Introduction

The Genitourinary Pathology Society (GUPS), established in 2018, is an international subspecialty organization that aims to advance the care of patients with urologic diseases by encouraging best practice, research, and education in urologic pathology. Renal neoplasms are increasingly recognized to be genetically complex, often requiring advanced morphologic, immunohistochemical, and/or molecular insights to distinguish distinct entities. Since the last edition of the World Health Organization (WHO) classification of renal cell tumors in 2016, significant advances have been made in the understanding of renal neoplasia [1].

In this project that included 42 urologic pathologists from 11 countries, GUPS focused on the new developments in the classification of renal neoplasia, resulting in: (1) proposals for the recognition of novel entities; (2) clarification and unification of nomenclature; (3) improved definitions and diagnostic criteria for some entities; and (4) integration of evolving molecular developments into routine practice. Working groups focusing on individual entities (group members are listed in Table 1) performed a critical literature review for the specific entities, where more significant developments occurred since the WHO 2016, and produced topical summaries. The groups also discussed and proposed resolutions for the open questions, a consensus terminology, and specific diagnostic criteria (if necessary), which were then discussed, modified, and ratified by consensus of the entire authorship group.

Table 1 Working groups.
Full size table

The project was organized into two parts: (1) New developments in currently recognized renal neoplasms (tumors currently listed in the WHO 2016 classification); and (2) new, emerging, and provisional entities. This article covers the first part, and the new, emerging, and provisional entities are presented in the companion article. A summary of the key findings for the currently recognized renal entities in the WHO classification covered in this paper is shown in Table 2.

Table 2 Summary of the key findings for the currently recognized renal entities in the WHO classification.
Full size table

Clear cell renal cell carcinoma (RCC)

Clear cell RCC is the most common RCC subtype, which accounts for over 60% of adult renal carcinomas and for most of the associated deaths. The majority of tumors are sporadic; however, a subset occurs in the setting of von Hippel Lindau (VHL) syndrome and rare cancer predisposition syndromes, including the recently described BAP1 (BRCA1 associated protein 1) tumor predisposition syndrome [2,3,4,5].

Although clear cell RCC is characterized by cells with clear cytoplasm and a delicate capillary network, its wide morphologic spectrum is being increasingly recognized. For example, such patterns and features include infiltrative growth, eosinophilic cytoplasm or globules, poorly differentiated adenocarcinoma-like morphology, rare papillary formation, giant multinucleated tumor cells, and sarcomatoid/rhabdoid morphology (Fig. 1A–D) [6, 7]. Some of these features have been associated with worse prognosis and may serve as biomarkers of prognosis [8,9,10,11]. Clear cell RCC with sarcomatoid differentiation appears to preferentially respond to immunotherapy (immune checkpoint inhibitors) over angiogenesis inhibitors [12]. Responsiveness to these drugs may also be affected by the presence of inflammatory cells or a vascular network in the tumor microenvironment. Although there is no entirely specific immunohistochemistry (IHC), clear cell RCC typically shows diffuse membranous positivity for carbonic anhydrase IX (CAIX), a hypoxia marker related to the HIF and dysregulated VHL pathway [13,14,15,16]. Although immunostaining for cytokeratin (CK) 7 is typically absent or focal, rare cases may show diffuse staining [17,18,19].

Fig. 1: Clear cell RCC with heterogeneous morphologic patterns.
figure 1

Such patterns may include: A syncytial-type giant tumor cells, B transition to poorly differentiated adenocarcinoma-like morphology (upper right), C features mimicking clear cell papillary RCC, or D abundant eosinophilic cytoplasm.

Full size image

Molecular studies of clear cell RCC characteristically show VHL inactivation, typically through an intragenic mutation (or less frequently promoter hypermethylation) of one allele and a large deletion of chromosome 3p that includes the second copy of VHL [20,21,22,23]. Deletions of 3p result in simultaneous loss of several other important tumor suppressor genes at this locus, specifically PBRM1, SETD2, and BAP1 [24,25,26]. PBRM1 and BAP1 are two-hit tumor suppressor genes that drive the tumor grade and prognosis in clear cell RCC [27,28,29]. Other implicated genes include histone-modifying genes (KDM5C and KDM6A), MTOR pathway genes (TSC1, TSC2, MTOR, PIK3CA, PTEN), and TP53 [2, 20, 21, 23, 24, 30,31,32]. In keeping with its microscopic appearance, clear cell RCC demonstrates profound genomic heterogeneity. Tumor progression is often associated with additional chromosomal alterations including frequent losses of 9p and 14q, and gain of 5q [23]. As tumors evolve and metastasize, distinct genomic subclones may develop, some of which may be more aggressive. For example, multiple “evolutionary” subclones were identified within each clear cell RCC deposit, associated with various behaviors, from indolent to metastatic [33]. Thus, sampling of multiple tumor areas per tissue block has been proposed to increase detection of potential aggressive clones, without significantly increasing the cost; if geographically separate tumor areas are sampled for molecular testing, they could be tested by massive parallel sequencing performed as one sample, potentially resulting in identification of aggressive/resistance mutations [34].

Of the most common RCC subtypes, clear cell RCC is an independent predictor of adverse outcome, with a significantly worse cancer-specific outcome than papillary and ChRCC [35]. Validated prognostic features within clear cell RCC include WHO/ISUP grade, AJCC pTNM stage, coagulative tumor necrosis, and rhabdoid and sarcomatoid differentiation [36,37,38,39,40,41,42,43,44]. Other features associated with adverse outcome include retrograde venous invasion and microvascular invasion [36, 45, 46]. Tumors >5 cm often extend into the renal sinus or vein branches and should be carefully examined, if these findings are not identified [47, 48].

There remain unresolved questions within clear cell RCC that require further study, including the prognostic significance of architectural patterns [8, 10], small-vessel and lymphatic invasion, perinephric fat vs. renal sinus invasion, as well as the interobserver variability of grading and assessing coagulative tumor necrosis. Additional work is also needed in the evaluation of histologic features and genomic alterations associated with response or resistance to therapy including immunotherapeutic agents.

Papillary renal cell carcinoma

With recognition of new RCC subtypes, many demonstrating at least focal papillary architecture, the diagnosis of papillary renal carcinoma (PRCC), the second most common RCC, is now restricted to a more precisely defined entity. PRCC shows a spectrum of clinical, morphological, and molecular features, ranging from low-grade to high-grade tumors. Extensive necrosis is common in PRCC and is not an adverse prognostic factor (in contrast to clear cell RCC and chromophobe RCC). The previous division into type 1 and type 2 subtypes [49] is not recommended for several reasons [50, 51]. Most importantly, this distinction shows poor interobserver reproducibility and lacks proven clinical significance. In practice, a large proportion of tumors demonstrate overlapping features (mixed type 1/type 2) [51, 52]. It is now recognized that the group of tumors previously classified as “type 2 PRCC” actually includes many previously unrecognized, but distinctive types of carcinomas, including fumarate hydratase (FH)-deficient RCC and microphthalmia-associated transcription factor (MITF) family translocation RCC. It has been also recently shown that WHO/ISUP grade and tumor architecture, including specific morphologic patterns, rather than the classic tumor types (PRCC classic type 1 vs. 2), better predict outcome [53]. Although a combined histologic and molecular classification that stratifies PRCC into 4 biologic subtypes has recently been proposed and is suggested to show clinical relevance [51], this proposal awaits further validation and is currently not endorsed by GUPS for routine clinical practice.

Molecularly, many low-grade tumors commonly harbor gains of entire chromosomes 7 and 17, loss of Y, as well as possible gains of 12, 16, and 20 as well as up to 15% of MET alterations. A smaller percentage of high-grade tumors also exhibit the classic gains of 7 or 17. Gains in part of chromosomes 17, 16, and 20 are more common, as well as a wide array of other chromosomal changes in some cases [54,55,56].

Although PRCC is not routinely subtyped, four new morphologic patterns have recently been described: “biphasic (alveolar/squamoid) PRCC”; “Warthin-like PRCC”; “solid PRCC” and “papillary renal neoplasm with reverse polarity” (Fig. 2A–D). It is important to recognize that these are currently considered part of the PRCC spectrum and not separate “subtypes”; additional studies are also needed to further validate their clinical significance.

