嗜酸性实性囊性肾细胞癌与 TFEB 扩增型肾细胞癌的临床病理及分子比较：15 例综合研究Clinicopathological and molecular comparison of eosinophilic solid and cystic renal cell carcinoma and TFEB-amplified renal cell carcinoma: a comprehensive study of 15 cases

Scope: PMC open-access full text (CC BY 4.0) — abstract, introduction, methods, results, discussion, author contributions and consent statement, with Figures 1–5 and Tables 1–5. Supporting information, conflict-of-interest and data availability statements, and references are not included (citation numbers are retained).

Abstract

Eosinophilic solid and cystic renal cell carcinoma (ESC‐RCC) and TFEB‐amplified renal cell carcinoma (TFEB‐amplified RCC, a subset of TFEB‐altered RCC) are rare eosinophilic renal cell tumours that can be difficult to distinguish on the basis of morphology and immunohistochemistry alone. In this study, we compared the clinicopathological and molecular features of these entities to identify distinguishing characteristics.

Nine cases of ESC‐RCC and six cases of TFEB‐amplified RCC from our institutional archives were evaluated for clinical, morphological, immunohistochemical and molecular features. Despite morphological overlap, each entity exhibited somewhat distinctive features. ESC‐RCC frequently exhibited solid and cystic architecture with tightly packed tubules, flocculent to coarsely granular cytoplasm with basophilic stippling and prominent admixed macrophages. In comparison, TFEB‐amplified RCC more frequently demonstrated solid growth of pseudopapillae and loose tubules, dense eosinophilic granular cytoplasm without basophilic stippling, focally apical nuclei and extracellular magenta globules. However, neither morphology nor immunohistochemistry was sufficiently specific to completely distinguish these entities. Ancillary molecular testing identified pathogenic TSC1 or TSC2 gene variants in ESC‐RCC and TFEB gene locus amplification in TFEB‐amplified RCC.

Although ESC‐RCC and TFEB‐RCC share overlapping morphological and immunohistochemical features, each demonstrates somewhat characteristic histological findings. Definitive distinction requires molecular testing to identify the underlying molecular alterations. As molecular testing assumes an increasingly important role in RCC diagnosis and classification, recognition of these rare entities and their overlapping and distinguishing features is essential for appropriate testing and accurate diagnosis.

Introduction

Over the past half century, the classification system of renal cell carcinoma (RCC) has changed multiple times to reflect our evolving understanding of renal cell tumour biology. ¹ , ² The most recent World Health Organization (WHO) classification defines six groups of renal cell tumours: clear cell, papillary, oncocytic and chromophobe, collecting duct, molecularly defined and other RCC. Clear cell, papillary and chromophobe RCC account for approximately 95% of renal cell tumour diagnoses. ³ , ⁴ These more common RCC subtypes have characteristic morphology and immunohistochemical (IHC) staining patterns, and are most often diagnosed without further ancillary testing.

In contrast, rarer RCC subtypes, especially eosinophilic or oncocytic renal tumours, may exhibit significant morphological and immunophenotypic overlap and frequently require advanced testing modalities for definitive diagnosis. ⁵ , ⁶ Though ancillary testing is often unavoidable in these cases, identifying specific morphological and IHC features of rare RCC subtypes may guide more effective work‐up and diagnosis of rare renal cell tumours, especially when diagnostic material is limited. Therefore, despite the expanding role of molecular testing in the diagnosis and classification of renal cell tumours, it remains critical to fully characterize and recognize specific features of rare RCC subtypes to improve diagnosis, treatment and outcomes for patients with these tumours. ⁷ , ⁸

