Abstract

Background

   Renal anastomosing hemangioma (RAH) is a rare benign renal tumor, and
   its clinicopathologic characteristics and genetic mutation spectrum
   related to its mechanisms of pathogenesis are unclear.

Methods

   We carried out whole-genome sequencing (WGS) on RAH samples to explore
   the genetic mutation spectrum and verified the results by Sanger
   sequencing. Immunohistochemical analysis was also performed to reveal
   the histopathological characteristics and the tumor microenvironment
   components. Moreover, a population-based study was conducted after
   searching the PubMed, EMBASE, and Ovid SP databases to systematically
   summarize the clinicopathologic features of patients with RAH.

Results

   WGS analysis revealed 10532 somatic single-nucleotide variants (SNVs),
   6705 somatic insertions and deletions (INDELs), and mutations in 32
   predisposing genes and 10 driver genes, among which the mutations in 8
   of the predisposing genes, CNTNAP2, NCOA2, FAT1, MET, TJP2, MAML2,
   SRGAP3, and CSMD3, and the mutation site in the driver gene HIP1 were
   confirmed by Sanger sequencing. Moreover, the immunohistochemical
   profile of the tumor microenvironment revealed that the expression
   content of tumor-associated macrophages (CD163, CD68) and fibroblasts
   (SMA) differs between cancerous and precancerous tissues which may
   regulate the disease development. On the basis of our population-based
   analysis, we summarized the clinicopathological features of 100
   patients with RAH and identified significant differences in age
   (p=0.001), tumor site (p<0.001), tumor focality (p<0.001), largest
   tumor diameter (p=0.001) and surgical approach (p=0.010) between
   patients with RAH with end-stage renal disease (ESRD) and those without
   ESRD.

Conclusions

   The distinct phenotypes of RAH may be associated with the different
   genetic mutation spectra identified in our study. The presence or
   absence of comorbid ESRD varies among patients with RAH. However,
   additional studies are required to validate our results.

   Keywords: renal anastomosing hemangiomas, whole-genome sequencing, gene
   mutations, immunoregulation, population-based study

1. Introduction

   Anastomotic hemangioma (AH) is a rare and histopathologically distinct
   benign vascular tumor that preferentially involves the genitourinary
   tract and paraspinal region. In 2009, Montgomery and Epstein first
   described AH as a distinct vascular lesion in the kidney and testis
   ([45]1). Renal anastomotic hemangiomas (RAHs) are listed in the 2016
   WHO tumor classification as a subtype of renal capillary hemangioma
   ([46]2). RAH can develop across a broad age spectrum, occurring from 10
   to 83 years of age (with an average age of 49 years), with a
   male-to-female ratio of 2:1 ([47]3). These hemangiomas are often
   observed in individuals with end-stage renal disease (ESRD), and
   approximately two-thirds of cases are associated with compromised renal
   function ([48]4). Although AHs are usually isolated, multifocal and
   bilateral lesions are also common in patients with ESRD ([49]3, [50]5).

   Most AHs are found incidentally on radiological evaluation for other
   purposes. Common symptoms of renal hemangiomas include abdominal pain,
   hematuria, and the presence of an abdominal mass ([51]3, [52]5). It is
   challenging to preoperatively diagnose renal hemangiomas by imaging
   because of their small size, nonspecific imaging characteristics, and
   difficulty in distinguishing them from other renal tumors ([53]6).
   Limited information is available regarding the imaging features of
   renal anastomotic hemangiomas. They typically appear as restricted
   lesions with high T2 signals on magnetic resonance imaging, peripheral
   or diffuse enhancement in endoarterial phases on dynamic CT, sustained
   enhancement in delayed phases, and centripetal filling with strong
   contrast enhancement from the periphery to the center as potential
   diagnostic clues ([54]7).

   Histologically, RAH tumors consist of splenic sinusoidal vascular
   channels that are frequently connected and lined with cuboidal
   endothelial cells ([55]5, [56]8–[57]12). Endothelial cells usually
   present as flat nails ([58]9–[59]11). Sometimes, eosinophilic clear
   spherules are present in the cytoplasm of tumor endothelial cells
   ([60]9, [61]11). The stroma is occasionally infiltrated with foamy
   macrophages or mast cells ([62]11). Extramedullary hematopoiesis is
   observed in vascular channels containing red lineage precursor cells
   and megakaryocytes ([63]5, [64]9, [65]11, [66]12).
   Immunohistochemically, almost all AHs exhibit diffuse positive staining
   for endothelial markers (including CD31, CD34, ERG, factor VIII, and
   FLI-1), supportive pericytes prominently express SMA, and lower
   endothelial cell proliferative activity is suggested given that most of
   the Ki-67 staining is < 5% ([67]5, [68]10). The prognoses of patients
   with AH, and cases of recurrence and death, have not been reported in
   previous studies, so AH is considered a benign tumor. Since
   preoperative clinical features are not sufficient for a definitive
   diagnosis of RAH and almost 90% of patients with renal RAH undergo
   total nephrectomy, precise preoperative percutaneous puncture is
   essential to prevent unnecessary surgical procedures ([69]4, [70]5).

   Regarding the genetic features of RAH, it was recently reported that
   90% of patients with AH present with the circulating activation of
   certain hotspot mutations, including GNAQ, GNA14, or GNA11, and that AH
   is a clonal tumor confirmed by these G proteins. However, these mutated
   G proteins are expressed in other vascular tumors. Moreover, these
   mutations are not found in angiosarcomas, which could provide evidence
   for distinguishing AH from AS ([71]13).

   Almost all previous genetic studies of patients with RAH have been
   based on NGS, and this study is the first in which whole-exome
   sequencing was performed on a patient with a pathological diagnosis of
   left renal anastomotic hemangioma, with new findings of somatic
   mutations, tumor susceptibility genes, and driver genes in several
   tumor samples. We also conducted a population-based study after
   searching the relevant literature to summarize the clinical and
   pathological features, possible prognoses, and treatment strategies for
   patients with RAH.

2. Materials and methods

2.1. Clinical analysis

   All relevant information was obtained from the Second Hospital of
   Dalian Medical University. The study protocol received approval from
   the Ethics Committee (approval number: KY2024-391-01) and adhered to
   the Helsinki Declaration of Ethical Principles for Human Medical
   Research. We obtained written informed consent from the patient.

