Abstract
OBJECTIVES
To explore the impact of immunotherapy on short-term outcomes and long-term survival rates, and the predictors of immunotherapy receipt in patients with glioblastoma (GBM) using NCDB.
METHODS
A total of 74245 GBM patients were derived from the NCDB from 2004 to 2014. Analyses include short-term outcomes (at least 6-month, 1-year, or 2-year after diagnosis) and long-term survivorship (at least 3-year). The Kaplan-Meier method and accelerated failure time (AFT) models were performed for survival analysis. The multivariable binary logistic regressions were applied to identify predictors of immunotherapy receipt. Random survival forest was conducted to validate the variable importance and decision tree as well.
RESULTS
A total of 766 (1.0%) GBM patients received immunotherapy as the first-line treatment. The multivariable binary logistic regressions identified the significant predictors related to immunotherapy receipt, including the recent years of diagnosis (especially 2013 and 2014), age < 65 years, higher income, private insurance, residence-hospital distance >50 miles, care transition, treatment at the facility located in South regions or academic facilities, and adjuvant therapy. After adjusting socio-demographics, facility characteristics, and clinical treatments (surgery and adjuvant therapy), patients received immunotherapy experienced the significantly prolonged OS compared to those who didn’t [OS (months): 16.0 vs. 9.8, log-rank test p-value: < 0.001; Time Ratio (TR): 1.26 vs. 1.00 (Ref.), multivariable AFT p-value: < 0.001]. In multivariable logistic regressions, compared to patients without immunotherapy, the likelihoods of 6-month (OR: 3.89, p< 0.001), 1-year (OR: 2.38, p< 0.001), and 2-year (OR: 1.58, p=0.001) survival rates were significantly increased for those received immunotherapy. Regarding the long-term survivorship, immunotherapy was significantly associated with a 68% higher likelihood of 3-year survival rate (p=0.004).
CONCLUSIONS
Our findings demonstrated that immunotherapy in GBM patients, albeit small sample size, was significantly associated with improved short-term outcomes and 3-year survivorship after adjusting traditional clinical treatments and other covariates.