# Stereotactic radiosurgery for brain metastases: multistate and competing risk models of progression and survival from a single-centre 11-year experience

## TL;DR

This study examines the long-term outcomes of stereotactic radiosurgery for brain metastases, identifying factors that influence progression and survival.

## Contribution

The study introduces multistate and competing risk models to analyze progression and survival in brain metastases treated with SRS.

## Key findings

- Breast cancer patients had the longest time without disease progression, while gastrointestinal cancers had the shortest.
- Each additional metastatic lesion increased progression risk, but not death risk, while larger tumor volume increased death risk.
- New metastases after SRS increased death risk, but local recurrence did not.

## Abstract

The diagnosis of brain metastases has risen due to advances in systemic treatments prolonging survival and the use of MRI. Whilst surgical resection remains a valuable mean of acutely alleviating raised intracranial pressure, reducing seizure risk, and preventing progression of neurological deficit, stereotactic radiosurgery (SRS) delivers radiotherapy to precise target volumes, including the treatment of a larger number of metastases with high rates of local control, whilst avoiding the neurological complications of WBRT or the need for invasive resection. The aim of this study was to evaluate the 11-year experience of SRS at the Leeds Gamma Knife Centre, examining patient outcomes, and the association of patient-dependent factors and intracranial disease burden to survival.

A retrospective cohort study of 1,031 patients (2,836 metastases) treated with SRS at the Leeds Gamma Knife Centre between 2010 and 2020. Data included primary tumour, Karnofsky performance status, intracranial disease burden, and prior/subsequent treatment. Follow-up was at least 30 months. Kaplan-Meier analysis was used for 10-year survival curves. Multistate and competing risks models assessed progression/death predictors.

Lung (50.4%), breast (15.4%) and melanoma (9.9%) were the most common primary tumour. 364 (35.3%) patients had recurrence, and 875 (84.5%) died by the end of follow-up. Survival outcomes were worse in men compared to women in both lung (p = 0.003) and other (p = 0.02) metastases. Breast cancer patients spent the most time in the no progression state (rmean = 1.56 years) whereas gastrointestinal primary spent the least (rmean = 0.99 years). Each additional metastatic lesion at the time of treatment increased the risk of progression by 9.2% (p < 0.001), but risk of death was not influenced by the number of lesions (p = 0.30); conversely, total tumour volume increased the risk of death by 2.2% for each 1cm3 increase. Post-SRS new metastatic disease increased the risk of death by 31% (p = 0.02), whilst local recurrence did not (p = 0.26).

These findings add to the accepted prognostic factors for brain metastases, further informing clinical consultations with patients, but further research is required to address unanswered questions regarding the role of SRS in relation to other treatments.