Two-year quality of life after robot-assisted radical prostatectomy according to pentafecta criteria and cancer of the prostate risk assessment (CAPRA-S)

The quality of life (QoL) of men with optimal outcomes after robot-assisted radical prostatectomy (RARP) is largely unexplored. Thus we assessed meaningful changes of QoL measured with the EORTC QLQ-C30 24 months after RARP according to postsurgical Cancer of the Prostate Risk Assessment score (CAPRA-S) and pentafecta criteria. 2871 prostate cancer (PCa) patients with completed EORTC QLQ-C30 were stratified according to CAPRA-S, pentafecta (erectile function recovery, urinary continence recovery, biochemical-recurrence-free survival (BFS), negative surgical margins) and 90-day Clavien–Dindo-complications (CDC) ≤ 3a. Multivariable logistic regression analyses (LRM) aimed to predict improvement of EORTC QoL. Mean preoperative QoL values did not significantly differ between CAPRA-S low- (LR) vs. high-risk (HR, 75.7 vs. 75.2; p = 0.7) and pentafecta vs. non-pentafecta groups (75.6 vs. 75.2; p = 0.6). After RARP, stable QoL rates for CAPRA-S LR vs. HR and pentafecta were 30, 26 and 30%, respectively. Corresponding improved QoL rates were 44, 32 and 47%. In LRM, CAPRA-S and pentafecta criteria were independent predictors of improved QoL. We conclude that most favourable combined outcomes after RARP might confer stable or even improved QoL but up to one third of patients might experience deterioration. This warrants further investigation how to capture the underlying cause and to address and potentially solve these perceived negative effects despite successful RARP.

Radical prostatectomy (RP) remains a standard treatment option for localised prostate cancer (PCa) with well documented local and long-term PCa control [1][2][3] . RP patients ideally wish to recover quickly and to preserve a satisfactory long-term quality of life (QoL). Moreover, compared to open RP, robot-assisted radical prostatectomy (RARP) might confer added benefit of better perioperative outcomes [4][5][6] . Monitoring these outcomes and associated QoL is essential to improve preoperative patient counselling, RARP technique and post-RARP rehabilitation 7,8 . It is highly conceivable but largely unexplored that certain combinations of favourable pathological, functional or oncological outcomes of RARP are perceived with best QoL. Specifically, the pentafecta criteria includes a combination of biochemical recurrence (BCR)-free survival, recovery of urinary continence and of erectile function at a specific time-point, as well as negative surgical margin status and absence of surgical complications 9,10 . Another favourable combined outcome is the postsurgical Cancer of the Prostate Risk Assessment score low risk score (CAPRA-S) 11 . A low-risk CAPRA-S is based on favourable surgical and pathological characteristics, such as low pre-surgical PSA < 20 ng/ml, pathological Gleason pattern ≤ 4 + 4 and ideally negative surgical margin and localized disease (i.e. no extracapsular extension, seminal vesicle invasion and lymph node invasion 12 ).
For purpose of QoL assessment, the European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire Core 30 (EORTC QLQ-C30) is widely used in cancer patients, including PCa. It is a validated patient-reported outcome measurement (PROM) tool that covers functional scales, symptoms, single items and global QoL [13][14][15] . Accordingly, it is regularly utilized in clinical trials 16 that compare treatment methods, e.g. RP vs. radiotherapy 1,[17][18][19] . In daily clinical practice, as in our institution, it also serves for monitoring QoL after RARP over time 20 .
It is of note that previous series on EORTC QLQ-C30 and RP mainly relied on exploratory analyses, e.g. mean change of individual EORTC QLQ-C30 scales over time and did not stratify according to specific functional or oncological scenarios. Moreover, most series omitted multivariable analyses 19,[21][22][23][24] and relied on ORP patients 19,21,22 . In general, comprehensive data on RARP and EORTC QLQ-C30 is sparse. To date, none applied the EORTC QLQ-C30 Summary Score in the context of RARP, which incorporate most functional and symptom scales and single items. None applied minimal clinically important differences (MCIDs) after RARP to enable intuitive interpretation of QoL effects. Finally, none dedicatedly examined effect of specific favourable patientcentric combined outcomes such as pentafecta or low-risk CAPRA-S on QoL.
Thus, we examined the impact of CAPRA-S risk categories and pentafecta criteria on EORTC QLQ-C30 global QoL and Summary Score in patients treated with RARP and report relevant respective improvement vs. deterioration rates.

