Back

Brain Metastasis Screening Strategies in relation to Neurological Symptoms in Stage IV Lung Adenocarcinoma

Sayin, S. I.; Eklund, E. A.; Beischer, M.; Moe, T.; Ali, K. X.; Gunnarsson, K.; Xylander, M.; Ny, L.; Jakola, A. S.; Häggström, I.; Wiel, C.; Hallqvist, A.; Sayin, V. I.

2025-03-12 oncology
10.1101/2025.03.09.25323599 medRxiv
Show abstract

BackgroundBrain metastases (BM) are a major clinical challenge in metastatic lung adenocarcinoma (LUAD), affecting up to 50% of patients during disease progression. Current guidelines do not mandate brain imaging for all metastatic lung cancer patients at diagnosis unless there are neurological symptoms present. However, real-world data on the predictive value of neurological symptoms for BM detection remain scarce. MethodsThis retrospective multicenter study analyzed all consecutive patients diagnosed with stage IV LUAD with molecular assessment in western Sweden from 2016-2021 (n = 912). We extracted data from patient charts, imaging referrals, radiology reports and the Swedish National Lung Cancer Registry to determine diagnostic brain imaging (DBI) frequency and modality, presence of neurological symptoms, BM detection rates, size, number, location and overall survival (OS). ResultsAmong stage IV LUAD patients, 63% underwent DBI, and BM was detected in 23% of all patients (37% of those receiving DBI). Neurological symptoms prompted DBI in 63% of cases, yet 58% of these symptomatic patients had no BM on imaging. Conversely, 28% of asymptomatic patients who underwent DBI had BM. Patients with BM detected in the absence of neurological symptoms had smaller metastases. Neurological symptoms were associated with worse OS, irrespective of the presence of BM. ConclusionNeurological symptoms poorly predict BM in metastatic LUAD. Routine MRI-based brain imaging at diagnosis of metastatic disease, irrespective of neurological symptoms, may improve outcomes for this underserved patient population. These results provide real-world evidence supporting the need for reassessment of current BM screening recommendations.

Matching journals

The top 7 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.