Renal tumor classification has evolved in recent decades, as evidenced by the comparable complexity of the 2016 revision to the World Health Organization Classification of Tumours of the Urinary System and Male Genital Organs. A recent expansion of the knowledge base surrounding the cells of origin and evolutionary genomic characteristics of renal tumors has led to molecular characterization of novel entities and enriched understanding of established entities. This pace of research and its implementation into clinical practice has again begun to surpass that of our own classification schemata, with significant discoveries having been made since the introduction of the 2016 revision to the World Health Organization classification. In particular, biomarkers for renal tumor diagnosis and prognosis are in translation for future clinical application.
To provide a brief framework for clinical characterization of renal tumors rooted in morphologic assessment, to briefly review the current and future status of renal tumor biomarkers with an emphasis on practical use of these ancillary tools for accurate diagnosis, and to discuss the impact of emerging technologies and clinical trials relevant to renal cell carcinoma classification and biomarker development.
We review recent literature relevant to renal tumor classification (including established and proposed entities), focusing on molecular characterization and biomarker assessment.
Accurate renal tumor diagnosis requires an up-to-date understanding of renal tumor classification, including an awareness of morphologic clues that should stimulate consideration of molecularly defined entities, as well as the ancillary biomarker testing required to confirm diagnoses.
The 2016 update to the World Health Organization Classification of Tumours of the Urinary System and Male Genital Organs1–3 continued the recent trend of an evolutionary increase in our understanding of the biology, pathology, and clinical features of renal tumors; the updated classification includes established tumors as well as proposed emerging entities, and has gradually been incorporated into routine clinical practice. Meanwhile, parallel and subsequent work by The Cancer Genome Atlas (TCGA) Research Network,4,5 the TRACERx renal studies,6 and other groups have leveraged next-generation sequencing data to richly expand understanding of the molecular evolution and characterization of renal tumors in a way that both reinforces the validity of entities new to the 2016 World Health Organization classification and proposes novel entities with distinct genomic and clinical features. Several independent groups have characterized common and uncommon renal tumors with both sporadic and hereditary predispositions.7–12 Ongoing work will continue to add prognostic and therapeutic implications to these entities, and the development of practical biomarkers for diagnosis will progress as well.
Within diagnostic pathology, the field of renal tumor biomarkers is currently at an interesting intersection at which older or previously established markers are beginning to be supplemented with an exciting array of new assays; several of these new biomarkers show promise because of their strong discovery element and a tumor biology– and genomics-based tumor-specific expression (instead of a coincidental, nonspecific expression). There is a potential and promise to link the cells within the normal nephron to their associated neoplastic counterparts, as well as the underlying molecular underpinnings that might characterize such transitions. Single-cell sequencing of normal tissues and tumors, including their renal counterparts, have played an important role in making such knowledge possible and available.13–15 In contemporary literature and for the purposes of this review, a cancer-specific biomarker is defined as a gene expressed in a given cancer subtype with very low or no expression in any nephron segment; a lineage-specific biomarker, on the other hand, is defined as a gene expressed in both a given cancer subtype and certain nephron segments.12
The increasing availability of ancillary molecular testing, whether in the phase of data evolution in literature or having an established utility in clinical practice, requires that surgical pathologists be aware of the interlaboratory and interdisciplinary repercussions of rendered diagnoses. Increasingly, in today's surgical pathology practice, specific renal tumor classification may necessitate (1) morphologic suspicion, (2) immunohistochemical workup, (3) molecular testing, (4) communication with urologists and medical/radiation oncologists, (5) tumor and/or germline sequencing, and (6) genetic counseling. A contemporary approach to renal tumor categorization requires an updated and integrated understanding of these tools at the morphologic, clinical, and molecular level.
An accurate pathologic categorization of renal tumors requires recognition of both the current state of renal tumor classification and the variety of available (and upcoming) ancillary tools. In this review we hope to provide a practical and succinct review of renal tumor biomarkers as an aid to facilitating and delivering effective and accurate renal tumor classification in a day-to-day surgical pathology sign-out practice. We will briefly correlate diagnostic entities grouped by morphologic features via categorization of renal tumors with clear cytoplasm, papillary architecture, or eosinophilic (oncocytic) cytoplasm (Table 1) with up-to-date knowledge of biomarkers that have current and future clinical importance for diagnosis, prognosis, and therapy. We also present the current status and scope of emerging technologies and clinical trials relevant to renal cell carcinoma (RCC) classification.