Fig. 2: New morphologic patterns of papillary RCC.
figure 2

A Biphasic papillary RCC typically demonstrates a dual cell morphology and “glomeruloid/alveolar” architecture. The outer cell layer is formed of smaller cells, whereas the inner are larger, non-cohesive, and often contain emperipolesis (or cytophagocytosis). B Warthin-like papillary RCC contains eosinophilic tumor cells with admixed lymphocytes, mimicking Warthin tumor of the salivary glands. C Solid papillary RCC may be diagnostically deceptive, due to compression of the tubular and papillary structures into a solid mass. D Papillary renal neoplasm with reverse polarity is characterized by eosinophilic cells with low-grade nuclei aligned at the apex of the cells. Most of these tumors have been found to harbor KRAS mutations.

Full size image

Biphasic (alveolar/squamoid) PRCC is characterized by a dual cell population of larger eosinophilic and smaller amphophilic/basophilic cells and often shows solid growth [57,58,59] (Fig. 2A). Cyclin D1 is only expressed in the larger cells. Evidence of aggressive behavior has been documented in up to 15% of patients [57,58,59]. MET alterations have been reported recently in 60% of biphasic PRCCs [60].

Warthin-like PRCC shows dense lymphocytic infiltration in the stroma and in the papillary cores, high-grade nuclei (WHO/ISUP 3/4) (Fig. 2B). It can also exhibit aggressive clinical course [61].

Solid PRCC has solid growth due to compressed tubules and abortive papillae at least focally, with the cells showing scant cytoplasm and low-grade nuclei (Fig. 2C). It has an indolent behavior, similar to low-grade PRCC [62].

Papillary renal neoplasm with reverse polarity also referred to in the literature as “oncocytic low-grade papillary RCC” and “type 4 PRCC” appears to be a distinct subtype, with recognizable morphology and IHC profile (Fig. 2D) [51, 63,64,65]. These tumors are typically small (< 4.5 cm), often cystic, and are composed of oncocytic cells arranged in a tubulopapillary architecture. They exhibit low-grade nuclei (WHO/ISUP nuclear grade 1–2) with luminal linear arrangement, opposite the basement membrane. The neoplastic cells are uniformly reactive for GATA3, CK7, and L1CAM, with variable/weak reactivity for vimentin and AMACR. Most (80–90%) harbor KRAS mutations. They have a very favorable prognosis, based on 73 reported cases, compiled from 7 series (follow-up range from 1 to 222 months), all without reported progression [50, 54, 63, 64, 66,67,68]. Although GATA3 expression and KRAS mutations suggest that this may be a unique tumor type, this tumor is currently considered a subtype of PRCC. This conclusion is supported by its predominant papillary architecture, the documented gains of chromosomes 7 and 17 and loss of Y, as well as the fact that KRAS mutations have been found in other tumors (https://www.cbioportal.org/).

When considering a diagnosis of PRCC, other entities that may show papillary architecture must be carefully considered and excluded. For example, other carcinomas with papillary growth in the high-grade spectrum include FH-deficient RCC, MITF family translocation family RCC, renal medullary carcinoma, collecting duct carcinoma, and acquired cystic disease-associated RCC. Tumors in the low-grade spectrum that can mimic PRCC include clear cell papillary RCC, metanephric adenoma, and mucinous tubular and spindle cell carcinoma.

Chromophobe RCC (ChRCC)

ChRCC accounts for 5–7% of all RCC (Fig. 3). Classic and eosinophilic are the two distinct morphologic variants of ChRCC. Sarcomatoid features are found in 2–8% of ChRCC and in a cohort of metastatic ChRCCs, sarcomatoid differentiation was associated with an increased risk of presentation with metastatic disease and a shorter interval until metastasis [69]. According to TCGA and others, the cell of origin of ChRCC has an mRNA expression signature that overlaps with the distal nephron, in contrast to clear cell RCC, which has a proximal nephron signature [70,71,72].

Fig. 3: Chromophobe RCC (ChRCC).
figure 3

Classic ChRCC is composed of cells with abundant pale-to-eosinophilic cytoplasm, irregular nuclei, and perinuclear clearing, which are arranged in large trabeculae with an associated incomplete branching vascular network (A). KIT (CD117) labeling is present in a membranous fashion (B). ChRCC demonstrate high-level nuclear LINC01187 expression by RNA in situ hybridization (C). The eosinophilic variant of ChRCC demonstrates eosinophilic cytoplasm and subtle perinuclear clearing, arranged in small, tightly packed nests and acini (D). Sarcomatoid ChRCC demonstrates high-grade spindle cell areas (E) with loss of FOXI1 protein in the high-grade spindle cell areas (F), lower right). Distribution of the most frequent whole chromosomal copy number variations observed are shown in classic ChRCC (G) and eosinophilic ChRCC (H).

Full size image

In addition to KIT (CD117) and CK7 immunolabeling (diffuse in classic, may be rare/patchy in eosinophilic), ChRCC frequently demonstrates FOXI1 protein expression and long noncoding RNA LINC01187 expression in both primary and metastatic sites; consequently, FOXI1 can be used as a potential biomarker, especially in metastases [71, 73].

Although most ChRCCs typically have a good prognosis, an accurate prediction of the clinical behavior of ChRCC using histologic features may be challenging. Evaluation of a large ChRCC cohort showed that pT stage, tumor necrosis, and sarcomatoid change were independently predictive of aggressive behavior in multivariable analysis [74]. The largest single-institution morphologic study showed significant association of clinical outcome with tumor size, small-vessel invasion, sarcomatoid features, and microscopic necrosis [75]. In contrast to clear cell RCC and PRCC, grading of ChRCC using the WHO/ISUP system is currently not recommended [36, 76]; however, further study of ChrRCC not associated with either necrosis or sarcomatoid differentiation may provide additional insight into this issue.

Most ChRCCs demonstrate complete losses of 7 chromosomes (1, 2, 6, 10, 13, 17, and 21), distinguishing them from other RCCs and renal oncocytoma [77,78,79,80,81,82]. However, eosinophilic ChRCC harbors a significantly lower frequency of individual losses of the 7 chromosome panel, and is diploid in up to 41.7% cases [83]; this is in contrast to the chromosomal instability observed in classic ChRCC [80,81,82]. ChRCC also displays fewer hypermethylation events and harbors a lower mutational burden compared with both clear cell RCC and PRCC [2, 70, 84]. In addition to TP53 mutations (20–32%), PTEN mutations (6–9%), TERT promoter mutations/rearrangements (12%), and mitochondrial DNA alterations, other mutated genes identified in ChRCC include MTOR (up to 9%), TSC1/2 (up to 7%), and NRAS [84,85,86]. Recent studies evaluating the genetics of metastatic ChRCC found that high-risk features associated with poor survival included TP53 mutations, PTEN mutations, CDKN2A alterations, DNA hypermethylation, and imbalanced chromosome duplication (defined as duplication of ≥3 chromosomes) [2, 85].

