Radiotherapy impact on health-related quality of life in localized prostate cancer: results of an observational, non-interventional, multicenter study in Spain

Background: Health-related quality of life (HRQoL) is greatly affected by prostate cancer (PCa) and associated treatments. This study aimed to measure the impact of radiotherapy on HRQoL and to further validate the Spanish version of the 16-item Expanded Prostate Cancer Index Composite (EPIC-16) in routine clinical practice. Methods: An observational, non-interventional, multicenter study was conducted in Spain with localized PCa patients initiating treatment with external beam radiotherapy (EBRT) or brachytherapy (BQT). Changes from baseline in EPIC-16, University of California-Los Angeles Prostate Cancer Index (UCLA-PCI), and patient-perceived health status were longitudinally assessed at end of radiotherapy (V2) and 90 days thereafter (V3). Psychometric evaluations of the Spanish EPIC-16 were conducted. Results: Of 516 patients enrolled, 495 were included in the analysis (EBRT, n = 361; BQT, n = 134). At baseline, mean (standard deviation [SD]) EPIC-16 global scores were 11.9 (7.5) and 10.3 (7.7) for EBRT and BQT patients, respectively; scores increased, i.e., HRQoL worsened, from baseline, by mean (SD) of 6.8 (7.6) at V2 and 2.4 (7.4) at V3 for EBRT and 4.2 (7.6) and 3.9 (8.2) for BQT patients. Changes in Spanish EPIC-16 domains correlated well with urinary, bowel, and sexual UCLA-PCI domains. EPIC-16 showed good internal consistency (Cronbach’s alpha = .84), reliability, and construct validity. Conclusion: EPIC-16 scores worsened after radiotherapy in different HRQoL domains, regardless of patient-perceived health status. UCLA-PCI scores mostly recovered for EBRT patients at V3, while scores for BQT patients did not. The Spanish EPIC-16 questionnaire demonstrated sensitivity, strong discriminative properties and reliability, and validity for use in clinical practice.


Background
In Spain, prostate cancer (PCa) is the most common leading type of cancer in terms of incidence and the third most common cause of cancer death in men, with age-standardized incidence and mortality rates of 104.2 and 13.2 per 100,000 men, respectively, in 2018 [1]. There are various treatment options for localized PCa depending on disease stage, including radiotherapy, such as external beam radiotherapy (EBRT) and brachytherapy (BQT) [2]. A recent study showed that for clinically localized PCa in Spain, the majority of patients analyzed (~ 84%) received treatment, with one-third undergoing radiotherapy; ~ 86% were treated with intensity-modulated radiation therapy or 3D radiotherapy and ~ 39% received BQT [3]. However, these therapeutic interventions negatively impact patients in terms of physical and emotional symptoms [4]. Therefore, patient-reported outcome (PRO) measures are increasingly being used to assess long-term health-related quality of life (HRQoL) [5]. A well-established instrument that is frequently used to assess a patient's post-intervention-related HRQoL is the 50-item Expanded Prostate Cancer Index Composite (EPIC) questionnaire, which was developed based on the University of California-Los Angeles Prostate Cancer Index (UCLA-PCI) [6,7]. There are two shorter versions of the EPIC-50 questionnaire, the 26-item (EPIC-26) [8] and the 16-item (EPIC-CP) [9] questionnaires; the latter was speci cally designed to be administered in routine clinical practice. EPIC-CP, a one-page questionnaire, measures urinary incontinence, urinary irritation, bowel, sexual, and hormonal HRQoL domains in patients with clinically localized PCa to evaluate aspects of therapy that are most bothersome [9]. The EPIC questionnaire has already been validated and translated in various languages, including Spanish for EPIC-50 [10], Italian for EPIC-26 [11] and German for the shortened EPIC-CP [12].
The primary objective of this study was to measure the impact of radiotherapy on HRQoL. The secondary objectives included further validation and assessment of the measuring properties of the Spanish version of the EPIC-CP questionnaire in routine clinical practice in Spain to enable less time-consuming evaluation of treatment-related HRQoL.

Study design and participant selection
This observational, non-interventional, multicenter study was conducted in Spain. Data were collected between January 2016 and September 2017 in 41 radiation oncology departments.
Male patients aged ≥ 40 years with histopathologically con rmed, localized PCa who were initiating EBRT or BQT treatment were included. Patients were required to complete follow-up visit questionnaires. Patients with prior prostatectomy, prior radiotherapy or BQT treatment, or with N1 or metastatic tumor stage were excluded.
The study was approved by the independent ethics committee of participating centers. Patients gave their voluntary informed consent to take part. The decision on treatment was not affected by this study and was solely based on the investigator's criteria.

