Variation in use of targeted therapies for metastatic renal cell carcinoma: Results from a Dutch population-based registry

For patients with metastatic renal cell carcinoma (mRCC), targeted therapies have entered the market since 2006. The aims of this study were to evaluate the uptake and use of targeted therapies for mRCC in The Netherlands, examine factors associated with the prescription of targeted therapies in daily clinical practice and study their effectiveness in terms of overall survival (OS). Two cohorts from PERCEPTION, a population-based registry of mRCC patients, were used: a 2008–2010 Cohort (n = 645) and a 2011–2013 Cohort (n = 233). Chi-squared tests for trend were used to study time trends in the use of targeted therapy. Patients were grouped based on the eligibility criteria of the SUTENT trial, the trial that led to sunitinib becoming standard of care, to investigate the use of targeted therapies amongst patients fulfilling those criteria. Multi-level logistic regression was used to identify patient subgroups that are less likely to receive targeted therapies. Approximately one-third of patients fulfilling SUTENT trial eligibility criteria did not receive any targeted therapy (29 % in the 2008–2010 Cohort; 35 % in the 2011–2013 Cohort). Patients aged 65+ years were less likely to receive targeted therapy in both cohorts and different risk groups (odds ratios range between 0.84–0.92); other factors like number of metastatic sites were of influence in some subgroups. Amongst treated patients, there was a decreasing trend in sunitinib use over time (p = 0.0061), and an increasing trend in pazopanib use (p = 0.0005). Targeted therapies have largely replaced interferon-alfa as first-line standard of care. Nevertheless, many eligible patients in Dutch daily practice did not receive targeted therapies despite their ability to improve survival. Reasons for their apparent underutilisation should be examined more carefully.

guidelines as first-line therapies for patients with a good or intermediate prognosis in 2009 and 2010, respectively [6,8]. For patients with a poor prognosis, temsirolimus was recommended [6] following the results of a multicentre, phase III trial in mRCC patients without any restrictions in histologic type, showing an increase in OS from seven to 11 months compared to IFN-a [9]. Furthermore, a number of second-line therapies have been added to guidelines, such as sorafenib, everolimus and axitinib [6,10].
Obviously, full and swift implementation of guidelines into clinical practice is essential to maximise the benefits of new therapies. However, the adoption of innovations in cancer care is generally quite heterogeneous, and differs between countries, and regions within countries [11]. A study by Jonsson et al. showed widespread use of sunitinib in the eight of the countries they studied, despite small differences between countries [12]. Sorafenib was widely prescribed in France, while a very low uptake and use in the United Kingdom and the United States were found. Besides between-country variation, Jonsson et al. found within-country variation in Sweden and suggested that more detailed information is needed on the use of first-and second-line therapies, to determine the extent of potential under-and overconsumption in different regions and different patient populations [12].
The aims of this study were to evaluate the uptake and use of targeted therapies for mRCC in The Netherlands, examine factors associated with the prescription of targeted therapies in daily clinical practice and study their effectiveness in terms of OS.

Study population
A population-based registry (entitled PERCEPTION) was created to include patients with mRCC. The PER-CEPTION registry consisted of two parts; a retrospective study and a prospective study. In the retrospective study, eligible patients were selected from the Netherlands Cancer Registry (NCR), which maintains a cancer registration database of all cancer patients in The Netherlands. Inclusion criteria for the retrospective study comprised a diagnosis of mRCC (i.e. metastases at initial presentation) of any histological subtype. Patients diagnosed from January 2008 until December 2010 in 42 of 51 hospitals (both general and academic) in four regions, covering approximately half of the country, were included. All patients were followed for a minimum of three years or until death .
The prospective study was designed differently in order to measure additional aspects of the disease, such as health-related quality of life (not reported in this study). In the prospective study, patients with RCC (all stages) of any histological subtype diagnosed from 2011 until June 30 th 2013 in 25 of 32 hospitals (both general and academic) in three regions were included. In contrast to the 2008-2010 Cohort, this cohort also comprised patients with mRCC who were initially diagnosed with localised disease. Besides the NCR, the hospitals' financing systems were used to select eligible patients at an early phase (for quality of life measurements). All patients were followed until the end of 2013 or until death (2011-2013 Cohort).

