Wednesday, September 26, 2018 7:35 AM–9:00 AM ePosters: P20. Morphometrics predicts overall survival in patients with lung, breast, prostate or myeloma spine metastases, independent of histology

Abstract

BACKGROUND CONTEXT Creating valid, accurate, and reliable markers of overall survival in patients with cancer would lead to better treatment stratification for this population, whether chemotherapy, radiation, surgery, or palliation. This concept is especially relevant to patients who have spinal metastasis, who will have the lowest rate of survival with stage IV disease. Select patients who undergo surgery for their metastasis may have improved overall survival, neurological outcome, and pain control. However, these procedures are resource intensive, delay the definitive oncologic treatments of chemotherapy and radiotherapy, and potentially morbid, which can erase any survival benefit. Scoring systems have been developed to guide surgical decision making, but validity studies have shown inaccuracies in predicting postoperative morbidity and mortality. New objective measurements of fitness for surgery would have the potential utility to help stratify and aid in the selection of better surgical candidates for spine metastases surgery. One of the hallmarks of human senescence is that of frailty, which has been defined as a decreased reserve to physiologic stressors. Frail patients are known to have increased risk for postoperative morbidity and mortality, but the direct measurement of frailty is impractical. Sarcopenia (lack of muscle mass) has been used successfully as a surrogate for frailty, and a growing body of literature seeks to use sarcopenia to predict outcomes after surgery. PURPOSE In this study, we applied the sarcopenia or frailty paradigm in patients with lung, breast, prostate, or myeloma metastasis to the spine to predict overall survival. Our hypothesis is that sarcopenia can be used as a unique predictor of overall mortality. STUDY DESIGN/SETTING This study was conducted after Institutional Review Board approval (IRB 4370). From an index of patients who have undergone stereotactic body radiation therapy (SBRT) for metastatic spine cancer, we retrospectively identified lung, breast, prostate, or multiple myeloma patients from 2002 to 2012 at our hospital. OUTCOME MEASURES The primary outcome measure was overall survival, which was calculated from the date of the patient's CT scan to date of death or last follow-up. METHODS Using the most recent CT scan available, the total area of each patient's psoas muscle at the L4 vertebral body was measured and recorded. Psoas muscle sizes were divided into tertiles (smallest, middle, and largest) according to average psoas area. Other demographic variables such as age, gender, ethnicity, number of levels, type of cancer, spine surgery, and SBRT target volume were also considered. The median survival time (in days) along with the corresponding 95% confidence interval were computed using Kaplan–Meier estimates for all patients, as well as for the variables of interest. Cox proportional hazards regression analyses were done to estimate the hazard ratios and test for differences in the variables of interest. RESULTS There were 417 patients with spinal metastases from lung cancer, breast cancer, prostate cancer, or multiple myeloma. The average age was 66.3, with equal split of males and females and 52% of patients Caucasian, 40% African American, and 4% other. The median survival for all patients was 173days (95% CI=140–204 days). Patient age, gender, ethnicity, number of levels treated with SBRT, spine surgery done, and SBRT target volume did not affect overall survival. Overall survival was associated with primary tumor histology, with multiple myeloma patients having a significantly longer survival compared to the other three groups (p .77 for all comparisons). Average psoas size significantly predicted overall survival. After multivariate analysis, patients in the 1st tertile (smallest muscle area) for average psoas size had significantly shorter survival as compared to the 3rd tertile (largest muscle area): 115days versus 298, hazard ratio 1.92 (95% CI=1.46–2.53), p CONCLUSIONS Sarcopenia predicts survival in patients with lung cancer, breast cancer, prostate cancer, and multiple myeloma metastases to the spine. Survival prediction was irrespective of cancer histology. This information can assist in patient selection for surgery, as well as to tailor oncologic treatments.