Purpose:

Translocation renal cell carcinoma (tRCC) is a rare, aggressive renal cell carcinoma (RCC) subtype. There is currently limited understanding on the role of molecular alterations in the pathogenesis and progression of these tumors. We investigated the association between somatic alterations and clinical outcomes in two independent cohorts profiled using DNA sequencing.

Experimental Design:

Twenty-two tRCCs underwent targeted sequencing [Memorial Sloan Kettering Cancer Center (MSK)-IMPACT]; a subset was profiled using exome-sequencing and combined with exome data from The Cancer Genome Atlas (TCGA) for analysis. The prognostic value of specific somatic aberrations, tumor mutation burden (TMB), and fraction of copy-number–altered genome (FCNAg) was explored. In TCGA cases, neoantigen prediction and immune cell deconvolution were performed using RNA-sequencing and exome data. Overall survival estimates were computed using the Kaplan–Meier method; time-on-treatment was calculated for 14 MSK-IMPACT patients who underwent systemic therapy. Associations between molecular features and outcomes were evaluated using nonparametric testing.

Results:

Copy-number aberrant tRCCs were associated with poor overall survival (P = 0.03). Pediatric patients had tumors with lower FCNAg (P = 0.01). In one adult case with two chronologically distinct tumor samples sequenced, we confirmed that copy-number events occurred early during evolution. TERT promoter mutations were found exclusively in high-stage tumors. We found that tRCCs displayed distinct angiogenesis and PD-L1 gene expression profiles compared with other RCC subtypes.

Conclusions:

Tumors molecularly defined by increased copy-number variations were associated with aggressive disease in tRCC. A higher burden of genomic events in adults compared with pediatric cases likely reflects a more aggressive clinical course. The unique immunophenotypic characteristics of tRCC merit further exploration.

Translational Relevance

Translocation renal cell carcinoma (tRCC) is a rare, aggressive form of kidney cancer. Currently, there are no established biomarkers to stratify this disease by aggressiveness or to evaluate treatment response. Our study found that certain copy-number variations (CNVs) were associated with disease aggressiveness in tRCC. By pooling our results and previously published data, we could confirm that tumors bearing 9p loss, 17q gain, or a generically high burden of CNVs were associated with poor survival in these patients. In addition, three pediatric cases with an indolent disease course were found to have a lower copy-number burden. The molecular features of patients showing exceptional responses to treatment are highlighted. We identified microenvironmental characteristics of tRCC tumors distinguishing them from common subtypes of kidney cancer. Finally, we explored the genetic evolution of tRCC in a patient with temporally distinct tumor samples and observed an early incidence of CNVs with later mutational diversification.

Microphthalmia-associated transcription factor (MiTF)-family translocation renal cell carcinoma (tRCC) constitutes approximately 4% and 41.5% of renal cell carcinoma (RCC) cases in adults and children, respectively, with pediatric cases generally showing more favorable outcomes. The pathognomonic event in this disease is a fusion between transcription factors TFE3 or TFEB (on chromosomal loci Xp11.2 and 6p21, respectively) to a potentially more transcriptionally active locus. MiTFs are involved in the regulation of genes related to cell growth, metabolism, and lysosomal biogenesis, but the specific pathways and molecular mechanisms triggering oncogenesis and driving aggressiveness in tRCC remain unclear (1–4). There are currently no prognostic biomarkers, nor standard-of-care therapy for advanced disease, in this rare and aggressive RCC subtype.

Prior analyses of recurrent genomic alterations in tRCC have been limited to relatively small cohorts and predominantly focused on the analysis of copy-number variations (CNVs). In a previous report of 16 primary tRCC specimens, distinct somatic CNVs were reported to be associated with poor prognosis. The most frequent CNVs in this cohort were 17q gain and 9p loss, the latter also representing a known prognostic feature of clear cell RCC (ccRCC; refs. 5, 6). In addition, only a small number of patients with advanced tRCC have been included in clinical trials of systemic therapy for RCC. Therefore, there is currently not enough information about the efficacy of different treatments and, as a result, no established standard of care (7–9). In a recent retrospective, multi-institutional study of 24 patients undergoing immune checkpoint blockade (ICB) therapy, to date the largest report on outcomes to systemic therapy in tRCC, the authors reported a benefit in patients harboring mutations in bromodomain-containing genes. However, an evaluation of copy-number (CN) status was not performed (10).

Here, we investigated the presence of recurrent genomic alterations (mutations and CNVs) and their association with prognosis in 22 patients prospectively sequenced at our institution using a combination of targeted and exome-wide sequencing. In addition, we explored the frequency of these events in a subset of cases with response to systemic therapy. We evaluated the reproducibility of our findings in an independent cohort of 14 tRCC cases from The Cancer Genome Atlas (TCGA) consortium. We further carried out neoantigen prediction of mutations and fusion-generated peptides and investigated the tumor microenvironment using RNA-sequencing (RNA-seq) data. All of the underlying mutation and CN data associated with our analysis is provided in the Supplementary Material and Methods as a public resource for research into this rare entity.

Patient cohorts

The study was conducted in accordance with the Declaration of Helsinki and was approved by an institutional review board. All subjects or subjects' guardians gave written informed consent.

Memorial Sloan Kettering Cancer Center-IMPACT cohort (discovery)

After obtaining institutional review board approval, our internal database was queried for patients with a histologic diagnosis of tRCC. Diagnoses were made by dedicated genitourinary pathologists and were based on positive IHC or the confirmation of an MiTF gene fusion by either FISH, anchored multiplex PCR assay (ARCHER), or next-generation sequencing (NGS). Because false-positive and false-negative results have been reported for IHC-based diagnoses, patients who did not have cytogenetic proof of a translocation were excluded from further analysis (Supplementary Fig. S1; ref. 11).

Memorial Sloan Kettering Cancer Center-exome cohort

A subset of the Memorial Sloan Kettering Cancer Center (MSK)-IMPACT cohort (n = 10) underwent additional whole-exome sequencing (WES). Sequencing methods are detailed in the relevant section below. An additional tumor sample, derived from the same patient as TCGA sample BQ-5887, was also submitted for exome sequencing.

TCGA-exome cohort (validation)

Non-MSK patients with tRCC were identified from TCGA pan-kidney (TCGA-KIPAN) cohort, with available RNA-seq and exome sequencing data. Clinical information for these cases was obtained from a publicly available dataset (www.firebrowse.org, TCGA data version 2016_01_28). MiTF gene fusion was confirmed using two independent datasets of transcript fusions across TCGA samples (12, 13). Baseline characteristics of these patients and sample metadata are available in Supplementary Tables S1 and S2.

MSK-IMPACT targeted sequencing

Tumor and matched-normal specimens were sent for targeted NGS using our previously validated sequencing panel (MSK-IMPACT), which targets 341, 410, or 468 actionable genes based on the platform version (14). In brief, DNA from tumor and matched blood normal specimens from each patient were extracted and sheared to create barcoded DNA libraries. Using captured DNA, all coding exons of cancer genes and a subset of polymorphic loci (for CN analysis) were sequenced. Deep sequencing was performed at an average of 500–1,000× coverage. After alignment to the reference human genome, somatic alterations (missense mutations, small insertions and deletions, structural rearrangements, and DNA CN changes) were identified using a bioinformatics pipeline described in detail previously (14). To assess oncogenicity, mutations were annotated using the oncology knowledge database OncoKB (15).

