# Validation of the Lung-Mol Graded Prognostic Assessment (GPA) System for the Prognosis of Patients Receiving Radiotherapy for Brain Metastasis From Non-small Cell Lung Cancer

## TL;DR

This study validates a scoring system to predict survival in lung cancer patients with brain metastases undergoing radiotherapy.

## Contribution

The Lung-mol GPA system is validated for predicting survival in NSCLC patients with brain metastases undergoing radiotherapy.

## Key findings

- The Lung-mol GPA system showed significant association with survival in NSCLC patients with brain metastases.
- The C-indices for survival prediction were 0.65 for adenocarcinoma and 0.69 for non-adenocarcinoma in the first radiotherapy course.
- The system may be applicable to predict median survival in some patients with adenocarcinoma in the second radiotherapy course.

## Abstract

Purpose: The Lung-mol graded prognostic assessment (GPA) system predicts the prognosis of patients with brain metastases (BM) from non-small cell lung cancer (NSCLC) separately for adenocarcinoma and non-adenocarcinoma. This study aimed to validate the Lung-molGPA system using a cohort of patients in our institution who received radiotherapy for BM.

Materials and methods: Three hundred and thirty-nine patients with NSCLC who received their first course of radiotherapy for BM were included in the analysis. Among them, 65 received their second course of radiotherapy for BM. Data on sex, age, Karnofsky performance status (KPS), extracranial metastases (ECM), number of BM, histological type, and gene mutations were collected according to the Lung-molGPA system. We examined the validity of the scores assigned to the factors included in the Lung-molGPA system, separately for adenocarcinoma and non-adenocarcinoma. In addition, we validated the Lung-molGPA system to predict survival during both the first and second courses of radiotherapy.

Results: The factors in the Lung-molGPA were significantly associated with survival, except for age in non-adenocarcinoma with marginal significance. Regarding discrimination ability, the C-indices were 0.65 and 0.69 for adenocarcinoma and non-adenocarcinoma, respectively, in the first course of radiotherapy for BM, while those in the second course were 0.62 and 0.74, respectively. Survival prediction by Lung-molGPA was almost consistent with actual survival in the first course of radiotherapy, except for the score of 0-1.0 in both histologies and 2.5-3.0 in non-adenocarcinoma. In the second course of radiotherapy, median survival could be predicted for some patients with adenocarcinoma.

Conclusions: Our study confirms the validity of Lung-molGPA for the estimation of median survival based on patient characteristics at the time of initiation of radiotherapy for patients in the first course of radiotherapy and shows that it may be applicable to patients with adenocarcinoma in the second course of radiotherapy.