Recurrence patterns following nephrectomy for renal cell carcinoma in a Danish nationwide cohort

Abstract Objectives This study aimed to characterize the demographic and clinical features of patients with renal cell carcinoma (RCC) post‐surgery for localized or locally advanced disease in a national Danish cohort, with a specific focus on describing recurrence patterns in a subgroup aligned with the adjuvant KEYNOTE‐564 trial classification. Methods This was a retrospective analysis of the Danish Renal Cancer (DaRenCa) database. Eligible subjects were individuals with an RCC diagnosis between January 2014 and December 2017 who subsequently underwent radical or partial nephrectomy. Variables of interest were demographic and clinical characteristics, rates and sites of recurrence. Recurrence rates were also assessed in a subpopulation stratified using the risk classifications of the KEYNOTE‐564 trial. Results A total of 2164 RCC patients were identified. Most patients (84.8%) had non‐metastatic RCC (stage M0). A recurrence was observed in 250 of the M0 patients (13.6%). Patients with a recurrence were older, male, had a higher tumour stage, had undergone radical nephrectomy and had a higher Leibovich score. The majority (74.8%) of M0 patients had recurrence at distant metastatic sites. A total of 392 patients were stratified by the KEYNOTE‐564 risk classification: 335 intermediate‐high risk, 17 high risk and 40 M1 NED (metastatic with no evidence of disease). Recurrence was observed in 37.0%, 88.2% and 27.5% of these risk groups, respectively. Conclusions This study elucidates the rates and determinants of post‐surgical RCC recurrence in Denmark, underscoring the potential of adjuvant immunotherapy in refining therapeutic strategies, identifying suitable beneficiaries and minimizing overtreatment risks in RCC care.


| INTRODUCTION
Kidney cancer represents a significant health challenge globally, with the World Health Organization reporting over 430 000 new cases and approximately 175 000 deaths in 2020. 1,2Northern Europe, including Denmark, exhibits one of the highest incidence rates of this disease worldwide. 3,4Renal cell carcinoma (RCC), the most common form of kidney cancer, 5 often presents as a localized disease 6 amenable to surgical intervention. 7Despite high survival rates following surgery, the recurrence of RCC is not uncommon and poses a considerable clinical and socio-economic impact, as well as a less favourable outlook for patients. 8e risk of RCC recurrence can be estimated using various prognostic factors, 9,10 yet the medical community has not reached a consensus on a single, most accurate predictive algorithm among the numerous available.5][16] Pembrolizumab, a programmed death 1 (PD-1) inhibitor, has recently been approved in the United States and Europe for adjuvant treatment in RCC patients at increased risk of recurrence, based on the results of the KEYNOTE-564 trial showing significant delay in recurrence in patients treated with adjuvant pembrolizumab (hazard ratio for disease-free survival 0.68 [95% confidence interval, CI, 0.53-0.87]and 0.63 [95% CI 0.50-0.80]with 24 and 30 months of follow-up, respectively). 17,18nsidering the evolving treatment landscape for localized RCC, it is crucial to understand real-world recurrence patterns across different risk groups.This knowledge can help identify patients who may benefit most from emerging adjuvant therapies.Therefore, the objective of this study was to comprehensively describe the demographic and clinical characteristics of RCC patients following surgery for localized or locally advanced disease in a national Danish cohort, with particular attention to analysing recurrence patterns.Additionally, the study aimed to characterize a subgroup of patients fitting the risk categories utilized for patient inclusion in the adjuvant KEYNOTE-564 trial classification.

| Study design and data source
We conducted a retrospective analysis using the Danish Renal Cancer

| Study population
Eligible individuals were those diagnosed with RCC and had undergone radical or partial nephrectomy between 1 January 2014 and 31 December 2017.Exclusion criteria included metastatic disease at diagnosis (except where specified in order to include the 'M1 no evidence of disease [NED]' risk category from KEYNOTE-564), other primary cancers before nephrectomy and insufficient follow-up or incomplete data.

| Study variables and outcomes
The primary outcome was the recurrence of RCC, categorized as either locoregional or distant metastasis.In alignment with the guidelines established by the Danish Renal Cancer Group, recurrence was defined as radiological or pathological evidence of RCC that manifests more than 120 days post-surgery.This specific time frame is utilized to differentiate between residual disease not completely addressed by surgery and the true recurrence of the disease.Patients presenting with signs of metastatic disease within the first 120 days post-surgery were considered to have primary metastatic disease and were excluded from the recurrence analysis.

| Statistical analysis
We described the demographic and clinical characteristics of all RCC patients (N = 2164) and performed a detailed analysis of the group with non-metastatic disease at baseline (n = 1835).Recurrence rates were calculated using Kaplan-Meier analysis at 3 and 5 years post-surgery.A separate analysis was conducted focusing on the patients meeting the KEYNOTE-564 trial risk criteria (n = 392), assessing recurrence rates and characteristics within defined risk subgroups.

| Description of the RCC cohort
Upon applying the inclusion criteria, we identified 2164 patients with RCC (Table 1).The median age was 66.6 years (interquartile range [IQR] 57.9-72.9),with 65.6% being male and 65.0% having an Eastern Cooperative Oncology Group (ECOG) performance status of 0. A substantial proportion of patients (n = 1835; 84.8%) had non-metastatic (M0) disease at diagnosis.Within this M0 cohort, 69.4% had an ECOG performance status of 0, 71.7% were diagnosed at tumour stage T1, 78.8% had node stage Nx and 74.4% had a clear cell RCC subtype (Table 2).The median Fuhrman grade among the M0 cohort was 2 (IQR 2-3), and the majority of patients (58.7%) had a Leibovich score of 0-2, indicating low risk of recurrence.