WES, processing, and mutation analysis

Alignment of FASTQ-formatted raw sequencing data to the human reference genome (b37) was performed using the Burrows–Wheeler aligner (BWA v.0.7.10). The genome analysis toolkit was used to perform local realignment (GATK v4.1.4.1; ref. 16). Duplicate reads were excluded using Picard v.2.13. Four different variant calling tools were used for mutation calling: MuTect2 (part of GATKv4.1.4.1; ref. 16), Strelka2 v2.9.10 (17), Varscan v2.4.3 (18), and Platypus (19). To pass filtering, mutations were required to be detected by at least two caller algorithms. To obtain highly accurate variant calls, ancillary filters were used. These filters included a base coverage of a minimum of 10 reads in the tumor, at least five reads supporting the mutation, and an allelic frequency below 2% in the matched blood sample. Furthermore, single-nucleotide variants identified at a frequency higher than 1% in dbSNP (20) or 1000 Genomes Project (21) were excluded. Mutation calls from both MSK-IMPACT and exome sequencing have been provided (Supplementary Table S3).

Allele-specific CN analysis

Allele-specific CN analysis was performed using our institutional bioinformatics algorithm FACETS (22). After genome segmentation, each segment was categorized as being diploid or nondiploid according to its CN integer values (ie, segments with a total CN ≠ 2 or minor CN ≠ 1, according to the FACETS outputs). In cases where FACETS failed to produce any estimates, the values from the closest adjacent segment (on the same chromosomal arm) were used instead. To compute the fraction of copy-number–altered genome (FCNAg), the length of each autosomal genomic segment (in base pairs) was determined. FCNAg was then calculated by adding the nondiploid segments and comparing their total length to the sum of all segments. Copy-neutral segments showing allelic imbalance were also considered to be aberrant (ie, total CN = 2 and lower CN = 0). Gains were called if the total CN of a segment was greater than 2 (diploid). Losses were called if the minor CN of an arm segment was equal to 0 (corresponding to LOH), regardless of the total number of copies. No arm-level CNVs were found to have both LOH and a total CN greater than 2 (Supplementary Table S4). Arm-level events were defined as any gain or loss occurring in an autosome that involved at least 10% of the arm. We used this relatively low arm fraction threshold both to remove potential artifacts from the assay, such as drops in coverage around centromeric or telomeric regions (which could be captured as losses), and at the same time capture any somatic CN events regardless of their breadth. In general, very few focal CN events were observed, except for some focal losses in 9p21.3 involving CDKN2A/2B loci (which were considered 9p losses in the analysis). Copy-number aberrant (CNA) tumors were classified as such when they satisfied at least one of the following criteria: (i) 9p loss/focal loss of CDKN2A/2B, (ii) 17q gain, or (iii) FCNAg greater than the median in the relevant cohort. All 9p and 17q calls were manually reviewed using the raw FACETS output, and in cases where it was clear that FACETS had misclassified a subclonal gain/loss at these loci, this was corrected and recorded. All CN calls, including notation of manual review, are available in Supplementary Table S4.

Calculation of tumor mutation burden

All nonsilent exonic mutations were considered for computation of tumor mutation burden (TMB) and this number was corrected for the total genomic length covered in the hybridization assay (1 Mb and 1.1 Mb for cases of MSK-IMPACT versions 5 and 6). Because the MSK and TCGA whole-exome assays covered different lengths of the genome, appropriate corrections were made (38.9 Mb for MSK-exome and 36.8 Mb for TCGA-exome). Mutation calling was not completed on two TCGA samples sequenced with the SOLiD sequencing platform, and therefore TMB for these samples was determined using public TCGA mutation calls.

Cancer cell fraction analysis

Mutations with a cancer cell fraction of ≥0.9 were defined as clonal, and all other mutations were considered subclonal. Using these definitions, an intratumoral heterogeneity (ITH) index (range 0, +∞) was computed by dividing the number of subclonal and clonal mutations (#subclonal/#clonal).

Neoantigen prediction and immune cell deconvolution

We used INTEGRATE, as described by Zhang and colleagues (23), for identifying gene fusions from RNA-seq data. Human leukocyte antigen (HLA) alleles were identified using POLYSOLVER (24). For the neoantigen prediction of the fusion proteins called by INTEGRATE, we used the INTEGRATE-neo pipeline (23), which reports any fusion-derived peptides found with binding affinity less than 500 nmol/L (default setting) for the given HLA allele through NetMHC v4.0 (25).

We further assessed differences in the expression of signatures associated with angiogenesis, PD1, PD-L1, and CTLA4 between tRCC cases and other RCC histologies. To do this, we performed gene set enrichment analysis (ssGSEA) using previously described gene signatures in single-sample RNA-seq data (26–28). We used the ssGSEA algorithm implementation through the R package “gsva” (29) to calculate the relative overexpression of gene sets of interest in each of the samples (30). Expression data were downloaded from the TCGA firebrowse website.

Statistical analysis

Cases with complete genomic information were included in the analysis (n = 22). Patient age was recorded at the time of diagnosis and pediatric cases were defined as patients diagnosed before 18 years of age. The association between FCNAg and clinicopathologic variables at baseline (ie, clinical stage and age) was tested using the Mann–Whitney U test. The same approach was used to perform pairwise TMB comparisons between the tRCC cohorts and other RCC subtypes in the TCGA (KIRC, KIRP, and KICH), as well as when assessing differences in FCNAg between two groups. Differences in PD-L1 gene expression were tested between tRCCs and other tumors within the corresponding TCGA subgroup (KIRC and KIRP).

The primary outcome was overall survival (OS), which was calculated from the time of pathologic diagnosis to the date of death or last follow-up. Survival estimates were computed using the Kaplan–Meier method and estimates were compared using the log-rank test. Two TCGA samples sequenced with SOLiD sequencing were excluded from survival analysis due to absence of CN data (Supplementary Table S2). To assess the response to systemic therapy, we measured time-to-treatment (TTT) and time-on-treatment (TOT). TTT was defined as the time interval from diagnosis to systemic therapy start; TOT was specified as the time from treatment start to discontinuation, either due to therapy-related toxicity or disease progression. A TOT > 12 months was used as a surrogate for response. All hypothesis tests were two-sided and statistical significance was defined as P < 0.05. All analyses were performed using the R platform v3.5.3.

Patient characteristics

We identified 37 patients treated for tRCC at our center between 1997 and 2019, with the diagnosis based on IHC, FISH analysis, anchored multiplex PCR, or NGS. We excluded 15 samples from further analysis due to lack of consent to sequencing or negative cytogenetic proof (ie, FISH, ARCHER, or NGS) of gene rearrangement (Supplementary Fig. S1). An overview of the clinicopathologic and molecular features of the MSK cohort who underwent NGS (n = 22) is provided in Supplementary Tables S1 and S2. A summary of the clinical and genomic features of the MSK-IMPACT cohort (n = 22) is also shown (Fig. 1). Gene fusions were found in both TFE3 (n = 20) and TFEB (n = 2). In addition, we identified an independent cohort from the pan-kidney (KIPAN) TCGA cohort of 14 primary tumors with a confirmed MiTF gene fusion and WES data (Supplementary Table S2), TFE3 and TFEB rearrangements were identified in 13 and one cases, respectively. Only 12 of the TCGA samples were included in most analyses, as two of them could not be processed because of our pipeline not being compatible with SOLiD sequencing (see Materials and Methods). The median follow-up time in the MSK-IMPACT and TCGA cohorts was 27.1 and 41.4 months, respectively.

