A challenging frontier – the genomics and therapeutics of non-clear cell renal cell carcinoma

Abstract

Purpose of review:

As molecular profiling of renal cell carcinoma (RCC) continues to elucidate novel targets for non-clear cell histologies, understanding the landscape of these targets is of utmost importance. In this review, we highlight the genomic landscape of non-clear cell RCC and its implications for current and future systemic therapies.

Recent findings:

Several genomic studies have described the mutational burden among non-clear cell histologies. These studies have highlighted the importance of MET in papillary RCC and led to several clinical trials evaluating the efficacy of MET inhibitors for papillary RCC. The success of immune checkpoint inhibitors, such as ipilimumab and nivolumab, in clear cell RCC has led to ongoing trials evaluating these novel therapeutics in non-clear cell RCC.

Summary:

Genomic profiling has allowed for the evaluation of novel targets for non-clear cell RCC. This evolving therapeutic landscape is being explored in promising, ongoing trials that have the potential for changing how non-clear cell RCC is managed.

Introduction

Approximately 76,000 new cases and 14,000 deaths due to kidney cancer are expected in 2021 in the United states alone [1]. Ninety percent of these cases are due to renal cell carcinoma (RCC), which arises from the nephron and contains many histologic subtypes. Most patients with RCC (75–80%) have clear cell histology (ccRCC), while other common histologies are collectively referred to as non-clear cell RCC (nccRCC; Table 1). Treatment of RCC presents a unique challenge as about 30% of patients initially present with metastatic RCC and an additional one-third of RCC patients have recurrence with distant metastases after surgical resection [2,3]. The recent successes of targeted and immunotherapies have shifted the paradigm in the management of metastatic ccRCC [4]. However, therapeutic advances in nccRCC have been limited by the lower incidence and limited clinical trial successes. In this review, we will highlight the mutational landscape of nccRCC based on genomic studies and recent treatment advances for specific subtypes of nccRCC based on clinical trial evidence.

Table 1:

Mutations within subtypes of nccRCC

Type	Incidence	Chromosomal mutation	Mutations
Papillary	10–15%	Type I Gain of Ch 7, 16, 17 Loss of 1p36 Type II Loss of Ch 9p21, 3p, 14p, 22q Gain of 5q, 7–16, 17	Type I MET Type II CDKN2A, SETD2, BAP1, PBRM1, NRF2-ARE, TFE3 fusions
Chromophobe	4–5%	Loss of Ch 1, 2, 6, 10, 13, 17, 21	TP53, PTEN, mTOR, TERT, FAAH2, PDHB, PDXDC1, ZNF765
Collecting duct	<1%	Loss of 8p, 16p, 1p, 9p Gain at 13q	mtDNA genome mutations
Translocation		Xp11.23, 6p21	BIRC7 expression
Renal medullary carcinoma			SMARCB1
Unclassified			NF2, SETD2, BAP1, KMT2C, MTOR, TSC1, TSC2, PTEN
Sarcomatoid†	5%		TP53, VHL, CDKN2A, NF2

^†can coexist with any RCC histology

Ch: chromosome; mtDNA: mitochondrial DNA

The Genomic Landscape of nccRCC

Over the last three decades the mutational landscape within ccRCC tumors has been well characterized through multiple large-scale genomic studies [5–8]. While ccRCC has a wide variety of mutations, the loss of chromosome 3p is reported in ~90% of sporadic ccRCC, with inactivating mutations and loss of heterozygosity in the VHL gene seen in >90% of ccRCC tumors [5,9,10]. Studying the consequences of VHL alterations has helped uncover aberrant degradation of hypoxia-inducible factor 1 and 2 alpha (HIF1-2α), which in turn promotes pathologic cell proliferation and angiogenesis via upregulation of vascular endothelial growth factor, platelet-derived growth factor, and transforming growth factor-α, which promote tumorigenesis. Discovery of these pathways has helped uncover a myriad of novel therapeutics for treating advanced and/or metastatic ccRCC [4,11].

