Eosinophilic Solid and Cystic Renal Cell Carcinoma: Morphologic and Immunohistochemical Study of 18 Cases and Review of the Literature

This study protocol was approved by the Research Ethics Committee of Tianjin Medical University Cancer Institute & Hospital (Tianjin, China) (bc2022091). Archive records showed 5763 RCCs received and processed between January 1, 2015, and June 30, 2021, at the Department of Pathology, Tianjin Medical University Cancer Institute & Hospital (Tianjin, China). All cases (except for one external consultation case) had IHC evaluation at the time of their diagnosis using a set including PAX8, CD10, carbonic anhydrase IX (CA-IX), vimentin, CK7, CD117, P504s, SDHB, and TFE3. Of these renal tumors, some cases were diagnosed as clear cell RCC, chromophobe RCC (ChrRCC), renal oncocytoma (RO), Xp11 translocation RCC, hybrid oncocytic tumor, succinate dehydrogenase (SDH)–deficient RCC, or others. Using keywords including oncocytic, eosinophilic, unusual, and unclassified, we identified a subset of 42 unclassified RCCs with eosinophilic cytoplasm and conducted IHC staining for CK20, TFEB, and Ki-67. Based on the available diagnostic criteria, 3 experienced pathologists reviewed independently the selected cases’ hematoxylin-eosin and IHC staining slides. Eventually, we identified 18 ESC RCCs and collected their data, including clinicopathologic features, treatments, and follow-up information (all patients were followed until June 2021), and made further study on tumor-related pathways including ALK, FH, NF2, phospho-4EBP1 (p-4EBP1), phospho-S6 (p-S6), and INI1.

Tumor tissues were fixed in 10% formalin and embedded in paraffin. The 4-μm-thick whole sections were used for IHC staining with an automated Ventana BenchMark XT system (Roche, Ventana Medical Systems Inc) for the following antibodies: CD117 (prediluted; YR145, MXB Biotechnologies), CK7 (prediluted; OV-TL12/30, MXB Biotechnologies), CD10 (prediluted; 56C6, MXB Biotechnologies), CAIX (1:1000; ab1508, Abcam), vimentin (prediluted; V9, MXB Biotechnologies), P504s (prediluted; 13H4, MXB Biotechnologies), PAX8 (1:100; 4H7B3, ProteinTech Group), CK20 (prediluted; ZA-0574, MXB Biotechnologies), TFE3 (1:300; SC-5958, Santa Cruz Biotechnology), TFEB (1:100; A303-673A-T, Thermo Fisher Scientific), SDHB (prediluted; MAB-0736, MXB Biotechnologies), p-4EBP1 (Thr37/46; 1:400; 236B4, Cell Signaling Technology), p-S6 (Ser235/236; 1:100; D57.2.2E, Cell Signaling Technology), NF2 (1:50; 21686-1-AP, Proteintech Group), FH (1:200; 11375-1-AP, Proteintech Group), ALK (prediluted; ZM-0248, MXB Biotechnologies), INI1 (prediluted; ZM-0173, MXB Biotechnologies), and Ki-67 (prediluted; MXB Biotechnologies). Positive and negative controls yielded appropriate results for each antibody.

Immunoreactivity was evaluated in a semiquantitative manner based on labeling intensity and the percentage of immunopositive tumor cells for all antibodies. The immunostaining was considered negative when less than 1% of cells were stained, focally positive when 1% to 50% of cells were stained, and positive when more than 50% of cells were stained.

We used the QIAamp DNA FFPE Tissue Kit (Qiagen) to extract DNA and used the Qubit dsDNA HS Assay Kit (Life Technologies) to quantify DNA. The gDNA library was captured using a customized 706-gene individually synthesized 5′-biotinylated DNA 120-bp oligonucleotide probe panel with xGen Hybridization and quantified using a Qubit dsDNA HS Assay Kit. RNA was extracted using a miRNeasy FFPE Kit (Qiagen) and quantified using a Qubit RNA HS Assay Kit (Thermo Fisher). The cDNA library was captured using a customized 649-gene individually synthesized 5′-biotinylated DNA 120-bp oligonucleotide probe panel with xGen Hybridization and Wash Kit (IDT). The captured libraries were sequenced on Illumina NovaSeq 6000 with 2 × 150-bp paired-end reads, following the manufacturer’s instructions (Illumina). A comprehensive genomic analysis of the sample was performed with a DNA + RNA cancer-related gene panel (YuanSu S, 706 DNA and 649 RNA gene panel, OrigiMed). DNA-seq reads were mapped to the hg19 reference sequence with BWA (version 0.7.12). PCR duplicates were removed by Pi-card (version 2.5.0) and recalibrated by the BaseRecalibrator tool from GATK (version 3.1.1). An in-house–developed algorithm was used for DNA fusion detection. RNA-seq reads were performed using STAR (version 2.5.3) algorithm for mapping and STAR-fusion (version 0.8) for fusion detection. Gene fusions were identified when the total number of supportive reads spanning the fusion junction was 5 or more.