Eosinophilic solid and cystic renal cell carcinoma (ESC‐RCC) and TFEB‐amplified renal cell carcinoma (TFEB‐amplified RCC, a subset of TFEB‐altered RCC) are rare eosinophilic RCC subtypes that are both associated with TSC/mTOR pathway dysregulation and can be difficult to distinguish on the basis of morphology and immunohistochemistry alone. ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³ Grossly, both entities present as well‐circumscribed tumours with a yellow cut surface with or without areas of haemorrhage and necrosis; histologically, these tumours are composed of large cells with abundant granular eosinophilic cytoplasm, large nuclei and prominent nucleoli. ¹⁴ Furthermore, though it has been proposed that these entities can be distinguished by differential IHC staining for Melan‐A and CK20, the reported utility of this strategy is variable. ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ As a result, both ESC‐RCC and TFEB‐amplified RCC may remain in the differential diagnosis for an eosinophilic tumour even after morphological and IHC work‐up. The distinction between these entities is critical, as ESC‐RCC is reported to have a better prognosis than TFEB‐amplified RCC and may have implications on clinical course and treatment strategy. ¹⁹ , ²⁰ , ²¹

Existing comparative studies of ESC‐RCC and TFEB‐amplified RCC remain limited due to the rarity of both entities, and due to the scarcity of papers that compare these entities directly using the same criteria. In this study, we compare the clinical, morphological, IHC and molecular profiles of nine ESC‐RCC cases and six TFEB‐amplified RCC cases from a single institution. In addition to expanding published knowledge of these uncommon renal cell tumours, we demonstrate the use of hybrid‐capture next‐generation sequencing (NGS) to identify the characteristic molecular alterations underlying each entity, including detection of TFEB gene locus amplification through copy‐number analysis. Using this integrated clinicopathological and molecular approach, we aimed to define features that facilitate recognition of these tumours while emphasizing the role of molecular testing in establishing a definitive diagnosis when morphological and IHC findings are equivocal.

Methods

Study Population

This study was approved by the Institutional Review Board at The University of Chicago under protocol IRB‐24‐1223. We retrospectively reviewed The University of Chicago Department of Pathology database from 2005 to 2025 and identified unclassified or unusual eosinophilic renal tumours suspicious for re‐classification as ESC‐RCC or TFEB‐amplified RCC, as well as tumours with established diagnoses of ESC‐RCC or TFEB‐amplified RCC. Seven cases included in this study (ESC‐RCC cases I, IIA, IIB, III, IV, V, and TFEB‐amplified RCC case II) have been reported previously. ¹³ , ¹⁴ ESC‐RCC cases were initially given diagnoses of ESC‐RCC (n = 1), papillary RCC (n = 1) and unclassified RCC (n = 7). TFEB‐amplified RCC cases were initially diagnosed as sarcomatoid RCC (n = 1), renal oncocytoma (n = 1) and unclassified RCC (n = 4). Clinical data were obtained through electronic medical record review.

Pathological Analysis and Immunohistochemical Staining

For each case, the original pathological diagnosis, haematoxylin and eosin (H&E)‐stained slides and IHC stains were reviewed by a genitourinary pathologist at The University of Chicago. IHC stains were performed in the clinical laboratory at The University of Chicago using autoimmunostainers (Leica BOND‐III; Leica Biosystems, Buffalo Grove, IL and BenchMark XT Ventana; Roche, Tucson, AZ) with appropriate controls. The following panel of IHC stains was performed: CK7 (COV‐TL 12/30; Dako, Carpinteria, CA), CK20 (KS20.8; Dako, Carpinteria, CA), Vimentin (V9; Dako, Carpinteria, CA), racemase (P504S, CPK 200 AK, BK, CK; Biocare Medical, Concord, CA), CD10 (56C6; Leica NCL, Buffalo Grove, IL), SDHB (21A11AE7; Abcam, Cambridge, MA), Melan‐A (A103; Dako, Carpinteria, CA) and PD‐L1 (22C3; Agilent, Santa Clara, CA).

