A study of non small cell lung cancer (NSCLC) patients with brain metastasis: A single centre experience

Abstract

BackgroundLung cancer is the leading cause of cancer death with the majority of cases being non-small cell lung cancer (NSCLC) [1]. A common complication of NSCLC is brain metastasis (BM) [2, 3], where the prognosis remains poor despite new treatments. Real world data complements data gained from clinical trials, providing information on patients excluded from prospective research [4]. However, information from patient notes may prove incomplete and difficult to extract. We developed an algorithm to identify patients in our clinical database with brain metastasis from the electronic health record (EHR). MethodsWe retrospectively extracted data from the EHR of patients managed at a large teaching hospital between 2007 and 2018. Using the ICD-10 code C34, for lung cancer, our algorithm used phrases associated with BMs to search the unstructured text of radiology reports. Summary statistics and univariant analysis was performed for overall survival. Results818 patients were identified as potentially having BM and 453 patients were confirmed on clinical review of their records. The median age of patients was 69 years, 50% were female and 66% had a performance status of >2. 12.2% had an identifiable mutation and 11.5% were identified as PD-L1 positive. In the first line setting, 65% of patients received symptomatic treatment, 23% received systemic anticancer therapy (SACT), 6.1% surgery and 10% radiotherapy, of which 6.5% had external beam and 3.5% stereotactic radiosurgery. Regarding those treated with SACT, 35% had an intracranial response to treatment (3% had complete response, 32% had a partial response). Median survival was 2 months (1.9 – 2.4 months 95% CI). ConclusionThe real-world prognosis for NSCLC patients with BMs is poor. By using an algorithm, we have reported outcomes on a comprehensive cohort of patients which helps identify those for whom an active treatment approach is appropriate.