Amir Jamaludin1, Meelis Lootus1, Timor Kadir2, Andrew Zisserman1, Jill Urban3, Michele C Battié4, Jeremy Fairbank5,6,7, Iain McCall8. 1. Department of Engineering Science, University of Oxford, Oxford, UK. 2. Mirada Medical, Oxford, UK. 3. Department of Physiology, Anatomy and Genetics, University of Oxford, Oxford, UK. 4. Faculty of Rehabilitation Medicine, University of Alberta, Edmonton, Canada. 5. Botnar Institute of Musculoskeletal Sciences, Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences (NDORMS), University of Oxford, Oxford, UK. jeremy.fairbank@ndorms.ox.ac.uk. 6. Nuffield Orthopaedic Centre, Oxford University Hospitals NHS Trust, Oxford, UK. jeremy.fairbank@ndorms.ox.ac.uk. 7. St Lukes Hospital, Latimer Road, Oxford, OX3 7PF, UK. jeremy.fairbank@ndorms.ox.ac.uk. 8. Spinal Studies, RJAH Orthopaedic Hospital Foundation Trust and ISTM (Keele University), Oswestry, UK.
Abstract
STUDY DESIGN: Investigation of the automation of radiological features from magnetic resonance images (MRIs) of the lumbar spine. OBJECTIVE: To automate the process of grading lumbar intervertebral discs and vertebral bodies from MRIs. MR imaging is the most common imaging technique used in investigating low back pain (LBP). Various features of degradation, based on MRIs, are commonly recorded and graded, e.g., Modic change and Pfirrmann grading of intervertebral discs. Consistent scoring and grading is important for developing robust clinical systems and research. Automation facilitates this consistency and reduces the time of radiological analysis considerably and hence the expense. METHODS: 12,018 intervertebral discs, from 2009 patients, were graded by a radiologist and were then used to train: (1) a system to detect and label vertebrae and discs in a given scan, and (2) a convolutional neural network (CNN) model that predicts several radiological gradings. The performance of the model, in terms of class average accuracy, was compared with the intra-observer class average accuracy of the radiologist. RESULTS: The detection system achieved 95.6% accuracy in terms of disc detection and labeling. The model is able to produce predictions of multiple pathological gradings that consistently matched those of the radiologist. The model identifies 'Evidence Hotspots' that are the voxels that most contribute to the degradation scores. CONCLUSIONS: Automation of radiological grading is now on par with human performance. The system can be beneficial in aiding clinical diagnoses in terms of objectivity of gradings and the speed of analysis. It can also draw the attention of a radiologist to regions of degradation. This objectivity and speed is an important stepping stone in the investigation of the relationship between MRIs and clinical diagnoses of back pain in large cohorts. LEVEL OF EVIDENCE: Level 3.
STUDY DESIGN: Investigation of the automation of radiological features from magnetic resonance images (MRIs) of the lumbar spine. OBJECTIVE: To automate the process of grading lumbar intervertebral discs and vertebral bodies from MRIs. MR imaging is the most common imaging technique used in investigating low back pain (LBP). Various features of degradation, based on MRIs, are commonly recorded and graded, e.g., Modic change and Pfirrmann grading of intervertebral discs. Consistent scoring and grading is important for developing robust clinical systems and research. Automation facilitates this consistency and reduces the time of radiological analysis considerably and hence the expense. METHODS: 12,018 intervertebral discs, from 2009 patients, were graded by a radiologist and were then used to train: (1) a system to detect and label vertebrae and discs in a given scan, and (2) a convolutional neural network (CNN) model that predicts several radiological gradings. The performance of the model, in terms of class average accuracy, was compared with the intra-observer class average accuracy of the radiologist. RESULTS: The detection system achieved 95.6% accuracy in terms of disc detection and labeling. The model is able to produce predictions of multiple pathological gradings that consistently matched those of the radiologist. The model identifies 'Evidence Hotspots' that are the voxels that most contribute to the degradation scores. CONCLUSIONS: Automation of radiological grading is now on par with human performance. The system can be beneficial in aiding clinical diagnoses in terms of objectivity of gradings and the speed of analysis. It can also draw the attention of a radiologist to regions of degradation. This objectivity and speed is an important stepping stone in the investigation of the relationship between MRIs and clinical diagnoses of back pain in large cohorts. LEVEL OF EVIDENCE: Level 3.