In the era of “histo-molecular” tumor classification, the spectrum of “eosinophilic ChRCC” has evolved. A strong contributing factor for this development is the lack of stringent diagnostic criteria in the WHO 2016 classification, which only stipulates that eosinophilic ChRCC should be “almost purely eosinophilic” [87]. Eosinophilic ChRCC typically shows compact acinar architecture, abundant eosinophilic cytoplasm, focal nuclear wrinkling, and subtle perinuclear clearings/halos. Low-grade oncocytic tumor (LOT), a recently described entity (covered in detail in the companion paper on new renal entities), has emerged from spectrum of difficult to classify oncocytoma-eosinophilic ChRCC tumors [88]. LOT typically lacks significant nuclear irregularities and is negative for KIT (CD117) immunohistochemistry [88]. A subset of these tumors has likely been previously considered “eosinophilic ChRCC” [70]. Finally, for the tumors with overlapping morphologic features of renal oncocytoma and eosinophilic ChRCC, previously most commonly designated as “hybrid oncocytic/chromophobe tumor” (HOCT), we now propose the term “oncocytic renal neoplasm of low malignant potential, not further classified”. This category should be specifically reserved for solitary, sporadic cases with overlapping features of oncocytoma and eosinophilic ChRCC, but lacking adequate criteria to diagnose either of these unequivocally. The term “hybrid oncocytic tumor” should be restricted to hereditary cases, when such tumors are multiple and/or bilateral (as in Birt–Hogg–Dubé (BHD) syndrome) [87]. Due to significant morphologic and IHC overlap of the tumors in the oncocytoma-eosinophilic ChRCC spectrum, additional studies are required, correlating the genomic profiling with a careful morphologic evaluation, to more reliably categorize such tumors [89]. For example, a subset of RCCs exhibiting TSC/MTOR-aberrations have shown morphologic similarity to eosinophilic ChRCC, some in the TSC setting [90] or sporadically [91]. Most eosinophilic ChRCCs share the expression of the recently described biomarkers, LINC1187 and FOXI1, as in the classic ChRCC; however, a subset of tumors considered eosinophilic ChRCCs that have lower or absent biomarker expression harbor MTOR gene mutations [71, 92].

Oncocytoma and the spectrum of “difficult to classify” low-risk tumors

Renal oncocytoma is a benign renal tumor with distinctive gross appearance, architectural growth patterns, cytoplasmic, and nuclear characteristics [93,94,95]. Strict adherence to most typical features allows accurate diagnosis of oncocytoma in the majority of cases. The morphologic spectrum of oncocytoma also includes features such as vascular/renal vein involvement, involvement of perinephric adipose tissue, small cells (“oncoblasts”), and localized collections of pleomorphic cells (so-called degenerative-type atypia). These features have not been found to alter its benign behavior [93, 94, 96, 97]. Presence of coagulative necrosis, well-formed papillary formations, and brisk or atypical mitotic activity, should deter from a diagnosis of oncocytoma. Immunohistochemical labeling for CK7 is usually restricted to rare, scattered cells, typically accompanied by diffuse reactivity for KIT (CD117), and negative reactivity for vimentin. Molecularly, oncocytoma usually harbors limited chromosomal abnormalities compared to ChRCC, ranging from a normal karyotype to loss of chromosome 1 and/or Y or rearrangement of 11q13 (likely corresponding to rearrangement of CCND1 in most cases) [20].

One of the most challenging areas in the routine clinical practice remains the diagnosis of tumors with overlapping or equivocal morphology between oncocytoma and ChRCC, as well as differentiating oncocytoma and ChRCC from eosinophilic/oncocytic tumors that do not completely fit the diagnostic criteria for these two tumors (Fig. 4). Such “borderline” tumors lack the classic cytologic features of ChRCC (for example, irregular “raisinoid” nuclei), but show a variable constellation of atypical features that may include greater nuclear size variability or nuclear shape irregularity, with cytologic atypia dispersed throughout the tumor, as opposed to limited clusters of cells with degenerative-type atypia, seen in oncocytoma. When referring to such tumors that do not strictly fit either oncocytoma or ChRCC, various diagnostic terms have been used, such as hybrid oncocytic tumor, hybrid oncocytic chromophobe tumor (HOCT), “oncocytic renal neoplasm” accompanied by additional modifier terms including “low-grade,” “borderline features,” “unclassified,” “low malignant potential,” “uncertain malignant potential,” and others [98,99,100,101]. Given a lack of consensus on nomenclature and diagnostic criteria for such tumors, we recommend a standardized approach: (1) follow very strict definitions for prototypical oncocytoma and ChRCC, (2) utilize a borderline diagnostic category for cases that do not precisely fit, and (3) exclude other well-described mimics (for example, epithelioid angiomyolipoma and succinate dehydrogenase (SDH) deficient RCC). A survey of urologic pathologists showed that 82% use an “intermediate” diagnostic category when encountering tumors with borderline morphology and 76% supported creation of an official designation [101]. Therefore, for such “borderline” cases we propose the term “oncocytic renal neoplasm of low malignant potential, not further classified”, which should be reserved for solitary, sporadic tumors with overlapping features. It should be understood that this is primarily a “clinical management” category that likely includes a heterogeneous group of renal neoplasms. Importantly, the collective experience from large renal cancer centers strongly suggests that these “difficult to classify” tumors have exceedingly low risk for developing metastatic disease. Use of a standardized term for such cases will facilitate further study and comparisons between institutions. At present, there is no compelling evidence to suggest that immunophenotypic or molecular testing aid in predicting biologic risk for tumors in this spectrum. The term “hybrid oncocytic tumor” has been used variably in the literature. We suggest that this term be reserved for hereditary cases, which are often characterized by scattered clusters and individual cells with clear cytoplasm, exhibiting a “checkerboard” mosaic pattern (as in BHD syndrome) [74, 93, 100, 101].

Fig. 4: Oncocytic neoplasm with borderline features between oncocytoma and eosinophilic ChRCC.
figure 4

For such tumor, characterized by nuclear size variation (A) or mostly compact architecture with minor or focal perinuclear clearing (B), we propose the designation “oncocytic neoplasm of low malignant potential, not otherwise classified”.

Full size image

The diagnosis of these neoplasms on renal core biopsy can be problematic due to regional heterogeneity of tumors. For example, ChRCC (or other subtypes) may have foci indistinguishable from oncocytoma, which may lead to a sample that is not diagnostically representative [74]. Although some pathologists are willing to render a diagnosis of oncocytoma in a biopsy sample, provided the morphology is supported by a typical immunohistochemical pattern that excludes other mimics, GUPS recommends doing so with caution. An alternate approach is to diagnose neoplasms with the appearance of oncocytoma on biopsy as “oncocytic renal neoplasm”, accompanied by a comment: “If this biopsy sample is representative of the entire lesion, the appearances would be consistent with an oncocytoma”.

Clear cell papillary renal cell carcinoma (CCPRCC)

CCPRCC, also referred to as clear cell tubulopapillary RCC, was included in the 2016 WHO classification as a novel RCC subtype RCC [102,103,104]. It is the fourth most common subtype of RCC, occurring both sporadically and in the setting of end-stage renal disease [105]. It accounts for ~3% of early-onset renal cell carcinomas [106, 107]. CCPRCC exhibits macrocystic and microcystic patterns with papillary infoldings, branching acini and tubules, and occasionally solid and collapsed acinar patterns (Fig. 5). CCPRCC shows diffuse CK7 expression, “cup-like” CAIX expression, and reactivity for high molecular weight cytokeratin, whereas AMACR and CD10 are negative [104].

Fig. 5: Clear cell papillary RCC.
figure 5

This tumor often demonstrates a branching tubular pattern (A) with nuclei oriented in a linear array opposite the basement membrane (B). Papillary formations may be limited, manifesting only as small papillae in cystic spaces (C). Immunohistochemistry is consistently positive for cytokeratin 7 (D, top) and often for high molecular weight cytokeratin (D, bottom).

Full size image

CCPRCC does not show loss of chromosome 3p, mutations in VHL, or VHL promoter hypermethylation. Cases previously reported as CCPRCC or resembling CCPRCC that arise in the setting of VHL syndrome or associated with VHL mutations probably represent misclassified clear cell RCCs mimicking CCPRCC [19, 108,109,110,111]. Although CCPRCC can sometimes exhibit more conspicuous smooth muscle stroma, recent evidence suggests that it is a different entity from the RCC with fibromyomatous (or leiomyomatous) stroma [59, 112, 113].

CCPRCC has not been shown to have any disease-specific genetic abnormalities, as detected by genome and exome sequencing, and characteristically has a low mutational burden compared to clear cell RCC [108, 114, 115]. A recent study found a gain of chromosome 18 as well as TET2 gene mutations, both in 3 of 6 cases [112]. Although CCPRCC is “quiet at the genomic level”, without a clear genetic driver for tumorigenesis, evidence suggests that the principal cellular alteration is metabolic, owing to disruption in the metabolism of sugar alcohols, and resulting in markedly elevated tumor sorbitol levels, compared with other RCCs and normal tissue. This phenomenon appears to be related to marked depletion of mitochondrial DNA, rather than abnormalities in the nuclear genome [115].