BIOMARKERS IN RENAL TUMORS WITH CLEAR CYTOPLASM
Not every RCC with clear cell features is a conventional (clear cell) RCC (CCRCC), and not every RCC with clear cell features (Table 1) can be easily categorized. Conventional (clear cell) RCC may be recognized via characteristic morphologic features, with nests of cells with clear cytoplasm and abundant delicate fibrovascular septations (Figure 1, A); however, morphology may be less classic in higher-grade lesions with eosinophilic cytoplasm and papillary and/or pseudopapillary formation (Figure 1, B through E). Diffuse membranous carbonic anhydrase 9 (CA-IX) protein immunohistochemical expression with CK7 negativity (except in cystic or fibrotic areas) is supportive of an interpretation of CCRCC in diagnostically challenging situations; however, currently there remains no cancer- or lineage-specific biomarker for CCRCC. Also complicating the diagnostic scenario may be the often-seen peri-necrotic accentuation of CA-IX expression in high-grade RCCs (including those not of the conventional CCRCC subtype), which needs to be carefully distinguished from the uniform and diffuse pan-membranous expression seen in CCRCC.
Data emerging from TCGA characterization of CCRCC and other groups have shown that mutations in BRCA-associated protein 1 (BAP1), often demonstrated by loss of nuclear staining by BAP1 immunohistochemistry (Figure 1, F), were associated with poor prognosis for patients with these tumors.4,16 Interestingly, BAP1-mutated papillary RCC (PRCC) and chromophobe RCC (CRCC) samples did not show this prognostic association, revealing differences in the effect of BAP1 somatic aberration in the context of histologic subtypes of RCC17 ; these findings need to be further validated in larger patient cohorts. Incorporation of BAP1 immunohistochemical expression in clinical urology practice is of potential promise; however, the application of this marker in the RCC patient management algorithm still requires specific utilitarian characterization. Specifically, practical application may include patients with small renal masses who might be candidates for active surveillance.
Although multigene-based diagnostic assays have long been used for breast cancer patients,18,19 there are convincing new data indicating that such assays may have clinical utility in renal cancer patient management. Based on our experience in application of a multigene signature (the cell cycle proliferation score) in patients with RCC after radical nephrectomy, the cell cycle proliferation score was found to be independently and significantly associated with risk of recurrence and cancer-specific survival.20 This specific assay holds promise for meaningful application in renal cancer management.
Clear cell papillary RCC (CCPRCC), another renal tumor demonstrating cytoplasmic optical clarity or clear cell features, typically demonstrates low-grade apically oriented nuclei; papillary and cystic areas may or may not be readily evident in such tumors. When necessary, a panel of biomarkers may be helpful in confirming the diagnosis such that in contrast to CCRCC, CCPRCC shows strong CK7 expression (often seen in more than 90% of the neoplastic cells), and CA-IX shows expression in a cuplike fashion (Figure 2, A through C). CCPRCC also often demonstrates positivity for CK34BE12 expression; other immunohistochemical markers that may be helpful include negative AMACR expression. Although CCRCC often cytogenetically demonstrates 3p loss, CCPRCC is considered to lack this and other von Hippel–Lindau tumor suppressor gene (VHL)–associated mutations or polysomies (chromosomes 7 or 17). Analogous to CCRCC, there is currently no cancer- or lineage-specific biomarker for CCPRCC.
Another renal tumor that may demonstrate some clear cell features, transcription elongation factor B polypeptide 1 (TCEB1)–mutated RCC, is a newly described entity and a distinct subtype of RCC with clear cytoplasm, tubulopapillary architecture, and unique fibromuscular stromal bands imparting the tumor a multinodular appearance (Figure 2, D). TCEB1-mutated RCC is defined at the molecular level by characteristic mutations in TCEB1, a gene that contributes to the VHL complex in order to ubiquitinate hypoxia-inducible factor (Figure 2, E).21 Occurrence of a TCEB1 mutational hotspot at tyrosine 79 position has been reported.22 The following example also highlights some of the difficulties encountered in the histologic subtyping of tumors in this molecular subgroup: of the total 4 TCEB1-mutated tumors sequenced by the TCGA, 3 were histologically judged prior to sequencing as CCRCC and 1 (TCGA-5P-A9JV) as PRCC.