Clinicopathologic features and follow-up of ESC RCCs are summarized in Table 1. All of our diagnoses were based on surgical resection of specimens and there was not any preoperative biopsy done on them according to our Chinese diagnosis and treatment guidelines. There were 11 men and 7 women, with a median age of 57 years (mean, 49.6 years; range, 17–73 years). Of the 18 cases with available data from our hospital, 8 were incidentally detected during clinical workup for urinary tract symptoms/infection or during annual routine physical examinations. The other 10 patients presented with symptoms such as abdominal pain or hematuria. Eight patients had a partial nephrectomy, and the other 10 patients had a radical nephrectomy. The tumors were less frequently located in the left than in the right kidney (left, 6; right, 12). The tumor sizes ranged from 1.5 to 7.5 cm in maximal dimension (median, 4.1 cm; mean, 5.0 cm). No patients had any previous significant medical history. Thirteen tumors were staged as pT1 (4 as pT1a and 9 as pT1b), 2 tumors were staged as pT2a, and 1 case was staged as pT3 as result of spreading to the perirenal fat, according to the American Joint Committee on Cancer 8th TNM system.²  Follow-up was available for 18 patients, ranging from 3 to 73 months (median, 20 months; mean, 29.8 months). During the follow-up period, one case (case 12) had metastasized into the collarbone. Recurrences or metastases were not documented in other patients.

Grossly, almost all the tumors were well marginated and solid, with tan-mahogany, gray, or dark brown, soft, and slightly heterogeneous cut surfaces (Figure 1, A). Only one case had the typical mixed macrocystic and solid appearance (Figure 1, B). The tumors mostly lacked a thick fibrous peripheral capsule. Necrosis was identified in 4 cases.

Microscopically, low-power views showed that all tumors were well circumscribed and predominantly consisted of solid or nested patterns mixed with cysts containing trabeculae of various thicknesses (Figure 1, C). The cystic component was lined by neoplastic cells with abundant eosinophilic cytoplasm, assuming a hobnail appearance (Figure 1, C). In some cases, tumor cells formed tubular architectures. As seen in RO, well-delineated hypocellular areas were frequently found in 5 tumors. There were thick-walled vessels scattered within or at the periphery of the tumors in 11 cases, and focal hemorrhage was noted in many cases. High-power microscopic examination showed that the tumor cells varied considerably in size. The oncocytic neoplastic cells had readily identifiable coarse, basophilic to purple cytoplasmic granules and eosinophilic globules (Figure 1, D) and exhibited indistinct to slightly distinct cell membranes (Supplemental Figure 1, A; see supplemental digital content containing 3 figures and 1 table at https://meridian.allenpress.com/aplm in the October 2024 table of contents): some cases had only a few entrapped microcysts (Supplemental Figure 1, B). No perinuclear halos were seen. Intracytoplasmic vacuoles could be observed in all cases (Supplemental Figure 1, C). The tumor cells presented round nuclei frequently with fairly conspicuous, prominent nucleoli, corresponding to WHO/International Society of Urological Pathology grade 3 (Supplemental Figure 1, D), and some tumors presented with occasionally atypical-looking or irregular nuclei (Supplemental Figure 1, E and F). Additionally, architectural and cytologic variations were focally presented in some cases. For example, scattered intratumoral lymphocytes were seen in 10 tumors (Supplemental Figure 2, A). Multinucleated tumor cells were present in 6 cases (Supplemental Figure 2, B). Scattered calcifications, including psammoma bodies, were present in 5 cases (Supplemental Figure 2, C and D), small aggregates of histiocytic infiltration could be seen in 3 cases (Supplemental Figure 2, E), and conspicuous crystallike spaces could be seen in 1 case (Supplemental Figure 2, F). Overall, the ESC RCC histology was morphologically heterogeneous.

