Prostate Cancer Patients with a High Polygenic Risk of Rheumatoid Arthritis have Increased Radiotherapy Toxicity

Abstract

<h3>Purpose/Objective(s)</h3> Overlapping genes are involved with rheumatoid arthritis (RA) and DNA repair pathways. Therefore, we hypothesized that patients with a high polygenic risk score (PRS) for RA will have an increased risk of radiotherapy (RT) toxicity. <h3>Materials/Methods</h3> Data from 1681 men with prostate cancer were available from the multicenter perspective REQUITE study. Germline DNA was genotyped using a sequencing platform and imputed using the 1000 Genomes Project Phase 3. PRS and weighted-PRS (wPRS) were calculated for RA (101 SNPs) and analyzed as continuous variables and using a >95^th percentile cut-off. Univariable and multivariable linear regression models analyzed relationships with: 1) acute and late standardized total averaged toxicity (STAT) scores; and 2) individual toxicity end points. Acute toxicity was defined as max reported toxicity within 90 days of RT (14 rectal, bladder and bowel endpoints) and late toxicity defined as max reported after 90 days for 5 endpoints. Multivariable analyses included: age, diabetes, previous surgery, hormone therapy and prescribed dose (converted to biologically equivalent dose). Residuals from multivariable linear regression models for acute and late STAT were calculated (rSTAT). Individual SNP gene doses were analyzed using linear regression against rSTAT to identify individual SNPs associated with worse reported toxicity. Bonferroni correction accounted for multiple testing with p<0.0005 considered significant. <h3>Results</h3> Men with >95^th percentile PRS had a significantly worse STAT_acute (p=0.02 in multivariable analysis). Age, previous surgery and hormones were also associated with worse STAT_acute. For individual acute toxicity end points, hematuria (p=0.05) and dysuria (p=0.003) were higher in patients with >95^th percentile PRS. PRS (p=0.02) and wPRS (p=0.03) were associated with worse STAT_late on multivariable analysis. Patients with >95^th percentile PRS also had worse STAT_late on multivariable analysis (p=0.008). Age, previous surgery, hormones and dose were also associated with worse late STAT. For individual late toxicity endpoints, hematuria (p=0.01) and urinary frequency (p=0.006) showed a significant association with wPRS. Analysis of individual SNP gene doses against rSTAT_acute identified an association with rs138193887 (p=0.0005) located in <i>ATM</i>, a gene linked with radiosensitivity. <h3>Conclusion</h3> Men with prostate cancer and a high PRS for RA have an increased risk of acute and late toxicities after radiotherapy. Analysis of individual toxicity endpoints have identified bladder toxicities as most strongly associated with the RA PRS. Results will be validated in further cohorts, but suggest an actionable target for further bladder sparing for these men.