ESTRO 2021 Abstract Book

S566

ESTRO 2021

Conclusion The proposed method utilized radiomic features to train a model for brain metastases survival estimation. Results showed that the model produces accurate and meaningful survival predictions. Our findings suggest that brain metastases radiomic features might be applicable for predicting survival in patients with brain metastases. PD-0736 Interval between MRI and SRS is associated with local control in lung squamous cell brain metastasis J.W. Ho 1 , L. Aznar-Garcia 2 , J. Christian 1 , P. Patel 3 1 Nottingham University Hospitals, Oncology and Radiotherapy, Nottingham, United Kingdom; 2 Nottingham University Hospitals , Oncology and Radiotherapy, Nottingham, United Kingdom; 3 Nottingham University Hospitals, Oncology and Radiotherapy , Nottingham, United Kingdom Purpose or Objective Limited brain metastases (BM) can be treated with stereotactic radiosurgery (SRS) with the aim for long term local control and improvement in survival. MRI head is often acquired for SRS planning; planning often takes several days or weeks to complete resulting in an interval between MRI and SRS treatment. Nicholl et al reported that 60% of BM planned for SRS underwent at least 3mm interval increase during the MRI-SRS interval. Interval increase was significantly associated with higher distant intracranial recurrence. On the contrary we did not observe any difference in neurological progression or survival stratified by MRI-SRS interval of within and beyond the median of 21-days. The purpose of this study is to investigate the effect of MRI-SRS interval on time to local progression (LP) of SRS-treated BM. We hypothesize that large volume BM and certain primary cancer may be at greater risk of LP with increasing MRI-SRS interval Materials and Methods We analysed patients treated at our centre between July 2017 and December 2020. Analysis was performed with SPSS Version 26. Survival was calculated by Kaplan-Meier method and hazard of event was calculated by Cox regression. Results A total of 258 patients with 432 SRS-treated BM were included in this study. Median MRI-SRS interval was 26 days (range 7 – 47 days). As a continuous variable, MRI-SRS interval was not significantly associated with LP (p = 0.513). Time to LP was not significantly different between below median (<26 days) and above median (>26 days) MRI-SRS interval (541 days vs 444 days, p = 0.482). BM volume was grouped by according to volume (<1ml, 1-5ml etc). MRI-SRS interval was not associated with time to LP when analysed by BM volume group. When analysed by primary cancer, only colorectal cancer was associated with higher risk of LP (HR = 2.587, 95% CI 1.185 – 5.646, p = 0.017). However, in colorectal cancer, there was no difference in LP time between MRI-SRS interval <26 days vs >26 days (p = 0.331). MRI-SRS interval was not associated with time to LP in any primary cancer except for lung squamous cell carcinoma (SqC). For lung SqC, MRI-SRS interval <26 days was significantly associated with longer time to LP compared to >26 days (904 days, 95% CI 321 – 1486 vs 94 days, 95% CI 73 – 115 days, p = 0.06), while a 1-day increase in MRI-SRS interval was significantly associated with 13.9% increased risk in LP (HR 1.139, 95% CI 1.020 – 1.272, p = 0.021) Conclusion Our findings suggest that minimising MRI-SRS interval might be critical in achieving durable local control for lung SqC. Our study is limited by the small case numbers nonetheless our study has generated hypothesis for further research. We plan to re-analyse in the future with a larger case number and to adjust for other parameters. MRI-SRS interval does not appear to be associated with time to local progression among other primary cancer

PD-0737 Diagnosis-Specific Graded Prognostic Assessment in Patients Receiving Stereotactic Irradiation