Immunohistochemically, CA-IX is diffusely positive in a membranous, boxlike pattern in TCEB1-mutated RCC, whereas CK7 can be patchily positive, precluding simple classification. Unfortunately, TCEB1 mutation–specific testing is not currently clinically available for diagnostic use for patients who might harbor such tumors. When TCEB1-mutated RCC is morphologically suspected, the authors of this review suggest signing out such cases with a differential diagnosis that includes CCRCC and CCPRCC in addition to TCEB1-mutated RCC, favoring that the tumor “is clinically/biologically similar to a CCRCC.” Specifically, the implicated TCEB1 mutations involve VHL-binding sites, and therefore although TCEB1-mutated RCCs do not technically involve the loss of VHL commonly observed in conventional CCRCC, they harbor a similar overall dysfunction in the VHL pathway, and are thought to be clinically and biologically similar to CCRCC.21
MIT gene family aberration–associated RCC, often generally referred to as translocation-associated RCC, consists of a morphologically diverse category of tumors that may show clear cell features. In particular, clear cell features without otherwise classically diagnostic morphology or immunohistochemistry may require exclusion of a TFE3-translocated RCC (Figure 1, C inset) and/or TFEB-translocated RCC. For brevity, these tumors will be discussed in further depth below among tumors with papillary features. See Table 2 for recommended immunohistochemical and cytogenetic testing and results to assist in identifying such cases.
BIOMARKERS IN RENAL TUMORS WITH PAPILLARY ARCHITECTURE
Papillary architecture is not unique to PRCC (Table 1); that said, PRCC is the most common RCC with a papillary architecture. PRCC accounts for 13% to 18% of renal epithelial neoplasms and has traditionally been split into types 1 and 2. Although type 1 PRCC frequently demonstrates cells with cuboidal cytoplasm that may be basophilic, amphophilic, eosinophilic, or relatively clear, type 2 PRCC is a diagnosis of exclusion. Morphologically, type 2 PRCC is characterized by cells with abundant eosinophilic cytoplasm, nuclear anaplasia, and nuclear pseudostratification. It is very important to exclude other entities from the differential diagnoses before rendering a diagnosis of PRCC type 2; such entities include (but are not limited to) TFEB-amplified RCC, hereditary leiomyomatosis and RCC (HLRCC)–associated RCC, acquired cystic disease (ACD)–associated RCC, and collecting duct carcinoma.
In general, PRCC continues to be a morphologic diagnosis supported by CK7 and AMACR positivity by immunohistochemistry (when necessary); type 2 PRCC demonstrates more variable CK7 and/or AMACR expression when compared with type 1. As elucidated and confirmed in comprehensive molecular characterization of PRCC by the TCGA research network, a subset of type 1 PRCC is associated with mutations in MET, a receptor tyrosine kinase gene; gain of chromosome 7 (Figure 3); and predictable mRNA expression. Type 2 PRCC, on the other hand, has a heterogenous molecular profile with at least 3 identifiable subtypes.5 Therefore, type 2 PRCC remains a debatable diagnosis of exclusion, especially given the evolving understanding of newer diagnostic entities discussed below in this section. Currently, there are no established cancer- or lineage-specific biomarkers known for PRCC. Few candidates such as AKR1B10, a target of the activated KEAP1-NRF2 pathway, have been proposed23 as biomarkers for PRCC type 2; however, further evaluation is necessary, as KEAP1 epigenetic loss was also reported in CCRCC.24
Papillary RCC can have overlapping histologic and immunohistochemical features (CK7+, AMACR+) with another renal tumor that may demonstrate focal papillary features known as mucinous tubular and spindle cell carcinoma (MTSCC), an often comparably favorable diagnosis. Mucinous tubular and spindle cell carcinomas are morphologically characterized by epithelioid cells in tubules and low-grade spindle cell patterns embedded in a myxoid or mucoid background (Figure 4, A and B); the occurrence of rare high-grade MTSCC and MTSCC with high-grade sarcomatoid features has been reported. In particular, PRCC with low-grade spindle cell foci may resemble low-grade spindle cell areas of MTSCC, especially with limited specimens or sampling, and hence these entities present a diagnostic conundrum.25 Mucinous tubular and spindle cell carcinoma is characterized by distinct cytogenetic copy number alterations not present in PRCC (Figure 