Data collection
Data on baseline demographics, clinical and laboratory factors were retrospectively collected from individual patient records by using uniform case report forms to ensure consistent data collection. Furthermore, data on treatment schemes and treatment endpoints (e.g. survival) were collected. Laboratory factors, such as haemoglobin and corrected calcium levels, were standardised according to routinely used reference values. Data were collected by personnel of the NCR and data collection stopped at the end of 2013.

Statistical analyses
To study differences in the proportion of patients receiving targeted therapy per half a year chi-squared tests were used. Exact tests were used to study possible time trends in the use of different therapies amongst treated patients. Additionally, chi-squared tests for trend were conducted.
Then, the use of targeted therapies within risk groups was studied. Risk groups were created using a slightly modified version of the MSKCC risk score [3]; a time from initial diagnosis to metastatic diagnosis of less than one year was used as a risk factor instead of a time from initial diagnosis to initiation of treatment of less than one year, since many patients in the study population did not receive any targeted therapy, thereby making it impossible to calculate the time to treatment. Additionally, the WHO performance status was used instead of Karnofsky performance status.
Furthermore, patients were grouped based on the eligibility criteria of the SUTENT trial [4], the trial that led to sunitinib becoming standard of care, to investigate the use of targeted therapies amongst patients fulfilling those criteria. Patients who had a clear-cell subtype, a WHO performance status of 0 or 1 and no brain metastases were classified as fulfilling the SUTENT trial eligibility criteria.
To identify patient subgroups that are less likely to receive targeted therapies in daily clinical practice among patients fulfilling SUTENT trial eligibility criteria, multilevel mixed-effects logistic regression was used to account for between-hospital variance. At the patient-level, patient and disease characteristics were taken into account including baseline demographics, clinical and laboratory factors [13,14]. Backward selection was used to select the covariates for the models; any non-significant covariates were excluded from the models one at a time.
OS was calculated from the start of therapy until death from any cause or the date of last follow-up, whichever came first, using the Kaplan-Meier method. For patients not receiving any targeted therapy, OS was calculated from the date of diagnosis.
Missing data regarding baseline characteristics were handled using multiple imputations by chained equations. This method generated imputations based on a set of imputation models, one for each variable with missing values [15].  Table 1 shows the patient and disease characteristics for this cohort. Median age was 66 years (range 23-93) and the majority of patients was male (66 %). The distribution of patients according to the MSKCC risk score showed a high proportion of patients (58 %) with a poor prognosis (versus 42 % with an intermediate prognosis). Since all patients in the 2008-2010 Cohort presented with metastatic disease, none of them had a favourable prognosis (i.e. time from initial diagnosis was less than one year). Additional file 1: Table S1 provides the observed patient and disease characteristics (without imputations). There is evidence of a difference between the half-year periods in the proportion of patients receiving targeted therapy (p = 0.041), but the chi-squared test for trend did not yield a significant result. Furthermore, no shift was found in the use of first-line therapies amongst treated patients.
Of the 336 patients receiving first-line therapy, 101 patients (30 %) also received a second-line therapy, with everolimus being the most common (40 %), followed by sorafenib (28 %). There was an increasing trend in everolimus use over time (p < 0.0001) and a decreasing trend in sorafenib use (p < 0.0001); from 2010 onwards, everolimus largely replaced sorafenib.