Authors: Ron I Riesenburger; Mina G Safain; Richard Ogbuji; Jackson Hayes; Steven W Hwang Journal: J Clin Neurosci Date: 2014-10-22 Impact factor: 1.961
Authors: Daniel Steffens; Mark J Hancock; Leani S M Pereira; Peter M Kent; Jane Latimer; Chris G Maher Journal: Eur Spine J Date: 2015-09-02 Impact factor: 3.134
Authors: James F Griffith; Yi-Xiang J Wang; Gregory E Antonio; Kai Chow Choi; Alfred Yu; Anil T Ahuja; Ping Chung Leung Journal: Spine (Phila Pa 1976) Date: 2007-11-15 Impact factor: 3.468
Authors: Theo Vos; Abraham D Flaxman; Mohsen Naghavi; Rafael Lozano; Catherine Michaud; Majid Ezzati; Kenji Shibuya; Joshua A Salomon; Safa Abdalla; Victor Aboyans; Jerry Abraham; Ilana Ackerman; Rakesh Aggarwal; Stephanie Y Ahn; Mohammed K Ali; Miriam Alvarado; H Ross Anderson; Laurie M Anderson; Kathryn G Andrews; Charles Atkinson; Larry M Baddour; Adil N Bahalim; Suzanne Barker-Collo; Lope H Barrero; David H Bartels; Maria-Gloria Basáñez; Amanda Baxter; Michelle L Bell; Emelia J Benjamin; Derrick Bennett; Eduardo Bernabé; Kavi Bhalla; Bishal Bhandari; Boris Bikbov; Aref Bin Abdulhak; Gretchen Birbeck; James A Black; Hannah Blencowe; Jed D Blore; Fiona Blyth; Ian Bolliger; Audrey Bonaventure; Soufiane Boufous; Rupert Bourne; Michel Boussinesq; Tasanee Braithwaite; Carol Brayne; Lisa Bridgett; Simon Brooker; Peter Brooks; Traolach S Brugha; Claire Bryan-Hancock; Chiara Bucello; Rachelle Buchbinder; Geoffrey Buckle; Christine M Budke; Michael Burch; Peter Burney; Roy Burstein; Bianca Calabria; Benjamin Campbell; Charles E Canter; Hélène Carabin; Jonathan Carapetis; Loreto Carmona; Claudia Cella; Fiona Charlson; Honglei Chen; Andrew Tai-Ann Cheng; David Chou; Sumeet S Chugh; Luc E Coffeng; Steven D Colan; Samantha Colquhoun; K Ellicott Colson; John Condon; Myles D Connor; Leslie T Cooper; Matthew Corriere; Monica Cortinovis; Karen Courville de Vaccaro; William Couser; Benjamin C Cowie; Michael H Criqui; Marita Cross; Kaustubh C Dabhadkar; Manu Dahiya; Nabila Dahodwala; James Damsere-Derry; Goodarz Danaei; Adrian Davis; Diego De Leo; Louisa Degenhardt; Robert Dellavalle; Allyne Delossantos; Julie Denenberg; Sarah Derrett; Don C Des Jarlais; Samath D Dharmaratne; Mukesh Dherani; Cesar Diaz-Torne; Helen Dolk; E Ray Dorsey; Tim Driscoll; Herbert Duber; Beth Ebel; Karen Edmond; Alexis Elbaz; Suad Eltahir Ali; Holly Erskine; Patricia J Erwin; Patricia Espindola; Stalin E Ewoigbokhan; Farshad Farzadfar; Valery Feigin; David T Felson; Alize Ferrari; Cleusa P Ferri; Eric M Fèvre; Mariel M Finucane; Seth Flaxman; Louise Flood; Kyle Foreman; Mohammad H Forouzanfar; Francis Gerry R Fowkes; Richard Franklin; Marlene Fransen; Michael K Freeman; Belinda J Gabbe; Sherine E Gabriel; Emmanuela Gakidou; Hammad A Ganatra; Bianca Garcia; Flavio Gaspari; Richard F Gillum; Gerhard Gmel; Richard Gosselin; Rebecca Grainger; Justina Groeger; Francis Guillemin; David Gunnell; Ramyani Gupta; Juanita Haagsma; Holly Hagan; Yara A Halasa; Wayne Hall; Diana Haring; Josep Maria Haro; James E Harrison; Rasmus Havmoeller; Roderick J Hay; Hideki Higashi; Catherine Hill; Bruno Hoen; Howard Hoffman; Peter J Hotez; Damian Hoy; John J Huang; Sydney E Ibeanusi; Kathryn H Jacobsen; Spencer L James; Deborah Jarvis; Rashmi Jasrasaria; Sudha Jayaraman; Nicole Johns; Jost B Jonas; Ganesan Karthikeyan; Nicholas Kassebaum; Norito Kawakami; Andre Keren; Jon-Paul Khoo; Charles H King; Lisa Marie Knowlton; Olive Kobusingye; Adofo Koranteng; Rita Krishnamurthi; Ratilal