Radiation within 30 Days of Death: TEACHH Score and Ability to Complete Treatment

Abstract

Up to 8% of cancer patients will receive radiotherapy (RT) within their last 30 days of life, and a significant proportion of these individuals will spend between 30% and 50% of their final days undergoing RT, often with palliative intent. Life expectancy prediction tools such as the TEACHH model (Krishnan et al., Cancer 2013) may allow physicians to better tailor palliative treatment courses to prioritize quality at the end of life for advanced cancer patients. All patients treated with external beam RT to a site of metastatic disease between 2012 and 2016 at our institution were identified, and those who received RT within 30 days of death were included in this analysis. Mann-Whitney U and chi-squared tests were used to compare TEACHH scores and other potential predictors of life expectancy such as primary cancer and sites of metastatic disease in patients who did or did not complete their final RT course. One hundred thirty-four patients underwent RT within 30 days of death, corresponding to 13.4% of all patients who received radiation for metastatic disease. Median age at the time of final RT course was 63 years (interquartile range (IQR) 55-72 years) and median duration since diagnosis was 14 months (IQR 3-37 months). Median duration from the last radiation fraction to death was 14 days (IQR 8-21 days). The most common primary sites were lung (31%) and breast (13%). RT intent was palliation for bone metastases in 49% and for brain metastases in 40% of treated patients. Fractionation pattern was one, five, and ten fractions in 16%, 35%, and 31% of RT plans, respectively. Median prescribed treatment duration was significantly shorter for bone metastases than brain metastases (5 vs. 10 fractions, P < 0.001). Patients treated for bone metastases were more likely to complete treatment than patients treated for brain metastases although this was not statistically significant (24% vs. 38%, P = 0.16). Thirty percent of patients did not complete their prescribed RT course. Patients who did not complete their final treatment course were prescribed more fractions (median 9 vs. 5 fractions, P < 0.001). Average TEACHH scores were higher in patients who did not complete their final radiation course (4.1 vs. 3.7, P = 0.016). While median KPS scores were lower in patients who did not complete their radiation course (50 vs. 60, P < 0.001), no individual component of the TEACHH model was significantly associated with failure to complete treatment. Appropriate use of palliative RT at the end of life remains a daunting task. The fact that patients who did not complete RT were prescribed longer courses suggests our ability to estimate life expectancy is limited and that we may underuse shorter treatment courses. Use of a tool that encompasses multiple variables, such as the TEACHH model, may enable physicians to better prescribe treatment at the end of life. Further research is needed to assess the role of shorter treatment courses for patients with brain metastases and prospectively apply the TEACHH model in the palliative setting.