Use of targeted therapies amongst patients with an intermediate prognosis (2008-2010 Cohort)
Forty-two percent (269/645) of the patients in the 2008-2010 Cohort had an intermediate prognosis.
105/269 patients (39 %) received no targeted therapy. Some (n = 15) of these patients received a metastasectomy (combined with a nephrectomy) with a possible curative intention, making systemic therapy redundant. 40 of the remaining 90 patients (44 %) who were given neither targeted therapy nor a metastasectomy (combined with a nephrectomy) fulfilled the SUTENT trial eligibility criteria, indicating that they might have been eligible for treatment with sunitinib or another targeted therapy. 164/269 patients (61 %) received a first-line treatment; the majority was treated with sunitinib (145/ 164; 88 %). Of the 145 patients treated with sunitinib, 102 fulfilled the SUTENT trial eligibility criteria.
In patients fulfilling SUTENT trial eligibility criteria (including patients not receiving any targeted therapy and patients treated with sunitinib), patients with an abnormal neutrophil count (OR, 0.28; p = 0.045) were less likely to receive sunitinib, whereas patients with more than one metastatic site (OR, 3.35; p = 0.010) were more likely to receive sunitinib after adjustment for additional patient and disease characteristics (see frequencies in Table 3).
The median OS of eligible patients not receiving any targeted therapy was 18.6 months (95 % C.I. 8.4-33.7). Table 4 presents the median OS in subgroups of patients with an intermediate prognosis treated with first-line sunitinib. Median OS of eligible patients treated with sunitinib was 14.8 months (95 % C.I. 10.8-16.1). Note that a different starting point was used for the survival analysis (compared to the survival analysis in patients not receiving any targeted therapy). The mean time from diagnosis to start of first-line sunitinib was 4.3 months (standard deviation [SD] 6.0).
Median OS was 11.9 months (95 % C.I. 6.5-18.3) for ineligible patients treated with sunitinib, which was not significantly shorter than the OS of eligible patients treated with sunitinib. No significant differences were observed within the other subgroups.    Amongst patients fulfilling SUTENT trial eligibility criteria, older patients (OR, 0.90; p = 0.006) and patients with more than one comorbidity (OR, 0.26; p = 0.090) were less likely to receive sunitinib, whereas patients with more than one metastatic site (OR, 5.38; p = 0.034)  were more likely to receive sunitinib (see frequencies in Table 3). Furthermore, a significant association was found between hospital of diagnosis and prescription of sunitinib (p = 0.0059). Median OS of eligible patients not receiving any targeted therapy was 6.2 months (95 % C.I. 1.7-9.9). Table 4 shows the median OS in subgroups of patients with a poor prognosis treated with first-line sunitinib. Median OS of eligible patients treated with sunitinib was 6.8 months (95 % C.I. 5.3-10.7). The mean time from diagnosis to start of first-line sunitinib was 2.9 months (SD 5.5).
Median OS was significantly reduced in poor-prognosis patients treated with sunitinib but not fulfilling the SUTENT trial eligibility criteria (4.7 months, 95 % C.I. 3.3-6.9). Additionally, OS was significantly reduced in patients with brain metastases and patients with a WHO performance status of 2-4.  Table 1 shows the patient and disease characteristics of the patients with mRCC in this cohort. Median age was 66 years, and 73 % (170/233) of the patients was men. Metastatic disease was present in 77 % (179/233) of patients at the time of diagnosis, whereas 23 % was initially diagnosed with localised disease. In this cohort, 4 % of the patients with mRCC had a favourable prognosis, whereas 54 % and 42 % had an intermediate or poor prognosis, respectively. Table 2 shows the first-line therapies used in the 2011-2013 Cohort. During the follow-up period, 139/233 (60 %) patients received a first-line therapy; the majority (110, 79 %) was treated with sunitinib. The distribution of patients across first-line therapies over time (half-year periods) is presented in Fig. 2. There were no significant differences between the half-year periods in the proportion of patients receiving targeted therapies. However, amongst treated patients, there was a decreasing trend in sunitinib use over time (p = 0.0061) and an increasing trend in pazopanib use (p = 0.0005).

