Related Experiment Video
Updated: Sep 16, 2026
Profiling Sensitivity to Targeted Therapies in EGFR-Mutant NSCLC Patient-Derived Organoids
Published on: November 22, 2021
EGFR Mutation Subtype and Risk of Brain Metastases From Non-Small Cell Lung Cancer in the Osimertinib Era: EGFR
Ameya Patel1, Victor Lee2, Amin H Nassar3
1Choate Rosemary Hall, Wallingford.
Objectives:
In lung adenocarcinoma, EGFR gene mutations are common and targetable by tyrosine kinase inhibitors (TKIs). Many patients develop brain metastases that are challenging to treat with TKIs because of the blood-brain barrier. The EGFR L858R mutation is believed to carry a higher brain metastasis risk than the Exon 19 deletion (Ex19Del). We tested whether this remains true with more modern treatment with osimertinib, a newer brain-penetrant TKI.
Methods:
We used MSK-CHORD, a clinicogenomic database that applies natural language processing (NLP) to extract treatment and metastasis data from electronic health records. Among 7809 NSCLC cases, we studied 647 patients with EGFR mutations without brain metastasis at diagnosis, treated with either early-generation TKIs or osimertinib. Kaplan-Meier analysis and Cox proportional hazards models were used to evaluate brain-metastasis-free survival (BMFS) with hazard ratios (HR) and 95% CIs.
Results:
Overall, BMFS was shorter for patients with L858R versus Ex19Del mutations (51.4% vs. 59.7% at 5 y, HR: 1.38 [1.02-1.88], P=0.036). For patients receiving early-generation TKIs (n=249, 38.5%), BMFS was shorter for those with L858R versus Ex19Del mutations (46.2% vs. 64.5% at 5 y, HR: 1.91 [1.20-3.06], P=0.007). For patients on osimertinib (n=398, 61.5%), no significant association was observed between mutation subtype and BMFS (53.4% vs. 53.1%, HR: 1.13 [0.76-1.68], P=0.561).
Conclusions:
Differences in brain metastasis development between EGFR L858R and Ex19Del mutations appeared to be mitigated among patients undergoing osimertinib compared with early-generation TKIs. NLP models can transform complex medical records into data for studying disease dynamics, representing an important systems bioengineering approach.