Study assessments and data collection
Patients had three study visits where clinical variables were collected at baseline, i.e., before the beginning of treatment (EBRT or BQT) [visit 1], at the rst follow-up visit, i.e., the nal EBRT session or 1 month after the rst BQT session or seed implantation (visit 2), and a nal follow-up visit approximately 3 months after the end of treatment (visit 3).

Study questionnaires
Patients completed the following PRO measures at each study visit, with the investigator ensuring all sections were completed properly; the median application time for questionnaires used in this study was not measured. The 16-item EPIC-CP questionnaire, referred to as EPIC-16 in this manuscript, consists of ve PCa HRQoL domains and measures urinary, bowel, sexual, vitality, and hormonal health. These ve domains contain three questions each with a Likert numeric response scale (NRS) of 0-12 (best to worst HRQoL) for items 2-10, giving a total NRS of 0-60 (Table 1) [9]. The Spanish version of EPIC-50 by Ferrer, et al. was used to create EPIC-16 used in this study [10].
The UCLA-PCI is a 20-item questionnaire with six domains, with an NRS of 0-100 (worst to best HRQoL) and measures the function and degree of impairment in the urinary, intestinal, and sexual domains (Table  1) [7]. Additionally, patients completed the patient-perceived state of health measure, a one-item questionnaire in which patients evaluated their own general PCa-related health on that day as "Very good", "Fairly good", "Slightly good", "Neither good nor bad", "Slightly bad", "Quite bad", or "Very bad".

Study objectives
The primary objective was to measure the change in HRQoL between baseline and follow-up visits using EPIC-16 and UCLA-PCI. A secondary objective was to validate the Spanish version of the EPIC-16 questionnaire by analyzing its psychometric properties.

Psychometric validations of the Spanish version of the EPIC-16 questionnaire
The psychometric properties of the EPIC-16 questionnaire were evaluated for feasibility, reliability, construct, and longitudinal validity and sensitivity to change for the total number of patients included in the study at baseline versus the follow-up visits (see Additional le 1).

Statistical analyses
Descriptive statistics were used to summarize patient demographic characteristics, including age, disease stage, and radiotherapy treatment. Standard deviations of the mean values are reported in the gures as error bars either above or below the mean. The correlations between outcomes in EPIC-16 and UCLA-PCI domains were tested using Spearman rank correlation coe cients. Sensitivity to change was assessed using the Student t test for paired data. The questionnaire scores at baseline versus each follow-up visit and scores before and after the type of radiotherapy (EBRT or BQT) were compared using a statistical signi cance level of .05, which was used for all statistical tests performed.

Patient demographics and baseline characteristics
Overall, 516 patients were enrolled in this study at the baseline visit. Following the second visit, 21 (4.1%) patients were excluded. The remaining 495 patients (95.9%) eligible at visit 3 were included in the analyzed patient population (Fig. 1). A total of 361 patients (72.9%) received EBRT and 134 patients Baseline characteristics strati ed by EBRT or BQT treatment groups are summarized in Table 2. At baseline, median patient age (range) was 73.0 (48.0-84.0) years and 67.0 (48.0-82.0) years for the for EBRT and BQT groups, respectively. Most patients who received EBRT had a Gleason score of ≥ 7 (73.1%), while most patients who received BQT had a Gleason score of < 7 (84.3%). A higher proportion of patients received neo-adjuvant hormone therapy in the EBRT group (62.0%) than in the BQT group (6.0%) [ Table 2].
At baseline, the total EPIC-16 mean (standard deviation [SD]) scores were 11.9 (7.5) and 10.3 (7.7) for the EBRT and BQT groups, respectively (see Supplementary  Impact of radiotherapy on quality of life (QoL) EPIC-16 total scores increased by a mean (SD) of 6.8 (7.6) points at visit 2 and by 2.4 (7.4) points at visit 3 for patients in the EBRT group. For patients who received BQT, the scores increased by a mean (SD) of 4.2 (7.6) and 3.9 (8.2) points at visits 2 and 3, respectively.
Overall, scores increased after radiotherapy, indicating a worsened QoL, across all EPIC-16 domains (Fig.   2). At visit 3 particularly, patients recovered their baseline scores in the urinary incontinence, bowel, vitality/hormonal, and urinary irritation domains, with the exception of the bowel and urinary irritation domains in patients receiving BQT ( Fig. 2B and Fig. 2C). The sexual domain scores of patients in both groups worsened at both time points and did not recover at visit 3 compared to other domains. Patients who received concomitant hormone therapy, initiated prior to either EBRT or BQT, showed a higher mean score in the sexual domain between study visits 1 and 2 (mean [SD] visit 1: 6.79 [3.16] For the overall evolution of urinary problems, e.g., item 1 of EPIC-16 ("Overall, how much of a problem has your function been for you?"), responses were similar at baseline for both EBRT and BQT patients (see Supplementary Fig. 2, Additional le 1). However, more patients treated with EBRT increasingly perceived urinary function as a small-to-big problem at visit 2 (62.0%, n = 221) compared with patients treated with BQT (53.2%, n = 67). At visit 3, patients' perceptions returned to baseline, i.e., with urinary condition perceived as less of a problem compared with visit 2 (small-to-big problem: 35.5%, n = 126 for EBRT; 48.5%, n = 65 for BQT) [see Supplementary Fig. 2, Additional le 1].
Similarly, for patients who received EBRT, all UCLA-PCI scores showed a decrease at visit 2, indicating worsened QoL, especially in urinary bother, bowel function and bother, and sexual function domains (Fig.  3). At visit 3, scores had recovered QoL in the urinary function, urinary bother, and bowel function domains. For patients who received BQT, UCLA-PCI scores were decreased at visit 2 in the urinary bother, bowel bother, and sexual function domains, none of which recovered at visit 3 (Fig. 3).