Lalloo; Laura L Laslett; Tim Lathlean; Janet L Leasher; Yong Yi Lee; James Leigh; Stephen S Lim; Elizabeth Limb; John Kent Lin; Michael Lipnick; Steven E Lipshultz; Wei Liu; Maria Loane; Summer Lockett Ohno; Ronan Lyons; Jixiang Ma; Jacqueline Mabweijano; Michael F MacIntyre; Reza Malekzadeh; Leslie Mallinger; Sivabalan Manivannan; Wagner Marcenes; Lyn March; David J Margolis; Guy B Marks; Robin Marks; Akira Matsumori; Richard Matzopoulos; Bongani M Mayosi; John H McAnulty; Mary M McDermott; Neil McGill; John McGrath; Maria Elena Medina-Mora; Michele Meltzer; George A Mensah; Tony R Merriman; Ana-Claire Meyer; Valeria Miglioli; Matthew Miller; Ted R Miller; Philip B Mitchell; Ana Olga Mocumbi; Terrie E Moffitt; Ali A Mokdad; Lorenzo Monasta; Marcella Montico; Maziar Moradi-Lakeh; Andrew Moran; Lidia Morawska; Rintaro Mori; Michele E Murdoch; Michael K Mwaniki; Kovin Naidoo; M Nathan Nair; Luigi Naldi; K M Venkat Narayan; Paul K Nelson; Robert G Nelson; Michael C Nevitt; Charles R Newton; Sandra Nolte; Paul Norman; Rosana Norman; Martin O'Donnell; Simon O'Hanlon; Casey Olives; Saad B Omer; Katrina Ortblad; Richard Osborne; Doruk Ozgediz; Andrew Page; Bishnu Pahari; Jeyaraj Durai Pandian; Andrea Panozo Rivero; Scott B Patten; Neil Pearce; Rogelio Perez Padilla; Fernando Perez-Ruiz; Norberto Perico; Konrad Pesudovs; David Phillips; Michael R Phillips; Kelsey Pierce; Sébastien Pion; Guilherme V Polanczyk; Suzanne Polinder; C Arden Pope; Svetlana Popova; Esteban Porrini; Farshad Pourmalek; Martin Prince; Rachel L Pullan; Kapa D Ramaiah; Dharani Ranganathan; Homie Razavi; Mathilda Regan; Jürgen T Rehm; David B Rein; Guiseppe Remuzzi; Kathryn Richardson; Frederick P Rivara; Thomas Roberts; Carolyn Robinson; Felipe Rodriguez De Leòn; Luca Ronfani; Robin Room; Lisa C Rosenfeld; Lesley Rushton; Ralph L Sacco; Sukanta Saha; Uchechukwu Sampson; Lidia Sanchez-Riera; Ella Sanman; David C Schwebel; James Graham Scott; Maria Segui-Gomez; Saeid Shahraz; Donald S Shepard; Hwashin Shin; Rupak Shivakoti; David Singh; Gitanjali M Singh; Jasvinder A Singh; Jessica Singleton; David A Sleet; Karen Sliwa; Emma Smith; Jennifer L Smith; Nicolas J C Stapelberg; Andrew Steer; Timothy Steiner; Wilma A Stolk; Lars Jacob Stovner; Christopher Sudfeld; Sana Syed; Giorgio Tamburlini; Mohammad Tavakkoli; Hugh R Taylor; Jennifer A Taylor; William J Taylor; Bernadette Thomas; W Murray Thomson; George D Thurston; Imad M Tleyjeh; Marcello Tonelli; Jeffrey A Towbin; Thomas Truelsen; Miltiadis K Tsilimbaris; Clotilde Ubeda; Eduardo A Undurraga; Marieke J van der Werf; Jim van Os; Monica S Vavilala; N Venketasubramanian; Mengru Wang; Wenzhi Wang; Kerrianne Watt; David J Weatherall; Martin A Weinstock; Robert Weintraub; Marc G Weisskopf; Myrna M Weissman; Richard A White; Harvey Whiteford; Steven T Wiersma; James D Wilkinson; Hywel C Williams; Sean R M Williams; Emma Witt; Frederick Wolfe; Anthony D Woolf; Sarah Wulf; Pon-Hsiu Yeh; Anita K M Zaidi; Zhi-Jie Zheng; David Zonies; Alan D Lopez; Christopher J L Murray; Mohammad A AlMazroa; Ziad A Memish Journal: Lancet Date: 2012-12-15 Impact factor: 79.321
Authors: W Katherine Tan; Saeed Hassanpour; Patrick J Heagerty; Sean D Rundell; Pradeep Suri; Hannu T Huhdanpaa; Kathryn James; David S Carrell; Curtis P Langlotz; Nancy L Organ; Eric N Meier; Karen J Sherman; David F Kallmes; Patrick H Luetmer; Brent Griffith; David R Nerenz; Jeffrey G Jarvik Journal: Acad Radiol Date: 2018-03-28 Impact factor: 3.173