Material and methods
Overall, 2871 consecutive PCa patients with complete pathological, surgical and complications data who received RARP at our institution between 5/2006 and 12/2018 were identified. Analyses were restricted to men with complete data on EORTC QLQ C30 at baseline and 24 months after surgery, who were preoperatively continent and who had corresponding follow-up information of urinary continence recovery and information on baseline erectile function. Patients with neoadjuvant androgen deprivation therapy, suspected metastases or with previous local therapy of the prostate were excluded. RARP specimen grading was performed according to 2014 International Society of Urological Pathology standards 25 .

EORTC-QLQ-C30 and minimum clinically important differences. The questionnaire includes 30
items that can be summarized into 15 scales, i.e. five functional scales (physical, role, emotional, cognitive and social), three symptom scales (fatigue, nausea/vomiting and pain), six singe items (dyspnea, insomnia, appetite loss, constipation, diarrhea, financial impact of disease) and a global QoL (QL).
For our study we utilized the QL and the EORTC QLQ-C30 Summary Score. Latter is calculated by the sum of 13 scales (QL and financial impact are excluded), divided by the number 13 11,[26][27][28] . For calculation, the symptom scales are reversed to obtain a uniform direction of every scale.
Outcomes. The CAPRA-S is a scoring system based on pre-surgical PSA, pathological Gleason score, surgical margin status, extracapsular extension, seminal vesicle invasion and lymph node invasion 12,30 . The Patients were stratified according to CAPRA-S into low-(LR), intermediate-(IR) and high-risk (HR) group, based on respective score ranges of 0-2, 3-5 and ≥ 6 12 .
Preoperative potency and recovery of erectile function were defined as combination of International Index of Erectile Function (IIEF-5) score ≥ 18 31 and/or score ≥ 3 at the second question 32 . Recovery of urinary continence was defined as fulfilling following criteria: up to one pad usage within 24 h (safety pad) or score of ≤ 2 at the International Consultation on Incontinence Questionnaire-Urinary Incontinence Short Form Questions 1 and 2 ("How often do you leak urine?" and "How much urine do you usually leak (whether you wear protection www.nature.com/scientificreports/ or not)?") 33 or finally, the International Continence Society male questionnaire score of ≤ 1 at each of the three questions I2, I3 and I4. For preoperative continence, no preoperative pad usage was allowed. BCR was defined as consecutive PSA rise ≥ 0.2 ng/ml. Those with recovery of erectile function-and urinary continence and BCR-free-survival within 24 months were classified as having met the trifecta criteria, i.e. ideal functional and oncological outcomes 24 months post-RARP. Those men with preoperative erectile dysfunction, were still modeled as having met the trifecta-criteria if urinary continence recovery and BCR-free-survival within 24 months was observed. Complications were coded according to the Clavien-Dindo classification (CDC) 34 . Finally, those, who met the combined criteria of negative surgical margins, CDC ≤ 3a and 24-months trifecta were classified as pentafecta group (vs. non-pentafecta).

Statistical analyses.
Chi-square test was used for categorical and t-test for continuous variables. All tests were two-sided with a statistical significance set at p < 0.05. Analyses were performed with the statistical package for R (R foundation for Statistical Computing, version 3.2.2).
Ethics approval. The institutional review board at the St. Antonius-Hospital, Gronau, approved the retrospective study design and access to the patients' medical records. All methods were carried out in accordance with the Declaration of Helsinki. Written informed consent was obtained from individual participants in the study.
Compliance with ethical standards. All authors of this research paper have directly participated in the planning, execution, or analysis of the study. All authors of this paper have read and approved the final version submitted. The contents of this manuscript have not been copyrighted or published previously. The contents of this manuscript are not under consideration for publication elsewhere. Table 1 demonstrates the baseline characteristics. Median age of the total cohort was 63 years (IQR 59-67) and median PSA was 7.3 ng/ml (IQR 5.4-10.7). RARP ISUP Gleason Grades 1, 2, 3 and ≥ 4 were 36, 29, 22 and 13%, respectively. Respective proportions of pT2, extracapsular extension, seminal vesicle invasion, positive surgical margin and lymph node invasion were 65, 25, 7.7%, 11 and 6.5%. Based on aforementioned clinicopathological metrics, patients were classified as CAPRA-S low-, intermediate and high-risk in 53, 32 and 15% respectively.