2.2. Immunohistochemical analysis

   A section of a renal mass was obtained by the laboratory of our
   hospital. The sample was fixed in 10% formaldehyde, embedded in
   paraffin, sliced into 4 mm sections, placed on slides, and some of the
   slices were stained with hematoxylin and eosin (HE) before being
   photographed under a light microscope. The other slices were stained
   with various antibodies to stain the markers for diagnosis including
   CD31, CD34, podoplanin (D2-40), EMA, cytokeratin (AE1/AE2), melanoma
   gp100 (HMB45), Ki67, FLI1, ERG, GLUT-1, and herpesvirus 8(HHV8) and the
   markers to detect components of the tumor microenvironment including
   PD-L1, markers of tumor-associated fibroblasts (SMA), markers of TAMs
   (Tumor-associated macrophages, CD163 and CD68). The detailed
   information of antibodies was presented in the [72]Supplementary Table
   S1 .

2.3. Whole-genome sequencing

   The data presented in the study are uploaded in the NCBI database SRA
   repository under accession numbers SRR28364768 and SRR28364767
   ([73]http://www.ncbi.nlm.nih.gov/sra/?term=SRR28364768,
   [74]https://www.ncbi.nlm.nih.gov/sra/?term=SRR28364767).

2.3.1. DNA extraction

   Following the manufacturer’s instructions, DNA was extracted from the
   FFPE tumor sample and matched peripheral blood sample via the GeneRead
   DNA FFPE Kit (Qiagen, Germany). Whole-genome sequencing (WGS) was
   performed on the samples. Using agarose gel electrophoresis and the
   Qubit^® DNA Assay Kit in a Qubit^® 3.0 Fluorometer (Invitrogen, USA),
   DNA quality was assessed. To construct libraries, DNA samples totaling
   0.2 µg and with concentrations greater than 20 ng/µL were utilized.

2.3.2. Library preparation and sequencing

   Sequencing libraries were created using the NEBNext^® UltraTM DNA
   Library Preparation Kit (NEB, USA), and each sample was given an index
   code. After the genomic DNA samples were divided into 350 bp fragments
   using Covaris sonication, the DNA fragments were polished, A-tailed,
   and ligated to full-length adapters. This was completed before
   additional PCR amplification. The AMPure XP system (Beckman Coulter,
   Beverly, USA) was used to purify the PCR products. A Qubit^® 3.0
   Fluorometer (Invitrogen, USA) was then used to measure the quantities
   of DNA, and an NGS3K/Caliper was used to analyze the size distribution
   of the libraries and quantify them using real-time PCR (3 nM). The
   index-coded samples were clustered using the cBot Cluster Generation
   System with the Illumina PE150 Cluster Kit (Illumina, USA).

2.3.3. Quality control

   The raw data were filtered to remove reads with adapters, more than 10%
   unidentifiable base information in single-end sequencing reads, and
   low-quality bases (below the 5th level) accounting for more than 50% of
   the read length. Reads that were more than half of the length were
   eliminated. The quality of the sequencing data was ensured to be
   primarily distributed at a Q30 of not less than 80%, and the sequencing
   error rate for each base location was less than 1%.

2.3.4. Bioinformatics analysis

   To produce the BAM format for the first alignment results, validated
   sequencing data were aligned using BWA ([75]14) against the reference
   genome (GRCh37/hg19/GRCh38). After that, SAMtools ([76]14) was used to
   score the results, and Sambamba was used to identify duplicate reads.
   Next, Sambamba was used to identify duplicate reads. Ultimately, the
   depth and coverage statistics of the identified duplicate reads were
   computed.

2.3.5. Variant detection, somatic mutation calling and functional annotation

   We identified somatic variants, including SNPs, INDELs, copy number
   variants (CNVs) and structural variants (SVs). SAMtools ([77]15) was
   used to identify SNPs, count the number of SNPs in different genome
   regions, determine the number of different types of SNPs in coding
   regions, determine the distribution of conversions and reversals,
   determine the number of SNPs, and determine the genotype distribution
   on the basis of the initial results (BAM files). INDELs were identified
   using SAMtools, which calculates the number of INDELs in different
   genomic areas and the variety of INDELs in coding regions. The variety
   of CNV occurrences was determined, and Control-FREEC ([78]16) was
   utilized to detect increases and decreases in CNVs. Lumpy ([79]17) was
   used to identify SVs and determine the variety of SV events that could
   occur. Somatic SNVs, INDELs, CNVs, and SVs were detected using Mutect
   ([80]18), Strelka ([81]19), Control-FREEC ([82]16), and Lumpy ([83]17).
   ANNOVAR was used for functional annotation of the identified gene
   variations ([84]20).

2.3.6. Identification of potential predisposing genes and driver mutations

   Using the CGC database (Cancer Gene Census,
   [85]http://cancer.sanger.ac.uk/cancergenome/projects/census/), FACD
   (Familial Cancer Database, [86]http://www.familialcancerdatabase.nl/),
   intOGen ([87]https://intogen.org/search) database, and the reported
   genes summarized in the Nature literature, we used SAMtools software to
   detect germline mutations (SNPs, INDELs) in patient normal tissues.
   Using muTect software, we analyzed somatic mutations (SNVs and INDELs)
   in tumor tissues. We subsequently searched the CGC database, Bert
   Vogelstein’s significant mutated genes (SMGs) ([88]21), synthesis
   ([89]22), and known driver genes.

2.3.7. Analysis of tumor purity, tumor ploidy and targeted drug prediction

   We calculated the purity of the tumor sample (the ratio of tumor cells
   to total cells), ploidy (the average copy number of the sample), and
   cancer DNA fraction (the ratio of tumor DNA to total DNA) using
   ABSOLUTE to guarantee the quality of the analyses ([90]23) We compared
   the identified somatic mutations with NovoDR medication databases, such
   as the My Cancer Genome, FDA, and Pharmacogenomics Knowledge Base
   Database (PharmGKB) ([91]24), to evaluate potential targeted drugs.

2.3.8. Sanger sequencing

   Sanger sequencing was used to validate the susceptibility and driver
   genes containing mutant bases identified by NGS. Following the
   manufacturer’s instructions, genomic DNA was extracted, and polymerase
   chain reaction (PCR) was carried out. Once the PCR products were
   purified via agarose electrophoresis, an Applied BiosystemsTM 3730xl
   sequencer was used to sequence the products.

2.3.9. Structure of the protein–protein interaction network of mutant genes

   To construct the protein–protein interaction (PPI) network, we searched
   the genes confirmed by Sanger sequencing with the Search Tool for the
   Retrieval of Interacting Genes/Proteins (STRING, version 12.0)
   database, set the minimum required interaction score at 0.40 and the
   maximum number of interactors at 50, and then visualized the PPI
   network with Cytoscape software (version 3.9.1).