Figure 1.

A, Oncoprint displaying recurrent somatic mutations (genes mutated in >1 sample) and frequent CNVs in the MSK-IMPACT cohort. CNVs are mostly found in cases with a high AJCC stage. The most frequent MiTF fusion gene was TFE3. B, TMB in merged MSK-TCGA exomes compared with other TCGA-RCC cohorts. C, In MSK-IMPACT cohort, TMB of tRCC is centered around a median value of 0.8 nonsynonymous mutations per megabase, with two outlier cases. Mann–Whitney U tests were used for pairwise TMB comparisons between RCC subtypes.

Figure 1.

A, Oncoprint displaying recurrent somatic mutations (genes mutated in >1 sample) and frequent CNVs in the MSK-IMPACT cohort. CNVs are mostly found in cases with a high AJCC stage. The most frequent MiTF fusion gene was TFE3. B, TMB in merged MSK-TCGA exomes compared with other TCGA-RCC cohorts. C, In MSK-IMPACT cohort, TMB of tRCC is centered around a median value of 0.8 nonsynonymous mutations per megabase, with two outlier cases. Mann–Whitney U tests were used for pairwise TMB comparisons between RCC subtypes.

Close modal

Somatic driver mutations, TMB, intratumoral heterogeneity, and immunogenicity of fusion peptides in tRCC

Previous analyses of tRCC tumors were limited to comparatively small cohorts, and therefore were limited in their power to detect recurrent somatic mutations. We therefore first focused on the analysis of somatic mutations in the MSK-IMPACT cohort. After excluding TFE fusions and CN alterations, 15 of 22 (68.2%) patients demonstrated somatic single-nucleotide variations or indels, with putatively oncogenic driver mutations in SWI/SNF and DNA damage repair pathways in five of 22 (22.7%) samples. We identified activating TERT promoter mutations (C228T) in three patients, all of which occurred in high American Joint Committee on Cancer (AJCC) stages, suggesting that activation of TERT may be a novel marker of aggressive tRCC. Where possible, we quantitatively determined the clonality of putatively oncogenic mutations using allele-specific CN analysis. We determined that one of two ATM mutations and one of one SMARCA4 mutations for which allele-specific CN data were available were clonal (Supplementary Table S3). Of these, only the clonal ATM mutation was accompanied by LOH of the complementary allele, suggesting that this lesion may be a truncal-driving event in the pathogenesis of the tumor.

High TMB has been used as a biomarker of favorable response to ICB in many solid cancers (31, 32). Because accurate quantification of TMB in low mutation burden tumors is difficult using targeted NGS panels, we assessed TMB from exome sequencing of 10 patients in the MSK-exome and 12 patients from the TCGA-exome cohort, for which single-nucleotide variations were available from the public repository (see Materials and Methods). We computed a median TMB of 0.80 mut/Mb for tRCCs, different from all other TCGA-RCC cohorts (Fig. 1). TMB was consistent between the TCGA and MSK-exome cohorts (net difference −0.12, Mann–Whitney U test; P = 0.47). Furthermore, the TMB of tRCC was characteristically different than that of other RCC subtypes: lower compared with clear cell [TCGA KIRC, −0.76 (−0.49 to −1.02), Mann–Whitney U test; P < 0.001) and papillary [KIRP, −1.17 (−0.77 to −1.54), Mann–Whitney U test; P < 0.001], but higher compared with chromophobe RCC [KICH, 0.16 (0.02–0.34), Mann–Whitney U test; P = 0.03]. No putatively explanatory mutations were identified in the outlier tRCC cases of exceptionally high TMB (Fig. 1). We summarized the clonality of somatic mutations using a metric for overall ITH (ITH-index, see Materials and Methods). The median ITH-index across all tRCC exome samples considering all mutations was 0.82 [interquartile ranges (IQR), 0.55–1.36], indicating a higher rate of clonal events relative to subclonal events. This is consistent with a model for the evolution of tRCC, whereby early somatic mutations occur which are succeeded later by a small number of subclonal mutations. Taken together, these results indicate that, from a mutational perspective, tRCC tumors exhibit (1) relatively low mutation burden, (2) relatively few canonically oncogenic mutational events aside from MiTF translocation, and (3) in cases with loss-of-function mutations to tumor suppressors, the wild-type allele infrequently undergoes concomitant LOH. Mutation annotation files for the MSK-IMPACT and TCGA-exome cohorts are provided in Supplementary Table S3.

Given the recent report of immunogenic fusion–derived neoantigens in metastatic head and neck cancer (33) and the concept of immunogenicity in tumors with relatively low mutational burden, we assessed immunogenicity and the composition of the tumor microenvironment in TCGA-tRCC samples for which exome and RNA-seq data were available. We identified a significant HLA-epitope binding affinity of fusion-derived peptides in eight of 14 (57.1%) cases, with a median affinity of 98.5 nmol/L (range 58.6–496.8). The relevant data, including the HLA type as well as the epitope peptide sequence, are provided in Supplementary Table S2. We also used immune deconvolution techniques to test for differences in gene signatures for angiogenesis and three immunotherapeutic targets in ccRCC: PD1, PD-L1, and CTLA4. The angiogenesis gene expression signature was higher in tRCCs compared with papillary RCC (Mann–Whitney U test; P = 0.007) and lower but almost equivalent with respect to clear cell tumors (Mann–Whitney U test, P = 0.053). tRCCs also showed a higher PD-L1 expression signature compared with the KIRC (P = 0.01) and KIRP cohorts (P = 0.002). However, we found no significant differences in CTLA4 (P = 0.2 and P = 0.71, respectively) or PD1 signatures (P = 0.74 and P = 0.7, respectively; Supplementary Fig. S2).

Prognostic implications of CN alterations

In prior analyses of tRCC, loss of chromosome arm 9p and gain of arm 17q were identified as common and, in the case of 17q gain, potential prognostic biomarkers (5). In contrast to earlier studies, our approach used allele-specific CN analysis that could more sensitively detect small gains and copy-number-neutral LOH events (see Materials and Methods). We therefore set out to evaluate the association between gross CN events and both clinical features and OS in tRCC. We began by calculating the fraction of the genome exhibiting any CN alteration (FCNAg). The median FCNAg was 0.16 (IQR: 0.09–0.32) in the MSK-IMPACT cohort and 0.09 (IQR: 0.04–0.33) in the TCGA cohort. High AJCC stage (III/IV) was found to be associated with higher FCNAg in the TCGA cohort (Mann–Whitney U test; P = 0.006) but this finding did not reach statistical significance in the MSK-IMPACT cohort (P = 0.09; Fig. 2).

Figure 2.