3).26,27 Given similar immunohistochemical expression patterns (including positive CK7 and AMACR expression), differentiation of MTSCC from PRCC can be challenging. By combining and comparing RNA-seq data from an institutional cohort of MTSCC as well as specific TCGA cases, our group identified VSTM2A (V-set and transmembrane domain–containing 2A) as a cancer-specific biomarker for MTSCC; VSTM2A overexpression as identified by RNA in situ hybridization is both sensitive and specific for a diagnosis of MTSCC.12 This assay is currently under clinical development at the University of Michigan (Ann Arbor) (Figure 4, B) to help in differentiating MTSCC and PRCC (and other entities). The relatively rapid translational investigation, development, and validation of this biomarker assay highlight the utility of leveraging next-generation sequencing data for accelerating the investigation and understanding of novel diagnostic entities.
HLRCC is caused by germline mutations in the fumarate hydratase (FH) gene, and is associated with increased incidence of leiomyomas and a potentially aggressive variant of RCC (HLRCC-associated RCC). HLRCC-associated RCC (and the associated FH-deficient RCC in the absence of an identifiable germline predisposition) is defined by a loss of heterozygosity of the FH gene. Morphologic features of HLRCC-associated RCC are variable but often show enrichment for papillary architecture (admixed with other morphologic patterns), prominent nucleoli, and perinucleolar clearing; these tumors have likely been classified as type 2 PRCC in the past, but distinct tumor biology and pathogenesis, as well as emerging therapies for HLRCC and surveillance protocols for patients and affected family members, necessitate accurate diagnosis of these tumors (Figure 4, C through F). Very importantly, and not surprisingly, following accurate diagnosis, patients with HLRCC-associated RCC and their families often undergo genetic counseling and appropriate surveillance for development of new malignancies. Loss of FH expression (in the presence of appropriate controls) by immunohistochemistry is a very useful clinical aid; diffuse positivity for 2-succinocysteine (2-SC) immunostain can also be helpful but is currently not routinely clinically available.28,29 Patients with previously undiagnosed HLRCC who are thought to harbor FH-deficient RCC, especially those with increased 2-SC modifications, should be referred for genetic and clinical workup to assess for a possible syndromic association with HLRCC. Finally, chromosomal numerical aberration patterns in these tumors are common but variable, and thus, unfortunately, do not provide a useful tool for identifying these cases.30 Morphologic suspicion with immunohistochemical and ancillary genetic workup remains crucial for identifying this subtype.
A mixed architectural pattern with a variable combination of papillary, cribriform, tubular, cystic, and solid growth characterizes ACD-associated RCC, a distinct RCC subtype included as an established entity in the 2016 World Health Organization classification; these tumors are seen to occur in end-stage renal disease and specifically with acquired cystic kidney disease, a common morbidity associated with dialysis for renal replacement therapy.31,32 Acquired cystic disease–associated RCC imparts a relatively low risk of aggressive behavior, with 2 patients (6%) developing local recurrence and 2 patients (6%) developing metastasis in a previous cohort of 36 patients with available clinical follow up.32 Cells in ACD-associated RCC have eosinophilic granular cytoplasm with intracellular and intercellular lumina formation; papillary architecture may be seen in approximately 70% of cases (Figure 5, A through D). Immunohistochemistry is somewhat nonspecific, but the CK7 negative/AMACR positive phenotype may be helpful in diagnostically challenging situations.31–35 CD117 expression has been shown to be negative, allowing for distinction from other oncocytic neoplasms that may be on the differential diagnosis, especially with limited specimens.32 Genetically, ACD-associated RCC has been shown to demonstrate chromosomal abnormalities that overlap with those of PRCC (gain of chromosomes 3, 7, 16, and 17) as well as aberrations that are not enriched in PRCC (occasional loss of chromosomes 3 and 16 and gain of chromosomes 1, 2, 6, and 10).33,36,37 Although clinical, morphologic, and immunohistochemical clues are helpful, there remain no cancer- or lineage-specific biomarkers for ACD-associated RCC.