2.3.10. GO and KEGG pathway enrichment analyses

   We analyzed the biological functions of the PPI node genes through Gene
   Ontology (GO) ([92]25)and Kyoto Encyclopedia of Genes and Genomes
   (KEGG) ([93]26) pathway enrichment analyses with a statistically
   significant difference threshold of q value < 0.05 and visualized the
   data in plots constructed via R software (version 4.3.2).

2.3.11. Relationship between predisposing gene mutations and immune cell
infiltration in the tumor microenvironment

   We searched the mutated genes and surface antigens of immune cells in
   the tumor microenvironment and analyzed the combined score with the
   STRING database, setting the minimum required interaction score at
   0.15, and then visualized the PPI network with Cytoscape software
   (version 3.9.1).

2.4. Population-based study

   We searched the relevant literature in the PubMed, EMBASE, and Ovid SP
   databases for RAH patients before February 2025 using the keywords
   “kidney”, “renal” and “anastomosing hemangioma”. Only patients for whom
   complete information was available were included in our analysis. A
   total of 100 patients were included, as shown in the detailed flow
   chart in [94]Supplementary Figure S1 . We summarized their clinical
   characteristics, including age, sex, clinical manifestations, clinical
   history of end-stage renal disease (ESRD), preoperative diagnosis and
   surgical approach; pathological features, including tumor site and
   location; tumor focality; largest tumor diameter (mm); concurrent or
   history of other renal tumors; and immunohistochemistry results. We
   divided the patients into an ESRD group and a non-ESRD group, compared
   the continuous variables using Student’s t test or the Mann–Whitney U
   test, and analyzed the classification indicators by the chi-square test
   or Fisher’s exact test. The significance level was established as p <
   0.05, and all the statistical analyses were performed using SPSS
   version 26.0 (SPSS Inc., Chicago, IL, USA). Moreover, we searched the
   SEER database ([95]https://seer.cancer.gov/), which consists of 17
   registries covering more than 26.5% of the United States population
   between 2000 and 2021, for information on hemangiosarcoma patients.
   Patients with an International Classification of Disease (ICD)-O-3 code
   of 9120 with malignant behavior were screened. We included a total of
   53 renal hemangiosarcoma patients in our study after excluding patients
   lacking histological confirmation or with unknown survival times. The
   clinical and tumor characteristics, including age, sex, tumor site and
   tumor size, were obtained, and survival outcomes, including overall
   survival (OS) and cancer-specific survival (CSS), were also downloaded.
   We compared the continuous variables between RAH patients and renal
   hemangiosarcoma patients using Student’s t test or the Mann–Whitney U
   test and analyzed the classification indicators between RAH patients
   and renal hemangiosarcoma patients via the chi-square test or Fisher’s
   exact test. Survival analysis was performed using GraphPad.

3. Results

3.1. Clinical analysis

3.1.1. Case presentation

   A 71-year-old woman with a left renal pelvic mass was identified via
   physical examination at an outside hospital. Her medical records
   revealed type 2 diabetes, hypertension, and coronary artery disease.
   She refuted having any other medical history, including end-stage renal
   disease (ESRD). The patient had no symptoms of left-sided low back
   pain, hematuria to the naked eye, or an abdominal mass. Preoperatively,
   the patient’s urine was retained for fluid-based cytology, suggesting
   that no uroepithelial cells were present. Preoperative urologic
   ultrasound revealed a mixed mass in the left renal pelvis measuring
   approximately 4.0*3.3 cm with clear borders and regular morphology,
   which was suspected to be cystic nephrocalcinosis of the left kidney.
   To further clarify the nature of the left renal pelvis tumor, CTU
   examination revealed a cystic solid-occupying lesion in the left
   pararenal pelvis measuring 3.4 cm*2.6 cm. Enhanced scanning revealed
   obvious enhancement of the solid portion, with persistent enhancement
   in the delayed phase and unclear demarcation from the adjacent ureter,
   which was considered to indicate a possible low-grade malignant
   neoplastic lesion, as shown in [96]Figure 1 , and the lesion and
   surrounding anatomical structure were clearly visible in the
   reconstructed three-dimensional images. Owing to the clinical diagnosis
   of a left renal pelvic mass, malignancy was not excluded, so the
   patient underwent robotic-assisted laparoscopic partial left
   nephrectomy under general anesthesia. Postoperative pathology suggested
   renal anastomotic hemangioma. Regular postoperative telephone
   follow-ups were conducted for 27 months, with the last follow-up
   occurring in February 2025, and no tumor recurrence or metastasis was
   observed.

Figure 1.

   [97]Figure 1
   [98]Open in a new tab

   CT urography imaging of renal anastomosing hemangioma. Flat sweep (A)
   enhanced cortical phase; (B) enhancing draining phase; (C) and
   reconstructed three-dimensional images (D). It showed a cystic solid
   space-occupying lesion adjacent to the left renal pelvis measuring 3.4
   cm*2.6 cm. It is poorly demarcated from the adjacent ureter and is
   considered a possible low-grade malignant neoplastic lesion. The
   reconstructed three-dimensional clearly showed the lesion and
   surrounding anatomical structure.

3.1.2. Histopathological considerations

   The size of the macroscopic tissue sample of the left pararenal mass
   was 2.4 cm * 2.2 cm * 1. The mass section was cystic and soft, had a
   mass adjacent to the peeling edge, did not show obvious necrosis, and
   contained a large amount of free adipose tissue. The size of the mass
   was 5×4.5×1.5 cm. The grayish-yellow texture was soft, and the mass
   could be a palpable nodule. Microscopic observation (left renal tumor
   base): tumor cells are polygonal or ovoid, with clear borders,
   clustered or scattered distribution, and proliferation of interstitial
   fibrous tissue, blood vessels, and adipose tissue. A benign or
   low-grade malignant tumor was considered. Short spindle-shaped,
   lattice-like, slit-like, irregularly dilated large blood vessels were
   observed, acute inflammatory cells and plasma cells were absent, and
   adipose tissue was also observed. No fibrin-like microthrombi,
   extramedullary hematopoiesis or lymphocytic infiltration was observed
   in focal areas within the tumor. No definitive diagnosis was made, and
   a benign or low-grade malignant tumor was initially considered, with
   immunohistochemistry required for further diagnosis.

