Real-World Data on the Use of Nivolumab Monotherapy in the Treatment of Advanced Renal Cell Carcinoma after Prior Therapy: Interim Results from the Noninterventional NORA Study

Article info Article history: Accepted November 26, 2021 Associate Editor: Christian Gratzke


Introduction
The European Medicines Agency approved the immune checkpoint inhibitor (ICI) nivolumab in 2016 as the first antibody targeting PD-1 for the treatment of advanced renal cell carcinoma (aRCC) after prior therapy [1]. The decision was based on results from the pivotal phase 3 CheckMate 025 (CM025) randomized clinical trial (RCT) that enrolled patients with aRCC in the second to fourth line [2]. Long-term follow-up demonstrates a continuous benefit with nivolumab in comparison to everolimus, a prior second-line standard therapy. Nivolumab showed a significant improvement in overall survival (OS; 26 vs 20 mo; hazard ratio [HR] 0.73, 95% confidence interval [CI] 0.62-0.85), a significantly higher objective response rate (ORR; 23% vs 4%), and a better toxicity profile. Progression-free survival (PFS) also favored nivolumab (HR 0.84, 95% CI 0.72-0.99; p = 0.0331) [3]. According to European and German guidelines on RCC, nivolumab monotherapy is a standard of care (SOC) in second-line therapy after progression on treatment with a VEGFR tyrosine kinase inhibitor (TKI) [4][5][6].
However, the conditions in RCTs only partially reflect routine clinical practice, in which nivolumab is applied without strict selection criteria [1,7]. For CM025, only patients with predominant clear-cell (cc) tumor histology, Karnofsky performance score (KPS) 70, and a maximum of three previous lines of systemic treatment (one or two antiangiogenic therapies) were included [2]. In CM025, prior treatment with an mTOR inhibitor was prohibited, as well as any history of or current metastases of the central nervous system.
Real-world data may complement data obtained from RCTs and may demonstrate effectiveness in a broader population that includes patient subgroups that were poorly studied so far (eg, elderly patients and patients with nonclear cell [ncc] histology or KPS <70) and can thus provide support for therapeutic decisions. Moreover, assessment of safety data in the post-approval setting is of paramount importance.
The present noninterventional study (NIS) NORA (NivOlumab in Renal cell cArcinoma) was designed to capture the early period after market authorization approval for nivolumab. The study aim was to describe the outcomes, patient characteristics, safety, and treatment patterns among adults with aRCC starting a second-line or later systemic treatment with nivolumab overall and by subgroups of interest.

Objectives
The primary objective was OS estimation in the overall population and by relevant subgroup. Secondary objectives included estimation of progression-free survival (PFS), best overall response (BOR), the objective response rate (ORR) and disease control rate (DCR), and the duration of response (DOR) in the overall population and in relevant subgroups. questionnaires.

Statistical analysis
All patients fulfilling the study entry criteria were included in the data set for analyses. As NORA is of descriptive character, no formal hypotheses were tested and no comparative analyses assessing the effectiveness of other treatments were undertaken. Instead, patient and clinical characteristics were summarized using descriptive statistics to provide context for the observations for patients treated with nivolumab in real life.

Results
In total, 228 patients were eligible and enrolled in NORA and their data for effectiveness and safety were analyzed. The baseline patient characteristics are summarized in of the patients; nccRCC histologies included papillary type 1 (4.8%), papillary type 2 (5.7%), chromophobe (1.8%), and others (4.4%). A sarcomatoid fraction was noted in 4.4% of the patients. The most common metastases were in the lung (71%), bone (32%), and liver (28%). Notably, 4.8% of the patients had brain metastases.

Treatment characteristics
Nivolumab treatment characteristics are shown in Table 2. For 56% of the patients more than 12 mo had passed between their primary diagnosis and the start of the first systemic treatment. At the first nivolumab dose, the median time since initial RCC diagnosis was 31 mo (range 1.6-363).
The median time between progression on the prior systemic therapy and the start of nivolumab treatment was 0.9 mo (range 0-65). Some 78% of the patients received nivolumab as secondline therapy, 14% as third-line therapy, and 8.3% as a fourth or later line of therapy. The vast majority of patients had previously received pazopanib (43%) or sunitinib (41%) as first-line therapy. The most frequent prior second-line therapies were pazopanib, axitinib (31% each of 51 patients receiving nivolumab as 3rd line), and everolimus (20%). Treatments in prior third and later lines were more diverse.
The ORR was 20% in the overall population; three patients (1.3%) experienced a complete response. Another 26% of patients achieved stable disease, which corresponds to a DCR of 46%; 33% had progressive disease as their best response (Table 3). At data cutoff, the estimated median DOR was 28 mo (95% CI 16-not estimable; from 17 responders with a documented progressive disease date and without censoring; Fig. 3 Fig. 1B and Table 3). Furthermore, ORR was lower for later lines of treatment (2nd line 23% vs 3rd line 12%). However, the median duration of nivolumab treatment and the respective Kaplan-Meier plots were similar irrespective of the treatment line (data not shown). Of 11 patients with brain metastases, three had a partial response, one experienced stable disease, and five had progressive disease as their BOR (data missing for 2 patients). Further subgroup analyses for OS revealed that median OS was not reached (NR) for patients with favorable IMDC risk and was 24 mo for the intermediate group and 12 mo for the poor risk group (Fig. 1C), and was longer for higher KPS, at 26 mo for KPS 70 and 6.1 mo for KPS <70 (Supplementary Fig. 1A). OS did not differ by patient age or the presence of bone metastases ( Supplementary Fig. 1B,C). PFS subgroup analyses revealed similar trends (Fig. 2B, Supplementary Fig. 2A-D).