IHC staining was implemented on all oncocytic renal tumors. The panel included PAX8, CA-IX, CK7, CD117, CD10, vimentin, P504s, SDHB, and TFE3. The IHC profiles of the 18 ESC RCCs are shown in Table 2. CK20 (Figure 2, A and B) was positive in 17 of 18 cases (we had made the CK20 stains on 3 different blocks that had sufficient tumor cells in all of the focal CK20⁺ cases), CD117 was negative in 17 of 18 tumors (Figure 2, C), CK7 was negative in 1518 cases (Figure 2, D), PAX8 was positive in 18 of 18 cases, all cases were negative for CAIX, CD10 was positive in 8 of 17 tumors, vimentin was positive in 11 of 17 tumors, AMACR was positive in 12 of 17 tumors, all tumors retained SDHB, and all tumors were negative for TFE3 and TFEB. Ki-67 indices were below 10% in all cases, except for 30% of the hot area in the metastatic case 12.

Interestingly, as for the 17 CK20-positive ESC RCCs, heterogeneous staining patterns were observed with the CK20 antibody. Five cases showed positive cytoplasmic staining (Figure 2, A), 6 cases showed diffuse or focal staining with membranous accentuation (Figure 2, B), and 6 cases had both focal cytoplasmic and membranous reactivity.

Moreover, we performed a series of initial analyses of tumor-related pathways on 18 ESC RCC cases by IHC, which covered the recurrent molecular change in 5 pathways, including hyperactive mechanistic target of rapamycin complex 1 (mTORC1) signaling, dysregulated Hippo-YAP pathway, FH deficiency, ALK translocation, and SMARCB1 mutations. Group 1 had positive test results for diffuse or focal intense staining of p-4EBP1 (Supplemental Figure 3, A) and p-S6 (Supplemental Figure 3, B) in 8 cases by IHC, which can imply hyperactive mTORC1 signaling (in 1 of the 8 cases the hyperactive mTORC1 signaling was not found by molecular analysis) (7 of 18; 38.5%). Group 2 probably harbored dysregulated NF2 protein level in 4 cases (4 of 18; 23%) in which NF2 was negative (Supplemental Figure 3, C). The tumors in groups 1 and 2 were mutually exclusive. Group 3 contained the remaining 7 tumors (7 of 18; 38.5%), in which there were no recurrent molecular features identified (although case 12 was positive for p-4EBP1 and p-S6). Of note, all of our tested cases had no evidence of FH deficiency, supported by the positive staining of FH (Supplemental Figure 3, D). There was no evidence of ALK translocation, as supported by the negative staining of ALK (Supplemental Figure 3, E), or SMARCB1 mutations, which was demonstrated by the positive staining of INI1 (Supplemental Figure 3, F). The findings shown in Figure 3 demonstrate 2 groups with distinct features.

The result of next-generation target sequencing showed that CDKN2A deletion was detected in the metastatic RCC (case 12).

Oncocytic renal neoplasms are tumors with eosinophilic cytoplasm with a broad spectrum of histologic presentations. Recently, more reports have been published regarding the histology, IHC, and molecular and clinical features of the existing and newly accepted and emerging renal cell tumors.^1–15  However, there is still a need to report and characterize those neoplasms, especially among the Asian patient population. We described a group of renal tumors composed of oncocytic tumor cells forming predominantly solid and cystic growth patterns, consistent with ESC RCC reported recently, which has been defined in the 2022 WHO classification of urogenital tumors.²  To our knowledge, until now, there have not been many reports in the English-language literature to summarize the morphologic, IHC, and molecular characteristics of these tumors in an Asian population. As a result, our work aimed to provide more details about this provisional renal entity.

The distinct renal tumors were first introduced as an emerging/provisional entity in 2016, following the initial description by Trpkov et al.³  Initially, 16 oncocytic tumors were reported, which were recognized almost exclusively in female patients without tuberous sclerosis complex (TSC). Since this initial report, other cases of ESC RCC have been identified and characterized in various studies and case reports.^5–16  A summary of the critical features of ESC RCC studies is shown in the Supplemental Table. Most reported ESC RCCs are sporadic, with a broad age range distribution, including pediatric patients. However, about 10% or less of these tumors with virtually identical features have been documented in TSC patients.^1,3,17–20  Most ESC RCCs exhibit indolent behavior, but 4 cases have been reported with metastases, making it necessary to continue follow-up of these patients.^1,3,6,8  The current molecular evidence of these tumors by next-generation sequencing (NGS) has demonstrated recurrent and mutually exclusive somatic biallelic loss or mutations of the TSC gene family.^5,8,11,21 

