1361P - The Stereotactic Radiosurgery-Brain Prognostic Index (SRS-BPI): A novel prognostic index for lung cancer patients with brain metastases eligible for SRS

Background

The existing prognostic scores for lung cancer patients with brain metastases (BM) have certain limitations and have not been solely designed for lung cancer patients eligible for central nervous system (CNS) stereotactic radiosurgery (SRS). This study aims to develop a novel prognostic index for this population.

Methods

This retrospective cohort study originated from a database of 673 cancer patients with BM. Eventually, the study population comprised 431 patients who received SRS and were treated for lung cancer at Karolinska University Hospital, Stockholm, Sweden from 2009 to 2020 (all comers from the Stockholm region). Demographics and clinicopathological parameters were extracted from electronic medical records (EMRs). We collected all data necessary in order to calculate the Score Index for Radiosurgery (SIR), the Graded Prognostic Assessment (GPA), and the Basic Score for BM (BSBM). Predictors of overall survival (OS) were identified and incorporated into our novel prognostic score using a penalised Cox regression model. The internal validation of our results was performed using the Efron-Gong optimism bootstrap analysis. The predictive performances of the aforementioned prognostic indices were compared to the performance of the newly introduced prognostic score.

Results

A new prognostic index, termed SRS-Brain Prognostic Index (SRS-BPI), was developed, which included nine variables: age, Eastern Cooperative Oncology Group Performance Status (ECOG PS), histology, molecular status, BM volume, neurological symptoms, BM at diagnosis, extracranial metastases, and extracranial disease control. The in-sample C-index for 1-year survival was 0.689. The optimism-corrected C-index was 0.67 indicating minor overfitting. A nearly optimal calibration was achieved with a low mean absolute error for 1-year OS of 0.04. The SRS-BPI provided a continuous risk prediction and demonstrated superior discrimination compared to GPA, SIR, and BSBM.

Conclusions

We suggest the SRS-BPI score as a new prognostic tool for enhanced decision-making for lung cancer patients with BM eligible for CNS SRS.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

The authors.

Funding

1) European Society for Medical Oncology (ESMO), 2) Hellenic Society of Medical Oncology (HeSMO) and 3) Region Stockholm.

Disclosure

All authors have declared no conflicts of interest.