Patterns of Care for Brain Metastasis Radiotherapy (RT) in an Integrated Healthcare System: Does Increasing Utilization of Stereotactic Radiosurgery (SRS) Compared to Whole Brain RT (WBRT) Lead to Excessive Use at the End of Life (EOL)?

Abstract

Previous study at our institution showed a high rate of WBRT received near the EOL. Growing evidence of improved neurocognitive and quality of life (QOL) outcomes without compromise of local control has led to increased advocacy of SRS for brain metastasis compared to WBRT. We examine our trend in patterns of care for non-small cell lung cancer (NSCLC) patients, including palliative care/hospice enrollment, and whether rate of brain metastasis RT near EOL has changed with evolving practices.Central Cancer Registry identified 4765 patients diagnosed with NSCLC at 2 facilities from 2007-17. Patients with brain metastasis were identified using natural language processing of imaging reports and confirmed by image review. Patient clinical/treatment characteristics were obtained through database linkage and manual chart abstraction. Bivariate and multivariate logistic regression analyses were performed.Of 792 patients with brain metastasis, 537 received at least 1 fraction of RT. SRS as initial treatment increased over time, comprising 10% of cases in 2007-11, 22% in 2012-16, and 52% in 2017. Number of lesions initially treated with SRS also increased: among patients with 1-3 lesions, 15% in 2007-11, 33% in 2012-16, and 68% in 2017 received SRS upfront, while 3%, 13%, and 33% of patients with 4-9 lesions received it in the respective periods. SRS for patients with 10-19 lesions was minimal, though half received SRS upfront in 2017. The proportion of patients receiving WBRT near EOL was substantial and higher than for SRS. Of 244 patients with 1-3 lesions, 64 (26%) died within 30 days of RT, comprising 25% of patients receiving WBRT and 11% receiving SRS. Older age, male sex, increasing number of extracranial metastatic sites, and inpatient consultation were independent predictors of death within 30 days of RT. Patients enrolled in palliative care/hospice were less likely to die within 30 days of RT (P = 0.048) and less likely to die as inpatients (P < 0.01).SRS utilization increased from 2007-17, particularly among patients with 1-9 lesions. Despite this increase, the proportion of patients receiving SRS near EOL was low and much less than the proportion receiving WBRT, suggesting that patients were selected well for SRS but not WBRT. As SRS is associated with shorter treatment time and lower toxicity compared to WBRT, greater SRS utilization in those without poor predictive factors may maintain QOL without risking excessive use at EOL. Early palliative care and expanding indications for upfront SRS may be considerations for decreasing the risk of treatment near the EOL.A.D. Batra: None. S. Yang: None. C. Zheng: None. D. Jiang: None. J. Rahimian: None. M.R. Girvigian: None. M.K. Gould: None. J.J. Ryoo: None.