Development of Prognostic Models for Survival and Quality of Life in Surgically Managed Patients with Metastatic Spinal Cord Compression

Untreated Metastatic Epidural Spinal Cord Compression (MESCC) leads to irreversible neurologic deficits and worsened quality of life (QoL). While all existing Clinical Prognostic Models (CPMs) predict survival, most were derived from retrospective data from small patient samples using unclear methodology and there is no consensus related to their accuracy. Of note, to date, there is no CPM predicting QoL. Accurate CPMs for survival and QoL may assist clinicians and patients in the process of joint therapeutic decision-making and help patients form realistic expectations related to the outcome of surgery. We conducted a survey and a systematic literature review to isolate key preoperative predictors of survival and QoL. In accordance with The Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis (TRIPOD) guideline, we identified significant predictors of survival and performed an external validation of eight existing CPMs for survival. Using high-quality prospective data, we integrated evidence from the current literature, expert opinion and statistical findings to create CPMs of survival and QoL in accordance with the TRIPOD guideline. We identified the site of primary tumor, the presence of organ metastasis and the SF-36 physical component score (PCS) to be significant predictors of survival. For all eight CPMs evaluated, discrimination was possibly helpful and calibration was poor. Our CPM of survival includes five predictors (SF-36 PCS, type of primary tumor, presence of organ metastasis, MESCC treated with radiotherapy prior to surgery and male sex) while the CPM of QoL at 3 months after surgery consist of four predictors (KPS, living in NA, SF-36 PCS and SF36 mental component score). Our CPMs demonstrated possibly helpful discrimination and good calibration. In accordance with the TRIPOD guideline, we created and internally validated a CPM of survival and the first CPM of QoL at 3 months after surgery for patients with MESCC. We believe we have achieved the maximum prognostic ability using traditional predictors, methods and outcome measures. Future work will concentrate on further external validation of CPMs of survival and QoL, to examine other predictors of outcome, such as molecular diagnosis and biomarkers, and to employ disease-specific outcome measurement instruments.Ph.D