Association of FCNAg with AJCC stage and 9p loss. The broad line of the boxplots illustrate the median FCNAg, top and bottom hinges showing the IQR (25% and 75% of values) and the whiskers illustrating values within 1.5× the IQR in the MSK (A) and TCGA (B) cohorts. Loss of chromosome 9p was found to be associated with higher FCNAg in the MSK (C) and TCGA (D) cohorts. WT, wild-type.

Figure 2.

Association of FCNAg with AJCC stage and 9p loss. The broad line of the boxplots illustrate the median FCNAg, top and bottom hinges showing the IQR (25% and 75% of values) and the whiskers illustrating values within 1.5× the IQR in the MSK (A) and TCGA (B) cohorts. Loss of chromosome 9p was found to be associated with higher FCNAg in the MSK (C) and TCGA (D) cohorts. WT, wild-type.

Close modal

We next examined CN alterations in the MSK-IMPACT cohort at a more granular level. The most common arm-level CN alterations observed in the MSK-IMPACT cohort were loss of 9p in nine of 22 (41%) patients, gain of 17q in eight of 22 (36%) patients, and gain of 6p in eight of 22 (36%) patients. No single CN event was significantly associated with OS in the MSK-IMPACT cohort (Supplementary Table S5). However, we did note that three of three TERT mutations were identified in tumors harboring 9p loss, and that the presence of 9p loss was significantly associated with higher FCNAg in both cohorts (Fig. 3; Mann–Whitney U test; P = 0.008 and 0.042 in the MSK and TCGA cohorts, respectively).

Figure 3.

Proportion of patients with 9p loss and 17q gain in the MSK (A) and TCGA cohorts (B). Kaplan–Meier analysis of OS for patients with CNA tumors [ie, 9p loss, 17q gain, or high FCNAg (>within-cohort median)] in the MSK-IMPACT (C) and TCGA (D) cohorts. Log-rank test results from both cohorts were combined in a meta-analysis using Fisher method [χ2 (4d.f) = 10.52; P = 0.03]. WT, wild-type.

Figure 3.

Proportion of patients with 9p loss and 17q gain in the MSK (A) and TCGA cohorts (B). Kaplan–Meier analysis of OS for patients with CNA tumors [ie, 9p loss, 17q gain, or high FCNAg (>within-cohort median)] in the MSK-IMPACT (C) and TCGA (D) cohorts. Log-rank test results from both cohorts were combined in a meta-analysis using Fisher method [χ2 (4d.f) = 10.52; P = 0.03]. WT, wild-type.

Close modal

On the basis of the relatively low TMB of tRCC and the comparatively high frequency of CN alterations, we sought to evaluate the association between clinical outcomes and a summary measure of CN changes. We defined a copy-number-aberrant (CNA) cohort of patients which harbored either 9p loss, 17q gain, or an FCNAg higher than the within-cohort median. In the MSK-IMPACT cohort, the 15 patients with CNA tumors demonstrated worse OS (5-year OS 34.9% vs 100%; log-rank P = 0.03). In the TCGA cohort, this analysis was not statistically significant (log-rank, 0.18), however, the point survival estimates were very similar (5-year OS 57.1% vs 100%) suggesting that the analysis in the TCGA data may have been underpowered (with only n = 10 samples). Two tumors in the MSK-IMPACT cohort and two tumors in the TCGA cohort exhibited biallelic loss of CDKN2A/B; treating these tumors as a separate group for the purposes of survival analysis did not meaningfully affect our findings, suggesting that biallelic loss of this gene is not the sole driver of poor prognosis in CNA tumors. Therefore, we performed a meta-analysis on the results with the two cohorts, which demonstrated a statistically significant association between CNA tumor status and poor survival (P = 0.03; Fishers combined probability test).

Temporal evolution of a metastatic tRCC tumor

We retrospectively identified an additional tRCC tumor derived from a patient treated at our institution who had an earlier tRCC tumor sample sequenced as part of the TCGA consortium (only a single such overlapping patient was identified). This patient initially underwent a partial nephrectomy (at our institution, sample unavailable) approximately 20 years ago for a 4.5 cm renal mass that was diagnosed as a high-grade papillary RCC. After 8 years, a local recurrence developed in the same kidney where the original tumor was removed, leading the patient to undergo a complete radical nephrectomy with retroperitoneal lymph node dissection (TCGA sample BQ-5887). To note, all the lymph nodes obtained during this procedure were found to be free of tumor. This local kidney recurrence was found to have concordant morphologic features when compared with the previous specimen, as well as a TFE3 gene rearrangement; prompting a change in the original diagnosis to MiTF-family tRCC. The patient remained disease-free for about 8 more years, after which he developed severe abdominal and back pain due to extensive disease spread in the upper abdomen and retroperitoneum. One of the retroperitoneal nodes was biopsied and also found to bear a PRCC-TFE3 fusion (MSK-RP-5887 sample). At this point, the patient was diagnosed with metastatic disease and started on systemic therapy (Fig. 4). After an initial period of partial response to everolimus/bevacizumab, the abdominal lesions became stable and subsequently grew again (with a new disease site also evidenced in the thorax). At this point, the treatment was stopped and the patient was administered two additional lines of therapy, which resulted in mixed responses (ie, shrinkage of some lesions but progression in others). After progression on the third line of therapy [plus significant tyrosine kinase inhibitor (TKI)-related toxicity], the patient refused additional treatments and decided to undergo hospice care instead.

Figure 4.

Temporal evolution of tRCC in a single patient. A, Private and shared CN alterations and somatic mutations in two tRCC tumors from a single patient. B, Clinical course of patient with tRCC from which tumor samples were derived. C, Putative model for tRCC evolution in this case. Early loss of 9p and other CN alterations was followed by large-scale WGD. The majority of somatic mutations were acquired in the period following the appearance of the MRCA, which bore a high burden of CNVs.

Figure 4.

Temporal evolution of tRCC in a single patient. A, Private and shared CN alterations and somatic mutations in two tRCC tumors from a single patient. B, Clinical course of patient with tRCC from which tumor samples were derived. C, Putative model for tRCC evolution in this case. Early loss of 9p and other CN alterations was followed by large-scale WGD. The majority of somatic mutations were acquired in the period following the appearance of the MRCA, which bore a high burden of CNVs.

Close modal

The newly identified tumor was a pathologically confirmed lymph node metastasis with evidence of a PRCC-TFE3 fusion (confirmed by IHC and FISH, as well as MSK-IMPACT sequencing). To understand the clonal relatedness of these tumors, we completed exome sequencing on the nodal metastasis, and compared CN aberrations and mutations between MSK-RP-5887 and TCGA-BQ-5887. Interestingly, we noted that both samples exhibited whole-genome doubling (WGD) and extensive CN alterations across the entire genome and therefore an aberrantly high FCNAg compared with other tRCC tumors. In fact, the vast majority of arm-level events (78%) were shared between the two tumors. Notably, loss of chromosome 9p was evident in the two samples and, in both cases, it preceded the WGD event. These findings are consistent with CNVs being an early event in the development of the tumor. In contrast, only 23% of somatic mutations were shared between the two tumors (ITH-index, MSK-RP-5887: 5.3, TCGA-BQ-5887: 16), indicating that extensive mutational diversification occurred after divergence of the most recent common ancestor (MRCA). Together, this suggests that in this particular tRCC tumor, CN alterations occurred early during tumor evolution, whereas the acquisition of the bulk of the somatic mutations occurred later (Fig. 4). This result is not in agreement with the median ITH-index result reported above (of 0.82), and may reflect different evolutionary patterns in metastatic versus primary tRCC.

