Ponnada A Narayana1, Ivan Coronado1, Sheeba J Sujit1, Jerry S Wolinsky2, Fred D Lublin3, Refaat E Gabr1. 1. Department of Diagnostic and Interventional Imaging, McGovern Medical School, University of Texas Health Science Center, Houston, Texas, USA. 2. Department of Neurology, McGovern Medical School, University of Texas Health Science Center, Houston, Texas, USA. 3. Department of Neurology, Icahn School of Medicine at Mount Sinai, New York, New York, USA.
Abstract
BACKGROUND: The dependence of deep-learning (DL)-based segmentation accuracy of brain MRI on the training size is not known. PURPOSE: To determine the required training size for a desired accuracy in brain MRI segmentation in multiple sclerosis (MS) using DL. STUDY TYPE: Retrospective analysis of MRI data acquired as part of a multicenter clinical trial. STUDY POPULATION: In all, 1008 patients with clinically definite MS. FIELD STRENGTH/SEQUENCE: MRIs were acquired at 1.5T and 3T scanners manufactured by GE, Philips, and Siemens with dual turbo spin echo, FLAIR, and T1 -weighted turbo spin echo sequences. ASSESSMENT: Segmentation results using an automated analysis pipeline and validated by two neuroimaging experts served as the ground truth. A DL model, based on a fully convolutional neural network, was trained separately using 16 different training sizes. The segmentation accuracy as a function of the training size was determined. These data were fitted to the learning curve for estimating the required training size for desired accuracy. STATISTICAL TESTS: The performance of the network was evaluated by calculating the Dice similarity coefficient (DSC), and lesion true-positive and false-positive rates. RESULTS: The DSC for lesions showed much stronger dependency on the sample size than gray matter (GM), white matter (WM), and cerebrospinal fluid (CSF). When the training size was increased from 10 to 800 the DSC values varied from 0.00 to 0.86 ± 0.016 for T2 lesions, 0.87 ± 009 to 0.94 ± 0.004 for GM, 0.86 ± 0.08 to 0.94 ± 0.005 for WM, and 0.91 ± 0.009 to 0.96 ± 0.003 for CSF. DATA CONCLUSION: Excellent segmentation was achieved with a training size as small as 10 image volumes for GM, WM, and CSF. In contrast, a training size of at least 50 image volumes was necessary for adequate lesion segmentation. LEVEL OF EVIDENCE: 1 Technical Efficacy Stage: 1 J. Magn. Reson. Imaging 2020;51:1487-1496.
BACKGROUND: The dependence of deep-learning (DL)-based segmentation accuracy of brain MRI on the training size is not known. PURPOSE: To determine the required training size for a desired accuracy in brain MRI segmentation in multiple sclerosis (MS) using DL. STUDY TYPE: Retrospective analysis of MRI data acquired as part of a multicenter clinical trial. STUDY POPULATION: In all, 1008 patients with clinically definite MS. FIELD STRENGTH/SEQUENCE: MRIs were acquired at 1.5T and 3T scanners manufactured by GE, Philips, and Siemens with dual turbo spin echo, FLAIR, and T1 -weighted turbo spin echo sequences. ASSESSMENT: Segmentation results using an automated analysis pipeline and validated by two neuroimaging experts served as the ground truth. A DL model, based on a fully convolutional neural network, was trained separately using 16 different training sizes. The segmentation accuracy as a function of the training size was determined. These data were fitted to the learning curve for estimating the required training size for desired accuracy. STATISTICAL TESTS: The performance of the network was evaluated by calculating the Dice similarity coefficient (DSC), and lesion true-positive and false-positive rates. RESULTS: The DSC for lesions showed much stronger dependency on the sample size than gray matter (GM), white matter (WM), and cerebrospinal fluid (CSF). When the training size was increased from 10 to 800 the DSC values varied from 0.00 to 0.86 ± 0.016 for T2 lesions, 0.87 ± 009 to 0.94 ± 0.004 for GM, 0.86 ± 0.08 to 0.94 ± 0.005 for WM, and 0.91 ± 0.009 to 0.96 ± 0.003 for CSF. DATA CONCLUSION: Excellent segmentation was achieved with a training size as small as 10 image volumes for GM, WM, and CSF. In contrast, a training size of at least 50 image volumes was necessary for adequate lesion segmentation. LEVEL OF EVIDENCE: 1 Technical Efficacy Stage: 1 J. Magn. Reson. Imaging 2020;51:1487-1496.
Authors: Sayan Mukherjee; Pablo Tamayo; Simon Rogers; Ryan Rifkin; Anna Engle; Colin Campbell; Todd R Golub; Jill P Mesirov Journal: J Comput Biol Date: 2003 Impact factor: 1.479
Authors: Balasrinivasa Rao Sajja; Sushmita Datta; Renjie He; Meghana Mehta; Rakesh K Gupta; Jerry S Wolinsky; Ponnada A Narayana Journal: Ann Biomed Eng Date: 2006-03-09 Impact factor: 3.934
Authors: Geert Litjens; Thijs Kooi; Babak Ehteshami Bejnordi; Arnaud Arindra Adiyoso Setio; Francesco Ciompi; Mohsen Ghafoorian; Jeroen A W M van der Laak; Bram van Ginneken; Clara I Sánchez Journal: Med Image Anal Date: 2017-07-26 Impact factor: 8.545
Authors: Gabriel Chartrand; Phillip M Cheng; Eugene Vorontsov; Michal Drozdzal; Simon Turcotte; Christopher J Pal; Samuel Kadoury; An Tang Journal: Radiographics Date: 2017 Nov-Dec Impact factor: 5.333
Authors: Fred D Lublin; Stacey S Cofield; Gary R Cutter; Robin Conwit; Ponnada A Narayana; Flavia Nelson; Amber R Salter; Tarah Gustafson; Jerry S Wolinsky Journal: Ann Neurol Date: 2013-03-11 Impact factor: 10.422
Authors: Daniel García-Lorenzo; Simon Francis; Sridar Narayanan; Douglas L Arnold; D Louis Collins Journal: Med Image Anal Date: 2012-09-29 Impact factor: 8.545
Authors: Mitchell T Wallin; William J Culpepper; Jonathan D Campbell; Lorene M Nelson; Annette Langer-Gould; Ruth Ann Marrie; Gary R Cutter; Wendy E Kaye; Laurie Wagner; Helen Tremlett; Stephen L Buka; Piyameth Dilokthornsakul; Barbara Topol; Lie H Chen; Nicholas G LaRocca Journal: Neurology Date: 2019-02-15 Impact factor: 9.910
Authors: Francesco La Rosa; Erin S Beck; Josefina Maranzano; Ramona-Alexandra Todea; Peter van Gelderen; Jacco A de Zwart; Nicholas J Luciano; Jeff H Duyn; Jean-Philippe Thiran; Cristina Granziera; Daniel S Reich; Pascal Sati; Meritxell Bach Cuadra Journal: NMR Biomed Date: 2022-03-31 Impact factor: 4.478