Safety outcomes
A total of 301 trAEs (any grade) were reported by 105 patients (46%) in the overall cohort ( Table 4). The most frequent trAE was diarrhea (9.2%), followed by fatigue (6.6%), malignant neoplasm progression (6.1%), pruritus (5.7%), and rash (4.4%). Thirty-four patients (15%) had at least one grade 3-4 trAE, with malignant neoplasm progression (3.1%) and diarrhea (1.8%) being the most common. One grade 5 trAE occurred (liver failure). Of the 301 trAEs, 11% were managed with treatment interruption. Treatment was discontinued for another 12% of trAEs of any cause; however, treatment discontinuation due to a trAE was considered by the investigator to be nivolumab-related in only 4.4% of the patients. E U R O P E A N U R O L O G Y F O C U S X X X ( X X X X ) X X X -X X X

Quality of life
The number of evaluable baseline questionnaires completed by the 228 NORA patients was 201 for the EQ-5D Visual Analog Scale (VAS) and 213 for FKSI-19. Up to 18 mo, more than half of the overall population were eligible for PRO assessments. The proportion of EQ-5D VAS and FKSI-19 questionnaires completed by patients with ongoing treatment was 66% and 77% at 6 wk, 57% and 62% at 6 mo, and 48% and 53% at 12 mo, respectively. PRO results are Please cite this article as: Marc-Oliver Grimm reported for time points up to 36 mo (Fig. 4), but were only interpreted up to 12 mo because of the low number of patients with ongoing treatment and the small proportion of questionnaires completed at later time points. EQ-5D VAS revealed no change in quality of life (QoL) during the first year of nivolumab treatment. RCC-specific PROs reported with FKSI-19 also remained stable during the first year of nivolumab treatment.

Discussion
NORA recorded real-world data for nivolumab monotherapy in aRCC. In comparison to the pivotal CM025 study [2], NORA patients were older (median age 70 vs 62 yr), had lower KPS (14% with KPS 70 vs 5.9% in CM025) and more patients had metastases of the bone (32% in NORA vs 19% in CM025). Of all patients, 4.8% had brain metastases and 17% had ncc histology, which were both excluded from CM025. In addition, 8.3% of patients had received three or more prior tumor therapies in NORA, while in CM025 a maximum of two prior antiangiogenic therapies were allowed. The distribution of IMDC risk categories also differed, with fewer patients having poor risk disease in the NORA population (15% vs 23% in CM025; information missing for 12% in NORA). In total, only 46% of the patients in NORA would have been eligible for CM025, the majority because of ncc histology, poor performance status, and prior treatments (data not shown). Nonetheless, our study suggests that the effectiveness of nivolumab under real-world conditions is comparable to that in the pivotal RCT, supporting similar data from a real-world study by the Italian Nivolumab Renal Cell Cancer Early Access Programme (IEAP) group. The IEAP population was comparable to the NORA cohort, with more patients with favorable IMDC risk (20% favorable, 29% intermediate, and 11% poor risk), fewer patients with nccRCC (6.7%), and more patients with brain metastases (8.2%) [11]. Median OS was consistent between NORA (24 mo, 95% CI 19-NR) and CM025 (25 mo, 95% CI 22-NR); the 1-yr and 2-yr survival rates were 65% (95% CI 59-71%) and 51% (95% CI 44-58%) in NORA, compared to 76% and 52% for CM025, respectively (minimum follow-up 26 mo) [2,12]. The IEAP study reported a similar 1-yr survival rate of 63% (median follow-up 12 mo) [11]. Median PFS was similar across NORA, CM025, and IEAP (5.3 vs 4.2 vs 4.5 mo). The numerical difference in the 24-mo PFS landmark favors NORA (21% vs 14% in CM025) [13]. The differences in imaging assessments and evaluation of progression may contribute to these outcomes. ORR varied numerically among the studies, but remained comparable (NORA 20%, CM025 26%, IEAP 23%). Complete response rates were similar across the studies (NORA 1.3%, CM025 1.0%, IEAP 0.8%) [2,11].   Table 4 -Treatment-related adverse events among patients treated with nivolumab monotherapy for advanced renal cell carcinoma after at least one prior treatment line: reported for all grades (if incidence 1%) and grade 3-4 (all events), n=228 a Patients, n (%)