Papillary features are among the diverse morphologic spectrum of findings in RCC with MiT family aberrations, often referred to as translocation-associated RCCs. Previous work by our group and others highlights some of the morphologic and clinical features associated with TFE3 translocation, TFEB translocation, and TFEB amplification;38,39 morphologic features can often overlap with CCRCC, CCPRCC, and PRCC. Although morphologic suspicion is key, these tumors have undergone rich characterization during the past several years, and fluorescence in situ hybridization assays for clinical diagnosis are available, mostly at tertiary centers (Figure 6, A through F). Table 2 summarizes the pertinent ancillary testing for these tumors. Immunohistochemical positivity for Melan-A and HMB-45 is more frequently observed in TFEB-altered RCCs. Fluorescence in situ hybridization for TFE3 and TFEB gene aberrations exhibits a high degree of diagnostic sensitivity and specificity for RCC with MiT family aberrations and is clinically used and available at the University of Michigan.39
Finally, papillary features in the rare and benign metanephric adenomas of the kidney may sometimes be difficult to distinguish from malignant epithelial tumors, especially on limited specimens. Immunohistochemical interrogation for the BRAF V600E–mutant epitope is a valuable tool that demonstrates reliable positivity in metanephric adenomas in cases for which the morphologic impression is equivocal.40
BIOMARKERS IN RENAL TUMORS WITH ONCOCYTIC CYTOPLASM
Renal oncocytic neoplasms (Table 1) are thought to frequently emanate from the intercalated cell and hence demonstrate positivity for CD117, known to be constitutionally expressed within a subset of cells in the distal nephron. Distinguishing ChRCC from other oncocytic neoplasms is clinically essential but can be difficult, especially in the case of the eosinophilic variant of ChRCC. A morphologic assessment is often most helpful in distinguishing renal oncocytoma from the classic and eosinophilic variants of ChRCC, with immunohistochemical markers needed in occasional cases to support morphologic impression. Along with positive CD117 expression, ChRCCs classically express diffuse and uniform CK7, which can be of diagnostic utility. The eosinophilic variant of ChRCC, on the other hand, demonstrates rare or patchy CK7 expression, in a manner similar to that seen in renal oncocytomas and other low-grade oncocytic tumors of the kidney. Based on recently presented data, ChRCC can demonstrate long noncoding RNA marker LINC01187 expression in primary and metastatic sites, and this may be a useful biomarker for this tumor type, especially in the metastatic setting.41 Currently, apart from those in research and development (and presented as abstracts at conferences), there are no clinically available cancer- or lineage-specific biomarkers for renal oncocytoma, ChRCC (including the eosinophilic variant), or other related oncocytic tumors.
Oncocytic morphologic features also require careful consideration of succinate dehydrogenase (SDH)–deficient RCC. Morphologic features of SDH-deficient RCC include cells with vacuolated or bubbly cytoplasm; characteristic cytoplasmic inclusions containing pale material and neuroendocrine-like nuclei are helpful but not always present (Figure 7, A through D). Loss of SDHB by immunohistochemistry is of great diagnostic assistance and is associated with tumors resulting from germline mutations of SDHB, SDHA, SDHC, SDHD, and SDHAF2, which can also be associated with extrarenal manifestations such as paragangliomas and gastrointestinal stromal tumors.42–44 Recognizing this subtype is essential for clinical identification of patients for whom surveillance, family genetic counseling, or other interventions may be indicated.
Eosinophilic solid and cystic RCC is a morphologically, clinically, and immunohistochemically distinct tumor that occurs predominantly in female patients. Characteristic eosinophilic and voluminous cells comprising solid and cystic architectural patterns typically stain positive for CK20 and negative for CD117 expression (Figure 8, A through C). Molecular interrogation by our group and others has demonstrated frequent loss of heterozygosity and copy number gains at multiple sites, including somatic mutations involving TSC1 and TSC2 (in the absence of germline defects), which may guide future therapeutic strategies.10,45,46 Eosinophilic solid and cystic RCC is a remarkable example of an entity that has been described well only in the last few years with clarification of its morphologic, clinical, and molecular identity, thus providing evidence and support for further modifications to the current 2016 World Health Organization renal tumor classification. Although the morphologic features and molecular underpinnings for these tumors have now been well described, there are currently no clinically available cancer- or lineage-specific biomarkers for eosinophilic solid and cystic RCC.