CCPRCC appears to be extremely indolent. To date, only one putative metastatic case has been reported [116]. Other cases reported as metastatic CCPRCC included some higher-grade areas, suggestive of clear cell RCC, and showed overlapping immunophenotypes between the two entities (18, 19). Thus, it could be postulated that these cases actually represent misclassified clear cell RCCs. Based on current evidence, CCPRCC may be a candidate for reclassification as a tumor of low malignant potential. Therefore, we recommend using strict diagnostic criteria, requiring both classic CCPRCC morphologic features and a typical immunohistochemical profile, as well as exercising caution when limited diagnostic material is available, as in biopsy specimens.

Tubulocystic renal cell carcinoma

Tubulocystic renal cell carcinoma (TC-RCC) represents <1% of all RCCs [117, 118]. TC-RCC was included as a novel entity in the WHO 2016 classification [119] (Fig. 6A). Initial reports considered tumors with a tubulocystic pattern admixed with papillary, solid, poorly differentiated, or sarcomatoid components to be part of the TC-RCC spectrum [119,120,121,122,123,124]. However, many such tumors had poor outcomes [124,125,126], contrasting the indolent behavior in pure TC-RCC [117, 127, 128]. In fact, many tumors comprising tubulocystic and other patterns represent FH-deficient RCC. Clues to the diagnosis of FH-RCC include the heterogeneous architecture that often includes a TC-RCC-like pattern, and their high-grade features [120, 129,130,131,132,133]. Therefore, the diagnosis of TC-RCC should be currently restricted to tumors with exclusive tubulocystic growth. A low threshold for performing FH (Fig. 6B) and/or 2-succino-cysteine (2SC) IHC is recommended, especially in limited samples where conservative therapy may be considered [134, 135].

Fig. 6: Tubulocystic renal cell carcinoma (TC-RCC).
figure 6

TC-RCC shows variably sized cysts (A) lined by flattened, cuboidal, or hobnail-shaped cells with eosinophilic cytoplasm and prominent nucleoli (inset). Strictly diagnosed, pure TC-RCCs tend to show an indolent course. On the other hand, due to the frequent presence of a pattern simulating TC-RCC in aggressive, high-grade FH-deficient RCCs, use of immunohistochemistry such as FH or 2SC is recommended. Strong, diffuse FH expression (B) (i.e., retained FH expression) in a morphologically pure tumor supports TC-RCC.

Full size image

The early molecular studies have generally supported the notion that TC-RCC is a unique entity [117]. Although a close relationship to PRCC was postulated, based on the frequent gains of chromosomes 7 and/or 17, and reactivity for vimentin, CD10, CK7, and AMACR in TC-RCC [121, 133, 136,137,138], others have not found chromosome 7 and 17 gains [127], or have only found gains of chromosome 17 [129, 139]. Recent molecular studies increasingly support the conclusion that morphologically pure TC-RCC represents a distinct entity [139, 140]. Targeted next-generation sequencing and FISH studies have identified recurrent losses of chromosome 9 (and Y, in males) and gains of chromosome 17, without mutations characteristic of the other RCC types [139]. In contrast, mutations of KMT2C and KDM5C chromatin-modifying genes were identified in a subset of TC-RCCs [139]. Similarly, noncoding RNA sequencing studies have identified distinct patterns of small nucleolar RNA and mature and pre-miRNAs, as well as frequent (>60%) mutations of ABL1 and PDGFRA [140]. While the lack of overlap between the observed mutations in these recent series implies molecular diversity and requires further study, these distinct mutational features support the classification of TC-RCC as a separate entity [139, 140].

MITF family translocation-associated renal cell carcinomas

MITF family translocation RCCs harbor gene fusions involving members of the MITF subfamily of transcription factors, predominantly TFE3 and TFEB. Although onset in childhood and younger age is common, there is no evidence of a hereditary predisposition.

TFE3 rearranged RCC (Xp11 translocation RCC)

Xp11 translocation RCC harbors gene fusions involving TFE3. The most distinctive histologic pattern of Xp11 translocation RCC is that of a papillary neoplasm with clear cells cytology and scattered psammoma bodies [141, 142]. Xp11 translocation RCCs can also mimic other renal neoplasms, including clear cell RCC, PRCC, CCPRCC, TCEB1/ELOC mutated RCC, multilocular cystic renal neoplasm of low malignant potential, oncocytoma, and epithelioid angiomyolipoma [143, 144]. The three most common translocations include:

  • t(X;1)(p11.2;q21), between PRCC and TFE3 genes [145];

  • t(X;17)(p11.2;q25), between ASPSCR1 and TFE3 genes (the same one found in alveolar soft part sarcoma) [146]; and

  • t(X;1)(p11.2;p34), between SFPQ and TFE3 genes, also common in Xp11 translocation perivascular epithelioid cell tumors (PEComas)/melanotic Xp11 translocation renal RCCs.

Widespread utilization of RNA sequencing has accelerated identification of many TFE3 gene fusion partners, including NONO (previously known as P54NRB), RBM10 [147, 148], DVL2 [149], PARP14 [150], GRIPAP1 [151], MED15 [151], KATA6A [152], NEAT1 [152], EWSR1 [153], and CLTC [154]. Because the TFE3 fusion protein is overexpressed relative to the native TFE3 protein, diffuse and strong nuclear immunoreactivity in a clean background, using an antibody to the C-terminal portion of TFE3, is a sensitive and specific marker for Xp11 translocation RCC; however, experience with this antibody in clinical practice has been mixed [155]. In contrast, TFE3 break-apart FISH assays have generally proven to be more useful than automated TFE3 IHC, because they are more reliable in variably formalin-fixed tissue [144, 156, 157]. The key exceptions are the Xp11 translocation neoplasms resulting from paracentric inversions involving Xp11 which are frequently unable to be confirmed by conventional TFE3 break-apart FISH. These inversions include RBM10-TFE3, GRIPAP1-TFE3, RBMX-TFE3 [158], and NONO-TFE3. Although the diagnosis of Xp11 translocation RCC can be suspected when a neoplasm demonstrates typical morphologic features, along with strong diffuse TFE3 and/or cathepsin K immunoreactivity [159, 160], the morphologic spectrum of Xp11 translocation RCCs with various fusion partners can be quite heterogeneous [141, 143, 144, 149, 161, 162].

The overall survival of patients with Xp11 translocation RCC is similar to that in patients with clear cell RCC, but it is significantly worse than in patients with PRCC [163]. In multivariate analysis, only distant metastasis and older age at diagnosis independently predicted death [164], but locally advanced stage in pediatric patients with positive regional lymph nodes did not predict adverse outcomes [165]. Xp11 translocation RCCs in pediatric patients have been recently shown to have a lower burden of genomic alterations than in adults, which may help explain their better prognosis [166].

Xp11 translocation PEComas [167] and melanotic Xp11 translocation carcinomas (the latter containing melanin pigment) [168] have a morphologic and immunohistochemical phenotype that overlaps more with conventional PEComa than RCC. PEComa is a term now used to describe a group of entities arising from perivascular epithelioid cells, including a subset of tumors previously knowns as “epithelioid angiomyolipomas”. These tumors demonstrate a mostly nested pattern and lack overt epithelial structures (tubules or papillae). By IHC, Xp11 translocation PEComas/melanotic Xp11 translocation carcinomas are reactive for cathepsin K and HMB45, but not for cytokeratins or renal tubular markers like PAX8. Although their phenotype most closely fits PEComa, they exhibit several features that distinguish them from typical PEComas [169]. Xp11 translocation PEComas often affect younger patients but are not associated with tuberous sclerosis complex (TSC), and there is no evidence of a hereditary basis. They almost always demonstrate an exclusively epithelioid clear cell morphology, but minimal to no immunoreactivity for muscle markers, and also do not exhibit inactivation of TSC1 or TSC2 genes. Most of these neoplasms harbor the SFPQ-TFE3 gene fusion [168]. Approximately half of these tumors have metastasized [167,168,169].

