David Choi1, Federico Ricciardi2, Mark Arts3, Jacob M Buchowski4, Cody Bunger5, Chun Kee Chung6, Maarten Coppes7, Bart Depreitere8, Michael Fehlings9, Norio Kawahara10, Yee Leung11, Antonio Martin-Benlloch12, Eric Massicotte9, Christian Mazel13, Bernhard Meyer14, Cumhur Oner15, Wilco Peul16, Nasir Quraishi17, Yasuaki Tokuhashi18, Katsuro Tomita10, Christian Ulbricht19, Jorrit-Jan Verlaan15, Mike Wang20, Alan Crockard1. 1. Department of Neurosurgery, National Hospital for Neurology and Neurosurgery, University College London, London, United Kingdom. 2. Department of Statistical Science, University College London, London, United Kingdom. 3. Department of Neurosurgery, Medical Center Haaglanden, Haaglanden, the Netherlands. 4. Department of Orthopaedic Surgery, Washington University School of Medicine, St. Louis, Missouri. 5. Department of Orthopedic Surgery, University Hospital of Aarhus, Aarhus, Denmark. 6. Department of Neurosurgery, Seoul National University Hospital, Seoul, Republic of Korea. 7. Department of Neurosurgery, University Medical Center Groningen, Groningen, the Netherlands. 8. Division of Neurosurgery, University Hospital Leuven, Leuven, Belgium. 9. Division of Neurosurgery and Spinal Program, University of Toronto and Toronto Western Hospital, Toronto, Ontario, Canada. 10. Department of Orthopedic Surgery, Kanazawa Medical University Hospital, Kanazawa, Japan. 11. Department of Orthopaedics, Musgrove Park Hospital, Taunton, United Kingdom. 12. Spinal Unit, University Hospital Dr Peset, Valencia, Spain. 13. Department of Orthopedic Surgery, Institut Mutualiste Montsouris, Paris, France. 14. TUM School of Medicine, Technische Universität München, Munich, Germany. 15. Department of Orthopedic Surgery, University Medical Center Utrecht, Utrecht, the Netherlands. 16. Department of Neurosurgery, Leiden University Medical Centre, Leiden, the Netherlands. 17. Centre for Spine Studies and Surgery, Queens Medical Centre, Nottingham, United Kingdom. 18. Department of Orthopaedic Surgery, Nihon University School of Medicine, Tokyo, Japan. 19. Department of Neurosurgery, Charing Cross Hospital, London, United Kingdom. 20. Department of Neurosurgery, University of Miami Hospital, Miami, Florida.
Abstract
STUDY DESIGN: A prospective multicenter cohort study. OBJECTIVE: To assess the clinical accuracy of six commonly cited prognostic scoring systems for patients with spinal metastases. SUMMARY OF BACKGROUND DATA: There are presently several available methods for the estimation of prognosis in metastatic spinal disease, but none are universally accepted by surgeons for clinical use. These scoring systems have not been rigorously tested and validated in large datasets to see if they are reliable enough to inform day-to-day patient management decisions. We tested these scoring systems in a large cohort of patients. A total of 1469 patients were recruited into a secure internet database, and prospectively collected data were analyzed to assess the accuracy of published prognostic scoring systems. METHODS: We assessed six prognostic scoring systems, described by the first authors Tomita, Tokuhashi, Bauer, van der Linden, Rades, and Bollen. Kaplan-Meier survival estimates were created for different patient subgroups as described in the original publications. Harrell's C-statistic was calculated for the survival estimates, to assess the concordance between estimated and actual survival. RESULTS: All the prognostic scoring systems tested were able to categorize patients into separate prognostic groups with different overall survivals. However none of the scores were able to achieve "good concordance" as assessed by Harrell's C-statistic. The score of Bollen and colleagues was found to be the most accurate, with a Harrell's C-statistic of 0.66. CONCLUSION: No prognostic scoring system was found to have a good predictive value. The scores of Bollen and Tomita were the most effective with Harrell's C-statistic of 0.66 and 0.65, respectively. Prognostic scoring systems are calculated using data from previous years, and are subject to inaccuracies as treatments advance in the interim. We suggest that other methods of assessing prognosis should be explored, such as prognostic risk calculation. LEVEL OF EVIDENCE: 3.
STUDY DESIGN: A prospective multicenter cohort study. OBJECTIVE: To assess the clinical accuracy of six commonly cited prognostic scoring systems for patients with spinal metastases. SUMMARY OF BACKGROUND DATA: There are presently several available methods for the estimation of prognosis in metastatic spinal disease, but none are universally accepted by surgeons for clinical use. These scoring systems have not been rigorously tested and validated in large datasets to see if they are reliable enough to inform day-to-day patient management decisions. We tested these scoring systems in a large cohort of patients. A total of 1469 patients were recruited into a secure internet database, and prospectively collected data were analyzed to assess the accuracy of published prognostic scoring systems. METHODS: We assessed six prognostic scoring systems, described by the first authors Tomita, Tokuhashi, Bauer, van der Linden, Rades, and Bollen. Kaplan-Meier survival estimates were created for different patient subgroups as described in the original publications. Harrell's C-statistic was calculated for the survival estimates, to assess the concordance between estimated and actual survival. RESULTS: All the prognostic scoring systems tested were able to categorize patients into separate prognostic groups with different overall survivals. However none of the scores were able to achieve "good concordance" as assessed by Harrell's C-statistic. The score of Bollen and colleagues was found to be the most accurate, with a Harrell's C-statistic of 0.66. CONCLUSION: No prognostic scoring system was found to have a good predictive value. The scores of Bollen and Tomita were the most effective with Harrell's C-statistic of 0.66 and 0.65, respectively. Prognostic scoring systems are calculated using data from previous years, and are subject to inaccuracies as treatments advance in the interim. We suggest that other methods of assessing prognosis should be explored, such as prognostic risk calculation. LEVEL OF EVIDENCE: 3.
Authors: Nicolas Dea; Anne L Versteeg; Arjun Sahgal; Jorrit-Jan Verlaan; Raphaële Charest-Morin; Laurence D Rhines; Daniel M Sciubba; James M Schuster; Michael H Weber; Aron Lazary; Michael G Fehlings; Michelle J Clarke; Paul M Arnold; Stefano Boriani; Chetan Bettegowda; Ilya Laufer; Ziya L Gokaslan; Charles G Fisher Journal: Neurosurgery Date: 2020-08-01 Impact factor: 4.654
Authors: Tamás Mezei; Anna Horváth; Péter Pollner; Gábor Czigléczki; Péter Banczerowski Journal: Int J Clin Oncol Date: 2020-01-28 Impact factor: 3.402
Authors: Arthur Wagner; Elena Haag; Ann-Kathrin Joerger; Philipp Jost; Stephanie E Combs; Maria Wostrack; Jens Gempt; Bernhard Meyer Journal: Sci Rep Date: 2021-04-12 Impact factor: 4.379