Papillary RCC (pRCC) type I and II and chromophobe RCC (chRCC) represent 10–15% and 4–5%, respectively, of all RCC. Other nccRCC histologies such as medullary, translocation, collecting duct, and unclassified types, while rare, also constitute nccRCC histologies. Similar to ccRCC, specific chromosomal rearrangements and mutational profiles have been described for various histologies of nccRCC (Table 1) and many are associated with hereditary RCC syndromes (Table 2) [12–17]. While detection of these mutations has been important for classification and diagnosis, it has been difficult to translate these discoveries for therapeutic purposes.

Table 2:

Hereditary RCC syndromes with nccRCC histologies

Syndrome	Gene	Inheritance	RCC Histology	Clinical Features
Von Hippel-Lindau	VHL	AD	Clear cell	Retinal angiomas, bilateral multifocal RCC, Brain, spine, or retinal hemangioblastomas
Hereditary papillary renal carcinoma	MET	AD	Type I papillary	Bilateral multifocal RCC
Birt-Hogg-Dubé	FLCN	AD	Chromophobe	Cutaneous fibrofolliculoma Pulmonary cysts Spontaneous pneumothorax
Tuberous sclerosis complex	TSC1 TSC2	AD	Angiomyolipoma Clear cell Cystic oncocytoma	Angiofibromas Shagreen patches Retinal nodular hamartomas Lymphangioleiomyomatosis Multiple renal cysts
Hereditary leiomyomatosis and RCC	FH	AD	Type II papillary	Leiomyomas of uterus and skin Adrenal adenoma Aggressive renal tumors
BAP1 tumor predisposition syndrome	BAP1	AD	Clear cell Chromophobe	Melanoma RCC Mesothelioma
Hereditary paraganglioma-pheochromocytoma syndrome	SDHA SDHB SDHC SDHD	AD	Clear cell Chromophobe Type II papillary Oncocytoma	Head and neck paraganglioma Adrenal or extra-adrenal pheochromocytoma GIST tumors Lung lesions
Cowden syndrome	PTEN	AD	Papillary Chromophobe Clear cell	Mucocutaneous lesions Breast cancer Follicular thyroid cancer Endometrial cancer Others

AD: autosomal dominant; RCC: renal cell carcinoma

Papillary RCC

pRCC is the second most common RCC histology and is subdivided based on morphologic features into types I and II [18]. The initial work on characterizing pRCC identified mutations in the MET proto-oncogene among patients with the hereditary (HPRC) and sporadic type I pRCC and mutations in FH in type II pRCC [19,20]. Following this, evaluation of 220 samples of sporadic pRCC by Albiges et al. found that MET expression was significantly elevated across all types of pRCC and that copy number alterations were present in 81% and 46% of type I and II pRCC, respectively [21]. Additional characterization of nccRCC demonstrated that MET, NF2, SLC5A3, PNKD, and CPQ were mutated in pRCC [13]. Approximately 15% of pRCC tumors in this study contained mutations in MET. The Cancer Genome Atlas subsequently performed the largest multidimensional genomic mapping study with whole exome sequencing, copy-number analysis, mRNA and miRNA sequencing, methylation analysis, and proteomic analysis on 161 pRCCs, which was expanded to include 274 pRCCs from among a larger pool of 843 RCCs representing a comprehensive study of all the major RCC subtypes [9]. This study confirmed differences in types I and II pRCC and identified 4 distinct molecular groupings with progressively worse survival (C1, C2a, C2b, and C2c) [12]. C1 comprised of predominantly type I pRCC tumors, which had a gain of chromosome 7 and 17. Among type I pRCC tumors, somatic mutations in MET were present in 17% of these tumors.

C2a was predominantly type II pRCC and was associated with early-stage tumors and a distinct DNA methylation pattern, whereas C2b consisted exclusively of type II pRCC, presenting with later stage (III and IV) tumors and mutations in SETD2. Tumors in the C2c group were associated with CpG island methylator phenotype (CIMP) and had the worse overall survival among all pRCC tumors. This study also identified other mutations found in type II pRCC, including the deletion of CDKN2a, gene fusions involving the microphthalmia family (MiTF) members TFE3 and TFEB, mutations in chromosomal remodeling genes, activation of the NRF2 oxidative stress pathway, and mutations in FH which are seen in both sporadic Type II pRCC and hereditary leiomyomatosis and renal cell cancer (HLRCC) [22–24].