All grades
Grade 3-4 Number of treatment-related adverse events 301 49 Patients with at least one treatment-related adverse event 105 (46) 34 (15) Blood and lymphatic system disorders Since the NORA real-world population features less favorable characteristics, it is not surprising that the results differ compared to the pivotal CM025 RCT. One such feature is the presence of ncc histology, associated with worse prognosis in comparison to cc histology [14] and accounting for 17% of NORA patients, while nccRCC was excluded from CM025. Although the results are numerically lower than outcomes for ccRCC patients, nivolumab treatment was effective in nccRCC patients, as shown by median OS, median PFS, and ORR data. Of the NORA nccRCC population, 11% had chromophobe tumor histology, which probably contributed to the lower effectiveness of nivolumab in this subpopulation in comparison to the ccRCC group and the entire cohort. This assumption is based on observations for pembrolizumab, another PD-1 ICI: in the first-line setting, an ORR of 26% was observed for the overall nccRCC population, but only 9.5% of patients with chromophobe RCC responded [15]. Our data are also consistent with recently published results from a phase 3b/4 safety study on nivolumab monotherapy (CheckMate 374) that suggest somewhat inferior outcomes for nccRCC in comparison to ccRCC [16,17]. However, since other available treatments such as VEGFR TKIs are also less effective in nccRCC, nivolumab remains an effective and well-tolerated treatment option [18].
NORA patients with poor performance status (KPS <70, excluded from CM025) had notably worse OS outcomes compared to those with KPS 70. However, the number of patients with KPS <70 was low. Brain metastases, accounting for 4.8% of NORA patients and excluded from CM025, are also considered unfavorable and have been associated with lower ICI antitumor activity [19].
A noteworthy proportion of patients in NORA were aged 75 yr (38% vs 8% in CM025) [2]. In our study, OS and PFS did not differ by patient age, demonstrating that elderly patients also benefit from nivolumab. This extends and sup- E U R O P E A N U R O L O G Y F O C U S X X X ( X X X X ) X X X -X X X ports a subgroup analysis for CM025 that revealed that patient groups aged <65 yr and 65 years benefit similarly from nivolumab in comparison to everolimus [20]. In addition, ORR, DCR, and estimated OS curves in the IEAP study were similar for elderly patients and the overall population [21].
Interestingly, the presence of bone metastases in NORA had hardly any influence on OS outcomes, again in line with the IEAP data [11]. This finding is of particular interest, as cabozantinib, an alternative to nivolumab in the pretreated aRCC setting, has been suggested as a treatment option for these patients [22].
An association was found between OS and IMDC risk, an accepted prognostic factor in aRCC [9]. As expected, favorable risk was associated with the highest survival probabilities, consistent with results from CM025 and IEAP data [11,20].
The proportion of patients with trAEs was much lower in NORA than in CM025 (all grades, 46% vs 79%; grade 3-4, 15% vs 19%) [2]. Differences in the reporting system for a real-world NIS compared to the strict intervals for interrogation and reporting in RCTs are likely to be the cause of this discrepancy. However, no new safety signals emerged and, similar to CM025, only 11% of trAEs in NORA led to permanent discontinuation of nivolumab (CM025: 8%) [2]. Overall, our data support the low toxicity of nivolumab in a broad real-world population.
In line with the safety profile observed, at least during the first 12 months with nivolumab, QoL is maintained. Long-term analysis of PROs in NORA is hampered by the limited number of questionnaires completed. However, our data are consistent with PROs from CM025 demonstrating stable or even improved QoL up to 3 yr after treatment initiation [3]. This constitutes an important aspect of disease management, especially considering the sometimes long treatment duration.
Our study has some limitations, mainly related to data capture. The lack of central pathology review, unstructured imaging modalities, and the limited utilization of RECIST evaluations in the real-world setting are key limitations. Nonetheless, our data suggest that nivolumab has a robust treatment effect and favorable tolerability profile.

Conclusions
The real-life setting of the NORA study demonstrates the effectiveness, safety, and tolerability of nivolumab monotherapy in a broad aRCC patient population after one or more prior systemic therapies. The outcomes are generally consistent with those reported in the pivotal phase 3 CM025 trial. NORA extends the evidence of effectiveness for nivolumab to subpopulations excluded from CM025.