Abstract Book

S247

ESTRO 37

to report HRQoL during the first 3 months after 2 different hypofractionated radiotherapy schedules. Material and Methods Patients were randomly assigned (1:1) to arm A (56 Gy in 16 fractions) or arm B (67 Gy in 25 fractions). Randomization was by computer-generated permuted blocks, stratified on prior transurethral resection of the prostate and presence of a dominant intraprostatic lesion. The primary outcome of this trial is acute toxicity. An interim analysis was planned on 160 patients to prove safety of both treatment regimens. Here we report HRQoL outcomes, as secondary endpoint within the main trial. HRQoL was assessed with the EORTC questionnaires QLQ-C30 and PR25 at baseline, end of treatment, after 1 and 3 months. Analyses of the longitudinal data were performed with linear mixed-effects models with an adjusted significance level of 0.004. Results So far 160 patients (77 in arm A vs. 82 in arm B) were prospectively enrolled between 26/6/2013 and 18/6/2016. Compliance rate for filling in the questionnaires was >90% at every time-point. No significant differences in change of HRQoL during the first 3 months after radiotherapy were observed between arm A and B. However, in both arms a significant worsening of urinary and bowel symptoms, diarrhea, sexual activity and functioning from baseline to end of treatment was observed. All patients had significant improvements of urinary symptoms and diarrhea after 3 months and 1 month, respectively. Bowel symptoms improved after 1 month in arm A and after 3 months in arm B. Sexual activity and functioning showed no improvements from end of treatment to month 3 and hormone-related treatment effects showed significant worsening from baseline to end of treatment, month 1 and month 3, irrespective of treatment arm. Additionally, arm A showed significant worsening of dyspnea from end of treatment to month 3. Arm B showed a significant increase in fatigue from baseline to end of treatment and also an improvement in emotional functioning from baseline to month 3. Clinically relevant changes over time in HRQoL, defined as changes of ≥10 points, are presented in figure 1.

Symposium: Radiogenomics

SP-0481 Radiogenomics - knowing the landscape to be explored C.N. Andreassen 1 , L.M.H. Schack 1 , J. Alsner 1 1 Aarhus University Hospital, Department of Experimental Clinical Oncology, Aarhus C, Denmark Abstract text Since the turn of the millennium, more than 150 original papers have addressed possible associations between genetic sequence alterations and risk of normal tissue toxicity after radiotherapy. The vast majority of these studies have investigated SNPs using a candidate approach. Only 7 studies were genome-wide association studies (GWASs). Despite these efforts, the number of compelling associations is still very limited. Approximately half of these associations were established using the candidate gene approach and the other half using GWASs. The SNPs identified by GWASs were typically not related to genes involved in radiobiology in a narrow sense. Furthermore, they were usually associated with specific types of normal tissue toxicity rather than radiosensitivity in general terms. As of November 2017, more than 3200 GWASs have provided in excess of 52,000 unique SNP-trait associations for various bio-medical phenotypes (https://www.ebi.ac.uk/gwas/home). Some of these studies included more than 100,000 subjects. These studies provide important insights into complex trait genetics. The studies showed that the typical effect size of SNPs is usually small with relative risks well below 1.2. Furthermore, the identified SNPs were often not located in genes anticipated to be of major importance for the trait. In fact, almost 90% of the SNPs were not located in genes but in non-coding sequence with regulatory or unknown functions. For some of the traits (e.g. height, breast and prostate cancer susceptibility) more than 100 trait-associated SNPs have been identified. Despite this, the identified SNPs usually only explained a limited proportion of the expected heritability for the trait. This raises the question as to where the ‘missing heritability’ is hiding. It seems likely that some of the missing heritability is accounted for by other types of sequence alterations than SNPs (i.e. various rare sequence alterations) or by SNPs with effect sizes below the detection threshold of the current GWASs. In order to adjust for multiple testing, a significance threshold of 5 x 10 -8 is often used in GWASs. It takes very large study cohorts to maintain statistical power under such stringent conditions. Complete sequencing will usually be required to identify rare sequence alterations with potential impact on phenotype. The cost of sequencing is continuously falling but genome-wide sequencing is still approximately 10 times more expensive than assessment of the genome-wide SNP genotype. Furthermore, certain statistical challenges relate to genome-wide sequencing. First, the number of rare alterations in the genome is likely to exceed the number of SNPs. Second, statistical power dramatically drops with decreasing population frequency of the sequence alteration. Thus, very large study cohorts will be needed in order to identify a substantial number of sequence alterations using unconstrained genome-wide approaches. Large cohorts of patients with detailed recordings of treatment parameters and normal tissue outcome after radiotherapy are relatively rare. Therefore, the above- mentioned factors pose particular challenges in the field of ‘radiogenomics’. Consequently, alternatives to unconstraint genome-wide approaches should be considered. Opposed to many other traits, normal tissue radio-

Figure 1: clinically relavant changes from baseline in HRQoL(*) Conclusion Although no significant difference in HRQoL was found between both hypofractionated radiotherapy regimens (56 Gy in 16 fractions or 67 Gy in 25 fractions), treatment-related side effects during RT occur frequently, hampering their HRQoL. However, most side effects improved during the first 3 months after RT. Funding A grant was provided by Stichting tegen kanker, a non- profit organization.