Results
The 90-day CDC rates ≤ 3a of the total cohort were 97.6%. At 24 months post-RARP, rates of erectile function and urinary continence recovery, as well as BCR-free survival of the total cohort were 63, 96 and 86%, respectively. Based on negative surgical margin and aforementioned metrics, 24-month pentafecta rates were 51%. Table 2 demonstrates pre-vs. post-operative mean QL and Summary Score and respective MCID proportions, stratified according to CAPRA-S risk groups and pentafecta vs. non-pentafecta groups. Before RARP there was no statistically significant difference in the QL (75.7, SD 19.4 vs. 75.2, SD 19.4; p = 0.7) nor in Summary Score (91.0, SD 9.2 vs. 91.2 SD 9.2; p = 0.8) between CAPRA-S LR vs. HR. Conversely, 24 months after RARP, there was a statistically significant difference in the QL (79.3, SD 19.3 vs. 73.7, SD 20.7; p < 0.001), as well as Summary Score (90.2, SD 11.5 vs. 87.6, SD 13.4; p < 0.001).
At 24 months after RARP, between pentafecta vs. non-pentafecta groups, QL remained stable in 30 vs. 28% and improved in 47 vs. 34% (p < 0.001). These amounted to 78 vs. 62% without any deterioration in QL. Similarly, the Summary Score remained stable in 60 vs. 50% and improved in 20 vs. 15% (p < 0.001). These amounted to 80 vs. 65% without any deterioration in the Summary Score. Table 3 shows the multivariable logistic regression model for prediction of improvement of QL and Summary Score at 24 months after robot-assisted radical prostatectomy. LRM included CAPRA-S risk groups and the remaining criteria (i.e. trifecta and CDC) that are utilized for the pentafecta criteria, as well as the preoperative mean score of either QL or Summary Score. Specifically, LRM for prediction of improved QL at 24 months after RARP, yielded following independent predictors: higher preoperative score ( www.nature.com/scientificreports/ Finally, LRM for prediction of at least stable or improved QL (or Summary Score) at 24 months, which represents a broader definition of a desirable outcome than just improvement, yielded comparable results (Supplemental Table 1).

