# Evaluation of the updated 2022 lung-GPA in NSCLC adenocarcinoma patients with brain metastases: analysis of prognostic factors in a German clinical cohort

## TL;DR

This study validated the 2022 NSCLC-GPA in German patients with lung adenocarcinoma and brain metastases, finding it effective for predicting survival.

## Contribution

The study confirms the updated NSCLC-GPA's applicability in a German cohort and identifies independent prognostic factors like low KPS and ECM.

## Key findings

- Lower GPA scores correlated with significantly worse overall survival in patients with brain metastases.
- Low KPS and extracranial metastases were independently significant prognostic factors in multivariate analysis.
- PD-L1 expression did not show a significant association with overall survival in this cohort.

## Abstract

Lung cancer is the most frequent source of brain metastases (BMs), with 20-40% of patients with non-small cell lung cancer (NSCLC) developing BMs during the course of disease. The NSCLC-GPA (Graded Prognostic Assessment), developed by Sperduto et al., aims to estimate overall survival (OS) in lung cancer patients with BMs based on multiple prognostic factors. This study aimed to validate the applicability and prognostic relevance of the updated 2022 NSCLC-GPA in a German clinical cohort, with particular attention to programmed death-ligand 1 (PD-L1) expression and treatment-related outcomes.

We retrospectively analyzed patients with NSCLC and BMs treated between 2020 and 2022 at a German university hospital. GPA scores were calculated using established parameters: age, Karnofsky Performance Status (KPS), number of BMs, presence of extracranial metastases (ECM), EGFR and ALK mutation status, and PD-L1 expression. Due to the low number of non-adenocarcinoma (NAC) cases in our cohort, only patients with adenocarcinoma (AC) were included. Univariate and multivariate Cox regression models, log-rank tests, and Kaplan–Meier curves for illustrative purposes were used to evaluate associations between prognostic factors and OS.

A total of 110 AC patients met the inclusion criteria. Median OS was 10 months (range: 6-11). Patients with GPA scores of 0-1 and 1.5-2 had significantly worse outcomes compared with the reference group (GPA 3.5-4; median OS: 36 months), with median OS times of 3 and 8 months, respectively (HR: 8.34, 95% CI: 2.53-27.5, p = 0.0005 and HR: 5.24, 95% CI: 1.584-17.33, p = 0.0067). Patients with a GPA of 2.5-3 had a median OS of 22.5 months (HR: 1.6, 95% CI: 0.469-5.45, p = 0.453), which was not statistically significant. Older age (≥ 70 years; HR: 1.95, 95% CI: 1.23-3.11, p = 0.0047), low KPS (KPS ≤ 70; HR: 3.88, 95% CI: 2.33-6.46, p < 0.0001), a higher number of BMs (HR: 1.58, 95% CI: 1.03-2.41, p = 0.035), and the presence of ECM (HR: 2.62, 95% CI: 1.66-4.13, p < 0.0001) were all significantly associated with decreased OS. In the multivariate analysis, low KPS and ECM remained independently significant prognostic factors. PD-L1 expression showed no significant association with OS.

Despite shorter OS outcomes in our cohort, the 2022 NSCLC-GPA proved to be a valuable prognostic tool, with lower scores consistently associated with poorer outcomes. In addition, older age, low KPS, ECM, and a higher number of BMs were identified as prognostic factors, with low KPS and ECM remaining independent prognostic factors in the multivariate analysis.

The online version contains supplementary material available at 10.1186/s13014-026-02805-0.