Author:

Keywords:

Science & Technology, Life Sciences & Biomedicine, Oncology, Spine surgery, Risk, Survival, Metastasis, Tumour, Outcome, PROGNOSTIC SCORING SYSTEMS, CARDIOVASCULAR RISK, SURVIVAL, ACCURACY, CANCER, VALIDATION, EVENTS, COHORT, Databases, Factual, Follow-Up Studies, Humans, Neoplasms, Neurosurgical Procedures, Predictive Value of Tests, Prospective Studies, Risk Assessment, Severity of Illness Index, Spinal Neoplasms, 1112 Oncology and Carcinogenesis, 1117 Public Health and Health Services, Oncology & Carcinogenesis, 3211 Oncology and carcinogenesis

Abstract:

AIM: Surgery for spinal metastases can improve symptoms, but sometimes complications can negate the benefits. Operations may have different indications, complexities and risks, and the choice for an individual is a tailor-made personalised decision. Previous prognostic scoring systems are becoming out of date and inaccurate. We designed a risk calculator to estimate survival after surgery, to inform clinicians and patients when making management decisions. METHODS: A prospective cohort study was performed, including 1430 patients with spinal metastases who underwent surgery. Of them, 1264 patients from 20 centres were used for model development using a Cox frailty model. Calibration slope, D-statistic and C-index were used for model validation based on 166 patients. Follow-up was to death or minimum of 2 years after surgery. Pre-operative indices (examination findings, pain, Karnofsky physical functioning score, and radiology) were assessed. RESULTS: An algorithm to predict survival was constructed including the tumour type, ambulatory status, analgesic use, American Society of Anesthesiologists score, number of spinal metastases, previous radiotherapy or chemotherapy, presence of visceral metastases, cervical or thoracic spine involvement, as predictors. An Internet-based risk calculator was developed based on this algorithm, with similar or improved accuracy compared to other validated prognostic scoring systems (C-index, 0.68; 95% confidence interval, 0.63--0.73, and calibration slope, 1.00; 95% confidence interval, 0.68--1.32). CONCLUSION: A large, prospective, surgical series of patients with symptomatic spinal metastases was used to create a validated risk calculator that can help clinicians to inform patients about the most appropriate treatment plan. The calculator is available at www.spinemet.com.