Response of patients with tRCC to systemic therapies in the MSK-IMPACT cohort

We assessed treatment response in 14 patients of the MSK-IMPACT cohort who underwent, in their majority, multiple lines of systemic treatment. Five of 14 (35.7%) patients showed a prolonged TOT (>12 months) in the first- or second-line of therapy. We observed no difference in the median FCNAg when comparing responders with nonresponders (Mann–Whitney U test; P = 0.55). We did not find any association between response to any therapy and genomic features. Notably, a patient with a prolonged response to ICB therapy harbored missense mutations in two genes of the PBAF chromatin-remodeling complex, SMARCA4 and PBRM1; variants in both genes have recently been linked to immunogenic features (34, 35). The treatment course and genomic characteristics of tRCC patients who showed prolonged responses to first- or second-line systemic therapy are shown in Fig. 5.

Figure 5.

A, Swimmer plot displaying a heterogeneous pattern of substances and response, a few cases showing an exceptionally prolonged TOT. B, Molecular features identified in patients with TOT > 12 months. AB, antibody; GLS1, glutaminase 1; Inh, inhibitor.

Figure 5.

A, Swimmer plot displaying a heterogeneous pattern of substances and response, a few cases showing an exceptionally prolonged TOT. B, Molecular features identified in patients with TOT > 12 months. AB, antibody; GLS1, glutaminase 1; Inh, inhibitor.

Close modal

Pediatric tRCC cases in the MSK-IMPACT cohort

Finally, we assessed the presence of molecular alterations in patients under 18 years of age (n = 3). Their sites of disease and their therapeutic course are displayed in Fig. 6. Two patients had undergone prior chemotherapy: one for retinoblastoma, who also had evidence of a germline RB1 mutation (P-0025629-T02-IM6, germline status identified from sequencing of matched normal blood per standard MSK-IMPACT sequencing), and the other for neuroblastoma (P-0024809-T02-IM6). None of the pediatric cases showed 9p loss or 17q gain and, consistently, they were also found to have a significantly lower FCNAg when compared with their older counterparts (Mann–Whitney U test; P = 0.02; see Supplementary Table S3).

Figure 6.

Three panels with pediatric cases in our MSK-IMPACT cohort displaying a favorable disease course after surgery as the only intervention (created with BioRender.com).

Figure 6.

Three panels with pediatric cases in our MSK-IMPACT cohort displaying a favorable disease course after surgery as the only intervention (created with BioRender.com).

Close modal

We present a comprehensive genomic analysis of 22 patients with tRCC utilizing high-depth–targeted sequencing and high-breadth exome sequencing, combining the results with additional exome-sequencing data from 14 TCGA tRCC cases. To serve as a resource for future studies, we report the genomic and clinical features found in each tumor in both cohorts. Our results suggest that tRCCs are driven primarily by copy-number aberrations, rather than by somatic mutations in tumor suppressors and oncogenes. We observed distinct RNA-seq deconvolution scores for angiogenesis and PD-L1 in TCGA-exome cases compared with clear cell and papillary tumors from the TCGA; as well as, a prediction of significant HLA-binding affinity in 57.1% of the fusion-derived neoantigens, suggesting that perhaps influencing the immune system to direct it against these clonal antigens may represent a viable treatment strategy for tRCC. The smaller proportion of copy-number-altered genome in pediatric tRCC cases, compared with adults, is in-line with a less aggressive tumor biology. However, these results must be viewed with caution due to the small patient numbers.

Gene rearrangements are the key oncogenic event in tRCC, and prior work has associated certain fusion partners with worse outcomes (36). In addition, recurrent somatic mutations and CNVs in tRCC have previously been explored. An earlier genomic study in 16 patients with tRCC found heterogeneous molecular profiles, with 17q gain and 9p loss identified as the most frequent CNVs. In that study, 17q gain was an essential feature of poor outcome, but presence of 9p loss was not significantly associated with poor OS (5). These results are partially consistent with our own findings, which, however, showed a significant association in the MSK-IMPACT cohort with poor outcome in a CNA group of tumors defined by the presence of either 9p loss, 17q gain, or elevated FCNAg. 9p loss was previously associated with a worse outcome in ccRCC and papillary RCC (6, 37–39). Although the processes underlying this association are unclear, it is notable that CDKN2A and CDKN2B, important cell-cycle regulators, which have been shown to interact with hypoxia-inducible factor-1α and cause VEGF upregulation (40), are located on 9p21 and were lost (sometimes focally) in a significant percentage of cases.

TERT promoter mutations have also previously been associated with an aggressive disease course in more common variants of RCC (41), but the implications of these mutations in tRCC have not been described. In the MSK-IMPACT cohort, we detected clonal activating hotspot mutations in the TERT promoter of three tumors. All of these cases consisted of high-stage tRCC, and exhibited 9p loss as well as a high burden of somatic CNVs (Fig. 1). This data collectively suggests that TERT mutations are associated with an aggressive phenotype in tRCC, however, it is impossible at this point to determine its role as a driver of aggressiveness.

In our cohort a first- or second-line treatment with a TKI led to extended TOT in three cases. The response to TKIs and mTOR inhibitors was previously evaluated in a retrospective study, which included 21 patients with tRCC. In that study, an objective response rate (ORR) of 33% could be observed, similar to ccRCC. Importantly, only four of 21 cases underwent cytogenetic confirmation of tRCC (9). The ORR for patients with tRCC in a large non-ccRCC (nccRCC) trial was comparable (29%; ref. 42). Notably, 1 patient who underwent a combined mTOR inhibitor/anti-VEGF regimen in our cohort showed a prolonged TOT of 37 months.

Our analysis here also provides an initial characterization of the tRCC tumor microenvironment, relative to clear cell and papillary RCC. We observed overexpression of PD-L1 in tRCC relative to KIRC and KIRP, which is consistent with findings from a large study of nccRCC in which 90% of tRCC cases showed PD-L1 overexpression in tumor-infiltrating mononuclear cells (43). Despite these apparent microenvironmental phenotypes, patients with tRCC demonstrate poor response rates to TKI and ICB, suggesting that emerging biomarkers for response to TKI and immunotherapy in ccRCC may not readily translate to tRCC. In this context, it is also notable that our analysis of RNA-seq data in the TCGA-exome cohort revealed fusion-derived neoantigens with high binding affinity to the HLA class-I molecules in 57% of the cases. In light of recent data describing the immunogenicity of fusion-derived neoantigens in head and neck cancers (33), this suggests that neoantigens produced by TFE3/TFEB translocation may be clinically meaningful. Taken together, these results argue for further exploration of the immunogenomic characteristics of tRCC, with special focus on T-cell response to ICB, as well as potential neoantigen-directed vaccines (44). Furthermore, we observed a notable case of prolonged response to ICB harboring SMARCA4 and PBRM1 variants. An immunogenic effect of SMARCA4 mutations, via PD-L1 upregulation, has previously been postulated for small-cell ovarian cancer (34). PBRM1 mutations were recently reported to have an increased clinical benefit in patients with metastatic ccRCC undergoing ICB (35). We further investigated potential genomic biomarkers of response to ICB. Prior literature described an inverse association between aneuploidy and response to ICB in cohorts of patients with metastatic melanoma and non–small cell lung cancer (45, 46). In our limited cohort, we did not find a statistically significant association between FCNAg and treatment response.