3.1.3. Immunohistochemical profile for diagnosis

   [99]Figure 2 shows diffuse staining for endothelial markers via tumor
   immunohistochemistry. The endothelial cell markers CD31, CD34, ERG, and
   FLI-1 were diffusely expressed in the AH cells. Negative expression of
   D2–40 differentiated the tumor from lymphoid-derived tumors. GLUT-1
   negativity excluded juvenile hemangiosarcoma. Negative HMB-45
   differentiated the tumor from Kaposi’s sarcoma. Tumors expressing SMA
   and Ki-67 markers can be distinguished from hemangiosarcoma, which has
   a relatively high Ki-67 proliferation index but no pericyte SMA
   expression.

Figure 2.

   [100]Figure 2
   [101]Open in a new tab

   Immunohistochemical microscopic images of renal anastomosing hemangioma
   tumor cells. Representative immunohistochemical images of cluster of
   differentiation (CD) 31, cluster of differentiation (CD) 34,
   cytokeratin AE1/AE 3, lymphatic vessel marker D2-40, epithelial
   membrane antigen EMA, endothelial cell markers ERG, FLI-1, glucose
   transporter GLUT-1, glycoprotein HMB-45, Ki-67, and smooth muscle actin
   SMA, which were in tumor cells were positive in tumor cells expressing
   endothelial cell markers.

3.1.4. Immunohistochemistry for detecting components of the tumor
microenvironment

   For immunohistochemistry to detect components of the tumor
   microenvironment, we analyzed the distributions of PD-L1, markers of
   tumor-associated fibroblasts (SMA), markers of TAMs (CD163 and CD68),
   in the tumor and adjacent normal tissue. As shown in [102]Figure 3 ,
   higher expression of PD-L1, CD163, and SMA was detected in the tumor
   tissue than in the adjacent nontumor tissue, indicating that TAMs and
   fibroblasts may participate in the pathogenesis of RAH.

Figure 3.

   [103]Figure 3
   [104]Open in a new tab

   Immunohistochemical microscopic images of smooth muscle actin SMA,
   cluster of differentiation (CD) 168, cluster of differentiation (CD) 68
   and PDL-1 for the tumor micro-environment analysis in the tumor sample
   and adjacent normal tissue.

3.2. Whole-genome sequencing study

3.2.1. Identification of SNPs and INDELs

   The average Q30 of the sequenced samples was 92.74%, and the average
   error rate was 0.03%, suggesting that the sequencing data were high
   quality and complied with the analytical specifications. The normal
   sample had a mean of 179,075,896 raw reads, and the tumor sample had a
   mean of 123,689,224 raw reads. The tumor sample and matched normal
   sample included 3,636,612 and 3,678,071 SNPs, respectively. The
   correctness of the SNP dataset is reflected in the
   transformation/inversion ratio, which is 2.05 for the tumor sample and
   2.04 for the normal sample. In total, 1,053,180 and 1,061,622 INDELs
   were confirmed in the tumor specimen and matched normal sample,
   respectively. The SNP and INDEL information were displayed in
   [105]Figures 4 , [106]5 and [107]Supplementary Tables S2–S5 .

Figure 4.

   [108]Figure 4
   [109]Open in a new tab

   The distribution of SNPs in tumor and normal specimens is depicted in
   the pie chart. The left and right images display the number of SNPs in
   various genomic regions and the coding regions in tumor and normal
   specimen.

Figure 5.

   [110]Figure 5
   [111]Open in a new tab

   The distribution of INDELs in tumor and normal specimens is depicted in
   the pie chart. The left and right images display the number of INDELs
   in various genomic regions and the coding regions in tumor and normal
   specimen.

3.2.2. Identification of Somatic SNVs and INDELs

   Somatic mutations play a major role in the pathogenesis and evolution
   of tumors. A total of 6705 somatic INDELs and 10532 somatic SNVs were
   found, which were mainly grouped in intronic and intergenic regions.
   [112]Supplementary Tables S6, S7 contain comprehensive data on somatic
   SNVs and INDELs.

3.2.3. Identification of SVs and CNVs

   We identified 5 duplications, 35 deletions, 187 interchromosomal
   translocations, and 20 intrachromosomal translocations among the
   somatic structural variants (SVs) of the tumor sample. There were 68
   somatic copy number variants (CNVs), comprising 60 gain and 8 loss
   counts. [113]Figure 6 shows the Circos plot of somatic variation.

Figure 6.

   [114]Figure 6
   [115]Open in a new tab

   Somatic genomic variation Circo. The chromosome number, sequencing
   coverage map, SNPs and INDELs density, CNVs and SVs results are
   represented by a five-layered structure that extends from the outside
   to the inside.

3.2.4. Identification of Predisposing and Driver Mutating Genes

   Mutations in cancer predisposition genes (CPGs) dramatically increase
   individuals’ risk of disease development. We identified 32 predisposing
   genes, such as CNTNAP2, NCOA2, FAT1, MET, TJP2, PCSK5, MAML2, SRGAP3,
   and CSMD3, the details of which are provided in [116]Supplementary
   Table S8 , the mutational heatmaps of predisposing genes was presented
   in [117]Figure 7 . Understanding the pathophysiology of cancer requires
   the identification of driver mutations and the genes responsible for
   these abnormalities, which provide tumors with a selective growth
   advantage. Ten driver genes mutation sites were identified, as
   indicated in [118]Supplementary Table S9 : CHD3, PRKCB, SPEG, CHD7,
   KAT6B, HIP1, ERBB4, RUNX1T1, NIPBL, and CHD3, presenting the mutational
   heatmaps of driver genes in [119]Figure 8 .

Figure 7.

   [120]Figure 7
   [121]Open in a new tab

   The mutational heatmaps of predisposing genes.

Figure 8.

   [122]Figure 8
   [123]Open in a new tab

   The mutational heatmaps of driver genes.

