Assessing suspected spinal cord compression: a multidisciplinary outcomes analysis of 342 episodes

The object of this work was to evaluate the assessment and document the outcomes of cancer patients with suspected spinal cord compression (SCC). In a retrospective cohort study of 342 episodes of suspected SCC in cancer patients evaluated by computed tomography (CT) of the spine, a multidisciplinary team of neurologists, radiologists, and oncologists assessed the impact of varying the anatomical criterion for SCC and including new SCC diagnosed shortly after definitive radiographical imaging. We developed a logistic regression model to identify independent clinical predictors of SCC, including the natural history of the underlying cancer as well as neurological and radiological risk factors. Management of suspected SCC infrequently involved neurology consultation (21% of episodes). The frequency of SCC increased more than four-fold when the definition was expanded to include epidural cancer rather than spinal cord displacement only (36% vs. 8%), and 90-day clinical follow-up identified few new lesions not evident on definitive imaging studies. Clinical information about the course of cancer (documentation and duration of metastatic cancer) added independent predictive information to that yielded by neurological assessment and prior imaging studies in a multiple regression model. The a priori predicted risk of SCC, which ranged from 4% to 87% in this study, may vary enough to affect treatment strategies, although our population may have excluded very-low-risk patients. Consistent anatomical definitions of SCC, clinical follow-up of definitive imaging studies and the addition of information on the natural history of cancer to traditional neurological and radiographical evaluation may all improve clinical assessment of suspected SCC in cancer patients.