Uptake of targeted therapies and their use in daily clinical practice (2011-2013 Cohort)
Thirty-seven patients also received a second-line therapy within the follow-up period. The majority was treated with everolimus (57 %), but a decreasing trend in everolimus use over time was observed (p = 0.0020).
Use of targeted therapies amongst patients with a favourable or intermediate prognosis  136/233 patients (58 %) had a favourable or intermediate prognosis. 52/136 patients (38 %) did not receive any targeted therapy within the follow-up period. However, 12 of these 52 patients received a metastasectomy (combined with a nephrectomy). 25 of the remaining 40 patients (63 %) who were given neither targeted therapy nor a metastasectomy (combined with a nephrectomy) fulfilled the SUTENT trial eligibility criteria. In addition, 45 of the 66 patients treated with sunitinib fulfilled the SUTENT trial eligibility criteria.
Amongst patients fulfilling SUTENT trial eligibility criteria, males (OR, 0.12; p = 0.020) and older patients (OR, 0.92; p = 0.011) were less likely to receive sunitinib after adjustment for additional patient and disease characteristics (see frequencies in Table 3).
Median OS of eligible patients not receiving any targeted therapy was 20.9 months (95 % C.I. 7.4-not reached [NR]). Table 4  Median OS was 10.9 months (95 % C.I. 2.7-NR) for patients treated with sunitinib but not fulfilling SUTENT trial eligibility criteria. No significant differences were observed within subgroups. Of patients fulfilling SUTENT trial eligibility criteria, older patients (OR, 0.84; p = 0.012) were less likely to receive sunitinib (see frequencies in Table 3). The unadjusted model showed a significant association between hospital of diagnosis and the prescription of sunitinib, but this association disappeared after adjustment for demographics, clinical and laboratory factors.
Median OS of eligible patients not receiving any targeted therapy was 3.4 months (95 % C.I. 0.8-NR). Table 4 shows the median OS in subgroups of patients with a poor prognosis treated with first-line sunitinib. Median OS of eligible patients treated with sunitinib was 6.6 months (95 % C.I. 3.8-NR). The mean time from diagnosis to start of first-line sunitinib was 1.9 months (SD 1.8).
Median OS was significantly reduced in patients not fulfilling the SUTENT trial eligibility criteria (3.5 months, 95 % C.I. 1.3-7.8). Additionally, as in the 2008-2010 Cohort, median OS was significantly reduced in patients with brain metastases and patients with a WHO performance status of 2-4. OS was also significantly reduced in older patients.

Discussion
Since 2006, several new targeted therapies for mRCC have entered the market and randomised controlled trial (RCTs) have shown that these therapies improve survival [4,5,9,[16][17][18][19][20][21][22][23][24][25][26][27]. This study examined the uptake and use of targeted therapies in The Netherlands. Not unexpected, targeted therapies, sunitinib in particular, have largely replaced IFN-a as first-line standard of care. Few patients were treated with bevacizumab (combined with IFN-a) or temsirolimus in the 2008-2013 period, even though these therapies were added to the ESMO guidelines in 2009 [6], and to Dutch guidelines in 2010 [6]. Pazopanib has only been recommended since 2010 [8], which partly explains why an increase in its use was only seen from 2012. Furthermore, there was a shift in the use of second-line therapies, where sorafenib was replaced by everolimus as the most frequent choice from 2010 onwards.
The median OS of patients with an intermediate prognosis treated with sunitinib in Dutch daily practice and fulfilling the SUTENT trial eligibility criteria was shorter than the median OS of patients in the SUTENT trial with an intermediate prognosis, i.e. 14.8 months (95 % C.I.  [5]. The median OS of patients with an intermediate prognosis treated with sunitinib in Dutch daily practice (regardless of their SUTENT trial eligibility status) was shorter than the OS in the expanded-access trial [7]. Median OS of patients with a poor prognosis was in line with the results of the expanded-access trial. The median OS of patients with an intermediate prognosis treated with sunitinib in Dutch daily practice was also shorter than the OS in a retrospective, non-interventional study in Australia [28]. These findings may indicate that the patients in the PERCEPTION registry with an intermediate risk had a worse prognosis than the patients with an intermediate risk in other studies.
While previous studies suggest that patients fulfilling SUTENT trial eligibility criteria have a survival benefit from first-line sunitinib [5], many eligible patients did not receive sunitinib (or any other targeted therapy) in daily practice. This was also seen in England where one in three patients with mRCC eligible for either sunitinib or pazopanib did not receive the drug [29]. Patients aged 65+ years were less likely to receive targeted therapy than younger patients after adjustment for other factors. This age factor was found in patients with an intermediate prognosis (2011-2013 Cohort) and in patients with a poor prognosis (2008-2010 Cohort and 2011-2013 Cohort). There are several explanations for this association, including medical contraindications, other grounds for physician reluctance, and patient refusal. Additionally, patients with one metastatic site were less likely to receive sunitinib (according to the 2008-2010 Cohort results), which might be explained by patients with low volume but unresectable metastases whose targeted therapy is delayed. Nevertheless, most of these patients died within the follow-up period without receiving targeted therapy at any point in time. The reasons for apparent underutilisation of targeted therapies should be examined more carefully. While hospital-level factors may also affect utilisation and lead to between-hospital variation, we found no significant differences in the prescription of targeted therapy between hospitals, except for the patients with a poor prognosis in the 2008-2010 Cohort. However, the sample size per hospital was small and the statistical power to show a difference was therefore limited.