3.2.5. Analysis of tumor purity, ploidy, and targeted drug prediction

   The tumor purity was 0.18, the tumor ploidy was 1.03, and the cancer
   DNA fraction was 10%. On the basis of the NovoDR drug database, we
   identified sixteen mutation sites associated with sixty-three targeted
   drugs, and the detailed information is presented in [124]Supplementary
   Table S10 . There were twenty-four drugs (levonorgestrel,
   spironolactone, flutamide, oxandrolone, testosterone, nilutamide,
   fludrocortisone, drostanolone, nandrolone phenpropionate, bicalutamide,
   fluoxymesterone, drospirenone, danazol, testosterone propionate,
   boldenone, calusterone, flufenamic acid, dihydrotestosterone,
   methyltrienolone, cyproterone, methyltestosterone, nandrolone
   decanoate, androgen receptor modulators and antagonists and pi3k
   inhibitors) for AR; one drug (phosphatidylserine) for ATP8A1; one drug
   (phosphatidylserine) for CASP1; one drug (purvalanol a) for CSNK1G3;
   one drug (imatinib) for DDR1; four drugs (flt3 inhibitors, mek
   inhibitors, jak2 inhibitors, dot1l inhibitors) for DEK; fifteen drugs
   (lorazepam, temazepam, clobazam, alprazolam, chlordiazepoxide,
   clorazepate, midazolam, flurazepam, diazepam, oxazepam, triazolam,
   clonazepam, bromazepam, nitrazepam and venlafaxine) for GABRQ, one drug
   (niacin) for HCAR2, one drug (pyridoxal phosphate) for IGSF10, one drug
   (marimastat) for MMP14, two drugs (adenine and formycin) for MTAP, one
   drug (NADH) for NADUFS6, one drug (platinum compounds) for PPP1R13L,
   two drugs (vitamin E and dexmedetomidine) for PRKCB, three drugs
   (staurosporine, phosphonothreonine and phosphonoserine) for PRKCQ, and
   four drugs (flt3 inhibitors, MEK inhibitors, jak2 inhibitors and dot1l
   inhibitors) for RUNX1T1.

3.2.6. Sanger sequencing

   To verify the mutations in the predisposing and driver genes,
   we carried out Sanger sequencing. As a result, we confirmed
   the presence of mutations in predisposing genes: CNTNAP2 at chr7:
   146818068 C>T, NCOA2 at chr8: 71033538 C>T, FAT1 at chr4: 187539340
   G>T, MET at chr7: 116340087 C>T, TJP2 at chr9: 71851954 G>A, MAML2 at
   chr11: 96074675 C>T, SRGAP3 at chr3: 9121722 C>T, and CSMD3 at chr8:
   113347624 C>T ([125] Figure 9 ), while the mutation in the driver gene
   HIP1 at chr7: 75186080 C>T was also confirmed by Sanger sequencing
   ([126] Figure 10 ).

Figure 9.

   [127]Figure 9
   [128]Open in a new tab

   Sanger sequencing mutations electropherograms of the CNTNAP2 at chr7:
   146818068 C>T (A), NCOA2 at chr8: 71033538 C>T (B), FAT1 at chr4:
   187539340 G>T (C), MET at chr7: 116340087 C>T (D), TJP2 at chr9:
   71851954 G>A (E), MAML2 at chr11: 96074675 C>T (F), SRGAP3 at chr3:
   9121722 C>T (G), CSMD3 at chr8: 113347624 C>T (H).

Figure 10.

   [129]Figure 10
   [130]Open in a new tab

   Sanger sequencing mutations electropherograms of the HIP1 at chr7:
   75186080 C>T (A) peripheral blood as control; (B) tumor.

3.2.7. Protein–protein interaction network structure and functional
enrichment

   The PPI network we constructed contained 59 nodes and 176 edges ([131]
   Figure 11 ). According to the GO analysis, the node genes in the PPI
   network were significantly enriched in cell−cell junction organization
   and cell junction assembly in the BP category, apical junction complex
   in the CC category, and transcription coactivator activity in the MF
   category, and the genes encoding proteins involved in forming tight
   junctions were strongly enriched in the KEGG analysis ([132] Figure 12
   ).

Figure 11.

   Figure 11
   [133]Open in a new tab

   Protein–protein interaction network (PPI). The size and color of the
   nodes represents the degree of the nodes.

Figure 12.

   [134]Figure 12
   [135]Open in a new tab

   Enrichment analysis results. (A) The BP category of GO analysis. (B)
   The CC category of GO analysis. (C) The MF category of GO analysis. (D)
   KEGG pathway enrichment analysis.

3.2.8. Analysis of immune cell infiltration alterations in the tumor
microenvironment

   The PPI network of the mutated genes and surface antigens of immune
   cells contained 67 nodes and 1466 edges, with a PPI enrichment p value
   <1.0e-16 ([136] Figure 13 ). The detailed results of the
   protein–protein interaction analysis between the mutant genes and
   surface antigens of immune cells (NK/T cells, mast cells, M1
   macrophages, M2 macrophages and plasmacytes) are presented in
   [137]Table 1 .

Figure 13.

   [138]Figure 13
   [139]Open in a new tab

   Protein–protein interaction network (PPI) of mutated genes and surface
   antigens of immune cells. The size and color of the nodes represents
   the degree of the nodes.

Table 1.

   Protein-protein interactions analysis between mutant gene and surface
   antigens of immune cells in String database.
   Node1 Node2 Node1 string id Node2 string id Co-expression
   Experimentally determined interaction Automated text mining Combined
   score
   CNTNAP2 NCAM1(CD56) 9606.ENSP00000354778 9606.ENSP00000480132 0.099
   0.069 0.509 0.552
   CSMD3 PDCD1 9606.ENSP00000297405 9606.ENSP00000335062 0 0 0.166 0.165
   CSMD3 CTLA4 9606.ENSP00000297405 9606.ENSP00000497102 0 0 0.201 0.201
   MET XBP1 9606.ENSP00000317272 9606.ENSP00000216037 0 0.047 0.156 0.161
   MET NT5E 9606.ENSP00000317272 9606.ENSP00000257770 0.179 0 0.272 0.376
   MET TPSB2 9606.ENSP00000317272 9606.ENSP00000482743 0 0.181 0 0.181
   MET SELL (CD62L) 9606.ENSP00000317272 9606.ENSP00000498227 0 0.045
   0.178 0.182
   MET MRC1 (CD206) 9606.ENSP00000317272 9606.ENSP00000455897 0 0.045
   0.181 0.185
   MET NOS2 (iNOS) 9606.ENSP00000317272 9606.ENSP00000327251 0.104 0.045
   0.169 0.227
   MET NCAM1 9606.ENSP00000317272 9606.ENSP00000480132 0 0.088 0.313 0.347
   MET SDC1 (CD138) 9606.ENSP00000317272 9606.ENSP00000370542 0.109 0
   0.369 0.413
   MET PDCD1 9606.ENSP00000317272 9606.ENSP00000335062 0 0 0.432 0.432
   [140]Open in a new tab