TFEB-rearranged renal cell carcinoma (typically t(6;11) translocation)

The t(6;11) RCCs typically show a distinctive biphasic morphology, made up of larger epithelioid cells, and smaller cells often clustered around nodules of basement membrane material. The larger cells have clear to eosinophilic cytoplasm, and their nested architecture is similar to clear cell RCC. The smaller cells characteristically cluster around basement membrane material resembling the Call-Exner bodies of adult granulosa cell tumor [170]. Although these tumors grossly appear well delineated, microscopically they entrap renal tubules at their periphery. t(6;11) RCC can overlap morphologically with the Xp11 translocation RCC, and vice versa [171]. When the smaller cell population is limited or absent, the differential diagnosis of the t(6;11) RCC is broader and also includes clear cell RCC and various oncocytic renal neoplasms.

The t(6;11) translocation fuses the TFEB gene with MALAT1 (formerly Alpha), an untranslated gene of unknown function, resulting in overexpression of native TFEB. As a result, nuclear immunoreactivity for TFEB has been proposed as specific for the t(6;11) RCC [170]. However TFEB break-apart FISH is the preferred test for establishing the diagnosis, because it is less affected by fixation [172]. Alterative TFEB fusion partners have also been reported, including COL21A1, CADM2, EWSR1, CLTC, PPP1R10, and KHDRBS2 [173, 174]. The t(6;11) RCC are more indolent than the Xp11 translocation RCC, with fewer than 10% of cases resulting in patient death. Necrosis has been found to correlate with aggressive behavior [174, 175].

TFEB-amplified renal cell carcinoma

More recently, RCCs with amplification of the TFEB gene have been described that also results in overexpression of TFEB [175,176,177,178,179]. Several features distinguish the TFEB- amplified RCC from the t(6;11) RCC. TFEB-amplified RCCs occur in older patients (median age 65 years) compared to unamplified t(6;11) RCC (median age 31 years), and frequently present with advanced local stage or metastatic disease. The morphology of the TFEB-amplified RCC is less distinctive, and frequently demonstrates high-grade poorly differentiated RCC morphology and frequent oncocytic and papillary features (Fig. 7). Aberrant melanocytic marker expression and cathepsin K labeling are also less consistently seen in TFEB-amplified RCC, typically in about half of cases. Approximately 50 TFEB-amplified RCCs have been documented in the literature, but precise diagnostic criteria remain to be defined. Although the initial study required high-level amplification and downstream effects of TFEB amplification (aberrant melanocytic marker expression) to make this diagnosis [176], other studies have included low-level TFEB-amplified cases [179]; the significance of these cases is currently unclear. Recent work showed that TFEB gene expression is increased in these tumors, although not as much as in t(6;11) RCC, raising the possibility that other genes at the 6p21, such as VEGFA or CCND3, or others, may be responsible for the aggressive behavior [180].

Fig. 7: TFEB-amplified RCC.
figure 7

This tumor type may show a wide range of histologic appearances, including solid nested architecture with clear and eosinophilic cells, mimicking high-grade clear cell RCC (A); prominent nucleoli may raise the possibility of fumarate hydratase deficient RCC (B); papillary architecture may suggest papillary RCC (C); and nonspecific high-grade rhabdoid features may suggest RCC with rhabdoid differentiation (D). Immunohistochemical labeling for melan A (E) or cathepsin K (F) can suggest the correct diagnosis.

Full size image

MITF rearranged renal cell carcinoma

Two RCCs with MITF rearrangements have been reported to date, one with an ACTG1-MITF fusion [84] and another with PRCC-MITF [181]. The illustrated morphology was similar to that seen in the TFE3 and TFEB-rearranged RCC, including rosette-like architecture, psammoma bodies, and cathepsin K immunoreactivity.

Acquired cystic disease-associated renal cell carcinoma

Acquired cystic disease-associated renal cell carcinomas (ACD-RCC) is a unique subtype of RCC, found exclusively in the setting of end-stage renal disease. It is characterized most often by cribriform (“sieve-like”) architecture, eosinophilic cells, and intratumoral calcium oxalate crystals [104]. Recent studies have found that the duration of hemodialysis and male gender are associated with increased occurrence of ACD-RCC [182,183,184]. Most ACD-RCCs are indolent, but some are associated with adverse prognosis and death [182, 183]. Sarcomatoid and rhabdoid features are considered adverse prognostic factors [184]. Intratumoral coagulative necrosis has also been reported as a possible adverse parameter, but further confirmatory studies are needed [183]. Some tumors showed an unusual pattern of invasion with cysts extending into the renal sinus [185]. Metastatic ACD-RCC may show cyst formation (Fig. 8A) and calcium oxalate crystal deposition, a unique feature amongst RCCs [183, 185].

Fig. 8: Acquired cystic disease RCC (ACD-RCC).
figure 8

ACD-RCC is often nonaggressive, but metastases can rarely be documented. This metastatic ACD-RCC in a lymph node (A) includes multiple cysts with hemorrhage. Cysts resembling ACD-RCC are thought to be the precursor lesion of this tumor (B). This cyst is lined by multilayered epithelium with sieve-like growth and may show short papillary projections.

Full size image

In the setting of acquired cystic kidney disease, multiple cysts can also be found lined by either single-layered eosinophilic or clear cell epithelium [104]. Both in the acquired cystic kidney disease and in the normal kidneys, some renal cysts may also show proliferative changes with multilayered epithelium or short papillations that have been designated “atypical renal cysts” [186, 187]. These atypical renal cysts are thought to be precursors of ACD-RCC and harbor cytogenetic alterations [188]. When such cysts are multilocular or clustered, it may be difficult to distinguish between an atypical renal cyst and early ACD-RCC, and therefore a solid nodular growth within the cyst is required to diagnose ACD-RCC (Fig. 8B) [183, 189].

There are limited data regarding the genetics of ACD-RCC, derived mainly from FISH and comparative genomic hybridization analysis in case reports and small series [190, 191]. Most recently, recurrent mutations in KMT2C (4 of 5 cases) and TSC2 (3 of 5 cases) have been detected by NGS analysis [184]. Notably, ACD-RCC cases with KMT2C mutation exhibited the classic “sieve-like” morphology, whereas one case without KMT2C mutation showed PRCC morphology and lacked cribriform architecture. Pathogenic mutations in the TSC2 often coexisted with the KMT2C mutations [184].

SMARCB1-deficient renal cell carcinoma: Renal medullary carcinoma, “unclassified renal cell carcinoma with medullary phenotype” and dedifferentiated renal cell carcinoma with secondary SMARCB1 loss

Renal medullary carcinoma (RMC) is a high-grade adenocarcinoma, arising in the distal nephron. RMC most frequently affects younger black males (median age, third decade) [192,193,194], harboring sickle cell trait, disease, or related hemoglobinopathy [194]. The presence of a hemoglobinopathy is now considered an essential diagnostic criterion [195, 196]. These aggressive tumors show exceptionally poor survival, with only a small minority of patients surviving 2 or 3 years [194, 197]. Morphologically, RMC is a high-grade, highly infiltrative adenocarcinoma, with tubular, cribriform, and reticular growth patterns [133, 192], and frequent rhabdoid features [198]. RMC also shows prominent desmoplastic myxoinflammatory stromal response and frequent necrosis [192, 193]. Sickled erythrocytes can be seen in associated vascular spaces.