| DISCUSSION
Our analysis provides a comprehensive characterization of Danish patients undergoing surgery for RCC and describes the distribution of variables in patients with and without disease recurrence.By employing the risk criteria utilized to include patients in the adjuvant pembrolizumab trial (KEYNOTE-564), we stratified a subgroup of patients and demonstrated an increased rate of recurrence corresponding to higher prognostic risk levels.
A significant proportion of recurrences in our cohort occurred at distant sites (74.8%), with the remainder being local.This distribution is consistent with some previous studies, 19,25 although others have reported a higher incidence of distant recurrences, closer to 90%. 8,22 found that recurrence was more frequent in patients with a higher tumour stage and Leibovich score, male patients and patients who underwent radical nephrectomy, which was in line with findings from earlier research. 9,12,26key aspect of our study was the application of risk stratification criteria from the KEYNOTE-564 trial, allowing us to assess whether our data were consistent with the trial's findings.The majority of our patients fell into the intermediate-high-risk category (85.5%), similar to the trial's composition (86.9% in the placebo group). 17Our descriptive analyses revealed a trend of increasing recurrence rates with higher risk levels among primary M0 patients, mirroring both the pattern observed and absolute recurrence rates in the placebo group of KEYNOTE-564 for the intermediate-high-and high-risk patients. 17tients with pT3 disease exhibited a wide range of recurrence rates, with higher recurrence observed with increasing Fuhrman grade.This has been observed previously, although our study reported lower overall 3-year recurrence rates for pT3/N0/M0 patients compared with those reported from the US SEER database Our findings indicate that the tendency for RCC recurrence post-surgery is a significant concern, with a notable proportion of patients experiencing disease return despite the surgical approach taken.While radical and partial nephrectomies are associated with different recurrence rates, the focus of our study is not on the surgical technique as a risk factor but on the broader patterns of recurrence that emerge following surgery.This distinction is crucial for guiding post-operative management and surveillance strategies.Understanding the trends in recurrence can help clinicians identify patients who may benefit from more intensive monitoring or adjuvant therapies, potentially improving long-term outcomes.
The M1 NED group in our study had a lower 3-year recurrence rate compared with KEYNOTE-564 (23% vs. > 65%). 17This discrepancy could be attributed to the high mortality risk in M1 NED patients, leading to censorship in our analysis and potentially underestimating the true recurrence risk.Given the small sample size of M1 NED patients in our study (n = 40), these results should be interpreted cautiously.
Our study's strengths include its multicentre design and comprehensive inclusion of Danish RCC patients, with complete followup data from registries and detailed information from medical records.However, we acknowledge limitations such as the retrospective nature of the study, the lack of adjustment for partial/ radical nephrectomy and confounders like socio-economic status.
Additionally, the absence of certain histological data, such as programmed cell death ligand 1 (PD-L1) expression, limits further risk stratification.
In conclusion, our real-world data analysis highlights the potential of adjuvant immunotherapy in managing primary non-metastatic RCC post-surgery.By identifying factors that are consistent with those delineated in clinical trials such as KEYNOTE-564, we can better identify patients who are likely to benefit from this treatment approach, thereby enhancing therapeutic precision and reducing the risk of overtreatment.These insights affirm the role of adjuvant immunotherapy as a valuable strategy for managing RCC recurrence risks.

AUTHOR CONTRIBUTIONS
Goran Bencina: Writing-original draft; writing-review and editing.

(
DaRenCa) database, a comprehensive national clinical quality database.The DaRenCa database integrates data from multiple Danish health registries and is supplemented by manual reviews of medical records.It includes demographic, clinical and treatment information for patients with RCC diagnosed from 2014 to 2017.Follow-up data collection extended through August 2022, with a median follow-up of 4.95 years for non-metastatic patients at diagnosis.The study was approved by the Danish Patient Safety Authority (3-3013-2902/1) and complied with the Danish Data Protection Agency regulations (REG-041-2021).

F I G U R E 2
Recurrence in patients stratified by the KEYNOTE-564 classification.(A) Distribution of 392 patients among the various KEYNOTE-564 risk categories.The categories are described in detail in Section 2. (B) Rates of recurrence within each risk category.Int, intermediate; NED, no evidence of disease; RCC, renal cell carcinoma.F I G U R E 3 Kaplan-Meier plot of recurrence in the patients stratified by KEYNOTE-564 risk classification (n = 392).The green line represents the 392 patients who fit the KEYNOTE-564 risk criteria.The blue line represents the remaining 1443 patients who did not meet the KEYNOTE-564 risk criteria.by Sundaram et al. 8 This may reflect differences in patient demographics and the centralized nature of nephrectomy procedures in Denmark compared with the United States.

F I G U R E 4
Kaplan-Meier plot of recurrence in the KEYNOTE-564 subpopulation of pT3/N0/M0 patients stratified by Fuhrman grade.
Clinical characteristics of patients with primary nonmetastatic renal cell carcinoma.
Note: Values are presented as n (%) unless otherwise specified.Abbreviations: ECOG, Eastern Cooperative Oncology Group; IQR, interquartile range; SD, standard deviation.TA B L E 2