In our practice, we have also encountered a morphologically distinct group of sporadic RCC with uniform features compatible with ESC RCC in our institution, the fourth largest tertiary oncology center in China. From 2015 to 2021, our laboratory processed 5763 renal tumors, of which 18 ESC RCCs were identified by searching the LIS data retrospectively. The estimated prevalence of ESC RCC is 0.3% (18 of 5763) among RCC diagnoses, which is relatively comparable to the incidence (0.2%) reported by Trpkov et al.⁵  In this group of cases, 6 were initially diagnosed as ESC RCC, following the criteria outlined by Trpkov et al³  in 2016. The other 12 cases were processed before the paper was published and were given different diagnoses. Among these cases, 8 were originally diagnosed as unclassified RCC, 2 as RO, 1 as ChrRCC, and 1 as hybrid oncocytic/chromophobe tumor (HOCT).

Kidney tumors featuring eosinophilic cytoplasm can usually lead the examining pathologist down a difficult path to correctly diagnose the entity. Morphologically, many of these masses share multiple similarities other than eosinophilic cytoplasm. The renal ESC RCC discussed in the article also shows overlapping morphologic characteristics with RO and eosinophilic ChrRCC as well as other oncocytic renal cell neoplasms such as HOCT.

As for RO, it is a benign neoplasm. Tumor cells are arranged in nest, solid, and/or tubular patterns without the cystic appearance that is the typical trait of ESC RCC. Besides, RO is typically positive for CD117 and shows negative or very focal scattered positivity for CK7.²²  With these features taken together, we can distinguish ESC RCC from RO very well.

Considering ChrRCC as a malignant epithelial tumor, it displays polygonal cells along with prominent plantlike thick cell membranes, and nuclei are irregularly featured, often with perinuclear halos. Binucleation is common, and ChrRCC is typically CD117 positive and diffusely positive for CK7.²³  In summary, we can also separate ESC RCC from ChrRCC based on the above several points of ChrRCC and the presentation of all the ESC RCCs we have observed.

Moreover, an entity with overlapping features of oncocytoma and chromophobe carcinoma is the group of HOCTs, which are currently considered either subtypes or peculiar variants of ChrRCC, according to the 2016 WHO classification of renal tumors.¹⁶  HOCT is believed to be a CD117-positive tumor that shows some histologic features of RO and some of ChrRCC.²⁴  Unfortunately, there is no description of patterns of CK20 expression. Among our cases, although some cases were CD117 positive and/or CK7 positive, all of them exhibited conspicuous typical mixed macrocystic and solid appearances as well as distinct positive CK20. As a result, it is easy to get the correct diagnosis.

Interestingly, we found that our cases demonstrated different CK20 staining patterns. Four tumors showed positive cytoplasmic staining, 6 tumors showed diffuse or focal staining with membranous accentuation, and 6 tumors had both focal cytoplasmic and membranous reactivity. We believe this is the first time the different CK20 staining patterns have been reported. Trpkov et al³  performed CK20 IHC in 81 ROs using tissue microarray methodology; only one case was positive with the membranous staining. One of 35 ChrRCCs showed diffuse positive membranous staining. We further conducted CK20 IHC studies on 30 distinct ChrRCCs, and none of these cases showed the membranous accentuation. Although it was unclear whether there was a significant distinction between CK20 cytoplasmic reactivity and membranous reactivity, it seems the CK20 IHC in our cohort demonstrated heterogeneity in terms of its patterns. In addition, it is necessary to mention that the frequent and diffuse CK20 positivity is not restricted to ESC RCC, but that a new subset of aggressive tumors labeled TFEB-amplified RCCs is also typically diffusely positive for CK20²⁵ ; these related studies/abstracts were also presented at the 2023 United States and Canadian Academy of Pathology meeting. Therefore, the diagnosis of ESC RCC should be based on a combination of gross and microscopic features, IHC, and molecular detection, rather than solely on CK20 positivity.