Pediatric tRCC cases commonly have better outcomes compared with adult variants, which was previously attributed to a smaller percentage of distant metastasis (47). In young patients, the disease is often manageable with surgery alone, which we could confirm for our three pediatric cases, all of whom were treated with surgery and were disease-free at the most recent follow-up. Furthermore, prior chemotherapy has been described as a risk factor for the development of tRCC in children (48). Two of our three pediatric patients had prior chemotherapy due to retinoblastoma and neuroblastoma. The occurrence of RCC as a secondary malignancy following neuroblastoma has been previously reported, with some of these tumors identified as tRCCs (49). RCC in neuroblastoma survivors is included as a provisional entity in the 2016 World Health Organization classification of renal tumors (49, 50). Notably, patient P-0025629-T02-IM6 is, to our knowledge, the first described case of an RCC following a diagnosis of retinoblastoma; however, the remaining RB1 copy was found to be unaffected in the tRCC sample. At the molecular level, fewer genomic events were previously reported in patients <18 years old (5). In our cohort, the favorable outcomes in pediatric cases were in-line with a lower FCNAg and a lack of “high-risk” CN events.

Limitations of our study include the heterogeneous composition of sequenced specimens, originating from both primaries and metastatic sites (ie, not paired). These samples may ultimately represent different snapshots of tumor evolution. Comparability with the TCGA cohort is limited as well, because TCGA samples were all obtained from untreated primary tumors. The rarity of tRCC imposes restrictions on the statistical power of our study; although we provided detailed clinical outcomes for all the patients included, another drawback is clearly the limited size of our cohort and of the TCGA cohort. However, our study only included cytogenetically confirmed tRCC and thus minimized contamination of the sample set with non-tRCC cases, which has not been routinely done in prior studies. Moreover, we acknowledge the fact that TFE3- and TFEB-associated tRCCs show distinct morphologic features. However, because of our limited sample size, both variants were included in our tRCC cohorts. Furthermore, we did not have RNA-seq data available for our institutional cases, thus a validation of the neoantigen prediction in TCGA cases could not be performed. With respect to the systemic therapy data, we used TOT as a surrogate of treatment response, which has potential interpretation issues. However, because RECIST scores were unavailable for these patients, the actual treatment courses were displayed graphically to provide an idea of the overall clinical effect.

Because of the low incidence of tRCC, future molecular studies should consider a multi-institutional pooling of cases to deepen our understanding of these tumors. In addition, the collection of a larger number of patients who underwent systemic therapy could enable the evaluation of potential relationships between genomic aberrations and treatment response. Additional transcriptomic analyses could shed further light on potential immunogenic or angiogenic effects of the aforementioned CN alterations. This is particularly interesting with regards to a recently opened multi-institutional prospective trial of an ICB/TKI combination therapy in patients with tRCC (NCT03595124), a study of an ICB/anti-VEGF combination regimen including patients with tRCC (NCT02724878), as well as potential trials involving neoantigen vaccines.

Conclusion

Copy-number variants appear to play a key role in the prognosis of tRCC. These aberrations are pervasive in tRCC, and appear to condition poor outcomes. The presence of TERT hotspot mutations in cases of advanced disease suggests an association with molecular progression in tRCC. An increased frequency of oncogenic events (including somatic mutations and CNVs) in adult patients could presumably be driving more aggressive disease phenotypes.

J. Marcon reports receiving other remuneration from the German Research Foundation - Deutsche Forschungsgemeinschaft. V. Makarov is listed as a co-inventor on a provisional patent application on determinants of cancer response to immunotherapy that is currently assigned to Sloan Kettering Institute for Cancer Research. T.A. Chan is an employee/paid consultant for Bristol-Myers Squibb, Illumina, AstraZeneca, and Gritstone Oncology, reports receiving commercial research grants from AstraZeneca and Illumina, and holds ownership interest (including patents) in Gritstone Oncology and an2H. R.J. Motzer is an employee/paid consultant for Pfizer, Genentech/Roche, Eisai, Novartis, Lilly Oncology, and Exelixis, and reports receiving commercial research grants from Pfizer, Exelixis, Eisai, and Bristol-Myers Squibb. M.H. Voss reports receiving other commercial research support from Pfizer, speakers bureau honoraria from Pfizer, Exelixis, Eisai, Calithera, and Corvus. No potential conflicts of interest were disclosed by the other authors.

Conception and design: J. Marcon, A. Sanchez, N.E. Benfante, P. Russo, A.A. Hakimi, E. Reznik

Development of methodology: J. Marcon, A. Sanchez, V. Makarov, A. Sandhu, T.A. Chan, A.A. Hakimi, E. Reznik

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J. Marcon, A. Sanchez, R.R. Kotecha, R. Mano, K.A. Blum, D.E. Nassau, N.E. Benfante, M.I. Carlo, V. Reuter

Analysis and interpretation of data (eg, statistical analysis, biostatistics, computational analysis): J. Marcon, R.G. DiNatale, A. Sanchez, R.R. Kotecha, F. Kuo, V. Makarov, A. Sandhu, A.W. Silagy, D.E. Nassau, N.E. Benfante, T.A. Chan, R.J. Motzer, M.H. Voss, J. Coleman, P. Russo, V. Reuter, A.A. Hakimi, E. Reznik

Writing, review, and/or revision of the manuscript: J. Marcon, R.G. DiNatale, R.R. Kotecha, S. Gupta, F. Kuo, R. Mano, A.W. Silagy, K.A. Blum, N.E. Benfante, M.V. Ortiz, T.A. Chan, R.J. Motzer, M.H. Voss, J. Coleman, P. Russo, V. Reuter, A.A. Hakimi, E. Reznik

Administrative, technical, or material support (ie, reporting or organizing data, constructing databases): R.G. DiNatale, F. Kuo, A. Sandhu, K.A. Blum, D.E. Nassau, J. Coleman