Next‐Generation Sequencing

Sufficient material was available for NGS in 5 of 9 ESC‐RCC cases (I, IIA, III, V and VIII) and 6 of 6 TFEB‐amplified RCC cases (I–VI). ESC‐RCC case IIB failed multiple attempts to sequence. ESC‐RCC case IV was not sequenced, as it occurred in a patient with tuberous sclerosis and displayed the typical morphological features and IHC phenotype of ESC‐RCC. From the cases that were sequenced, a representative formalin‐fixed, paraffin‐embedded block was selected for NGS on The University of Chicago Medicine OncoPlus (UCM‐OncoPlus) panel, a hybrid‐capture panel targeting 1005 cancer‐associated genes with 168 clinically reported genes. DNA extraction, DNA quantification, library preparation and NGS were performed as described previously. ²²

Data analysis was performed on a HIPAA (Health Insurance Portability and Accountability Act)‐compliant high‐performance computing system (Center for Research Informatics, University of Chicago) using an in‐house developed bioinformatics pipeline. Copy number results were calculated using a combination of CNVkit 14 software ²³ and additional in‐house intra‐run normalization to eliminate run‐specific artefacts, by comparison with a pooled cohort of clinical controls. Focal gene‐level and chromosome‐level changes were called using a combination of thresholds and manual interpretation, as described previously. ²² , ²³ , ²⁴ The specific thresholds used in this study were log₂ (fold change) ≥ 1.5 for amplification and log₂ (fold change) between 1.0 and 1.5 for equivocal amplification, consistent with thresholds used previously. ²⁵ Somatic variant calls were inspected using Integrated Genomics Viewer (IGV; Broad Institute, MIT Harvard, Cambridge, MA). Variant review was performed by molecular pathologists and included filters based on population frequencies (the 1000 Genomes Project, https://www.internationalgenome.org/), frequencies in cancer variant databases (COSMIC: Catalogue Of Somatic Mutations In Cancer, https://cancer.sanger.ac.uk/cosmic, and cBioPortal, https://www.cbioportal.org/), and coding effects.

Results

Clinical Demographics

Retrospective review and molecular profiling of previously unclassified or unusual eosinophilic renal cell tumours identified nine ESC‐RCC cases (n = 9) and six TFEB‐amplified RCC cases (n = 6) from our institutional archives (Table 1). Two ESC‐RCC cases (cases IIA and IIB) arose in a single patient who underwent right radical nephrectomy for a right‐sided kidney mass (ESC‐RCC case IIA), and 2 months post‐operatively was found to have a mass in the contralateral kidney (ESC‐RCC case IIB). One of the TFEB‐amplified RCC cases is from a patient who had presented with metastatic disease to the brain; approximately 15 years earlier, the patient had undergone resection of a primary renal mass that was diagnosed as ‘sarcomatoid RCC’. The metastatic lesion (TFEB‐amplified RCC case I) displayed a similar histological appearance to the primary renal tumour and was sequenced for diagnostic purposes. The ESC‐RCC cohort of 9 cases arose in 8 patients (6 female and 2 male) ranging in age from 33 to 68 years at diagnosis (mean: 50 years). The 6 cases of TFEB‐amplified RCC arose in 4 female patients and 2 male patients ranging in age from 53 to 76 years at diagnosis (mean: 66 years). One patient in the ESC‐RCC group had a history of tuberous sclerosis (ESC‐RCC case IV). In the TFEB‐amplified group, one patient had a history of Gaucher's disease, type 1 (TFEB‐amplified RCC case IV) and 1 patient was suspected to have multiple endocrine neoplasia, type 1, though confirmatory testing was not performed (TFEB‐amplified RCC case II). No additional patients were known or suspected to have an underlying hereditary syndrome.

Previously published cases.

AJCC Tumour Staging

Upon resection, TFEB‐amplified tumours were larger and more frequently higher in stage than ESC‐RCC tumours (Table 1). ESC‐RCC tumours ranged in size from 1.7 to 11.5 cm in greatest dimension (mean: 4.3 cm). Per the American Joint Committee on Cancer (AJCC), six ESC‐RCC resections revealed pathological stage T1 disease (n = 6) and the remaining three resections revealed pathological stage T3 disease (n = 3). Lymph nodes were recovered in two ESC‐RCC resections corresponding to a single patient (ESC‐RCC cases IIA and IIB); metastatic carcinoma was identified in >1 regional lymph node in each of the resection specimens. Resected TFEB‐amplified tumours ranged in size from 2.9 cm to 11.0 cm in greatest dimension (mean: 7.0 cm). At primary resection, 5 patients with TFEB‐amplified RCC had pathological stage T3 disease (n = 5) and 1 patient had pathological stage T1 disease (n = 1). No lymph nodes were recovered in the resection specimens for TFEB‐amplified RCC.

