Hui Xue1, Ethan Tseng1, Kristopher D Knott2, Tushar Kotecha3, Louise Brown4, Sven Plein4, Marianna Fontana3, James C Moon2, Peter Kellman1. 1. National Heart, Lung and Blood Institute, National Institutes of Health, Bethesda, Maryland, USA. 2. Barts Heart Centre, London, UK. 3. National Amyloidosis Centre, Royal Free Hospital, London, UK. 4. Department of Biomedical Imaging Science, Leeds Institute of Cardiovascular and Metabolic Medicine, University of Leeds, Leeds, UK.
Abstract
PURPOSE: Quantification of myocardial perfusion has the potential to improve the detection of regional and global flow reduction. Significant effort has been made to automate the workflow, where one essential step is the arterial input function (AIF) extraction. Failure to accurately identify the left ventricle (LV) prevents AIF estimation required for quantification, therefore high detection accuracy is required. This study presents a robust LV detection method using the convolutional neural network (CNN). METHODS: CNN models were trained by assembling 25,027 scans (N = 12,984 patients) from three hospitals, seven scanners. Performance was evaluated using a hold-out test set of 5721 scans (N = 2805 patients). Model inputs were a time series of AIF images (2D+T). Two variations were investigated: (1) two classes (2CS) for background and foreground (LV mask), and (2) three classes (3CS) for background, LV, and RV. The final model was deployed on MRI scanners using the Gadgetron reconstruction software framework. RESULTS: Model loading on the MRI scanner took ~340 ms and applying the model took ~180 ms. The 3CS model successfully detected the LV in 99.98% of all test cases (1 failure out of 5721). The mean Dice ratio for 3CS was 0.87 ± 0.08 with 92.0% of all cases having Dice >0.75. The 2CS model gave a lower Dice ratio of 0.82 ± 0.22 (P < 1e-5). There was no significant difference in foot-time, peak-time, first-pass duration, peak value, and area-under-curve (P > .2) comparing automatically extracted AIF signals with signals from manually drawn contours. CONCLUSIONS: A CNN-based solution to detect the LV blood pool from the arterial input function image series was developed, validated, and deployed. A high LV detection accuracy of 99.98% was achieved.
PURPOSE: Quantification of myocardial perfusion has the potential to improve the detection of regional and global flow reduction. Significant effort has been made to automate the workflow, where one essential step is the arterial input function (AIF) extraction. Failure to accurately identify the left ventricle (LV) prevents AIF estimation required for quantification, therefore high detection accuracy is required. This study presents a robust LV detection method using the convolutional neural network (CNN). METHODS: CNN models were trained by assembling 25,027 scans (N = 12,984 patients) from three hospitals, seven scanners. Performance was evaluated using a hold-out test set of 5721 scans (N = 2805 patients). Model inputs were a time series of AIF images (2D+T). Two variations were investigated: (1) two classes (2CS) for background and foreground (LV mask), and (2) three classes (3CS) for background, LV, and RV. The final model was deployed on MRI scanners using the Gadgetron reconstruction software framework. RESULTS: Model loading on the MRI scanner took ~340 ms and applying the model took ~180 ms. The 3CS model successfully detected the LV in 99.98% of all test cases (1 failure out of 5721). The mean Dice ratio for 3CS was 0.87 ± 0.08 with 92.0% of all cases having Dice >0.75. The 2CS model gave a lower Dice ratio of 0.82 ± 0.22 (P < 1e-5). There was no significant difference in foot-time, peak-time, first-pass duration, peak value, and area-under-curve (P > .2) comparing automatically extracted AIF signals with signals from manually drawn contours. CONCLUSIONS: A CNN-based solution to detect the LV blood pool from the arterial input function image series was developed, validated, and deployed. A high LV detection accuracy of 99.98% was achieved.
Authors: Peter D Gatehouse; Andrew G Elkington; Nicholas A Ablitt; Guang-Zhong Yang; Dudley J Pennell; David N Firmin Journal: J Magn Reson Imaging Date: 2004-07 Impact factor: 4.813
Authors: Julien Milles; Rob J van der Geest; Michael Jerosch-Herold; Johan H C Reiber; Boudewijn P F Lelieveldt Journal: IEEE Trans Med Imaging Date: 2008-11 Impact factor: 10.048
Authors: Andre Esteva; Alexandre Robicquet; Bharath Ramsundar; Volodymyr Kuleshov; Mark DePristo; Katherine Chou; Claire Cui; Greg Corrado; Sebastian Thrun; Jeff Dean Journal: Nat Med Date: 2019-01-07 Impact factor: 53.440
Authors: Alex Bratt; Jiwon Kim; Meridith Pollie; Ashley N Beecy; Nathan H Tehrani; Noel Codella; Rocio Perez-Johnston; Maria Chiara Palumbo; Javid Alakbarli; Wayne Colizza; Ian R Drexler; Clerio F Azevedo; Raymond J Kim; Richard B Devereux; Jonathan W Weinsaft Journal: J Cardiovasc Magn Reson Date: 2019-01-07 Impact factor: 5.364
Authors: George Koulaouzidis; Tomasz Jadczyk; Dimitris K Iakovidis; Anastasios Koulaouzidis; Marc Bisnaire; Dafni Charisopoulou Journal: J Clin Med Date: 2022-07-05 Impact factor: 4.964
Authors: Hui Xue; Rhodri H Davies; Louise A E Brown; Kristopher D Knott; Tushar Kotecha; Marianna Fontana; Sven Plein; James C Moon; Peter Kellman Journal: Radiol Artif Intell Date: 2020-10-21
Authors: Andreas Seraphim; Kristopher D Knott; Katia Menacho; Joao B Augusto; Rhodri Davies; Iain Pierce; George Joy; Anish N Bhuva; Hui Xue; Thomas A Treibel; Jackie A Cooper; Steffen E Petersen; Marianna Fontana; Alun D Hughes; James C Moon; Charlotte Manisty; Peter Kellman Journal: JACC Cardiovasc Imaging Date: 2021-05-19