In the context of sickle cell trait/disease and appropriate morphology, the absence of nuclear expression for SMARCB1 (INI-1) protein, in the presence of internal positive control in the non-neoplastic cells (Fig. 9A, B), supports the diagnosis of RMC [133, 199]. However, it is important to keep in mind that a secondary loss of SMARCB1 expression has been found in other RCC types [200, 201]. Strong nuclear OCT3/4 expression has also been found in most RMCs [202]. Recent molecular profiling studies have demonstrated that genetic loss of both SMARCB1 alleles on chromosome 22 accounts for loss of expression in >85% of cases; most frequent are the inactivating mutations with hemizygous (>50%), or biallelic (>15%) deletions [203,204,205]. Exceedingly rare “unclassified RCCs with medullary phenotype” (RCCU-MP) with SMARCB1 loss have also been recently described (Fig. 9C, D) [195]. RCCU-MP is essentially indistinguishable from RMC, by histology and IHC, except for the lack of hemoglobinopathy in the affected patients, and often older age at presentation [206]. Although the number of documented RCCU-MPs is limited [205, 207,208,209], additional molecular studies are warranted due to their aggressive behavior [206]. SMARCB1 loss has also been rarely reported in tumors that otherwise do not show medullary morphology [210]. These carcinomas are best designated as “unclassified RCCs with SMARCB1 loss”. We recommend a thorough sampling in such cases to potentially identify a recognizable RCC component.

Fig. 9: SMARCB1-deficient renal cell carcinoma and collecting duct carcinoma.
figure 9

Renal medullary carcinoma is found in patients with sickle cell trait or disease (A). Negative immunohistochemistry for SMARCB1 (INI1) supports the diagnosis (B). Rare unclassified RCCs with medullary phenotype (C) occur in the absence of sickle disease/trait, and are characterized by abnormal negative SMARCB1 (D), with positive internal controls of normal cells. Collecting duct carcinoma (E) should only be diagnosed when other considerations, particularly FH-deficient RCC, medullary carcinoma (normal SMARCB1 shown in F), and metastases from other primaries are excluded.

Full size image

Although RMC shows poor response to conventional therapy [204, 211], mechanistic studies of loss of SMARCB1 function have identified novel therapeutic options for targeting DNA replication stress [204], proteasome inhibition [212], and EZH2 inhibition [213].

Collecting duct carcinoma

Collecting duct carcinoma (CDC) is an aggressive carcinoma, usually found in older patients, and typically presenting at an advanced stage [192]. Morphologically, CDC is a high grade, predominantly tubular, medulla-centered adenocarcinoma, with a distinctive infiltrative growth and stromal desmoplasia, often histologically similar to RMC or RCCU-MP, but not associated with either hemoglobinopathy [192] or SMARCB1 abnormality [133]. The true incidence of CDC is difficult to ascertain, as many previously diagnosed CDCs can now be reclassified as other tumor types. Therefore, the diagnosis of CDC requires careful exclusion of other RCC types and urothelial carcinoma, particularly when limited tissue is available, as in biopsy specimens. In one large multi-institutional cohort, 25% of cases diagnosed as CDC were reclassified as FH-deficient RCC, using FH and 2SC immunostaining [133]; a similar rate of CDC reclassification was found in recent sequencing studies demonstrating FH and SMARCB1 mutations [214]. Other renal tumors that should be considered in the differential diagnosis include metastatic carcinoma (most commonly from a lung primary) [215, 216], urothelial carcinoma, including BK polyomavirus-associated upper tract urothelial carcinoma with prominent glandular differentiation [217, 218], and mucinous tubular and spindle cell carcinoma with high-grade transformation [219].

Recent molecular studies support the notion that when strictly diagnosed and based on gene expression, CDC represents a true entity, clustering separately from the other RCCs and upper tract urothelial carcinomas [220, 221]. Other molecular findings include frequent CDKN2A loss [222], as well as NF2 and SETD2 mutations [214]; the latter however seem to be more prevalent in unclassified RCCs [223]. Ultimately, true CDC is exquisitely rare, and the diagnosis should be made only after excluding other tumors, especially the recently described FH-deficient RCC and SMARCB1-deficient RCC (Fig. 9E, F), or urothelial carcinoma and metastases from other primaries.

Mucinous tubular and spindle cell carcinoma

Mucinous tubular and spindle cell carcinoma (MTSCC) is a rare entity with distinctive morphologic and cytogenetic features (Fig. 10A). It occurs more frequently in women (M:F= 1:3-4) and is usually indolent. Cytogenetically, recent studies have demonstrated a relatively specific pattern of multiple chromosomal losses that frequently involve chromosomes 1, 4, 6, 8, 9, 13, 14, 15, and 22 [224,225,226]. Additionally, biallelic alteration and dysregulation of the Hippo pathway have been suggested as a common molecular basis for the disease [225].

Fig. 10: Mucinous tubular and spindle cell carcinoma (MTSCC).
figure 10

The classic features include intermixed tubules and spindle cells in a mucinous stroma (A). MTSCC exhibiting WHO/ISUP grade 3 nuclei and adverse histologic features, such as increased mitoses, solid growth, and single cell infiltration, is associated with more aggressive behavior (B). A low-grade element is focally present (arrow).

Full size image

Although most MTSCC cases can be readily identified by morphology, it may be challenging to distinguish them from papillary RCC, particularly the solid variant, as the morphological features and immunohistochemical profiles may overlap [227, 228]. Moreover, a spindle cell predominant MTSCC may be misdiagnosed as a sarcomatoid RCC or a smooth muscle tumor, such as a myoid-predominant angiomyolipoma [229]. The diagnostic challenge is more significant in biopsies (particularly for metastatic disease) when different management options would be considered. Supported by the distinctive chromosomal alteration patterns seen in MTSCC vs. PRCC, a distinct, well-formed papillary area and spindled cell-lined, angulated/curvilinear tubules with irregular lumina are histologic clues for PRCC diagnosis, despite features mimicking MTSCC [226, 230]. Copy number analyses (e.g. SNP array) can help establish a definitive diagnosis in challenging cases or in biopsies. Novel biomarkers for MTSCC, such as VSTM2A overexpression by RNA in situ hybridization are also emerging, but require further validation [231].

A small subset of MTSCC has been reported to exhibit aggressive clinical behavior and/or high-grade transformation (Fig. 10B), including sarcomatoid changes and high-grade epithelial elements [229, 230, 232,233,234,235,236,237,238]. Additionally, rare cases of MTSCCs with typical low-grade morphology have also been reported to develop metastases [239,240,241]. In the few institutional series, metastasis has been reported in 10–27% of patients [235, 237, 238]. The molecular characterization of cases with high-grade features or aggressive behavior has been limited, and some studies identified either chromosomal alterations similar to typical cases, or variable cytogenetic patterns [229, 236, 237, 242]. A recent study comparing the morphologic and molecular features of locally advanced/metastatic MTSCC to clinically indolent cases, identified the following adverse features: necrosis, solid growth, single file infiltration, sarcomatoid transformation, lymphovascular invasion, and increased mitoses [219]. This study also provided new evidence that MTSCC with aggressive clinical behavior is part of the MTSCC category that progressed through clonal evolution, and these cases frequently demonstrated CDKN2A/B deletion, as well as additional complex genomic abnormalities [219].

Metanephric adenoma and related tumors

The metanephric family of tumors includes metanephric adenoma (MA), metanephric adenofibroma (MAF), and metanephric stromal tumor (MST) (Fig. 11). These benign, unencapsulated tumors are separated morphologically by the relative amounts of epithelial to stromal components [243]. MA is composed entirely of small, monotonous, round cells forming tubules and short papillae. MST is a stromal neoplasm that frequently demonstrates concentric growth around native tubules and vessels, the latter often showing angiodysplasia. MAF is a biphasic proliferation of epithelial cells similar to metanephric adenoma, and stromal cells similar to MST [243].

Fig. 11: Metanephric adenoma and related tumors.
figure 11

Metanephric adenoma is sharply circumscribed with tightly packed tubules composed of small round blue cells (A). Occasionally, mitotic figures may be present (B). Metanephric adenofibroma (C) is a biphasic proliferation that contains an epithelial component, similar to metanephric adenoma, and a stromal component, with spindle to stellate cells in a loose myxoid background. Metanephric stromal tumors are composed entirely of stromal elements exhibiting concentric growth around entrapped tubules and vessels (D).