Chromophobe RCC

Germline mutations have also been described with chRCC. For example, germline mutation in FLCN in Birt-Hogg-Dubé syndrome is associated with increased likelihood of chRCC [25]. Additionally, germline mutations in PTEN causes Cowden syndrome, in which 5–29% patients can develop various histologies of RCC, including the chromophobe subtype. Comprehensive profiling has been used by Davis et al. and Ricketts et al. to characterize somatic mutations among 66 and 81 primary chRCC tumors, respectively [9,14]. Loss of one copy of chromosomes 1, 2, 6, 10, 13, and 17 was seen in 86% of cases. Furthermore, mitochondrial DNA mutations were implicated in upregulation of oxidative phosphorylation among chRCC compared to ccRCC and normal kidney cells. Additionally, whole genome sequencing identified a pattern of localized hypermutation within the TERT promoter region that led to increased TERT expression, suggesting these genetic mutations may lead to immortalization and a mechanism for metabolic upregulation. Other profiling studies by Durinck et al. identified mutations in TP53, PTEN, FAAH2, PDHB, PDXDC1, and ZNF765 among the chRCC tumors [13].

Other rare nccRCC Histologies

Other rare nccRCC histologies include translocation RCC, collecting duct RCC, sarcomatoid RCC, and unclassified RCC. MiTF family translocation RCC is characterized by translocation involving Xp11.23 and 6p21, with tumors predominantly occurring in children and young adults [26]. Gene fusions of MiTF genes TFE3 and TFEB occur with different fusion partners. Collecting duct carcinoma is an extremely rare type of nccRCC with highly aggressive behavior. Frequent alterations of NF2 and SMARCB1 were initially identified and further studies highlighted mutations in CDKN2A and SLC7A11, which was associated with cisplatin-resistance [27,28]. Sarcomatoid RCC is a histologic subtype that can occur with any other type of RCC [29]. This histology is associated with poor clinical outcomes, but treatment with systemic immunotherapy has shown promising response rates [30,31].

Comparative genomic profiling of 26 sarcomatoid RCC against 56 ccRCC tumors and datasets from The Cancer Genome Atlas that included ccRCC, pRCC, and chRCC demonstrated that TP53, VHL, CDKN2A, and NF2 were most commonly mutated among the sarcomatoid RCC [32]. This study highlighted that sarcomatoid RCC could be subdivided based on mutations in TP53 or NF2. Additional profiling efforts by Wang et al. comparing sarcomatoid RCC to other RCC subtypes demonstrated that sarcomatoid RCC is less likely to have two-hit loss of VHL and PBRM1 and is more likely to have mutations in PTEN, TP53, and RELN, which are associated with worse prognosis [33]. Further profiling of metastatic sarcomatoid RCC using RNA-seq by Pal et al. demonstrated elevated aurora kinase A activity, which was associated with increased activity within the mammalian target of rapamycin (mTOR) pathway [34]. Unclassified RCC are comprised of aggressive nccRCC that have no standard therapy. A thorough molecular analysis of 62 unclassified RCC identified somatic mutations in 29 genes, which included NF2, SETD2, BAP1, KMT2C, and MTOR [35]. This study further demonstrated that 26% of unclassified RCC with loss of NF2 and dysregulated Hippo-YAP pathway were associated with worse survival, whereas 21% of unclassified RCC with mutations in MTOR, TSC1, TSC2, or PTEN, and increased mTOR signaling were associated with better clinical outcome.

