The First Survival Score for Patients Aged ≥80 Years Irradiated for Brain Metastases.

Abstract

Simple SummaryDecisions regarding personalized treatment for brain metastases should consider the patient’s lifespan, which can be estimated with survival scores. Since very elderly patients (≥80 years) are different from other patients, e.g., due to a higher comorbidity index and decreased organ functions, these patients likely benefit from specific scores. A survival score was generated in 94 patients aged ≥ 80 years who were irradiated (whole-brain radiotherapy) for metastases of the brain. The score achieved high accuracy with respect to the prediction of death up to 6 months and survival for ≥1 and ≥2 months following treatment. When compared to an existing tool, the new score was more precise regarding death ≤ 1 month following radiotherapy and survival (all time points). Therefore, the new score appears preferable.Survival scores facilitate personalized cancer treatment. Due to demographic changes, very elderly patients are more prevalent than in the past. A score was developed in 94 patients aged ≥ 80 years undergoing whole-brain radiotherapy for brain metastases. Dose fractionation, treatment period, age, sex, performance score (ECOG-PS), tumor type, count of lesions, metastases outside the brain, and interval tumor diagnosis to radiotherapy were retrospectively evaluated. Independent predictors of survival were used for the score. Based on individual scoring points obtained from 3-month survival rates, prognostic groups were designed. Additionally, the score was compared to an existing tool developed in patients ≥ 65 years. ECOG-PS, count of lesions, and extra-cranial metastases were independent prognostic factors. Three groups were created (7, 10, and 13–16 points) with 3-month survival of 6%, 25%, and 67% (p < 0.001), respectively. Positive predictive values (PPVs) regarding death ≤ 3 and survival ≥ 3 months were 94% and 67% (new score) vs. 96% and 48% (existing tool), respectively. PPVs for survival ≥1 and ≥2 months were 88% and 79% vs. 63% and 58%, respectively. Both tools were accurate in predicting death ≤2, ≤3, and ≤6 months. The new score was more precise regarding death ≤1 month and survival (all time periods) and appeared preferable. However, it still needs to be validated.