Full size image

The frequent presence of BRAF mutations unifies the metanephric family of tumors. Two MAF case reports identified BRAF V600E mutations in both the epithelial and the stromal components [244, 245]. BRAF V600E mutations were also identified in 6 of 7 MST cases in the seminal report by Argani et al. [246]. In another study, 11 of 17 (65%) MST cases were positive for BRAF V600E mutations using Sanger sequencing, without showing significant morphologic differences between the positive and the negative cases [247]. These findings raise the possibility that BRAF V600E senescence pathway contributes to the benign nature of conventional MA. BRAF positivity is frequently seen in metanephric adenomas and is rarely documented in papillary RCC [248,249,250].

BRAF V600E mutations are present in over 90% of MA and were initially considered a distinguishing feature of MA from Wilms tumor (WT) [249, 251, 252]. A recent study highlighted a subset of genetically confirmed MA with increased mitotic activity (up to 12 mitotic figures per 10 high power fields), demonstrating a morphologic overlap with epithelial WT. The same study also included 9 epithelial-predominant WT with foci that mimicked MA, including tubular pattern of well-differentiated epithelial cells within a myxoid stroma, with low mitotic rate [253]. Interestingly, 4 of 9 (44%) of these epithelial-predominant WT harbored BRAF V600E mutations, indicating that some epithelial-predominant WT genetically overlap with MA [253]. Recent case reports have also demonstrated the clinical utility of identifying BRAF mutations in WT, where a dual BRAF/MEK inhibitor was used to treat a patient with metastatic WT resulting in complete response [254, 255].

In addition to BRAF, other mutations identified by NGS DNA sequencing have recently been described in a cohort of MA (n = 11) [256]. One conventional MA showed BRCA2 pathogenic mutation. One MA with sarcomatoid features showed EIF1AX and TERT promoter hotspot mutations in both epithelial and sarcomatous components, whereas a deep deletion of CDKN2A and amplification of MYC were identified only in the sarcomatous component [256]. Another group identified mutations in NF1, NOTCH1, PTEN, SPEN, AKT2, APC, ATRX, and ETV4 in a series of 28 MA [257].

Some tumors with MA-like morphology have been recently reported to harbor ALK rearrangements, raising the question of whether MA can also show ALK rearrangement or whether these represent ALK-RCC mimicking MA [258].

Hereditary renal carcinoma syndromes

It has been estimated that 5–8% of all RCCs are hereditary, and many more arise in the setting of some form of genetic predisposition [259,260,261,262]. Clinical indications for referral for genetic assessment regardless of histology include the onset of RCC at a young age, variably defined as <50, <46, or <40 years [259, 263, 264], or an association with syndromic features suggestive of TSC (for example renal angiomyolipomas, skin lesions, or history of seizures), PTEN hamartoma syndrome (breast, thyroid or other tumors), BHD syndrome (skin fibrofolliculomas and trichodiscomas, spontaneous pneumothorax), hereditary leiomyomatosis and renal cell carcinoma (HLRCC) syndrome (uterine leiomyomas under the age of 40, cutaneous leiomyomas), SDH mutation (SDH deficient gastrointestinal stromal tumor (GIST), paraganglioma/phaeochromocytoma) or VHL syndrome (phaeochromocytoma, hemangioblastoma, pancreatic neuroendocrine tumor, or pancreatic serous cystadenoma) [259, 263, 264]. In addition, fumarate hydratase deficient RCC and SDH deficient RCC are so closely associated with germline mutations, and BHD syndrome-related renal neoplasia is usually so characteristic, that genetic assessment is indicated regardless of the age of onset and even in the absence of other features.

Fumarate hydratase deficient renal cell carcinoma

Autosomal dominant germline mutations in FH (fumarate hydratase, chromosome 1q43) cause HLRCC syndrome (also known as Reed syndrome), characterized by a distinctive and often aggressive type of renal carcinoma, cutaneous and uterine leiomyomas and, rarely, pheochromocytoma/paraganglioma and Leydig cell tumors of the testis [265,266,267]. Absence of staining for FH in the presence of a positive control in non-neoplastic cells is highly specific, but less sensitive for FH deficiency, that is biallelic inactivation of FH; whereas positive staining for 2SC is highly sensitive but less specific [129,130,131, 268]. Whilst FH-deficient uterine leiomyomas are commonly sporadic, particularly if occurring in older adults [266, 269], the overwhelming majority of FH-deficient RCCs appear to be associated with germline mutations or deletions [129, 131, 268]. As the germline mutation status is often unknown at the time of diagnosis, and not all cases are syndromic, the term “FH-deficient RCC” is preferred over “HLRCC-associated RCC” [129, 131], to encompass tumors with compatible histology and IHC evidence of FH-deficiency (FH loss and/or 2SC reactivity), but with uncertain clinical and family history.

On microscopy, FH-deficient RCCs typically demonstrate a variety of architectural patterns, usually a combination of papillary and tubulocystic, cystic, cribriform/sieve-like, and solid pattern (Fig. 12A–D). Recognition of these multiple patterns is an important clue to the diagnosis [20, 129,130,131,132,133,134, 268, 270, 271]. Although prominent inclusion-like nucleoli are a consistent feature in most FH-deficient RCCs, they are not specific [129, 131, 132, 268, 271]. Variant morphologies, including low-grade tumors and extensive cystic change, have also been reported, and therefore a low threshold for IHC is recommended in any difficult to classify renal carcinoma [129, 131, 268, 272,273,274].

Fig. 12: Hereditary renal carcinoma syndromes.
figure 12

Fumarate hydratase (FH) deficient renal cell carcinoma can demonstrate both a papillary (A) and cribriform/sieve-like (B) architecture. C There is abnormal negative fumarate hydratase expression (loss of expression), with positive granular cytoplasmic (mitochondrial) expression in the internal positive controls. D 2SC shows strong nuclear and cytoplasmic staining. E Succinate dehydrogenase deficient RCC is characterized by sheet-like growth pattern with a vaguely nested architecture. F In this example, the distinctive cytoplasmic inclusions are readily identified. G Interpretation of SDHB immunohistochemistry can be difficult. In this example, SDHB demonstrates strong granular cytoplasmic staining in the internal positive controls in the arterioles, which contrast to the weak cytoplasmic blush (that lacks distinct granularity) in the neoplastic cells. This is interpreted as a negative (abnormal staining). H In this case, SDHA immunohistochemistry was diffusely and strongly positive, essentially excluding SDHA mutation.

Full size image

Succinate dehydrogenase deficient renal cell carcinoma

Germline mutation of any of the four succinate dehydrogenase (SDH) genes (SDHA, SDHB, SDHC, and SDHD) causes an autosomal dominant tumor syndrome characterized by phaeochromocytoma/paraganglioma, a unique type of gastric gastrointestinal stromal tumor (GIST) termed “SDH deficient GIST”, pituitary adenoma, and RCC [20, 275,276,277,278,279].

The majority of RCCs arising in this syndrome demonstrate distinctive morphology, characterized by a sheet-like or compact nested growth of bland cuboidal cells, typically showing

eosinophilic, sometimes vacuolated cytoplasm, that lacks the fine granularity seen in most oncocytic tumors, with frequent microcysts and entrapped non-neoplastic tubules [280,281,282]. Cytoplasmic inclusions containing eosinophilic or pale flocculent material are a clue to the diagnosis but may be inconspicuous or absent (Fig. 12E–G). IHC for SDHB is negative whenever there is biallelic inactivation of any of the four SDH genes (SDHA, SDHB, SHDC or SDHD), which is almost always associated with germline mutation with a somatic second hit [20, 275,276,277,278,279]. The majority occur in the setting of SDHB mutation, with SDHC and SDHA mutation being uncommon but well reported, and SDHD mutation being very rare [281, 283]. Care is required in interpreting SDHB IHC. There is an absolute requirement for an internal positive control in the non-neoplastic cells before the neoplastic cells are considered negative. Furthermore, a weak diffuse cytoplasmic blush is still considered negative, if it contrasts with the darker and distinctly granular (mitochondrial) staining found in internal controls (Fig. 12G). Most cases are indolent, but high-grade transformation (found in up to one-third of cases), coagulative necrosis, and sarcomatoid transformation are associated with a high risk of metastasis—up to 70% [281]. High-grade cases may be unrecognizable by morphology, justifying a low threshold for IHC screening in any unusual or difficult to classify renal tumor, particularly with eosinophilic cytoplasm, or in patients <35 years old [272, 281]. Other important IHC features include lack of reactivity for CD117 and frequent absence, or weak and focal pan-cytokeratin and CK7 reactivity [281].