Comprehensive profiling efforts have helped uncover numerous genetic mutations among the various histologic subtypes of nccRCC, which are often overlapping. These discoveries are still nascent efforts towards better understanding the complex, intersecting pathways that promote tumorigenesis among nccRCC. Understanding the molecular underpinnings of nccRCC is of utmost importance in further categorization of nccRCC and improved targeting with novel therapeutic strategies.

nccRCC trials and tribulations

The therapeutic armamentarium available to treat advanced and metastatic ccRCC expanded rapidly over the last decade [4]. However, the same therapeutic strategies that have been successful in ccRCC have displayed muted responses in patients with nccRCC (Table 3). Therefore, for metastatic nccRCC the National Comprehensive Cancer Network (NCCN) guidelines recommend enrollment in clinical trials or treatment with the vascular endothelial growth factor receptor (VEGFR)-inhibitor sunitinib [36].

Table 3:

Clinical trials reporting data on locally advanced or metastatic nccRCC

Treatment	Clinical trial	Mechanism of action	Control	Cohort (Line of therapy)	Endpoint(s)
Sunitinib	SUPAP [44] NCT00541008	VEGFR & PDGFR inhibitor	N/A	pRCC (1st)	Type I Median PFS: 6.6 mo, 95% CI 2.8–14.8 mo Median OS: 17.8 mo, 95% 5.7–26.1 mo Type II Median PFS: 5.5 mo, 95% CI 3.8–7.1 mo Median OS: 12.4 mo, 95% 8.2–14.3 mo
Everolimus	RAPTOR [68] NCT00688753	mTOR inhibitor	N/A	pRCC (1st)	Median PFS: 4.1 mo, 95% CI 3.6–5.5 mo Median OS: 21.4 mo, 95% CI 15.4–28.4 mo
Sunitinib	ESPN [45] NCT01185366	VEGFR & PDGFR inhibitor	Everolimus	nccRCC Sarcomatoid RCC	Median PFS: 6.1 vs 4.1 mo, p=0.6 Median OS: 16.2 vs 14.9 mo, p=0.18
Sunitinib	ASPEN [46]	VEGFR & PDGFR inhibitor	Everolimus	nccRCC	PFS: HR1.41, 80% CI 1.03–1.92; p=0.16
Sunitinib (1st) Everolimus (2nd)	RECORD-3 [69] NCT00903175	VEGFR & PDGFR inhibitor mTOR inhibitor	Everolimus (1st) Sunitinib (2nd)	ccRCC (1st) nccRCC (1st)	PFS: HR 1.2, 95% CI 0.9–1.6 Median OS: HR 1.1, 95% 0.9–1.4
Sunitinib	NCT01219751 [70]	VEGFR & PDGFR inhibitor	N/A	nccRCC	Median PFS: 6.4 mo, 95% CI 4.2–8.6 mo 1-year PFS rate: 40%
Foretinib	NCT00726323 [47]	MET & VEGFR2 inhibitor	N/A	pRCC (1st/2nd)	ORR: 13.5% Median PFS: 9.3 mo
Crizotinib	CREATE [48] NCT01524926	MET, ALK, ROS1 inhibitor	N/A	Type I pRCC (1st)	MET+ ORR: 50%, 95% CI 6.8–93.2% 1-yr PFS: 75%, 95% CI 12.8–96.1% 1-yr OS: 75%, 95% CI 12.8–96.1% MET− ORR: 6.3%, 95% CI 0.2–30.2% 1-yr PFS: 27.3%, 95% CI 8.5–50.4% 1-yr OS: 71.8%, 95% CI 41.1–88.4%
Erlotinib	SWOG S0317 [71] NCT00060307	EGFR inhibitor	N/A	pRCC	ORR: 11%, 95% CI 3–24% Median OS: 27 mo, 95% CI 13–36 mo
Tivantinib Erlotinib	SWOG S1107 [72] NCT01688973	MET inhibitor EGFR inhibitor	Tivantinib	pRCC (1st/2nd)	RR 0% (Study closed early) Median PFS 2 mo vs 3.9 mo
Savolitinib	SAVOIR [49] NCT03091192	MET inhibitor	Sunitinib	MET-driven, pRCC (1st/2nd)	PFS: HR 0.71, 95% CI 0.37–1.36, p=0.31 OS: HR 0.51, 95% CI 0.21–1.17, p=0.11
Savolitinib Durvalumab	CALYPSO [52,53] NCT02819596	MET inhibitor PD-L1 inhibitor	N/A	pRCC (1st/2nd)	ORR: 27% Median PFS: 4.9 mo, 95% CI 2.5–12 mo
Atezolizumab Bevacizumab	NCT02724878 [73]	Anti-PD-L1 mAb Anti-VEGF mAb	N/A	nccRCC Sarcomatoid RCC	ORR: 26% nccRCC and 50% sarcomatoid Median PFS: 8.3 mo, 95% CI 5.7–10.9 mo
Ipilimumab Nivolumab	CheckMate 214 [55] Post hoc analysis	Anti-PD1 mAb Anti-CTLA4 mAb	Sunitinib	Sarcomatoid RCC	Median OS: HR 0.45, 95% CI 0.3–0.7, p=0.0004 PFS: HR 0.54, 95% CI 0.33–0.86, p=0.0093 ORR: 60.8% vs 23.1% CR: 18.9% vs 3.1%
Atezolizumab Bevacizumab	IMmotion151 [54] Subgroup analysis	Anti-PD-L1 mAb Anti-VEGF mAb	Sunitinib	Sarcomatoid RCC	PFS: HR 0.52, 95% CI 0.34–0.79 ORR: 49% vs 14% CR: 10% vs 3%