Morphological Features

Review of H&E slides demonstrated that both entities exhibited shared as well as somewhat distinctive morphological features. Both ESC‐RCC and TFEB‐amplified tumours were at least partially solid and composed of polygonal cells with abundant eosinophilic cytoplasm and focally large, high‐grade nuclei with prominent nucleoli (Table 2).

ESC‐RCC was further characterized by solid and cystic architecture, with solid areas populated by small, tightly packed tubules (Figure 1). The abundant eosinophilic cytoplasm of malignant cells in ESC‐RCC had coarse basophilic stippling and was occasionally vacuolated; alternatively, the cytoplasm was flocculent‐appearing (Figure 2). Additional features of ESC‐RCC included hob‐nailing of cells lining cystic spaces and conspicuous foamy macrophages admixed with or clustered within the tumour.

TFEB‐amplified RCC was predominantly solid with areas of alternating tubular and pseudopapillary growth (Figure 3). The tubules were generally larger than those observed in ESC‐RCC and neither macrocystic nor microcystic architecture was present; TFEB‐amplified tumour cells predominantly demonstrated dense granular eosinophilic cytoplasm. Basophilic stippling was not identified. Additional findings included apically placed nuclei within pseudopapillary areas, extracellular magenta globules and rare calcifications (Figure S1).

Overlapping morphological features were observed in selected tumours. ESC‐RCC case VIII exhibited focal pseudopapillary architecture with apically positioned nuclei resembling features more associated with TFEB‐amplified RCC (Figure 4A). Conversely, TFEB‐amplified RCC case III contained areas with small back‐to‐back tubules and clusters of macrophages reminiscent of ESC‐RCC (Figure 4B).

Immunohistochemistry

Immunohistochemical (IHC) staining was performed using antibodies against vimentin, CK7, CK20, P504S, CD10, SDHB, Melan‐A and PD‐L1 (Table 3). Staining was scored according to percentage of tumour cells demonstrating the appropriate positive staining pattern as follows: 0% (score 0), 1%–5% (score 1), 5%–50% (score 2) and >50% (score 3). Tumours with >5% positive cells were considered positive (corresponding to IHC score ≥2). Both ESC‐RCC and TFEB‐amplified RCC exhibited strong, diffuse staining for vimentin and P504S with retained SDHB. CK7 expression was absent or only focally present in both entities. CD10 staining was focal when present. PD‐L1 expression was variable in both cohorts. CK20 staining (>5% positive tumour cells) was identified in 7 of 9 ESC‐RCC cases (78%) and in 3 of 6 TFEB‐amplified RCC cases (50%) (p = 0.3287). Melan‐A staining (>5% positive tumour cells) was identified in 2 of 9 ESC‐RCC cases (22%) and in 4 of 6 TFEB‐amplified RCC cases (66%) (p = 0.1357) (Figures S2 and S3).

Scoring. 0%–0% staining. 1%–1% to 5%. 2%–5% to 50%. 3%– >50%.

ND, not done.

Next‐Generation Sequencing

Next‐generation sequencing (NGS) was successfully performed on 5 of 9 ESC‐RCC cases and all 6 TFEB‐amplified RCC cases included in this study (Table 4). All sequenced ESC‐RCC cases harboured pathogenic TSC1 or TSC2 variants. All TFEB‐amplified RCC cases demonstrated focal copy number gain involving chromosome 6p and encompassing the TFEB gene locus (Figure 5). Copy number alterations were not identified in any of the ESC‐RCC cases.