In addition to SDHB loss, IHC for SDHA is also negative when there is biallelic inactivation of SDHA [284]. SDHA-deficient RCCs more commonly show variant morphologies, higher nuclear grade, and a combination of architectures, including papillary, solid, cribriform and desmoplastic [285,286,287,288]. It is important to note that germline mutations in SDHA are found relatively frequently as an incidental finding (estimated to occur in up to 0.3% of the population) [284], with an extremely low lifetime penetrance (as low as 1.7%) [289]. Therefore, when SDHA mutation is identified as part of a broad sequencing panel, it may be an incidental finding unrelated to renal neoplasia [284, 290].

BHD syndrome

Autosomal dominant germline mutations in FLCN, encoding the protein folliculin, cause the BHD syndrome [5, 291]. BHD syndrome is characterized by multiple skin lesions, classified as fibrofolliculomas or trichodiscomas [292], lung cysts which predispose to pneumothorax (predominantly basal and often unrecognized without CT-scan) [293] and renal tumors [294]. BHD-related renal tumors demonstrate a range of morphologies that include a dominant eosinophilic/oncocytic differentiation [295, 296]. These tumors show a spectrum of morphologies resembling oncocytoma and ChRCC, but also “hybrid tumors” with mixed oncocytoma-chromophobe features that may not precisely fit into any type, as well as clear cell carcinoma [295]. Important diagnostic clues include a finding of multiple renal tumors, often bilateral and with slightly different histologies, as well as frequent microscopic foci of oncocytic cell clusters (“oncocytosis”) in the non-neoplastic kidney [5, 295]. Recent molecular analysis demonstrated broad uniparental disomy (also known as copy-neutral loss of heterozygosity) and some copy number variations in BHD-RCCs with variable histologies [297]. Interestingly, bilateral, multifocal renal oncocytomas in non-BHD patients were found to have disruptive mitochondrial DNA mutations, while renal tumors from BHD patients did not show such mutations [298].

Renal neoplasia occurring post-chemotherapy/radiation in pediatric patients

RCC associated with neuroblastoma, listed as a separate subtype in the 2004 WHO classification, was not included in the 2016 classification due to limited data [299]. Other renal tumors have also been described in association with childhood neoplasia, including both solid tumors and hematologic neoplasia [300]. The common risk factor for the majority of these patients included cytotoxic chemotherapy/radiation in the early childhood [301, 302]. Renal neoplasia occurring in this context has been documented between the first to the sixth decade of life, suggesting that these patients should receive extended clinical follow-up into adulthood. Some renal neoplasia occurring post-chemotherapy/radiation have been documented secondary to inherited germline alterations of various genes [300, 303, 304]. Some studies have also documented RCCs with “oncocytoid” morphology occurring post-neuroblastoma; some of these tumors may be currently reclassified as eosinophilic solid and cystic RCC, a recently described entity [305, 306]. However, it is unclear if these post-neuroblastoma tumors occurred secondary to germline or somatic TSC1/TSC2 gene mutations [299].

A systematic assessment of whether these renal tumors are associated with specific histologic/molecular features has been challenging, because the majority of cases have been documented either as case reports or small series, and often lacked a more comprehensive evaluation using contemporary diagnostic techniques. More recent studies have also failed to identify a set of characteristic features to indicate a unique subtype of renal neoplasia occurring in this setting [299, 300]. A current review of the literature revealed that in cases with specified histologic subtype, clear cell RCC was the most common renal tumor, occurring after Wilms tumor (12 of 29, 41%) and neuroblastoma (17 of 46, 37%) [300]. It was previously suggested that MITF family translocation RCC are frequently identified in this setting; however, this is likely due to their higher incidence in pediatric patients [300, 301].

Although the current literature suggests that there is an increased risk of renal neoplasia in pediatric patients after chemotherapy/radiation, and particularly after neuroblastoma, there is insufficient data to support the existence of a unique renal neoplastic subtype in this setting. These cases represent a heterogeneous group of tumors, associated either with hereditary tumor syndromes, or representing MITF family translocation RCCs, or newly described entities.

Rhabdoid and sarcomatoid changes in renal cell carcinoma: molecular correlates and therapeutic strategies

Rhabdoid and sarcomatoid changes in RCC are associated with poor prognosis and are graded as WHO/ISUP grade 4 [36]. Rhabdoid change can arise in all types of RCC [307], and is typically associated with high-stage [39] and increased risk of death independent of other factors [44, 308, 309]. Similarly, sarcomatoid RCC change is a pattern of dedifferentiation associated with adverse outcome and poor cancer-specific survival, irrespective of the underlying RCC subtype [41, 310]. Accordingly, rhabdoid and sarcomatoid RCC changes do not represent distinct RCC types, but rather a clonal and morphologic progression to a high-grade biologic state.

A clonal relationship has been documented between lower grade clear cell RCC and rhabdoid areas, including loss at the locus for SMARCB1 (INI1), while loss of chromosome 11p was found to be relatively specific [311]. Others have confirmed that in addition to loss of SMARCB1, variable losses of at least one other SWI/SNF complex subunit (SMARCA2, SMARCA4, ARID1A, SMARCC1, and SMARCC2), documented by IHC, were present in two-thirds of rhabdoid and high-grade RCCs [200]. Although these findings were validated in both rhabdoid and sarcomatoid RCCs, significant study limitations included the lack of molecular confirmation, and the aggregation of cases with either reduced, or absent protein expression [200, 312]. In addition, NGS studies of macrodissected rhabdoid and sarcomatoid areas and corresponding lower grade RCC areas showed uniform, clonal VHL alterations; however, rhabdoid and sarcomatoid areas exhibited a distinct transcriptomic profile and a higher rate of BAP1/ SETD2 alterations on chromosome 3p, compared to PBRM1 [313, 314]. The common molecular alterations in sarcomatoid RCC included TP53, CDKN2A, PTEN, RELN, as well as the Hippo pathway alterations (including NF2) [314,315,316].

The most recent National Comprehensive Cancer Network (NCCN) guidelines increasingly adopt the use of immunotherapeutic agents in advanced RCC, requiring biomarkers for better patient selection to identify those most likely to respond [317,318,319]. Emerging data suggest that patients with advanced RCC, including sarcomatoid RCC, may show favorable response to these agents. Potential biomarkers of interest include: tumor cell PD-L1 expression, gene expression signatures, tumor mutation burden, PBRM1 status, and CD8+ T-cell density [318, 319]. Current evidence suggests that RCC with sarcomatoid differentiation may show significantly higher PD-L1 expression, with a subset of cases exhibiting constitutive expression of PD-L1 secondary to 9p24.1 amplifications [320,321,322,323]. In a most recent phase III clinical trial, patients with sarcomatoid RCC have shown significantly improved responses to immune modulator agents, compared to sunitinib alone [324]. Overall, conflicting data exist regarding tumor PD-L1 status as a potential biomarker, because current clinical trials have not evaluated patient responses considering the extent of tumor cell PD-L1 expression; these trials also demonstrated significant variability regarding the selection of therapeutic agents, PD-L1 antibody clones, as well as the methodology and cutoffs to assess PD-L1 status [318]. Future studies and powered clinical trials are therefore needed to correlate specific patterns of tumor PD-L1 expression (adaptive/constitutive) with response to therapy, as well as to identify biomarkers to assess response to both VEGF inhibitors and immune modulators in patients with advanced RCC, including those with sarcomatoid and/or rhabdoid changes.

Conclusions

In this collective work by the GUPS, we provide an update on the evolving histologic and molecular classification of existing renal cell entities. We particularly focused on the novel information, aiming to better define the diagnostic criteria of certain entities, and to integrate the evolving molecular developments pertaining to currently recognized entities. This update also provides a direction for future studies where knowledge gaps still exist. We hope that this update will provide more clarity and guidance on this complex topic and would facilitate future research efforts in this area.