PD-L1: programmed death-ligand 1; VEGFR2: vascular endothelial growth factor receptor 2; EGFR: epidermal growth factor receptor; PDGFR: platelet-derived growth factor receptor; mTOR: mammalian target of rapamycin; PD-1: programmed cell death protein 1; CTLA4: cytotoxic T-lymphocyte associated protein 4; mAb: monoclonal antibody; pRCC: papillary renal cell carcinoma; nccRCC: non-clear cell renal cell carcinoma; ccRCC: clear cell renal cell carcinoma; PFS: progression-free survival; OS: overall survival; ORR: overall response rate; CR: complete response; HR: hazard ratio

Clinical trials for nccRCC have included a large variety of nccRCC histologies, making interpretation and clinical implementation difficult. Conversely, histology-specific trials have faced a unique challenge of accrual given the lower incidence of certain histologies. As a result, current evidence is mostly comprised of retrospective analyses, subgroup analyses of expanded access program studies, or small phase II studies [37–40]. Furthermore, translating the aforementioned genomic targets into viable therapeutic targets or biomarkers for assessing treatment efficacy have compounded these issues. Nevertheless, efforts are underway to determine whether the success with targeted therapies and immune checkpoint inhibitors in treating advanced ccRCC can be extended to advanced nccRCC (Table 4); the results of such studies may potentially transform the management paradigm of nccRCC.

Table 4:

Ongoing clinical trials in locally advanced or metastatic nccRCC

Clinical trial	Phase	Experimental Arm	Control Arm	Cohort (Line of therapy)	Primary Endpoint(s)
PAPMET, SWOG S1500 NCT02761057	II	Cabozantinib Crizotinib Savolitinib Sunitinib	N/A	pRCC (1st/2nd)	PFS
NCT02019693	II	Capmatinib	N/A	pRCC (1st/2nd/3rd)	ORR
NCT01130519	II	Bevacizumab+Erlotinib	N/A	pRCC (1st/2nd/3rd) HLRCC (1st/2nd/3rd)	ORR
NCT02495103	I/II	Vandetanib+Metformin	N/A	HLRCC SDH-associated pRCC	ORR
CABOSUN 2 NCT03541902	II	Cabozantinib	Sunitinib	Non-clear cell (1st/2nd)	PFS
BONSAI NCT03354884	II	Cabozantinib	N/A	Collecting duct	ORR
NCT02915783	II	Lenvatinib+Everolimus	N/A	Non-clear cell (1st)	ORR
CYTOSHRINK NCT04090710	II	Ipilimumab+Nivolumab+SBRT	Ipilimumab+Nivolumab	Clear cell (1st) Non-clear cell (1st)	PFS
SUNIFORECAST NCT03075423	II	Ipilimumab+Nivolumab	Sunitinib	Non-clear cell (1st)	OS at 1 year
CA209-9KU NCT03635892	II	Nivolumab+Cabozantinib	N/A	Non-clear cell (1st/2nd)	ORR
NCT03170960	I/II	Atezolizumab+Cabozantinib	N/A	Clear cell (1st) Non-clear cell (2nd)	MTD ORR
ANZUP1602, UNISoN NCT03177239	II	Nivolumab (1st) followed by Ipilimumab+Nivolumab (2nd)	N/A	Non-clear cell (1st/2nd)	ORR
CheckMate-920 NCT02982954	IIIb/IV	Ipilimumab+Nivolumab	N/A	Clear cell (1st) Non-clear cell (1st)	IMAE
KEYNOTE 427 NCT02853344	II	Pembrolizumab	N/A	Clear cell (1st) Non-clear cell (1st)	ORR
CheckMate 374 NCT02596035	IIIb/IV	Nivolumab	N/A	Clear cell (1st/2nd/3rd) Non-clear cell (1st/2nd/3rd)	IMAE
OMNIVORE NCT03203473	II	Ipilimumab+Nivolumab	N/A	Clear cell Non-clear celll	PR/CR at 1 year after discontinuing nivolumab PR/CR at 1 year after adding ipilmumab
AREN1721 NCT03595124	II	Axitinib+Nivolumab	Axitinib only Nivolumab only	Translocation RCC	PFS
PROSPER RCC NCT03055013	III	Perioperative Nivolumab and Nephrectomy	Nephrectomy alone	Biopsy proven RCC ≥T2Nx or TanyN+ or M1*	EFS

^*M1 disease has to be resected/definitively treated at the same time or within 12-weeks of surgery such that patient is considered “No Evidence of Disease”

DLT: dose-limiting toxicity; RCC: renal cell carcinoma; ccRCC: clear cell RCC; pRCC: papillary RCC; HLRCC: hereditary leiomyomatosis and RCC; PFS: progression-free survival; OS: overall survival; ORR: overall response rate; MTD: maximum tolerated dose; IMAE: incidence of high-grade immune-mediated adverse events; EFS: event-free survival; PR: partial response; CR: complete response; N/A: not applicable

Treatment landscape of nccRCC

The therapeutic strategies in nccRCC have predominantly been adapted from the successes of various agents in the management of ccRCC (Table 3 & 4) [4]. With the approval of multiple VEGF and mTOR inhibitors for metastatic ccRCC in the early 2000s [41–43], investigators consequently performed post-hoc analysis of sunitinib, sorafenib, and temsirolimus in nccRCC patients. [37–40]. nccRCC patients had less robust responses to these systemic therapies compared to their ccRCC counterparts. However, the studies collectively demonstrated objective response rates of 5–15% and provided support for targeting the mTOR and VEGF pathways. Subsequently, a prospective phase II trial, SUPAP, demonstrated median progression-free survival (PFS) of 5–7 months in patients who received sunitinib as first-line therapy for metastatic type I and II pRCC [44]. Two phase II trials, ESPN and ASPEN, evaluated sunitinib versus everolimus in metastatic nccRCC [45,46]. The ESPN trial included patients with a mix of nccRCC histology and ccRCC harboring >20% sarcomatoid histology. The trial did not demonstrate a significant difference in median PFS (6.1 vs 4.1 mo, p=0.6) or OS (16.2 vs 14.9 mo, p=0.18) between sunitinib and everolimus [45]. While the ASPEN trial showed a significant increase in median PFS in patients treated with sunitinib compared to everolimus (8.3 vs 5.6 mo, p=0.16), this analysis was done with a prespecified Type I error of 20% compared to the conventional 5% and did not translate to significant increase in OS (HR 1.12, 95% CI 0.7–2.1, p=0.60) [46]. Patients with good risk disease (HR 2.9, 80% CI 1.5–5.7) and papillary histology (HR 1.6, 80% CI 1.1–2.3) had a longer PFS with sunitinib compared to everolimus.