Discussion
To date, postoperative QoL metrics in combination with MCID are still rarely reported in context of specific functional or oncological profiles of PCa patients, which were treated with RARP. Interpretability of QoL change is furthermore limited due averaged data, e.g. observed mean EORTC QLQ-C30 values before and after surgery. This requires rectification since QoL metrics serve as important patient counselling tool before surgery, monitoring after surgery and benchmarking tool for comparison with other institutions 35,36 . Thus, we compared the impact of specific pathological PCa profiles and of the widely adopted pentafecta criteria on key measurements of the EORTC QLQ-C30.
Our study revealed important findings. First, median preoperative QL and Summary Score values were virtually identical between CAPRA-S groups, as well as pentafecta vs. non-pentafecta groups. Such preoperative balance indicates validity of our MCID findings, since MCIDs are calculated as the difference between preoperative vs. postoperative values. Moreover, our preoperative values are highly consistent with randomized trials and indicate a healthy baseline functioning in the majority of our RARP patients and adequate patient selection [17][18][19] .
Second, despite balanced and healthy baseline values, we are the first to report substantial proportions of RARP patients who improve upon their preoperative global QoL (47%) and Summary Score (20%). Reassuringly, depending on scale, the vast majority of pentafecta patients, 77% (QL) to 80% (Summary Score) do not show any deterioration after RARP, but either remain stable or improve further.
Third of those pentafecta patients reporting deterioration, 8-10%, about a third, are only classified as minor. It is important to note, that deterioration might occur in a specific scale but be absent in another. This is evidenced by the mean change of QL, which is positive in CAPRA-S LR and pentafecta patients and negative in the Summary Score even in the same patients. This is consistent with the general observation of perceived deterioration in cancer patients despite successful treatment and favourable prognosis 37 .
Fourth, this phenomenon of discrepant perception in scales is further substantiated by multivariable analyses. Despite congruent patterns of the respective impact of CAPRA-S LR and pentafecta profiles on individual QL and Summary Score scales, LRM clearly demonstrated that CAPRA-S risk profiles are independent of the pentafecta criteria. For example, an individual with CAPRA-S high-risk profile and simultaneous pentafecta criteria  www.nature.com/scientificreports/ might perceive certain scales as deteriorated due to greater weight on PCa risk than outcomes. In consequence, pathology and functional outcomes, should be considered in patients counselling.
It is important to note that a high-risk CAPRA-S score strongly correlates with adjuvant/salvage androgen deprivation therapy and/or radiotherapy, which is expected 12,30 . Accordingly, in context of QoL, an increasing CAPRA-S should also be interpreted as higher risk of biochemical recurrence and administration of additional therapy after RARP that potentially further contributes to a negative impact on QoL 1,18,32,38-40 .    www.nature.com/scientificreports/ With regard to functional outcomes, erectile aid such as phosphodiesterase type-5 inhibitors or intraurethral/ intracavernosal application of alprostadil and urinary continence treatments such as male sling procedures or artificial urinary sphincter implantation ideally promote recovery of erectile function or urinary continence, respectively. These might translate to stable or improved QoL. This notion is substantiated by about 44% of preoperatively potent patients in our pentafecta group, who used erectile aids at time of erectile function recovery.
Finally, it is also important to point out that CAPRA-S LR vs. HR represents a distinct contrast, whereas pentafecta vs. non-pentafecta does not. Specifically, the difference is highly variable and might only contain minor differences compared to the very strict concept of pentafecta. However, this notion strengthens our findings of significant differences between pentafecta vs. non-pentafecta effects on QoL.
Our study has limitations. First, despite regular adoption of EORTC QLQ-C30 in randomized cancer trials, most MCIDs are not validated yet for PCa. We utilized the widely accepted MCID definition by Osoba et al. 29 Although it was based on patients with breast and small-cell lung cancer, several subsequent series confirmed 10 point changes or more as being clinically significant after collating results from various studies and various cancer sites including PCa 41,42 . Our approach of MCIDs enables superior interpretation compared to simply reporting mean value as the majority of preceding series.
Second, we are also the first to report the validated EORTC QLQ-C30 Summary Score for RARP patients to complement global QoL 11,[26][27][28] . Specifically, the Summary Score was validated as prognostic factor for survival 11 . Moreover, we aimed to improve interpretability and robustness of the scales/items in concatenated fashion as opposed to presenting each individual scales/items. Since the Summary Score represents a mean of aforementioned components, MCIDs could be applied accordingly. However, MCIDs are not validated yet for the Summary Score. Nonetheless, the direction and extent of change is transparently demonstrated in our study.
Third, since pentafecta outcomes are composed of time-to-event variables it is mandatory to focus on a specific time point after RARP, e.g. 24 months, for purpose of examining the relationship between the pentafecta outcome and QoL. This represents an inherent flaw of the pentafecta concept. For example, longer post-RARP follow-up will be associated with higher oncological failure rates that lead to pentafecta exclusion vs. increased recovery rates that lead to pentafecta inclusion. At 24 months post-RARP, QoL is likely more influenced by functional recovery than BCR. Conversely, post-RARP QoL might further improve or deteriorate after 24 months, since previous series indicate that QoL values stabilize after up to 6 years in absence of further interventions 20 .
Finally, it is important to note that surgical expertise is essential to achieve favorable combined outcomes and that different learning curves are needed for specific outcomes 43,44 . Specifically, previous series reported a shorter learning curve for achieving satisfactory outcomes for continence, followed by a longer curve for potency and then for surgical margins 45 . This is evidenced by a significantly higher surgical expertise in the pentafecta group (median 767, IQR 174-212 vs. 552, IQR 140-1750; p < 0.001). In the same context, generalizability of our findings are limited since our department represents a specialized prostate cancer center with highly experienced robotic surgeons.

Conclusions
We confirm that favourable pathology, oncological and functional outcomes independently confer not only stability, but even positive change on specific QoL measurements in the vast majority of RARP patients. However, optimal outcomes after treatment do not guarantee best perceived QoL, since up to one third of patients might experience deterioration. This warrants further investigation of how to capture the underlying cause and how to finally address and potentially solve these perceived negative effects of PCa despite successful RARP.

Data availability
Full availability.