Genetic biomarkers associated with pain flare and dexamethasone response following palliative radiotherapy in patients with painful bone metastases.

ABSTRACT

BACKGROUND: In patients who receive palliative radiation therapy (RT) for painful bone metastases, 40% experience a transient increase in pain known as a pain flare. Prophylactic dexamethasone has been shown to reduce pain flare incidence to 25%. We aimed to identify DNA biomarkers associated with pain flare and dexamethasone response. METHODS: Daily pain levels were recorded by 81 patients who received a single 8 Gy RT for painful bone metastases, of which 50 also received prophylactic dexamethasone. To identify single-nucleotide variants (SNVs), patient saliva samples obtained at day of RT were sequenced for 4,813 disease-associated genes, then filtered for genes associated with inflammation, radiation or immune response, and DNA damage. Significant SNVs (P<0.005) identified by the Cochran-Armitage trend test underwent the Penalized LASSO method with minimum Bayesian Information Criterion to select a multi-SNV model that jointly predicted pain flare, and pain flare despite prophylactic dexamethasone (dexamethasone response). The corresponding estimated effects of the multi-SNVs were used to drive the prognostic score of developing pain flare for each patient, who were divided into three risk groups of roughly equal sizes. RESULTS: Risk groups were significantly predictive of pain flare (P<0.0001) and dexamethasone response (P<0.0001). The high-risk patient groups had a 78% chance of developing pain flare, and pain flare despite dexamethasone [OR =24.6, 95% confidence interval (CI) 1.8-342.7, P=0.02]. The multivariable model for pain flare included 15 variants, with effect sizes ranging from -4.97 (NBPF1 rs3872309 C>T) to 5.54 (DNM2 10940838 A>C). The multivariable model for dexamethasone response included 6 variants, with effect sizes ranging from -1.03 (NBPF1 rs3872309 C>T) to 0.85 (TSEN54 rs62088470 C>G). CONCLUSIONS: Significant SNVs associated with pain flare were found in genes with functions in biosynthesis (DHODH, PECR), lipid excretion and metabolism (UGT2A1/2, VLDLR), and intracellular signalling (DNM2, SEC23A). Significant SNVs associated with dexamethasone response were from genes involved in extracellular matrix (HAS1, ADAMTS16) and cytoskeleton regulation (GAS2L2). Identification of SNVs predictive of pain flare and dexamethasone response enables targeted prophylactic therapy according to a patient's predisposed response.