Feasibility of EPIC-16
The oor effect, i.e., worst score of 12 for each domain, was present in < 1% of patients, with the exception of the sexual domain for which 24 (4.9%) patients had a maximum score of 12 at baseline, which increased to 44 (9.1%) and 56 (11.5%) patients at visits 2 and 3, respectively (see Supplementary  Fig. 3, Additional le 1).
The ceiling effect, i.e., best score of 0 for each domain, was reached by a high percentage of men in urinary incontinence (n = 364; 65.6%), bowel (n = 345; 70.1%), and vitality/hormonal function (n = 223; 45.5%) domains at baseline (see Supplementary Fig. 3, Additional le 1). Generally, responses were maintained for most items during radiotherapy treatment; while some decreased at study visit 2, they recovered to baseline scores at visit 3. Only items related to sexual domain (7, 8, and 9) showed a greater number of problems and worsened after radiotherapy (see Supplementary Fig. 3, Additional le 1).

Construct validity of EPIC-16
Correlations between different EPIC-16 domains were modest (r < .50 for all Spearman's correlations at baseline), indicating that the EPIC-16 domains are conceptually distinct and merit independent measure (see Supplementary Table 1, Additional le 1).

EPIC-16 sensitivity to change
Compared to baseline, almost all EPIC-16 mean domain scores worsened at visit 2, regardless of the health-state change as perceived by the patient; similar results were reported at visit 3, with the exception of the urinary irritation/obstruction domain which improved (Fig. 4).
Patients who reported an improved health status at visit 2 compared with baseline generally had smaller effect sizes (i.e., .20 in magnitude) versus those who reported a worsened health status; similar results were also reported at visit 3 (Fig. 4).

Correlation between EPIC-16 and UCLA-PCI
Overall, the EPIC-16 questionnaire scores showed strong correlations with the UCLA-PCI questionnaire domains at study visit 3, with Spearman's correlations > .4 across all corresponding UCLA-PCI domains ( Table 3). The urinary, bowel, and sexual function domains in the EPIC-16 questionnaire were most greatly correlated with respective domains in the UCLA-PCI questionnaire, e.g., the UCLA-PCI urinary function domain had a Spearman's correlation coe cient with the EPIC-16 urinary incontinence domain of .713, the bowel domains with .579, and the sexual domains with .739.