3.3. Population-based study

3.3.1. Clinical characteristics and pathological features

   A total of 100 RAH patients were included in our study despite
   incomplete information on some patients, and the clinical
   characteristics, treatment, pathological features, and outcomes of the
   patients are presented in [141]Table 2 . With a male-to-female ratio of
   approximately 2:1, most patients were males, with ages ranging from
   0.75 to 94 years. The presence of a renal mass was incidentally
   detected in most patients on examination for other diseases, while a
   minority of patients presented with clinical symptoms, including
   abdominal/flank pain (n=11), hematuria (n=7) and abdominal bulge (n=1).
   Among the 100 RAH patients, 46 had comorbid ESRD when RAH was detected;
   the cause of ESRD in 20 patients was not reported in the literature,
   and SLE was the most frequent cause of ESRD (n=9). Only 11 patients
   exhibited bilateral kidney involvement; 40 patients had unilateral
   kidney involvement in the left kidney, and 31 patients had unilateral
   kidney involvement in the right kidney. Among the 91 patients for whom
   treatment data were available, 71 patients underwent radical
   nephrectomy, 20 patients underwent partial nephrectomy, and 24 patients
   had cooccurring or a history of other renal tumors, including renal
   cell carcinoma (RCC), papillary adenomas, papillary RCC,
   angiomyolipoma, metanephric adenoma, acquired cystic kidney disease
   (ACKD), ACKD-associated RCC and Wilms’ tumor; the largest diameter
   of the tumors ranged from 1 mm-140 mm. Microscopic examination revealed
   typical capillary-sized vessels with a spleen-like sinusoidal pattern
   of AH, an absence of diffuse infiltrative growth and no malignant
   cytological features of angiosarcoma. Immunohistochemical analysis
   revealed that CD31, CD34 and factor VIII were positive, with a low
   Ki-67 labeling index. Among the patients with available follow-up data,
   no patients experienced disease progression or death.

Table 2.