Renal cell carcinoma, unclassified, is a diagnostic category for tumors with features not readily fitting into recognized RCC subtypes; this group does not simply represent a pathologist's inability to correctly classify a renal tumor, and instead may reflect the unknown clinical and biological findings associated with a novel or rare renal tumor type. Renal cell carcinoma, unclassified, is a diagnosis of exclusion and requires excluding the above-described RCC types, including (but not limited to) HLRCC-associated RCC, the eosinophilic variant of ChRCC, and SDH-deficient RCC, as well as collecting duct carcinoma and renal medullary carcinoma. In work by Chen et al47 to molecularly characterize 62 cases of RCC, unclassified, recurrent somatic mutations in 29 genes, including NF2, SETD2, BAP1, KMT2C, and MTOR, were identified, and 76% of the cohort fit into distinct molecular subsets with differential clinical outcomes. In particular, aggressive clinical behavior was noted in RCC, unclassified, tumors with NF2 loss, often via biallelic inactivation of NF2 with concurrent NF2 mutation and loss of heterozygosity; this novel finding in RCC may help develop assays to more richly predict patient prognosis. In contrast to traditional thinking of the group of RCC, unclassified, as a “wastebasket,” these tumors do have characteristic genomic features, lending validity, for now, to this diagnostic category and encouraging additional work into subtyping.
SEQUENCING-BASED DISCOVERIES IN RCC WITH BIOMARKER IMPLICATIONS
To date more than 1500 kidney tumors have been characterized by various DNA sequencing (whole-genome and whole-exome sequencing), RNA (total and polyA RNA sequencing), microRNA sequencing, and high-throughput technologies, including those to monitor DNA methylation and protein/phosphoprotein expression by TCGA and the Clinical Proteomic Tumor Analysis and TRACERx consortia (https://portal.gdc.cancer.gov). These studies have fostered the development of several data/sample histology visualization portals, including the following, notable for their ease of use: https://portal.gdc.cancer.gov; http://ccrcc.cptac-data-view.org/; https://www.cbioportal.org/; http://maplab.imppc.org/wanderer/; and https://cancerimagingarchive.net/datascope/cptac/ (all websites accessed as of August 16, 2019). Some of these data portals allow simultaneous examination of sample histology slides alongside the molecular aberrations of a given case, a feature that will serve as a great comparative resource for pathologists. The ongoing Clinical Proteomic Tumor Analysis Consortium's in-depth proteomic and phospho-proteomic data assessments from their discovery and validation cohorts is likely to have a major impact on RCC (and other solid tumor) protein biomarker identification. Thus far the molecular results from such initiatives (TCGA and TRACERx) have been described in several publications, both individually according to RCC histologic subtype and in pan-RCC integrative analysis.17,48 Although the current review focuses mainly on RNA and protein biomarkers, the pan-RCC analyses have captured cancer-specific aberrations that occur at the genomic, transcriptomic, epigenomic, and proteomic levels, and this information should aid in both facilitating and prioritizing the development of novel proteo-genomic biomarkers for RCC. For instance, Chen et al48 have identified the presence of 9 genomic subtypes within the 3 common histologic RCCs: the CC-e.1, CC-e.2, CC-e.3 (CCRCC), P-e.1a, P-e.1b (PRCC type 1), P-e.2 (PRCC type 2), Ch-e, mixed RCC, and P.CIMP-e categories. Importantly, they also identified an 800-gene signature that distinguishes these subtypes. Because CpG island methylator phenotype RCC (mostly found among the PRCC samples that have been sequenced) and CC-e.3 samples were associated with the worst prognosis, developing specific and sensitive biomarkers for CpG island methylator phenotype RCC and CC-e3 subtypes is likely to help with future disease prognostication. A few cases of rare RCC samples, such as MTSCC, translocation-associated RCC, TCEB1-mutated RCC, NF2-mutated unclassified RCC, and others, have also been profiled by TCGA but remain embedded in the data sets. Thus, future analysis of sequencing data of rare RCCs from independent groups, if integrated with the common RCC data set generated by these consortia, will facilitate the identification of novel RCC subtype–specific biomarkers. Such an approach led to the identification of VSTM2A as a biomarker for MTSCC, as stated above, and could be expanded for other RCC histologic and molecular subtypes as well.