Study supervision: R.J. Motzer, J. Coleman, V. Reuter, E. Reznik

This research was funded in part through the NIH/NCI Cancer Center Support Grant P30 CA008748. J. Marcon obtained financial support from the German Research Foundation (Deutsche Forschungsgemeinschaft, grant no. MA 8008/01 and/02). E. Reznik was supported by the Geoffrey Beene Cancer Research Center, the Department of Defense Kidney Cancer Research Program W81XWH-18-1-0318, and Kidney Cancer Association Young Investigator Award.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Ross
H
,
Argani
P
. 
Xp11 translocation renal cell carcinoma
.
Pathology
2010
;
42
:
369
73
.
2.
Cajaiba
MM
,
Dyer
LM
,
Geller
JI
,
Jennings
LJ
,
George
D
,
Kirschmann
D
, et al
The classification of pediatric and young adult renal cell carcinomas registered on the children's oncology group (COG) protocol AREN03B2 after focused genetic testing
.
Cancer
2018
;
124
:
3381
9
.
3.
Magers
MJ
,
Udager
AM
,
Mehra
R
. 
MiT family translocation-associated renal cell carcinoma: a contemporary update with emphasis on morphologic, immunophenotypic, and molecular mimics
.
Arch Pathol Lab Med
2015
;
139
:
1224
33
.
4.
Kauffman
EC
,
Ricketts
CJ
,
Rais-Bahrami
S
,
Yang
Y
,
Merino
MJ
,
Bottaro
DP
, et al
Molecular genetics and cellular features of TFE3 and TFEB fusion kidney cancers
.
Nat Rev Urol
2014
;
11
:
465
75
.
5.
Malouf
GG
,
Monzon
FA
,
Couturier
J
,
Molinié
V
,
Escudier
B
,
Camparo
P
, et al
Genomic heterogeneity of translocation renal cell carcinoma
.
Clin Cancer Res
2013
;
19
:
4673
84
.
6.
Turajlic
S
,
Xu
H
,
Litchfield
K
,
Rowan
A
,
Chambers
T
,
Lopez
JI
, et al
Tracking cancer evolution reveals constrained routes to metastases: TRACERx renal
.
Cell
2018
;
173
:
581
94
.
7.
McKay
RR
,
Bossé
D
,
Xie
W
,
Wankowicz
SAM
,
Flaifel
A
,
Brandao
R
, et al
The clinical activity of PD-1/PD-L1 inhibitors in metastatic non-clear cell renal cell carcinoma
.
Cancer Immunol Res
2018
;
6
:
758
65
.
8.
Koshkin
VS
,
Barata
PC
,
Zhang
T
,
George
DJ
,
Atkins
MB
,
Kelly
WJ
, et al
Clinical activity of nivolumab in patients with non-clear cell renal cell carcinoma
.
J Immunother Cancer
2018
;
6
:
9
.
9.
Malouf
GG
,
Camparo
P
,
Oudard
S
,
Schleiermacher
G
,
Theodore
C
,
Rustine
A
, et al
Targeted agents in metastatic Xp11 translocation/TFE3 gene fusion renal cell carcinoma (RCC): a report from the Juvenile RCC network
.
Ann Oncol
2010
;
21
:
1834
8
.
10.
Boilève
A
,
Carlo
MI
,
Barthélémy
P
,
Oudard
S
,
Borchiellini
D
,
Voss
MH
, et al
Immune checkpoint inhibitors in MITF family translocation renal cell carcinomas and genetic correlates of exceptional responders
.
J Immunother Cancer
2018
;
6
:
159
.
11.
Green
WM
,
Yonescu
R
,
Morsberger
L
,
Morris
K
,
Netto
GJ
,
Epstein
JI
, et al
Utilization of a TFE3 break-apart FISH assay in a renal tumor consultation service
.
Am J Surg Pathol
2013
;
37
:
1150
63
.
12.
Gao
Q
,
Liang
W-W
,
Foltz
SM
,
Mutharasu
G
,
Jayasinghe
RG
,
Cao
S
, et al
Driver fusions and their implications in the development and treatment of human cancers
.
Cell Rep
2018
;
23
:
227
38
.
13.
Hu
X
,
Wang
Q
,
Tang
M
,
Barthel
F
,
Amin
S
,
Yoshihara
K
, et al
TumorFusions: an integrative resource for cancer-associated transcript fusions
.
Nucleic Acids Res
2018
;
46
:
D1144
9
.
14.
Cheng
DT
,
Mitchell
TN
,
Zehir
A
,
Shah
RH
,
Benayed
R
,
Syed
A
, et al
Memorial Sloan Kettering-integrated mutation profiling of actionable cancer targets (MSK-IMPACT): a hybridization capture-based next-generation sequencing clinical assay for solid tumor molecular oncology
.
J Mol Diagn
2015
;
17
:
251
64
.
15.
Chakravarty
D
,
Gao
J
,
Phillips
SM
,
Kundra
R
,
Zhang
H
,
Wang
J
, et al
OncoKB: a precision oncology knowledge base
.
JCO Precis Oncol
2017
;
2017
.
16.
DePristo
MA
,
Banks
E
,
Poplin
R
,
Garimella
KV
,
Maguire
JR
,
Hartl
C
, et al
A framework for variation discovery and genotyping using next-generation DNA sequencing data
.
Nat Genet
2011
;
43
:
491
8
.
17.
Kim
S
,
Scheffler
K
,
Halpern
AL
,
Bekritsky
MA
,
Noh
E
,
Källberg
M
, et al
Strelka2: fast and accurate variant calling for clinical sequencing applications
.
bioRxiv
2017
:
192872
.
18.
Koboldt
DC
,
Zhang
Q
,
Larson
DE
,
Shen
D
,
McLellan
MD
,
Lin
L
, et al
VarScan 2: somatic mutation and copy number alteration discovery in cancer by exome sequencing
.
Genome Res
2012
;
22
:
568
76
.
19.
Rimmer
A
,
Phan
H
,
Mathieson
I
,
Iqbal
Z
,
Twigg
SRF
WGS500 Consortium
et al
Integrating mapping-, assembly- and haplotype-based approaches for calling variants in clinical sequencing applications
.
Nat Genet
2014
;
46
:
912
8
.
20.
Sherry
ST
,
Ward
MH
,
Kholodov
M
,
Baker
J
,
Phan
L
,
Smigielski
EM
, et al
dbSNP: the NCBI database of genetic variation
.
Nucleic Acids Res
2001
;
29
:
308
11
.
21.
The 1000 Genomes Project Consortium.
A map of human genome variation from population-scale sequencing
.
Nature
2010
;
467
:
1061
73
.
22.
Shen
R
,
Seshan
VE
. 
FACETS: allele-specific copy number and clonal heterogeneity analysis tool for high-throughput DNA sequencing
.
Nucleic Acids Res
2016
;
44
:
e131
.
23.
Zhang
J
,
White
NM
,
Schmidt
HK
,
Fulton
RS
,
Tomlinson
C
,
Warren
WC
, et al
INTEGRATE: gene fusion discovery using whole genome and transcriptome data
.
Genome Res
2016
;
26
:
108
18
.
24.
Shukla
SA
,
Rooney
MS
,
Rajasagi
M
,
Tiao
G
,
Dixon
PM
,
Lawrence
MS
, et al
Comprehensive analysis of cancer-associated somatic mutations in class I HLA genes
.
Nat Biotechnol
2015
;
33
:
1152
8
.
25.
Andreatta
M
,
Nielsen
M
. 