   Clinicopathological features of renal Anastomosing Hemangiomas in the
   population-based study.
   Reference Age Gender Clinical Manifestation ESRD Cause of ESRD Diabetes
   Hypertension Site Location Focality Largest Tumor Diameter (mm)
   Preoperative diagnosis Therapy Cooccurrence/History of Renal tumor
   Immunohistochemistry Follow-up
   CD31 CD34 factor VIII Ki-67
   1 Abboudi H 2017 62 F Hematuria Yes AAV None/NA None/NA Bilateral Renal
   sinus/hilus/medulla Multifocal 27 NA Metachronous Nephrectomy None NA
   NA NA NA NA
   2 Al-Maghrabi HA 2014 55 F Left flank pain None/NA – Yes Yes Left Renal
   sinus/hilus/medulla Unifocal 20 RCC Partial nephrectomy A history of
   right partial nephrectomy three years earlier due to clear cell RCC
   (Fuhrman nuclear grade II) + + + NA NED (12 months)
   3 Aravind A 2025 28 M Abdominal pain and dyspepsia None/NA – None/NA
   None/NA Left Renal sinus/hilus/medulla and cortex Unifocal 40 Low
   aggressive potential lesion Partial nephrectomy None + + + 0.12-0.15
   NED
   4 Bean GR 2017 (1) 49 M NA None/NA – None/NA None/NA NA NA NA 12 NA NA
   NA NA NA NA NA NED (9 months)
   5 Bean GR 2017 (2) 53 M Incidentally None/NA – None/NA None/NA NA NA NA
   33 NA NA NA NA NA NA NA NED (84 months)
   6 Bean GR 2017 (3) 64 M Incidentally None/NA – None/NA None/NA NA NA NA
   7 NA NA NA NA NA NA NA NED (107 months)
   7 Berker NK 2017 (1) 24 F Incidentally Yes Membranoproliferative
   glomerulonephritis None/NA None/NA Bilateral Renal sinus/hilus/medulla
   Multifocal 30 NA Partial nephrectomy Papillary adenomas + + + NA NED
   (10 months)
   8 Berker NK 2017 (2) 57 F Incidentally Yes Diabetes Yes None/NA Left
   Renal sinus/hilus/medulla Unifocal 22 NA Radical nephrectomy NA NA NA
   NA NA NED (4 months)
   9 Brown JG 2010 (1) 56 M NA None/NA – None/NA None/NA Right NA NA 13 NA
   Partial nephrectomy None NA NA NA NA NA
   10 Brown JG 2010 (2) 33 F NA None/NA – None/NA None/NA Left NA NA 32 NA
   Radical nephrectomy None NA NA NA NA NA
   11 Brown JG 2010 (3) 21 M Incidently Yes None/NA None/NA None/NA Right
   NA NA 22 NA NA None NA NA NA NA NED (24 months)
   12 Brown JG 2010 (4) 44 F NA None/NA – None/NA None/NA Left NA NA 20 NA
   NA None NA NA NA NA NED (72 months)
   13 Brown JG 2010 (5) 83 F NA None/NA – None/NA None/NA Left NA NA 35 NA
   NA None NA NA NA NA NED (24 months)
   14 Buttner M 2013 (1) 33 M NA Yes NA None/NA None/NA NA Renal
   sinus/hilus/medulla and cortex Multifocal 7 NA Radical nephrectomy None
   NA NA NA NA NA
   15 Buttner M 2013 (2) 70 M NA Yes NA None/NA None/NA NA Renal
   sinus/hilus/medulla Unifocal 3 NA Radical nephrectomy Papillary adenoma
   NA NA NA NA NA
   16 Buttner M 2013 (3) 55 M NA Yes NA None/NA None/NA NA Renal cortex
   Unifocal 2 NA Radical nephrectomy None NA NA NA NA NA
   17 Buttner M 2013 (4) 42 M NA Yes NA None/NA None/NA NA Renal
   sinus/hilus/medulla and cortex Multifocal 1.5 NA Radical nephrectomy
   Papillary adenoma, cRCC NA NA NA NA NA
   18 Buttner M 2013 (5) 32 M NA Yes NA None/NA None/NA NA Renal cortex
   Multifocal 6 NA Radical nephrectomy Papillary RCC, angiomyolipoma NA NA
   NA NA NA
   19 Buttner M 2013 (6) 42 M NA Yes NA None/NA None/NA NA Renal
   sinus/hilus/medulla Unifocal 3 NA Radical nephrectomy None NA NA NA NA
   NA
   20 Buttner M 2013 (7) 45 F NA Yes NA None/NA None/NA NA Renal
   sinus/hilus/medulla and cortex Multifocal 25 NA Radical nephrectomy
   None NA NA NA NA NA
   21 Buttner M 2013 (8) 44 F NA Yes NA None/NA None/NA NA Renal
   sinus/hilus/medulla Unifocal 1 NA Radical nephrectomy Papillary RCC,
   metanephric adenoma NA NA NA NA NA
   22 Caballes AB 2019 10 M A firm and painless bulge at the left
   paraumbilical area None/NA – None/NA None/NA Left Renal
   sinus/hilus/medulla Unifocal 120 Wilms’ tumor Radical nephrectomy None
   NA + NA 0 NA
   23 Capinha MD 2023 70 M Incidentally None/NA – None/NA Yes Left Renal
   sinus/hilus/medulla Unifocal 20 RCC Radical nephrectomy None + NA NA NA
   NED (18 months)
   24 Cha JS 2016 43 M Incidentally None/NA – Yes None/NA Right Renal
   sinus/hilus/medulla Unifocal 43 RCC Radical nephrectomy None + + + NA
   NED (5 months)
   25 Chandran N 2019 36 M Incidentally Yes NA None/NA None/NA Bilateral
   Renal sinus/hilus/medulla Multifocal 26 RCC Radical nephrectomy None NA
   + NA NA NA
   26 Chen J 2024 59 M Incidentally None/NA – None/NA None/NA Left Renal
   sinus/hilus/medulla Unifocal 22 RCC Radical nephrectomy RCC + + NA
   0.02-0.05 NED
   27 Cheon PM 2018 40 M Incidentally None/NA – None/NA None/NA Left Renal
   sinus/hilus/medulla Unifocal 53 RCC Radical nephrectomy None + + + 0.2
   NED (1 months)
   28 Chou S 2014 (1) 50 F Incidentally Yes – None/NA None/NA Left Renal
   sinus/hilus/medulla Unifocal 10 NA Radical nephrectomy NA + + NA NA NA
   29 Chou S 2014 (2) 60 M Incidentally Yes – None/NA Yes Left Renal
   sinus/hilus/medulla Multifocal 18 NA Radical nephrectomy NA + + NA NA
   NA
   30 Chua WM 2022 32 F Incidentally Yes SLE None/NA None/NA Left Renal
   sinus/hilus/medulla Multifocal NA Neuroendocrine tumor NA None + NA NA
   NA NA
   31 Downes MR 2014 (1) 59 F Incidentally None/NA – None/NA None/NA Left
   Renal sinus/hilus/medulla Unifocal 45 RCC Radical nephrectomy NA + + +
   NA NA
   32 Downes MR 2014 (2) 28 M Incidentally None/NA – None/NA None/NA Right
   Renal sinus/hilus/medulla Unifocal 13 NA Ultrasound guided kidney core
   biopsy History of left RCC + + NA NA NA
   33 Faraz M 2025 (1) 52 M Incidentally None/NA – None/NA None/NA Left NA
   Unifocal 28 NA Partial nephrectomy None + NA + NA DWOD (81months)
   34 Faraz M 2025 (2) 71 F Incidentally None/NA – None/NA None/NA Left NA
   Unifocal 28 NA Radical nephrectomy None + NA NA NA NED (83 months)
   35 Faraz M 2025 (3) 69 M Incidentally None/NA – None/NA None/NA Left NA
   Unifocal NA Partial nephrectomy None NA NA NA NA NED (35 months)
   36 Faraz M 2025 (4) 94 M Incidentally None/NA – None/NA None/NA Right
   NA Unifocal NA Biopsy None NA NA NA NA NED (31 months)
   37 Faraz M 2025 (5) 47 F Incidentally None/NA – None/NA None/NA Left NA
   Unifocal NA Radical nephrectomy None NA + NA NA NED (161 months)
   38 Faraz M 2025 (6) 76 M Incidentally None/NA – None/NA None/NA NA NA
   Unifocal NA Partial nephrectomy None NA NA NA NA NA
   39 Faraz M 2025 (7) 45 M Incidentally None/NA – None/NA None/NA Left NA
   Unifocal NA Radical nephrectomy None + + NA NA NED (156 months)
   40 Faraz M 2025 (8) 59 M Incidentally None/NA – None/NA None/NA Left NA
