Variation in the management of cT1 renal cancer by surgical hospital volume: A nationwide study

Abstract Objectives To analyse variation in clinical management of cT1 renal cell carcinoma (RCC) in the Netherlands related to surgical hospital volume (HV). Materials and methods Patients diagnosed with cT1 RCC during 2014–2020 were identified in the Netherlands Cancer Registry. Patient and tumour characteristics were retrieved. Hospitals performing kidney cancer surgery were categorised by annual HV as low (HV < 25), medium (HV = 25–49) and high (HV > 50). Trends over time in nephron‐sparing strategies for cT1a and cT1b were evaluated. Patient, tumour and treatment characteristics of (partial) nephrectomies were compared by HV. Variation in applied treatment was studied by HV. Results Between 2014 and 2020, 10 964 patients were diagnosed with cT1 RCC. Over time, a clear increase in nephron‐sparing management was observed. The majority of cT1a underwent a partial nephrectomy (PN), although less PNs were applied over time (from 48% in 2014 to 41% in 2020). Active surveillance (AS) was increasingly applied (from 18% to 32%). For cT1a, 85% received nephron‐sparing management in all HV categories, either with AS, PN or focal therapy (FT). For T1b, radical nephrectomy (RN) remained the most common treatment (from 57% to 50%). Patients in high‐volume hospitals underwent more often PN (35%) for T1b compared with medium HV (28%) and low HV (19%). Conclusion HV is related to variation in the management of cT1 RCC in the Netherlands. The EAU guidelines have recommended PN as preferred treatment for cT1 RCC. In most patients with cT1a, nephron‐sparing management was applied in all HV categories, although differences in applied strategy were found and PN was more frequently used in high HV. For T1b, high HV was associated with less appliance of RN, whereas PN was increasingly used. Therefore, closer guideline adherence was found in high‐volume hospitals.


| INTRODUCTION
Renal cell carcinoma (RCC) represents 2%-3% of all cancers diagnosed worldwide. 1,2 In the Netherlands, the incidence of RCC has risen from approximately 1500 cases per year in 2000 to more than 2600 cases per year in 2020. 3 Widespread use of imaging has led to the increase in the incidence of small renal masses (renal tumours ≤4 cm) in the last decade, now representing 40%-50% of all new patients with RCC. 4,5 Partial nephrectomy (PN) has evolved as the standard treatment for cT1 tumours, although alternative nephron-sparing strategies are also used for cT1a tumours, such as active surveillance (AS) and focal therapy (FT). 6 For cT1 tumours, PN is the preferred treatment. 6 However, when PN is considered risky in frail patients or when technically not feasible, radical nephrectomy Although post-operative mortality is low, PN is recognised as a complex procedure with increased perioperative risk, such as bleeding, compared with RN. 9,10 In 2018, the Dutch Association of Urology (NVU) introduced the Dutch volume standard (DVS) in order to stimulate quality for hospitals performing PN and RN. 11 According to the DVS, the minimum number of RNs is 10 per year, and for PNs, at least 10 procedures per year are required, calculated as mean over a period of 3 years.
Our objectives were to analyse clinical variation over time in the management of cT1 renal cancers in the Netherlands and to investigate the adherence to the DVS of hospitals performing surgeries for RCC.

| MATERIALS AND METHODS
In this historic cohort study, all patients diagnosed with a cT1 RCC during 2014-2020 were identified in the Netherlands Cancer Registry (NCR), maintained by the Netherlands Comprehensive Cancer Organisation (IKNL). The NCR is a population-based registry composed of data on all newly diagnosed cancer patients in the Netherlands and has nationwide coverage since 1989. The main source of notifications of new cancers is the automated nationwide network and registry of histo-and cytopathology (PALGA). In addition, cases of non-pathology-proven tumours are supplied to the NCR by the Dutch Hospital Data (DHD). After notification, independent and trained data managers routinely extract patient-, tumourand treatment-related characteristics from medical records in all Dutch hospitals. Topography and morphology are coded according to the International Classification for Oncology (ICD-O) third edition 12 and disease stage according to the UICC Tumour-Node-Metastasis classification. 12,13 Treatment was categorised into five groups: RN, PN, FT, AS and others. It was assumed that patients with cT1 RCC without active treatment entered an AS protocol and were therefore classified as AS. Furthermore, PN, FT and AS were considered as nephron-sparing strategies. To allow comparison, HV categories were based on the BAUS HV categories. 7 It was not possible to use the exact BAUS HV categories because of smaller number of cases in the Netherlands.
Therefore, HV categories were defined as follows: Hospitals performing surgeries were grouped according to their annual number of (partial) nephrectomies and categorised into low volume    in 2020). A slightly increased use of PN was observed for cT1b, from 23.0% to 26.7%. AS was also applied more frequently over the years, from 12.5% to 17.0% ( Figure 1B).

