; Arun Sanyal; Rohit Loomba; Claude B Sirlin Journal: Gastroenterology Date: 2017-06-15 Impact factor: 22.682
Authors: Geraldine H Kang; Irene Cruite; Masoud Shiehmorteza; Tanya Wolfson; Anthony C Gamst; Gavin Hamilton; Mark Bydder; Michael S Middleton; Claude B Sirlin Journal: J Magn Reson Imaging Date: 2011-07-18 Impact factor: 4.813
Authors: Z Permutt; T-A Le; M R Peterson; E Seki; D A Brenner; C Sirlin; R Loomba Journal: Aliment Pharmacol Ther Date: 2012-05-03 Impact factor: 8.171
Authors: An Tang; Justin Tan; Mark Sun; Gavin Hamilton; Mark Bydder; Tanya Wolfson; Anthony C Gamst; Michael Middleton; Elizabeth M Brunt; Rohit Loomba; Joel E Lavine; Jeffrey B Schwimmer; Claude B Sirlin Journal: Radiology Date: 2013-02-04 Impact factor: 11.105
Authors: Lauren F Chun; Elizabeth L Yu; Mary Catherine Sawh; Craig Bross; Jeanne Nichols; Lynda Polgreen; Cynthia Knott; Alexandra Schlein; Claude B Sirlin; Michael S Middleton; Deborah M Kado; Jeffrey B Schwimmer Journal: J Pediatr Date: 2021-02-03 Impact factor: 6.314