Based on reported molecular analyses about ESC RCC, these tumors were characterized by somatic mutations of TSC2 or MTOR that led to hyperactive mTORC1 signaling. Protein synthesis is by far the best-characterized process controlled by mTORC1. mTORC1 directly phosphorylates the translational regulators eukaryotic translation initiation factor 4E–binding protein 1 (4E-BP1) and S6 kinase 1 (S6K1), which in turn promote protein synthesis; as a result, to determine if the TSC1/2 aberrations in ESC RCC lead to mTOR pathway activation, we can monitor the phosphorylation status of both S6K1 (at 2 sites) and 4EBP1 by IHC. Moreover, many studies have used the antibodies of p-4EBP1 and p-S6 to research the status of the mTOR pathway and verified the IHC feasibility in turn by NGS.¹¹  As such, we also set out to detect the condition of mTORC1 signaling in this study by IHC of p-S6 and p-4EBP1. Eventually, we found that 8 cases had hyperactive expression. In light of the work from Chen et al⁴  in 2016, the somatic mutation analysis of their aggressive unclassified renal cell cancer cohort demonstrated that potential mTORC1 pathway–activating mutations comprising MTOR (5), TSC1 (4), TSC2 (3), and PTEN (4) occurred mutually exclusively in 16 cases (26%). Shortly thereafter, they collected 7 retrospectively identified cases in 2019 and further demonstrated that these tumors were characterized by somatic mutations of TSC2 or MTOR that led to hyperactive mTORC1 signaling.²⁶  Most recently, Tjota et al²⁷  raised that TSC/MTOR-associated neoplasms were a distinct group exhibiting heterogeneous morphology and IHC staining, as well as clinical behavior. What is encouraging is that there are mTOR inhibitors currently available for the treatment of RCCs. In summary, it is imperative that further studies with molecular analyses be performed to figure out the relationship between ESC RCC and other renal entities to clarify the pathologic spectrum of sporadic RCC primarily driven by TSC1/TSC2 or MTOR alterations.

We found the loss of NF2 in our cohort of ESC RCCs, which were the first reported in ESC RCC. NF2, a tumor-suppressor gene, was first discovered as the principal genetic change regarding the human neurofibromatosis type 2 cancer predisposition syndrome.^28,29  In the Hippo signaling pathway, Merlin, which is encoded by NF2, is of particular clinical relevance.^30,–,32  As a result, the NF2 tumor suppressor has been the subject of intense investigation. Until now, NF2 mutations have also occurred in other types of cancer, including mesothelioma,³³  breast,³⁴  colorectal,³⁵  hepatic,³⁶  prostate,³⁷  and so on. However, there is only a limited amount of research about the NF2 status of renal neoplasms. Durinck et al³⁸  analyzed diverse genomic alterations of the 167 primary non–clear cell RCCs. They confirmed only 2 cases with mutations of NF2 (2 of 167); thus, the rate is very low. Later, Pal et al³⁹  found that 5 of 17 collecting duct carcinomas had genomic alterations of NF2, and that mTOR inhibitors may be of interest in patients with NF2 alterations. Simultaneously, Chen et al⁴  identified 11 aggressive unclassified renal cell cancers (11 of 62) with NF2 mutations, which exhibited a range of architectural patterns. In our study, NF2 alterations occurred in 4 ESC RCCs, mutually exclusive of the 8 tumors showing positive staining for p-4EBP1 and p-S6, indicating there might be another distinct heterogeneous subset.

In this study, we explored the clinicopathologic and genetic features of the metastatic RCC (case 12), which had metastasized into the collarbone 3 months after primary nephrectomy. We could see the imaging and morphologic features of the metastatic lesion in Figure 4, A through D. CDKN2A deletion was detected in this case. CDKN2A is a tumor suppressor and cell cycle regulator that has been elucidated in some cancers, such as melanoma,⁴⁰  lung cancer,⁴¹  colorectal cancer,⁴²  pancreatic cancer,⁴³  breast cancer,⁴⁴  and bladder cancer.⁴⁵  CDKN2A deletion has been reported in clear cell RCC,⁴⁶  papillary RCC,⁴⁷  and metastatic RCC.⁴⁸  Disruption of CDKN2A (deletion or methylation) has been reported to be a frequent event in tumorigenesis; it can affect the clinical characteristics and patient outcomes. Thus, the result of the CDKN2A mutation further explained the metastatic nature of our case.

Taken together, ESC RCC demonstrates readily recognizable solid morphology with or without cystic growth, oncocytic cytoplasm with granular cytoplasmic stippling, round or irregular nuclei, and different positive stain patterns of CK20. Based on our cases, it is tempting to speculate that mTORC1 hyperactivity and NF2 loss could play key roles in the tumorigenesis. Moreover, one case showed metastatic behavior, indicating the aggressive potential. In conclusion, in view of morphologic and IHC characteristics, these data clearly indicate that ESC RCC is a novel tumor entity that requires further study to be fully characterized. We hope that this study will attract more and more experts and scholars to pay attention to this rare and emerging tumor entity.

This work was supported by the Tianjin Medical University Cancer Institute and Hospital and National Clinical Research Center for Cancer.