OBJECTIVES: To evaluate the prognostic relevance of temporal muscle thickness (TMT) in brain metastasis patients. METHODS: We retrospectively analysed TMT on magnetic resonance (MR) images at diagnosis of brain metastasis in two independent cohorts of 188 breast cancer (BC) and 247 non-small cell lung cancer (NSCLC) patients (overall: 435 patients). RESULTS: Survival analysis using a Cox regression model showed a reduced risk of death by 19% with every additional millimetre of baseline TMT in the BC cohort and by 24% in the NSCLC cohort. Multivariate analysis included TMT and diagnosis-specific graded prognostic assessment (DS-GPA) as covariates in the BC cohort (TMT: HR 0.791/CI [0.703-0.889]/p < 0.001; DS-GPA: HR 1.433/CI [1.160-1.771]/p = 0.001), and TMT, gender and DS-GPA in the NSCLC cohort (TMT: HR 0.710/CI [0.646-0.780]/p < 0.001; gender: HR 0.516/CI [0.387-0.687]/p < 0.001; DS-GPA: HR 1.205/CI [1.018-1.426]/p = 0.030). CONCLUSION: TMT is easily and reproducibly assessable on routine MR images and is an independent predictor of survival in patients with newly diagnosed brain metastasis from BC and NSCLC. TMT may help to better define frail patient populations and thus facilitate patient selection for therapeutic measures or clinical trials. Further prospective studies are needed to correlate TMT with other clinical frailty parameters of patients. KEY POINTS: • TMT has an independent prognostic relevance in brain metastasis patients. • It is an easily and reproducibly parameter assessable on routine cranial MRI. • This parameter may aid in patient selection and stratification in clinical trials. • TMT may serve as surrogate marker for sarcopenia.