Xiaowei Zou1, Blaine L Hart2, Marc Mabray2, Mary R Bartlett3, Wei Bian4, Jeffrey Nelson5, Leslie A Morrison3, Charles E McCulloch6, Christopher P Hess1, Janine M Lupo1, Helen Kim7,8. 1. Department of Radiology and Biomedical Imaging, University of California, San Francisco, San Francisco, California, USA. 2. Department of Radiology, University of New Mexico, Albuquerque, New Mexico, USA. 3. Department of Neurology, University of New Mexico, Albuquerque, New Mexico, USA. 4. Department of Radiology, Stanford University, Stanford, California, USA. 5. Department of Anesthesia and Perioperative Care, University of California, San Francisco, 1001 Potrero Avenue, Box 1363, San Francisco, 94143, California, USA. 6. Department of Epidemiology and Biostatistics, University of California, San Francisco, California, USA. 7. Department of Anesthesia and Perioperative Care, University of California, San Francisco, 1001 Potrero Avenue, Box 1363, San Francisco, 94143, California, USA. Helen.Kim2@ucsf.edu. 8. Department of Epidemiology and Biostatistics, University of California, San Francisco, California, USA. Helen.Kim2@ucsf.edu.
Abstract
PURPOSE: Familial cerebral cavernous malformation (CCM) patients present with multiple lesions that can grow both in number and size over time and are reliably detected on susceptibility-weighted imaging (SWI). Manual counting of lesions is arduous and subject to high variability. We aimed to develop an automated algorithm for counting CCM microbleeds (lesions <5 mm in diameter) on SWI images. METHODS: Fifty-seven familial CCM type-1 patients were included in this institutional review board-approved study. Baseline SWI (n = 57) and follow-up SWI (n = 17) were performed on a 3T Siemens MR scanner with lesions counted manually by the study neuroradiologist. We modified an algorithm for detecting radiation-induced microbleeds on SWI images in brain tumor patients, using a training set of 22 manually delineated CCM microbleeds from two random scans. Manual and automated counts were compared using linear regression with robust standard errors, intra-class correlation (ICC), and paired t tests. A validation analysis comparing the automated counting algorithm and a consensus read from two neuroradiologists was used to calculate sensitivity, the proportion of microbleeds correctly identified by the automated algorithm. RESULTS: Automated and manual microbleed counts were in strong agreement in both baseline (ICC = 0.95, p < 0.001) and longitudinal (ICC = 0.88, p < 0.001) analyses, with no significant difference between average counts (baseline p = 0.11, longitudinal p = 0.29). In the validation analysis, the algorithm correctly identified 662 of 1325 microbleeds (sensitivity=50%), again with strong agreement between approaches (ICC = 0.77, p < 0.001). CONCLUSION: The automated algorithm is a consistent method for counting microbleeds in familial CCM patients that can facilitate lesion quantification and tracking.
PURPOSE:Familial cerebral cavernous malformation (CCM) patients present with multiple lesions that can grow both in number and size over time and are reliably detected on susceptibility-weighted imaging (SWI). Manual counting of lesions is arduous and subject to high variability. We aimed to develop an automated algorithm for counting CCM microbleeds (lesions <5 mm in diameter) on SWI images. METHODS: Fifty-seven familial CCM type-1 patients were included in this institutional review board-approved study. Baseline SWI (n = 57) and follow-up SWI (n = 17) were performed on a 3T Siemens MR scanner with lesions counted manually by the study neuroradiologist. We modified an algorithm for detecting radiation-induced microbleeds on SWI images in brain tumorpatients, using a training set of 22 manually delineated CCM microbleeds from two random scans. Manual and automated counts were compared using linear regression with robust standard errors, intra-class correlation (ICC), and paired t tests. A validation analysis comparing the automated counting algorithm and a consensus read from two neuroradiologists was used to calculate sensitivity, the proportion of microbleeds correctly identified by the automated algorithm. RESULTS: Automated and manual microbleed counts were in strong agreement in both baseline (ICC = 0.95, p < 0.001) and longitudinal (ICC = 0.88, p < 0.001) analyses, with no significant difference between average counts (baseline p = 0.11, longitudinal p = 0.29). In the validation analysis, the algorithm correctly identified 662 of 1325 microbleeds (sensitivity=50%), again with strong agreement between approaches (ICC = 0.77, p < 0.001). CONCLUSION: The automated algorithm is a consistent method for counting microbleeds in familial CCM patients that can facilitate lesion quantification and tracking.
Authors: Charlotte Cordonnier; Gillian M Potter; Caroline A Jackson; Fergus Doubal; Sarah Keir; Cathie L M Sudlow; Joanna M Wardlaw; Rustam Al-Shahi Salman Journal: Stroke Date: 2008-11-13 Impact factor: 7.914
Authors: Samuel R S Barnes; E Mark Haacke; Muhammad Ayaz; Alexander S Boikov; Wolff Kirsch; Dan Kido Journal: Magn Reson Imaging Date: 2011-05-14 Impact factor: 2.546
Authors: D Rigamonti; M N Hadley; B P Drayer; P C Johnson; K Hoenig-Rigamonti; J T Knight; R F Spetzler Journal: N Engl J Med Date: 1988-08-11 Impact factor: 91.245
Authors: Mohamed L Seghier; Magdalena A Kolanko; Alexander P Leff; Hans R Jäger; Simone M Gregoire; David J Werring Journal: PLoS One Date: 2011-03-23 Impact factor: 3.240
Authors: T L A van den Heuvel; A W van der Eerden; R Manniesing; M Ghafoorian; T Tan; T M J C Andriessen; T Vande Vyvere; L van den Hauwe; B M Ter Haar Romeny; B M Goraj; B Platel Journal: Neuroimage Clin Date: 2016-07-02 Impact factor: 4.881
Authors: Robert Shenkar; Changbin Shi; Douglas A Marchuk; Issam A Awad; Tania Rebeiz; Rebecca A Stockton; David A McDonald; Abdul Ghani Mikati; Lingjiao Zhang; Cecilia Austin; Amy L Akers; Carol J Gallione; Autumn Rorrer; Murat Gunel; Wang Min; Jorge Marcondes De Souza; Connie Lee Journal: Genet Med Date: 2014-08-14 Impact factor: 8.822