Automated segmentation of postsurgical resection cavities on magnetic resonance imaging in focal epilepsy: A Multicentre Epilepsy Lesion Detection study.
Authors
Affiliations (72)
Affiliations (72)
- UCL Great Ormond Street Institute of Child Health, London, UK.
- School of Biomedical Engineering & Imaging Sciences, King's College London, London, UK.
- Neurobiology Research Unit, Rigshospitalet, Copenhagen, Denmark.
- Department of Neurology and Epileptology, University of Tübingen, Tübingen, Germany.
- Department of Neuroscience, Monash University, Melbourne, Victoria, Australia.
- Department of Neurology, Alfred Hospital, Melbourne, Victoria, Australia.
- Department of Neuroscience, School of Translational Medicine, Monash University, Melbourne, Victoria, Australia.
- Amrita Advanced Center for Epilepsy, Amrita Institute of Medical Sciences, Amrita Vishwa Vidyapeetham, Kochi, India.
- Boston Children's Hospital, Boston, Massachusetts, USA.
- University of Iowa Health Care, Iowa City, Iowa, USA.
- Bristol Royal Hospital for Children, Bristol, UK.
- Beijing Tiantan Hospital, Capital Medical University, Beijing, China.
- Central Emergency Military Hospital, Bucarest, Romania.
- University Hospital of Wales, Cardiff, UK.
- Department of Neurology, Welsh Epilepsy Centre, University Hospital of Wales, Cardiff, UK.
- Center for Neuroscience Research, Children's National Research Institute, Children's National Hospital, Washington, District of Columbia, USA.
- Department of Neurology, George Washington University School of Medicine, Washington, District of Columbia, USA.
- Department of Pediatrics, George Washington University School of Medicine, Washington, District of Columbia, USA.
- CHU Sainte-Justine, Montreal, Quebec, Canada.
- Université de Montréal, Montreal, Quebec, Canada.
- Collégial International Sainte-Anne, Lachine, Quebec, Canada.
- Latin America Brain Health Institute, Universidad Adolfo Ibanez, Santiago, Chile.
- Department of Biophysics, School of Medicine, Istanbul Medipol University, Istanbul, Türkiye.
- School of Medicine, Finis Terrae University, Santiago, Chile.
- Biomedical Imaging Unit and Artificial Intelligence, Foundation for the Promotion of Health and Biomedical Research of the Valencia Region - Prince Felipe Research Center (FISABIO-CIPF), València, Spain.
- Department of Neuroradiology, University Hospital Freiburg, Freiburg, Germany.
- Epilepsy Center, Medical Center-University of Freiburg, member of European Reference Network EpiCARE, Faculty of Medicine, University of Freiburg, Freiburg, Germany.
- Research Department, Fundación Iniciativa Para las Neurociencias, Madrid, Spain.
- Department of Neurology, Epilepsy Program, Ruber International Hospital, Madrid, Spain.
- Child Neuropsychiatry Unit, member of European Reference Network EpiCARE, Istituto di Ricovero e Cura a Carattere Scientifico (IRCCS) Istituto Giannina Gaslini, Genoa, Italy.
- Department of Neuroradiology, IRCCS Istituto Giannina Gaslini, Genoa, Italy.
- Division of Neurosurgery, IRCCS Istituto Giannina Gaslini, Genoa, Italy.
- DINOGMI, University of Genoa, Genoa, Italy.
- IRCCS Istituto Giannina Gaslini, Genoa, Italy.
- Department of Neurosciences, Rehabilitation, Ophthalmology, Genetics, Maternal and Child Health, University of Genoa, Genoa, Italy.
- Department of Neurosurgery, Great Ormond Street Hospital, London, UK.
- Radiology Department, Great Ormond Street Hospital for Children, London, UK.
- Hospital for Sick Children, Toronto, Ontario, Canada.
- Department of Neuroradiology, Diagnostic Imaging Center, Barcelona, Spain.
- Fundació de Recerca Clínic Barcelona, Institut d'Investigacions Biomèdiques August Pi i Sunyer (FCRB-IDIBAPS), Barcelona, Spain.
- Kuopio University Hospital, full member of European Reference Network EpiCARE, Kuopio, Finland.
- University of Eastern Finland, Kuopio, Finland.
- Neuroscience and Human Genetics Department, Meyer Children's Hospital IRCCS, Florence, Italy.
- University of Florence, Florence, Italy.
- Department of Neurosurgery, National Center of Neurology and Psychiatry, Tokyo, Japan.
- Department of Radiology, National Center of Neurology and Psychiatry, Tokyo, Japan.