The pan-RCC integrative analysis studies have identified several important associations between molecular features and disease prognosis in RCC. Capturing such information using a comprehensive multigene panel/platform will be of essence in making informed diagnostic/therapeutic decisions in the future. Taking CCRCC as an example, TRACERx studies6,49,50 using whole-genome sequencing data showed a complex chromothripsis event as the mechanism of translocation between chromosomes 3 and 5 leading to 3p loss in nearly 40% of patients. Sequencing multiple tumor specimens from a given patient identified 7 evolutionary subtypes of CCRCC in which mutational heterogeneity, loss of chromosomes 9p or 14q, and genome instability were each associated with poor prognosis. More recently, Bihr et al51 used simultaneous PBRM1, BAP1, and histone H3K36me3 (a surrogate for SETD2 mutations) immunohistochemistry to characterize some of the TRACERx proposed evolutionary CCRCC subtypes in their institutional cohort. Because of limitations in the various personalized medicine gene panels, some of the genomic features mentioned above are not fully captured with currently used testing platforms. The challenge in the coming years will be to devise tailored methods to capture all these molecular features and to present and integrate them with histologic findings to generate a comprehensive pathology report or assessment.
Several laboratories, including ours, have embarked on characterizing the gene expression of benign kidneys, renal tissues from chronic kidney diseases, and histologically distinct malignant renal samples at the single-cell level.13–15 These single-cell mRNA sequencing data will result in the generation of so-called cell expression atlases, in which the gene signature of each benign cell type present in the kidney and how these signatures are altered during various renal pathologies can be studied. Such data sets, when integrated with bulk tissue sequencing data from the consortia described above, will uncover novel disease mechanisms and above all will provide a major thrust to the field of RCC biomarker discovery in the near future.
INCORPORATION OF NOVEL TECHNOLOGIES INTO WORKUP OF RCC
The advent and adoption of new technologies has increased the pace of discovery and characterization of novel entities as well as the discovery of biomarker assays. As mentioned above, the accessibility of next-generation sequencing for use on clinical specimens, as used in TCGA, TRACERx, MI-ONCOSEQ,52 MSK-IMPACT,53 and other institutional and multi-institutional efforts, has led directly and via hypothesis generation to a biologically reinforced restructuring of older categories of renal tumors; the clinical, prognostic, and therapeutic relevance within these discoveries cannot be understated given the relative prevalence, morbidity, and mortality associated with renal malignancies.
Beyond research applications, the increasing availability of ancillary testing within the routine clinical workflow has also had an impact on renal tumor diagnosis as well as on the clinical workflow itself. Although morphologic suspicion is still crucially important, the modern renal tumor workup requires the possible use of fluorescence in situ hybridization assays, a variety of immunohistochemical stains, and RNA in situ hybridization for the identification and confirmation of the entities discussed above. Figure 9 portrays a simplified diagram of the way in which biomarker discovery and validation can fit into the clinical workflow; in an ideal system, clinical work both benefits from and contributes to assay development. This efficient translational approach to assay development parallel to clinical practice is discussed in more depth in our prior work.54 Moreover, the adoption of clinically ordered cancer sequencing for identification of therapeutic targets (ie, MI-ONCOSEQ, MSK-IMPACT, and others) or prognostic information on a patient-by-patient basis is also progressing; responsible selection of patients for this kind of testing is and will continue to be important for managing utilization and demands for interpretation.