Discussion
This observational, non-interventional, multicenter study in Spain measured the change in HRQoL in men with localized PCa undergoing radiotherapy using the EPIC-16 and UCLA-PCI questionnaires, and further validated the Spanish EPIC-16 for routine clinical practice. Overall, EPIC-16 scores worsened after radiotherapy in different HRQoL domains, regardless of patients' perceptions of their health status, suggesting that patients did not perceive the change in functional domains as a global change in their health status. Similarly, the UCLA-PCI scores decreased with radiotherapy in both treatment groups (EBRT or BQT). However, 3 months after the end of radiotherapy, EBRT patients had recovered their scores in the urinary function, urinary bother, and bowel function domains. Conversely, BQT patients did not recover in urinary bother, bowel bother, and sexual function domains.
Similarly, in a long-term prospective QoL study in patients with localized PCa using EPIC-50, Ferrer, et al. reported that BQT treatment caused the least impact on QoL, except for moderate urinary irritativeobstructive symptoms, while sexual deterioration was observed for patients receiving EBRT [13].
The UCLA-PCI baseline values presented in this study were very similar to those reported by van de Poll-Franse, et al. in the Cancer of the Prostate Strategic Urologic Research Endeavor (CaPSURE) study, although the post-treatment scores of this study were smaller in magnitude compared to those of the CaPSURE study, which is likely due to the shorter period for the follow-up visits in this study (3 months) versus the 6-24-month follow-up period of CaPSURE [14].
Additionally, validation of the EPIC-16 questionnaire results obtained in this study was similar to results obtained by a number of EPIC-CP validation studies, especially those at study visit 2 [9,15]. However, differences exist in the sexual and vitality/hormonal domains between these studies, which may be due to differences in use of concomitant hormone therapy [15,16].
Even though the EPIC questionnaires were developed from UCLA-PCI [6,7], both instruments were used in this study to further assess the validity of the Spanish version of the EPIC -16 questionnaire. The previous EPIC-CP study by Chang, et al. [9] was validated in a smaller study cohort (N = 307) compared to this study, with 175 treated and 132 untreated patients. Furthermore, the current study also included evaluation of longitudinal validity, in addition to correlational analysis between EPIC-16 and UCLA-PCI domains. Hence, the results presented here further add to these previously reported results for the English EPIC-CP.
The results of the current study further highlight the importance of using appropriate PRO instruments to aid physicians treating patients newly diagnosed with localized PCa, enabling them to decide appropriate treatment strategies, to consider their potential adverse effects, and to incorporate individual patient preferences [17][18][19][20][21].
This study showed that the Spanish version of the EPIC-16 questionnaire demonstrated sensitivity to detect both PCa treatment-related effects and sensitivity for clinical improvement after radiotherapy. Furthermore, it was also shown that EPIC-16 has strong discriminative properties and reliability, demonstrating its validity for use in clinical practice and clinical trials to evaluate the effect of interventions, further expanding similar results obtained by Balbontin, et al. from a smaller patient population [22]. The EPIC-16 domains showed correlations with the respective functional domains in the UCLA-PCI questionnaire. Although these results show that both questionnaires are highly correlated in the urinary, bowel, and sexual function domains, they do not measure exactly the same information and are therefore complementary.
In future, the validated Spanish EPIC-16 could be utilized in a larger-scale analysis with more longitudinal components to further assess PROs comparing different types of radiation therapies, similar to what has been recently done for EPIC-26 by Nossiter, et al [23].
A main limitation of this study was its observational design. In addition, the smaller-than-planned sample for the BQT group affected the power of the hypothesis testing and resulted in large CIs and lack of statistical signi cance. It should be noted that the changes seen in BQT patients are not permanent; these patients were still on treatment due to the nature of the therapy. Additionally, the use of neoadjuvant and concomitant hormone therapy might impact data concerning sexual function [16]. Higher use of neo-adjuvant/androgen deprivation therapy in patients treated with EBRT versus those treated with BQT is a limitation for the interpretation of results and may have impacted differences observed in the sexual and hormonal/vitality domains, as reported previously [24]. Furthermore, despite several studies having suggested that neo-adjuvant/androgen deprivation therapy in addition to RT increased the incidence of urinary or rectal toxicity [25], other studies have not corroborated these ndings [26,27].

Conclusions
In this observational study conducted in Spain, men with localized PCa undergoing radiotherapy reported worsened scores in different HRQoL domains of the EPIC-16 and UCLA-PCI questionnaires immediately after radiotherapy treatment; however, a strong tendency towards recovery was seen at the 3-month follow-up visit. Regardless, patients did not perceive a global HRQoL change. Validation of the Spanish version of the EPIC-16 demonstrated sensitivity, strong discriminative properties and reliability, and validity. This shortened questionnaire is therefore suitable for use in routine clinical practice in Spain to measure urological HRQoL in this population.

Declarations
Ethics approval and consent to participate The study was approved by the independent ethics committee of participating centers. Patients gave their voluntary informed consent to take part.

Consent for publication
Not applicable.

Availability of data and materials
Researchers may request access to anonymized participant-level data, trial-level data, and protocols from Astellas sponsored clinical trials at www.clinicalstudydatarequest.com. For the Astellas criteria on data sharing, see: https://clinicalstudydatarequest.com/Study-Sponsors/Study-Sponsors-Astellas.aspx.
Competing interests AZ has received grants from Astellas during the conduct of the study and has received lecture fees from Astellas and Janssen and travel grants from IPSEN. XMP has received lecture fees from Astellas and Bayer and travel grants from IPSEN and Sano . AHM is a speaker and research collaborator for Astellas and Janssen. AGC and JLT report collaborations with Astellas for the current study. JPM, VMH, JLMG, and APE were investigators in this Astellas-sponsored study. PAA and CGJ are employees of Astellas.

Funding
This study was funded by Astellas Pharma Inc.