- Department of Psychiatry and Behavioral Science, Juntendo University Graduate School of Medicine, Tokyo, Japan.
- Department of Diagnostic Radiology, Institute of Science, Tokyo, Japan.
- CHU Lyon, Lyon, France.
- Department of Neuroradiology, Neurosurgery Institute Dr. A. Asenjo, Santiago, Chile.
- Neurophysiology Unit, Neurosurgery Institute Dr. A. Asenjo, Santiago, Chile.
- Pediatric Neurosurgery, Neurosurgery Institute Dr. A. Asenjo, Santiago, Chile.
- Medical Physics Unit, Bambino Gesù Children's Hospital, IRCCS, Rome, Italy.
- Neurosurgery Unit, Bambino Gesù Children's Hospital, IRCCS, Rome, Italy.
- Neurology, Epilepsy, and Movement Disorders, Bambino Gesù Children's Hospital, IRCCS, Rome, Italy.
- Diagnostic and Interventional Neuroradiology Unit, Bambino Gesù Children's Hospital, IRCCS, Rome, Italy.
- Department of Radiology, St. Luke's Hospital, Thessaloniki, Greece.
- Epilepsy Center, St. Luke's Hospital, Thessaloniki, Greece.
- Birmingham Women's and Children's NHS Foundation Trust, Birmingham, UK.
- Institute of Health and Neurodevelopment, Aston University, Birmingham, UK.
- UCL Queen Square Institute of Neurology, London, UK.
- National Hospital for Neurology and Neurosurgery, London, UK.
- University of Campinas, Campinas, Brazil.
- Brazilian Institute of Neuroscience and Neurotechnology, São Paulo, Brazil.
- Department of Neuropediatrics, University Children's Hospital Zurich, Zurich, Switzerland.
- MR-Research Center, University Children's Hospital Zurich, Zurich, Switzerland.
- University of Zurich, Zurich, Switzerland.
- Department of Neurology, University of Pennsylvania, Philadelphia, Pennsylvania, USA.
- Department of Radiology, University of Pennsylvania, Philadelphia, Pennsylvania, USA.
- Department of Biostatistics, Epidemiology, and Informatics, University of Pennsylvania, Philadelphia, Pennsylvania, USA.
- Department of Neurology, Washington University School of Medicine, St. Louis, Missouri, USA.
- Department of Radiology, Washington University School of Medicine, St. Louis, Missouri, USA.
Abstract
Quantitative assessment of extent of tissue resection following epilepsy surgery requires accurate delineation of the resection cavity on postoperative magnetic resonance imaging (MRI). Current methods for resection cavity masking are time-consuming and labor-intensive, and existing automated approaches exhibit variable segmentation accuracy, particularly on extratemporal resections. We developed MELD-PostOp, a deep learning tool trained and evaluated on a large, heterogeneous cohort to automatically segment resection cavities. The study included 1.5- and 3T postoperative three-dimensional T1-weighted MRI images from the Multicentre Epilepsy Lesion Detection (MELD) project (n<sub>subjects</sub> = 969, 27 centers) and from the EPISURG dataset (n = 133). The cohort included children and adults, alongside a range of resection locations, pathologies, and MRI characteristics. Resection cavities were individually segmented in 285 subjects and used to train an nnU-Net prototype model. The prototype model was used to generate an additional 680 resection masks, which were subsequently quality-controlled, edited, and combined with the original 285 to train the final MELD-PostOp model (n = 965). A Stratified Test Cohort (n = 50) and Independent Test Cohort (n = 87) were withheld for model evaluation. Performance was evaluated using Dice similarity coefficient (DSC), 95th percentile Hausdorff distance (HD95), number of predicted clusters, and inference runtime, and compared against established tools (Epic-CHOP, ResectVol, and RESSEG). MELD-PostOp achieved a median DSC of .85 and HD95 of 3.61 on the combined test cohort, outperforming Epic-CHOP (DSC .69, HD95 9.67), ResectVol (DSC .66, HD95 15.05), and RESSEG (DSC .43, HD95 32.67), with significant improvements seen in both temporal and especially extratemporal resections. The model detected 98.5% (135/137) of resection cavities. MELD-PostOp runtime was 17 s per MRI, compared to 612 s (ResectVol), 3205 s (Epic-CHOP), and 4 s (RESSEG). MELD-PostOp performance remained high across clinical and imaging subgroups (median DSC > .8). MELD-PostOp is an open-source research tool that provides an accurate, efficient, and generalizable solution for postoperative resection cavity segmentation using only postoperative MRI scans.