CLINICAL GUIDELINES, TRIALS, AND PREDICTIVE BIOMARKERS
The current National Comprehensive Cancer Network (NCCN) 2019 guidelines for RCC do not discuss the routine use of biomarkers for RCC.55 In fact, there is not a principles of pathology section, and thus guidelines are focused on standard American Joint Committee on Cancer staging and Memorial Sloan Kettering Cancer Center or International Metastatic RCC Database Consortium prognostic models to guide treatment. As we have discussed, this will be an important section to expand in current national guidelines to better identify the various subtypes of RCC, distinct genomic alterations, and subsequent unique treatment sensitivities as these data continue to be elucidated.
The landscape of clinical trials in RCC is changing rapidly away from single-agent tyrosine kinase inhibitors (TKIs) toward immunotherapy-based backbones. The programmed death receptor-1 (PD-1)/programmed death ligand-1 (PD-L1) pathway has been implicated in inhibition of the antitumoral immune response,56 and PD-L1 expression has been associated with poor prognosis in metastatic RCC.57,58 Clinical trials of immunotherapy with anti–PD-1 and inhibitors of cytotoxic T lymphocyte–associated protein 4 (anti–CTLA-4) demonstrate improved overall survival when compared with prior standard TKI therapy in patients with advanced clear cell RCC59,60 ; improved overall survival was observed in cases with PD-L1 expression (as determined by immunohistochemistry), as well as cases without PD-L1 expression, albeit to a lesser extent. Therefore, PD-L1 expression by immunohistochemistry can offer important prognostic information and may predict the degree of response to immunotherapy, but it is often not a prerequisite for checkpoint inhibitor immunotherapy for patients with advanced RCC. Even more recently, nivolumab plus ipilimumab was found to be superior to sunitinib in the CheckMate 214 randomized trial in first-line untreated advanced RCC, and now the combination of the PD-1 and CTLA-4 inhibitor is a Food and Drug Administration–approved standard.59 However, the identification of predictive biomarkers in immunotherapy for RCC continues to present a challenge for disease management teams, as only 42% of patients receiving dual checkpoint blockade have an objective response, and only 9% have a complete response to treatment.59 Identification of these patients is a critical unmet need.
In nonmetastatic RCC there is an increasing focus on the benefit of neoadjuvant and adjuvant approaches by combining immunotherapy with standard surgical removal of the kidney (eg, PROSPER NCT03055013). In contrast, since the publication of the CARMENA trial demonstrating no benefit, and potential harm, to cytoreductive nephrectomy in combination with sunitinib in men with metastatic RCC,61 there is interest in determining the contemporary role of cytoreductive nephrectomy in the setting of immunotherapy now that it is approved as first-line therapy. Furthermore, the use of stereotactic body radiotherapy is being tested to treat localized RCC (NCT02613819, NCT02853162), oligometastatic RCC (NCT03575611), and oligoprogressive RCC (NCT02019576), and there is interest in testing whether stereotactic body radiotherapy can serve as an immunomodulatory agent to attempt to elicit an abscopal effect.
Most clinical trials in RCC have been somewhat permissive of various histologic entities in order to maximize accrual. Thus, most of the high-quality subtype-specific and biomarker data to date have been generated from post hoc analysis of randomized trials.62–64 Fortunately, there are a select number of molecularly driven trials underway. One such trial includes HLRCC and sporadic PRCC patients, and requires an FH mutation for HLRCC. It is a phase II trial testing the use of bevacizumab and erlotinib in this subset of kidney cancers (NCT01130519). Another phase II trial is selecting patients with a MET mutation with metastatic type 1 PRCC to determine the response to crizotinib (NCT01524926). There is an urgent unmet need to better understand the biology of RCC and design and enroll molecularly informed and rational clinical trials to best understand who benefits from our growing arsenal of treatment options.
The development of clinically applicable biomarkers is a complex process that must match the pace of discovery as the morphologically diverse world of renal neoplasms is further molecularly characterized and specifically subtyped. A broad but up-to-date understanding of the currently known renal tumor entities is essential for diagnosing and prognosticating these neoplasms. In addition to continuing to update our classification system, as members of the genitourinary oncology management team, we must be aware of morphologic features that are suggestive of molecularly defined entities, as well as the ancillary tests available needed for confirmation or rule-out.
The authors have no relevant financial interest in the products or companies described in this article.
Presented in part at the New Frontiers in Pathology meeting; September 27–29, 2018; Ann Arbor, Michigan.