Following the availability of additional genomic data, therapeutic development for nccRCC focused on targets with high mutation rates that could be easily targeted. For example, the MET gene is mutated among 17% of pRCC [12,13]. A phase II trial looking at the efficacy of foretinib, a dual MET/VEGFR inhibitor, among patients with pRCC demonstrated median PFS of 9.3 months, and with 50%, 20%, and 9% partial response (PR) among patients with germline MET mutation, somatic MET mutation, and no MET mutation, respectively [47]. Despite these promising results, further development of this inhibitor was discontinued. Similar results were obtained when crizotinib, a MET/ALK/ROS1 inhibitor, was evaluated for type I pRCC in the CREATE trial [48]. In MET+ patients, 50% had PR with 1-year PFS rate of 75% compared to 6.3% PR and 1-year PFS rate of 27.3% in MET- patients. A potent MET inhibitor, savolitinib was recently evaluated in the phase III SAVOIR trial [49], building on a previous study that demonstrated significant improvement in PFS among patients with MET-driven pRCC [50]. Patients underwent baseline genomic profiling and those with MET-driven disease were randomized to sunitinib or savolitinib. Unfortunately, this study was discontinued due to poor accrual and did not demonstrate a statistical difference in PFS or OS between the two treatment arms.

Immune checkpoint inhibitors are effective in ccRCC, but data on their activity in nccRCC has been lacking until recently. A retrospective analysis of patients who had received nivolumab, an anti-PD1 antibody, for various nccRCC histologies demonstrated median PFS of 3.5 months and median OS that was not reached [51]. However, data from this study is difficult to interpret given several confounders such as different histologies, selection bias inherent to retrospective studies, and short follow up time.

More recently, preliminary results from a phase II study evaluating savolitinib in combination with durvalumab, an anti-PD-L1 antibody, for metastatic pRCC demonstrated ORR of 27% with median PFS of 4.9 months [52,53]. Notably patients receiving first-line therapy had a response rate of 33% and median PFS of 12.3 months. Interestingly, PD-L1 and MET expression were not associated with higher response rates or longer OS [53]. Further supporting the strong activity of the concept of dual inhibition, McGregor et al. performed a multicenter phase II trial to evaluate atezolizumab combined with bevacizumab in patients with nccRCC or sarcomatoid RCC; they demonstrated a remarkable ORR among nccRCC (26%) and sarcomatoid RCC (50%), with median PFS of 8.3 months.

This data further corroborates a subgroup analysis from the IMmotion 151 clinical trial, which evaluated these agents against sunitinib in metastatic ccRCC; this trial demonstrated higher median PFS (8.3 vs 5.3 mo) and ORR (49% vs 14%) among patients with sarcomatoid RCC [54]. Post-hoc analysis of patients with sarcomatoid histology from a similar trial, CheckMate 214, that evaluated the combination of nivolumab and ipilimumab against sunitinib for metastatic ccRCC demonstrated a higher OS (HR 0.45, 95% CI 0.3–0.7, p=0.0004) and improved PFS (HR 0.54, 95% CI 0.33–0.86, p=0.0093), with ORR of 60.8% in favor of nivolumab and ipilimumab [55]. Taken together, these results suggest that combination strategies may yield promising results in both nccRCC and sarcomatoid RCC. Further efforts to perform post-hoc analyses from pivotal phase III trials will help better tailor the use of targeted therapies and immunotherapies.

Ongoing clinical trials

The results of the aforementioned trials and other ongoing trials are leading to significant progress towards developing therapeutic regimens for metastatic nccRCC and sarcomatoid RCC. (Table 4). Notably, the SWOG S1500 (PAPMET) trial is prospectively assessing the role of targeting MET in patients with pRCC by randomizing an estimated 180 patients to either sunitinib, crizotinib, cabozantinib, or savolitinib [56]. The crizotinib and savolitinib arms were prematurely closed based on futility analysis and the final analysis, which is expected in 2021, will evaluate the efficacy of sunitinib and cabozantinib. Together with the CABOSUN II trial, this trial will help establish the role of cabozantinib, a MET/VEGFR/AXL inhibitor, in treating metastatic nccRCC, especially pRCC.