| Treatment variations
Variation in applied treatment was evaluated by HV. For cT1a tumours, approximately 85% were treated with nephron-sparing management, independent of the HV category. However, differences in the applied nephron-sparing strategy were observed between  Analysis of treatment patterns of referred and non-referred patients revealed that hospitals not adhering to the DVS referred patients mainly for PN while managing FT, AS and some RN in their own hospital. FT is not included in the DVS but was evaluated in our study for referral patterns (Figure 3). In addition, geographical distribution showed large regional differences for patients that received FT for cT1a RCC, ranging from 1.4% to 24.5%, based on the zip code of the patient at the time of diagnosis ( Figure 4).  Furthermore, HV seems to be related to the type of approach:

| Surgical treatment variation by HV
Most (partial) nephrectomies in high-volume hospitals were T A B L E 1 Characteristics of all surgically treated patients (n = 7120) diagnosed between 2014 and 2020 divided by surgical hospital volume category, and for cT1a and cT1b renal cell tumours. performed robot-assisted (67.0%), whereas in medium-and lowvolume hospitals, the majority of (partial) nephrectomies were performed laparoscopically. In addition, low HV performed more surger- ies with an open approach (27%), compared with 6.7% in high HV ( p < 0.01; Table 1).

| DISCUSSION
This study shows that high-volume hospitals showed closer guideline adherence compared with low and medium HV for T1 RCC. The EAU guidelines have recommended PN as preferred treatment for T1 RCC. 6 The majority of patients with cT1a RCC were treated with nephron-sparing options in all hospital categories, but variation in self-reported data, our data were retrieved from a nationwide registration, collected by trained and independent data managers.
Another interesting finding of the BAUS audit was decreased complication rates with increasing HV for all patients, including patients treated with PN. PN is known to be a complex procedure and has been associated with higher complication rates compared with RN. 9,14 Other studies showed that undergoing robot-assisted PN at higher volume hospitals has been associated with decreased risk of conversion, positive surgical margins and complication rates. 15,16 Arora et al. analysed outcomes after any PN in the United States in relationship with HV and attempted to identify an optimal HV threshold for performing PN. They found that decreased complication rates were associated with increasing annual HV, with plateauing seen at 35 to 40 PNs per year. In their study, robot-assisted PN showed a similar association, with plateauing seen at 18 to 20 PNs annually. 17 In our study, we could not analyse complication rates, as these data F I G U R E 5 Proportion (%) of patients who underwent surgery that was performed in hospitals that adhere to the Dutch volume standard (DVS). PN, partial nephrectomy; RN, radical nephrectomy.
were not available. Future work should therefore be focused on a national registration of specialised care and surgery for RCC to improve clinical outcomes, decrease variations in practice patterns and subsequently increase guideline adherence in the Netherlands.
In an earlier study from the Netherlands, Aben et al. 8  Nevertheless, our study is based on a large nationwide registry with important information on differences in the management of cT1 RCC based on HV. We observed variation in applied management between different hospital categorisations. As fewer RNs were performed in high-volume hospitals and PN was more often applied, it is questionable whether a volume standard with a minimum of 10 PNs per year is adequate, and therefore, an increase of the volume norms should be considered. With no data available on case-mix of cT1 tumours in the Netherlands, a nationwide registry could be the solution to further understand the current differences in the management of cT1 RCC in the Netherlands.