Gapped sequence alignment using artificial neural networks: application to the MHC class I system
.
Bioinformatics
2016
;
32
:
511
7
.
26.
Barbie
DA
,
Tamayo
P
,
Boehm
JS
,
Kim
SY
,
Moody
SE
,
Dunn
IF
, et al
Systematic RNA interference reveals that oncogenic KRAS-driven cancers require TBK1
.
Nature
2009
;
462
:
108
12
.
27.
Bindea
G
,
Mlecnik
B
,
Tosolini
M
,
Kirilovsky
A
,
Waldner
M
,
Obenauf
AC
, et al
Spatiotemporal dynamics of intratumoral immune cells reveal the immune landscape in human cancer
.
Immunity
2013
;
39
:
782
95
.
28.
Şenbabaoğlu
Y
,
Gejman
RS
,
Winer
AG
,
Liu
M
,
Van Allen
EM
,
de Velasco
G
, et al
Tumor immune microenvironment characterization in clear cell renal cell carcinoma identifies prognostic and immunotherapeutically relevant messenger RNA signatures
.
Genome Biol
2016
;
17
:
231
.
29.
Hänzelmann
S
,
Castelo
R
,
Guinney
J
. 
GSVA: gene set variation analysis for microarray and RNA-seq data
.
BMC Bioinformatics
2013
;
14
:
7
.
30.
Hakimi
AA
,
Voss
MH
,
Kuo
F
,
Sanchez
A
,
Liu
M
,
Nixon
BG
, et al
Transcriptomic profiling of the tumor microenvironment reveals distinct subgroups of clear cell renal cell cancer: data from a randomized phase III trial
.
Cancer Discov
2019
;
9
:
510
25
.
31.
Goodman
AM
,
Kato
S
,
Bazhenova
L
,
Patel
SP
,
Frampton
GM
,
Miller
V
, et al
Tumor mutational burden as an independent predictor of response to immunotherapy in diverse cancers
.
Mol Cancer Ther
2017
;
16
:
2598
608
.
32.
Chan
TA
,
Yarchoan
M
,
Jaffee
E
,
Swanton
C
,
Quezada
SA
,
Stenzinger
A
, et al
Development of tumor mutation burden as an immunotherapy biomarker: utility for the oncology clinic
.
Ann Oncol
2019
;
30
:
44
56
.
33.
Yang
W
,
Lee
K-W
,
Srivastava
RM
,
Kuo
F
,
Krishna
C
,
Chowell
D
, et al
Immunogenic neoantigens derived from gene fusions stimulate T cell responses
.
Nat Med
2019
;
25
:
767
75
.
34.
Jelinic
P
,
Ricca
J
,
Van Oudenhove
E
,
Olvera
N
,
Merghoub
T
,
Levine
DA
, et al
Immune-active microenvironment in small cell carcinoma of the ovary, hypercalcemic type: rationale for immune checkpoint blockade
.
J Natl Cancer Inst
2018
;
110
:
787
90
.
35.
Miao
D
,
Margolis
CA
,
Gao
W
,
Voss
MH
,
Li
W
,
Martini
DJ
, et al
Genomic correlates of response to immune checkpoint therapies in clear cell renal cell carcinoma
.
Science
2018
;
359
:
801
6
.
36.
Malouf
GG
,
Camparo
P
,
Molinié
V
,
Dedet
G
,
Oudard
S
,
Schleiermacher
G
, et al
Transcription factor E3 and transcription factor EB renal cell carcinomas: clinical features, biological behavior and prognostic factors
.
J Urol
2011
;
185
:
24
9
.
37.
Turajlic
S
,
Xu
H
,
Litchfield
K
,
Rowan
A
,
Horswell
S
,
Chambers
T
, et al
Deterministic evolutionary trajectories influence primary tumor growth: TRACERx renal
.
Cell
2018
;
173
:
595
610
.
38.
Brunelli
M
,
Eccher
A
,
Gobbo
S
,
Ficarra
V
,
Novara
G
,
Cossu-Rocca
P
, et al
Loss of chromosome 9p is an independent prognostic factor in patients with clear cell renal cell carcinoma
.
Mod Pathol
2008
;
21
:
1
6
.
39.
Klatte
T
,
Pantuck
AJ
,
Said
JW
,
Seligson
DB
,
Rao
NP
,
LaRochelle
JC
, et al
Cytogenetic and molecular tumor profiling for type 1 and type 2 papillary renal cell carcinoma
.
Clin Cancer Res
2009
;
15
:
1162
9
.
40.
Zhang
J
,
Lu
A
,
Li
L
,
Yue
J
,
Lu
Y
. 
p16 Modulates VEGF expression via its interaction with HIF-1alpha in breast cancer cells
.
Cancer Invest
2010
;
28
:
588
97
.
41.
Casuscelli
J
,
Becerra
MF
,
Manley
BJ
,
Zabor
EC
,
Reznik
E
,
Redzematovic
A
, et al
Characterization and impact of TERT promoter region mutations on clinical outcome in renal cell carcinoma
.
Eur Urol Focus
2019
;
5
:
642
9
.
42.
Martínez Chanzá
N
,
Xie
W
,
Asim Bilen
M
,
Dzimitrowicz
H
,
Burkart
J
,
Geynisman
DM
, et al
Cabozantinib in advanced non-clear-cell renal cell carcinoma: a multicentre, retrospective, cohort study
.
Lancet Oncol
2019
;
20
:
581
90
.
43.
Choueiri
TK
,
Fay
AP
,
Gray
KP
,
Callea
M
,
Ho
TH
,
Albiges
L
, et al
PD-L1 expression in nonclear-cell renal cell carcinoma
.
Ann Oncol
2014
;
25
:
2178
84
.
44.
Lee
C-H
,
Yelensky
R
,
Jooss
K
,
Chan
TA
. 
Update on tumor neoantigens and their utility: why it is good to be different
.
Trends Immunol
2018
;
39
:
536
48
.
45.
Rizvi
H
,
Sanchez-Vega
F
,
La
K
,
Chatila
W
,
Jonsson
P
,
Halpenny
D
, et al
Molecular determinants of response to anti-programmed cell death (PD)-1 and anti-programmed death-ligand 1 (PD-L1) blockade in patients with non-small-cell lung cancer profiled with targeted next-generation sequencing
.
J Clin Oncol
2018
;
36
:
633
41
.
46.
Davoli
T
,
Uno
H
,
Wooten
EC
,
Elledge
SJ
. 
Tumor aneuploidy correlates with markers of immune evasion and with reduced response to immunotherapy
.
Science
2017
;
355
:
pii:
eaaf8399
.
47.
Geller
JI
,
Argani
P
,
Adeniran
A
,
Hampton
E
,
De Marzo
A
,
Hicks
J
, et al
Translocation renal cell carcinoma: lack of negative impact due to lymph node spread
.
Cancer
2008
;
112
:
1607
16
.
48.
Argani
P
,
Laé
M
,
Ballard
ET
,
Amin
M
,
Manivel
C
,
Hutchinson
B
, et al
Translocation carcinomas of the kidney after chemotherapy in childhood
.
J Clin Oncol
2006
;
24
:
1529
34
.
49.
Falzarano
SM
,
McKenney
JK
,
Montironi
R
,
Eble
JN
,
Osunkoya
AO
,
Guo
J
, et al
Renal cell carcinoma occurring in patients with prior neuroblastoma
.
Am J Surg Pathol. Wolters Kluwer
; 
2016
;
40
:
989
97
.
50.
Udager
AM
,
Mehra
R
. 
Morphologic, molecular, and taxonomic evolution of renal cell carcinoma: a conceptual perspective with emphasis on updates to the 2016 World Health Organization Classification
.
Arch Pathol Lab Med
2016
;
140
:
1026
37
.