Follow‐Up

Four of the patients in this study were lost to follow‐up (n = 4; 2 patients from the ESC‐RCC group corresponding to cases IIA, IIB and III, and 2 patients from the TFEB‐amplified RCC group corresponding to cases I and II) (Table 5). One of the ESC‐RCC patients lost to follow‐up had documented concern for metastatic disease to the liver; of note, this patient (corresponding to ESC‐RCC cases IIA and IIB) had known lymph node metastases at the time of tumour resection. One patient with TFEB‐amplified RCC lost to follow‐up had documented brain and bony metastases (TFEB‐amplified RCC case I) and the other had known recurrent disease in the ipsilateral kidney (TFEB‐amplified RCC case II).

NED, no evidence of disease.

Of the remaining patients not lost to follow‐up, three ESC‐RCC patients were deceased at the time of analysis (n = 3; ESC‐RCC cases IV, V, VII). Two of these patients died due to causes unrelated to RCC (ESC‐RCC cases V and VII) and the other patient's cause of death was not documented in our records (ESC‐RCC case IV). The remaining three ESC‐RCC patients that were alive at the time of analysis had follow‐up periods ranging from 38 to 231 months (n = 3). These patients did not present with concern for recurrence or metastasis during their respective follow‐up periods. The four TFEB‐amplified RCC patients not lost to follow‐up had follow‐up periods ranging from 1 to 54 months (n = 4). One of these patients had confirmed metastases (TFEB‐amplified RCC case III) and 1 patient was under evaluation due to documented concern for metastatic disease to the lung (TFEB‐amplified RCC case IV). Neither of the remaining two TFEB‐amplified RCC patients not lost to follow‐up had evidence of disease at the time of this report.

Discussion

In this study, we compared the clinicopathological and molecular characteristics of nine ESC‐RCC cases and six TFEB‐amplified RCC cases from our institutional archives. Although this study was limited by the rarity of these entities and the resulting small cohort size, the clinical characteristics of our cases are consistent with prior reports and further expand the published experience with both entities. Patients with ESC‐RCC were generally younger (mean age, 50 years), demonstrated a female predominance and most presented with stage pT1 disease, consistent with the expected indolent clinical behaviour reported for this entity. ²⁵ , ²⁶ In contrast, patients with TFEB‐amplified RCC were older at diagnosis (mean age, 66 years), more frequently presented with more advanced stage tumours and were more likely to develop recurrent or metastatic disease during the follow‐up period, supporting previous observations that TFEB‐amplified RCC is an aggressive neoplasm. ²¹ , ²⁷ , ²⁸ , ²⁹

Although ESC‐RCC and TFEB‐amplified RCC demonstrated some morphological overlap, each entity demonstrated somewhat characteristic histological patterns. ESC‐RCC typically exhibited a solid and cystic pattern composed of tightly packed tubules, abundant eosinophilic cytoplasm with coarse basophilic stippling or flocculent appearance, hobnail cells lining cystic spaces and admixed foamy macrophages. In contrast, TFEB‐amplified RCC generally demonstrated predominantly solid growth with tubular and pseudopapillary architecture, dense granular eosinophilic cytoplasm lacking basophilic stippling, apically positioned nuclei within pseudopapillary areas, extracellular magenta globules and occasional calcifications.

Importantly, no single morphological feature was entirely specific for either entity. Selected tumours demonstrated overlapping features, including focal pseudopapillary architecture in one ESC‐RCC and focal tightly packed tubules with clustered macrophages in one TFEB‐amplified RCC. Furthermore, while basophilic stippling was present in all ESC‐RCC cases and absent in all TFEB‐amplified RCC cases, this finding has been reported in other eosinophilic renal neoplasms and therefore should be interpreted in the context of the overall and ancillary findings rather than as an isolated diagnostic feature. ³⁰ , ³¹ , ³² , ³³ Likewise, psammomatous calcifications, previously reported in up to 50% of ESC‐RCCs, were not identified in our cohort. ²⁷ , ²⁹

The IHC staining profiles of these entities also demonstrated considerable overlap. Both entities showed diffuse positive expression of vimentin and P504S with retained SDHB expression and were largely negative or only focally positive for CK7. Although ESC‐RCC more frequently exhibited diffuse CK20 expression with absent or patchy Melan‐A staining, TFEB‐amplified RCC more often exhibited the opposite pattern; these differences were not statistically significant. Therefore, CK20 and Melan‐A may provide supportive evidence but are insufficient to reliably distinguish these tumours in isolation.