The paradigm-shifting phase III clinical trials, including KEYNOTE 426, JAVELIN 101, and CheckMate 214, that have transformed the management of metastatic ccRCC have suggested a potential role for combination therapy with checkpoint inhibitors in nccRCC [57–59]. CheckMate 920 and CheckMate 374 are phase IIIb/IV trials to confirm the safety and efficacy of ipilimumab+nivolumab and nivolumab alone, respectively, and thereby may help expand the indications for these regimens to include nccRCC [60,61]. In addition, the SUNIFORECAST trial will evaluate the efficacy of ipilimumab+nivolumab as first-line therapy against sunitinib in nccRCC in a European multicenter phase II trial [62]. The combination of a checkpoint inhibitor and a tyrosine kinase inhibitor, nivolumab and cabozantinib, is also being assessed in the CA209–9KU trial.

Preliminary results from NCT01130519, a trial of bevacizumab and erlotinib in patients with advanced HLRCC or sporadic pRCC, demonstrated a median PFS of 14.3 months and overall response rates of 54.2%, 72%, and 35% in the entire, HLRCC, and sporadic pRCC cohorts, respectively [63]. This is the first prospective study in HLRCC and suggests encouraging activity of bevacizumab and erlotinib particularly in this cohort. Final results from the overall study are eagerly awaited.

One notable addition to the list of ongoing clinical trials is the PROSPER RCC trial; in this study patients with biopsy-proven RCC of any histology who have locally advanced or oligometastatic disease that can be completely resected via nephrectomy receive perioperative nivolumab versus nephrectomy alone [64,65]. This study is the first-of-its-kind by including any RCC histology and incorporating a checkpoint inhibitor earlier in the treatment course.

The multitude of ongoing trials for nccRCC is an important step towards developing a variety of biology-driven therapies for nccRCC. Genomic and metabolic studies have clearly demonstrated key differences between ccRCC and nccRCC and among the various subtypes of nccRCC [9]. Leveraging this vast space of genomic information to target nccRCC is seeing new light in the era of better targeted and checkpoint inhibitor therapies.

Conclusions

Comprehensive profiling techniques have revolutionized our understanding of the mutational burden that underlies various nccRCC. While these efforts have helped in understanding the basis of mutations underlying both RCC-associated syndromes and sporadic RCC (Table 2) [66], the studies have revealed an extraordinary number of mutations associated with nccRCC, underscoring the great diversity of nccRCC [12–14,67]. Validation of these targets has yet to be thoroughly performed, leaving a significant gap between target identification, role in pathogenesis and drug development. Using previous and prospective genomic profiling may help identify biomarkers that better point to which therapies are best suited for which histologic subtypes. Signaling pathways and complex tumor microenvironment interactions could undermine drug efficacy, especially in histologies with significant heterogeneity. Furthermore, selection pressures from multiple lines of therapy could also result in metabolic alterations and responses that might lead to decreased efficacy.

While our understanding of nccRCC continues to evolve, a robust platform for understanding the complexities of tumor biology is needed. The results of the SWOG S1500 trial will be met with enthusiasm as it will provide clarity on the clinical significance of MET inhibition in pRCC and the efficacy of cabozantinib in nccRCC. The expansion of clinical trials into the nccRCC space provides optimism as newer, proven ccRCC treatment strategies, such as checkpoint inhibition and combinatorial therapy, are being applied to nccRCC and sarcomatoid RCC. A concerted effort needs to be made to include multiple RCC histologies in future clinical trials with planned subgroup analyses in order to further advance the treatment of both clear cell and non-clear cell RCC.

Key Points.

• Non-clear cell RCC is comprised of various histologies with significantly different genetic mutations, posing a challenge for translating genomic mutations into clinically significant targets.

• Successes of therapeutics in clear cell RCC has not yielded similar results for non-clear cell RCC; thus, clinical trial participation should be considered for all advanced or metastatic non-clear cell RCC cases.

• Ongoing trials with novel combination therapies provide a promising future for treating patients with non-clear cell RCC.