2537 - External Validation of a Novel Prognostic Score for Patients with Brain Metastases Referred for Stereotactic Radiosurgery

D. Johnston¹, J. Tee¹, B. Ó. Kearney¹, O. McLaughlin², G. M. Walls^1,2, G. Calvert³, T. Flannery⁴, J. Harney¹, D. Conkey¹, and C. McGarry^1,2; ¹Northern Ireland Cancer Centre, Belfast Health & Social Care Trust, Belfast, United Kingdom, ²Patrick G. Johnston Centre for Cancer Research, Queens University Belfast, Belfast, United Kingdom, ³Queens University Belfast, Belfast, United Kingdom, ⁴Department of Neurosurgery, Royal Victoria Hospital, Belfast Health & Social Care Trust, Belfast, United Kingdom

Purpose/Objective(s): Stereotactic radiosurgery (SRS) has been adopted as a critical component of brain metastasis clinical management. Appropriate patient selection can prove challenging, especially in the era of targeted and immunotherapy. Factors commonly considered include the volume of metastatic disease and the anticipated survival, however prognostication is often difficult. Herein, SRS outcomes at a tertiary referral centre were analysed and sought to externally validate a novel prognostic tool, the Royal Marsden Hospital SRS Survival Score (RMH–SSS, PMID:38368229). Materials/

Methods: Patients treated with brain SRS at a regional referral centre between 2016 and 2023 were identified. Records were interrogated for details regarding demographics, primary tumour and systemic therapy, and SRS treatment planning. Median local control (mLC) and median overall survival (mOS) intervals were recorded as the time between SRS treatment and date of event (local relapse & death respectively). The RMH–SSS was calculated, based on age, performance status, number & volume of brain metastases, presence of extracranial disease, primary tumour histology and molecular subtype. Benchmarking against the primary RMH–SSS publication was undertaken using the 4 recommended score categories: 0–2, 3, 4–5 and 6–11. Kaplan-Meier time-to event analyses were undertaken according to prospectively defined clinically relevant strata, including gross tumour volume (GTV) and groups were compared using log-rank or log-rank for trend tests (R Studio).
Results: Outcomes for 225 patients with 397 metastases were analysed. 308/397 (78%) of metastases were treated with single fraction SRS and all treatments were delivered by frameless LINAC-based SRS. Eight patients had radiological evidence of radionecrosis identified on MRI imaging. mLC (all treated metastases) was 18.4 months and mOS 13.4 months. GTV was associated with mOS (29 months vs 9 months for <1.5 cm³ vs >1.5 cm³ (HR 0.52, 95%CI 0.36–0.76, p<0.001) and also primary histology (melanoma 36 months, breast 19 months, lung 10 months, other grouped primaries 10 months, p=0.0007). Higher RMH–SSS was associated with improved mLC, p=0.0002 (6–11 vs 0–2, HR 0.35, 95%CI 0.21–0.57), and mOS (score 6–11 vs 0–2, HR 0.23, 95%CI 0.14–0.37, p<0.0001). mOS values also were comparable with the primary RMH–SSS data (Table 1).
Conclusion: The RMH–SSS had prognostic value in this external, representative real world cohort comprising contemporary SRS treatment and therefore may inform patient selection for brain SRS in the clinic. Table 1: RMH-SSS of presented cohort versus the primary publication.

Survival Score	Presented Cohort		Primary Publication
	n	Median OS [95%CI]	n	Median OS [95%CI]
0-2	66	5.9 [4.2–8.6]	319	8.6 [7.5–11.0]
3	24	11.5 [5.8–NR]	267	12.8 [10.6–14.7]
4-5	79	15.2 [10.9–31.9]	304	20.7 [17.6–24.0]
6-11	56	47.2 [26.2–NR]	147	45.8 [33.3–NR]