PD‐L1 expression was variable in both entities, both among tumours and within individual cases. PD‐L1 expression is of particular interest in TFEB‐amplified RCC because of its aggressive clinical behaviour and the potential role of immune checkpoint inhibition. Experimental studies have shown that the TFEB protein promotes PD‐L1 expression, facilitating immune evasion and mTOR inhibition resistance. ³⁴ Consistent with previous reports demonstrating PD‐L1 expression in TFEB‐amplified tumours, ³⁵ , ³⁶ several tumours in our cohort exhibited strong PD‐L1 staining unrelated to the age of the specimen, despite literature reporting a negative correlation between specimen age and PD‐L1 immunoreactivity ³⁷ , ³⁸ Furthermore, a published case report describing a favourable clinical response of TFEB‐amplified RCC to anti‐PD‐L1 therapy suggests that immune checkpoint inhibition may represent a therapeutic strategy for selected patients, though additional clinical studies are needed. ³⁹

Molecular profiling confirmed the defining genomic alterations of both entities. All sequenced ESC‐RCCs harboured pathogenic TSC1 or TSC2 variants and every case of TFEB‐amplified RCC demonstrated focal amplification of the short arm of chromosome 6 involving the TFEB locus. ²⁹ The protein products encoded by TSC1/2 and TFEB participate in regulation of the PI3K‐AKT–mTOR pathway, which is recurrently altered and up‐regulated in renal tumorigenesis. ² , ⁴⁰ Pathogenic TSC1/2 variants resulting in loss of function lead to up‐regulation of mTOR signalling, while TFEB amplification results in increased TFEB function downstream of mTOR. ⁴¹ These convergent molecular mechanisms may partially explain the morphological overlap between these entities. ⁴²

Traditionally, TFEB amplification has been established using fluorescence in situ hybridization (FISH). Recent studies have proposed FISH thresholds of >5 copies of TFEB to define amplification, with 5–10 copies representing low‐level amplification and >10 copies representing high‐level amplification. ³⁶ In the present study, TFEB amplification was identified using copy‐number analysis generated from routine hybrid capture NGS. NGS panels are routinely used to evaluate copy number gains, ⁴³ , ⁴⁴ , ⁴⁵ and have been previously used to diagnose TFEB‐amplified RCC. This study demonstrates the feasibility of using NGS to simultaneously identify sequence variants characteristic of ESC‐RCC and focal amplification characteristic of TFEB‐amplified RCC.

In summary, morphology and immunohistochemistry remain essential for generating an appropriate differential diagnosis in eosinophilic renal cell tumours. However, due to substantial overlap between ESC‐RCC and TFEB‐amplified RCC, molecular characterization is frequently required to establish a definitive diagnosis. Accurate distinction between these entities has important prognostic and potentially therapeutic implications, as ESC‐RCC generally follows an indolent clinical course whereas TFEB‐amplified RCC is associated with more aggressive behaviour. As molecular testing becomes increasingly routine in renal cell tumour diagnosis, integrated clinicopathological and molecular evaluation will be essential for the accurate classification of these uncommon neoplasms.

Author contributions

Tatjana Antic and Hayley Zullow contributed to the study conception and design. Material preparation, data collection and analysis were performed by all authors (Hayley Zullow, Melissa Y. Tjota, Jung Woo Kwon and Tatjana Antic). The first draft of the manuscript was written by Hayley Zullow and all (Hayley Zullow, Melissa Y. Tjota, Jung Woo Kwon and Tatjana Antic) commented on previous versions of the manuscript. All authors (Hayley Zullow, Melissa Y. Tjota, Jung Woo Kwon and Tatjana Antic) read and approved the final manuscript.

Consent

Waived by the University of Chicago Institutional Review Board (IRB‐24‐1223).