The TEACHH model to predict life expectancy in patients presenting for palliative spine radiotherapy: external validation and comparison with alternate models.

INTRODUCTION: The TEACHH score was developed to identify patients with predicted short (< 3 months) and long (> 1 year) life expectancy. We aimed to validate this model in an independent group of patients presenting for palliative spine radiotherapy and to compare it to alternate prognostic models. METHODS: We retrospectively reviewed charts of 195 consecutive patients referred for palliative spine radiotherapy. Patients were grouped according to the number of risk factors from the TEACHH model, Chow model, and Oswestry Risk Index. RESULTS: One hundred and eighty patients with a median age of 65 years were included. Follow-up was 5.8 months in all patients and 31.8 months in living patients. For the TEACHH model, patients in groups 1, 2, and 3 had a median (95% CI) overall survival (OS) of 22.3 (15.7-36.1), 4.9 (3.8-6.6), and 1.5 (0.8-5.4) months, respectively. Wilcoxon pairwise comparisons showed statistically different survival between groups 1 and 2, and 1 and 3. In the Chow model, patients in groups 1, 2, and 3 had a median (95% CI) OS of 16.1 (10.0-22.3), 5.9 (3.8-9.2), and 1.9 (1.2-2.5) months, respectively. There was a significant difference between all groups. The Oswestry Risk Index identified five prognostic groups with median OS (95% CI) ranging from 22.2 (12.9-30.2) to 2.1 (0.8-4.0) months. Only group 1 was statistically different from the others. Although the effect of age was small, the TEACHH model performed best with the inclusion of all parameters. CONCLUSIONS: The TEACHH model is useful to identify patients with spinal metastases with predicted short, intermediate, and long LE. Its prognostic ability is similar to the Chow model.