   Unifocal NA Partial nephrectomy None NA NA NA NA NED (108 months)
   41 Faraz M 2025 (9) 56 F Incidentally None/NA – None/NA None/NA Left NA
   Multifocal NA Radical nephrectomy AMLEC, AML, cortical cysts NA NA NA
   NA NED (24 months)
   42 Faraz M 2025 (10) 67 F Incidentally None/NA – None/NA None/NA Right
   NA Unifocal NA Partial nephrectomy ccRCC + NA NA NA NED (9 months)
   43 Faraz M 2025 (11) 41 M Incidentally None/NA – None/NA None/NA Right
   NA Unifocal NA Partial nephrectomy None + NA NA NA NED (81 months)
   44 Faraz M 2025 (12) 62 M Incidentally Yes None/NA None/NA None/NA Left
   NA Multifocal NA Radical nephrectomy None + NA NA NA NED (46 months)
   45 Faraz M 2025 (13) 50 F Incidentally Yes None/NA None/NA None/NA Left
   NA Multifocal NA Radical nephrectomy None + NA NA NA NA
   46 Faraz M 2025 (14) 80 M Incidentally None/NA – None/NA None/NA Right
   NA Unifocal NA Partial nephrectomy None + NA NA NA NED (24 months)
   47 Faraz M 2025 (15) 70 M Incidentally None/NA – None/NA None/NA Right
   NA Unifocal NA Partial nephrectomy None + + NA NA NED (4 months)
   48 Faraz M 2025 (16) 33 F Left flank painand gross hematuria None/NA –
   None/NA None/NA Right NA Unifocal NA Radical nephrectomy None + + NA NA
   NA
   49 Faraz M 2025 (17) 34 M Incidentally Yes – None/NA None/NA Bilateral
   NA Multifocal NA Partial nephrectomy AML, transplant kidneys NA NA NA
   NA NED (5 months)
   50 Faraz M 2025 (18) 50 M Incidentally None/NA – None/NA None/NA Left
   NA Unifocal NA Partial nephrectomy None + NA NA NA NED (72 months)
   51 Gong C 2024 58 F Pain in the area of right kidney None/NA – None/NA
   None/NA Left Renal sinus/hilus/medulla Unifocal 25 NA Radical
   nephrectomy None NA NA NA NA NA
   52 Heidegger I 2014 56 M Incidentally None/NA – None/NA None/NA Right
   Renal sinus/hilus/medulla Unifocal 50 RCC Radical nephrectomy None + +
   NA <1% NED (128 month)
   53 Johnstone KJ 2020 70 M Incidentally None/NA – None/NA None/NA Right
   Renal sinus/hilus/medulla and cortex Multifocal 35 NA Radical
   nephrectomy None + + + Low NA
   54 Kim CS 2021 35 F Incidentally Yes SLE None/NA Yes Right Renal
   sinus/hilus/medulla Unifocal 17 RCC Radical nephrectomy None + + NA NA
   NA
   55 Kryvenko ON 2011 (1) 51 F Incidentally Yes None/NA None/NA None/NA
   Right Renal sinus/hilus/medulla Unifocal 10 NA Radical nephrectomy NA
   NA + NA NA NED (7 months)
   56 Kryvenko ON 2011 (2) 39 M Incidentally None/NA None/NA None/NA Yes
   Right Renal sinus/hilus/medulla and cortex Unifocal 50 NA Radical
   nephrectomy NA NA + NA NA NED (122 months)
   57 Kryvenko ON 2011 (3) 54 F Incidentally Yes None/NA None/NA None/NA
   Bilateral Renal sinus/hilus/medulla and cortex Multifocal 12 NA Radical
   nephrectomy NA NA + NA NA NED (3 months)
   58 Kryvenko ON 2014 (1) 68 F Gross haematuria Yes Diabetes Yes None/NA
   Right NA Multifocal 15 NA Radical nephrectomy None NA NA NA NA NA
   59 Kryvenko ON 2014 (2) 51 F Incidentally Yes NA None/NA None/NA Right
   NA Unifocal 10 NA Radical nephrectomy ACKD NA NA NA NA NA
   60 Kryvenko ON 2014 (3) 54 F Incidentally Yes Hypertension None/NA Yes
   Bilateral NA Multifocal 11 NA Radical nephrectomy ccRCC, papillary
   adenoma NA NA NA NA NA
   61 Kryvenko ON 2014 (4) 29 M Incidentally Yes Focal segmental
   glomerulosclerosis None/NA None/NA Bilateral NA Multifocal 13 NA
   Radical nephrectomy None NA NA NA NA NA
   62 Kryvenko ON 2014 (5) 40 M Incidentally Yes SLE None/NA None/NA Left
   NA Unifocal 2.5 NA Radical nephrectomy ACKD, Clear cell papillary RCC,
   ACKD- associated RCC, papillary adenoma NA NA NA NA NA
   63 Kryvenko ON 2014 (6) 34 M Abdominal pain, haematuria,
   retroperitoneal haematoma Yes SLE None/NA None/NA Right NA Multifocal
   13 NA Radical nephrectomy ACKD, ACKD- associated RCC NA NA NA NA NA
   64 Kryvenko ON 2014 (7) 62 M Incidentally Yes Hypertension None/NA Yes
   Left NA Unifocal 7 NA Radical nephrectomy ACKD, ACKD-associated RCC,
   ACKD-associated RCC precursor NA NA NA NA NA
   65 Kryvenko ON 2014 (8) 40 M Incidentally Yes SLE None/NA None/NA Left
   NA Multifocal 28 NA Radical nephrectomy None NA NA NA NA NA
   66 Kryvenko ON 2014 (9) 46 M Incidentally Yes IgA nephropathy None/NA
   None/NA Left NA Unifocal 16 NA Radical nephrectomy None NA NA NA NA NA
   67 Kryvenko ON 2014 (10) 60 M Incidentally Yes Hereditary nephritis
   None/NA None/NA Left NA Unifocal 12 NA Radical nephrectomy ACKD,
   ACKD-associated RCC precursor, papillary adenoma NA NA NA NA NA
   68 Kryvenko ON 2014 (11) 49 M Microscopic haematuria Yes Focal
   segmental glomerulosclerosis None/NA None/NA Right NA Unifocal 35 NA
   Radical nephrectomy None NA NA NA NA NA
   69 Kryvenko ON 2014 (12) 49 M NA Yes Hypertension None/NA Yes Left NA
   Unifocal 13 NA Radical nephrectomy ACKD NA NA NA NA NA
   70 Kryvenko ON 2014 (13) 66 M Flank pain Yes Hypertension None/NA Yes
   Right NA Unifocal 30 NA Radical nephrectomy ACKD, ACKD-associated RCC
   precursor, papillary adenoma NA NA NA NA NA
   71 Kryvenko ON 2014 (14) 15 M Incidentally Yes Focal segmental
   glomerulosclerosis None/NA None/NA Bilateral NA Multifocal 7 NA Radical
   nephrectomy ACKD NA NA NA NA NA
   72 Kryvenko ON 2014 (15) 0.75 M Incidentally Yes Congenital
   nephroticsyndrome None/NA None/NA Right NA Unifocal 10 NA Radical
   nephrectomy Wilms’ tumour NA NA NA NA NA
   73 Kryvenko ON 2014 (16) 17 M Abdominal pain, haematuria,
   retroperitoneal haematoma Yes FSGS secondary tominimal changedisease
   None/NA None/NA Bilateral NA Multifocal 28 NA Radical nephrectomy ACKD,
   Unclassified RCC, Papillary adenomas NA NA NA NA NA
   74 Lo CH 2021 84 M Incidentally None/NA None/NA None/NA Yes Left Renal
   sinus/hilus/medulla Unifocal 55 Cystic RCC Radical nephrectomy None +
   NA NA <10% Dead of aspiration pneumonia (2 weeks)
   75 Lobo J 2017 63 M Hematuria and dysuria None/NA None/NA None/NA
   None/NA Right Renal sinus/hilus/medulla Unifocal 50 Urothelial neoplasm
   Radical nephrectomy None NA NA NA <1% NA
   76 Manohar V 2020 40 F Vague upper abdominal pain None/NA None/NA
   None/NA None/NA Left Renal sinus/hilus/medulla Unifocal 140 RCC Radical
   nephrectomy None + + NA NA NED (24 months)
   77 Mehta V 2012 (1) 49 M NA Yes None/NA None/NA None/NA NA NA NA 20 NA
   Radical nephrectomy None NA NA NA NA NED (3 months)
   78 Mehta V 2012 (2) 55 M NA Yes None/NA None/NA None/NA NA NA NA 6 NA
   Radical nephrectomy Papillary adenomas NA NA NA NA NED (3 months)
   79 Mehta V 2012 (3) 45 M NA Yes None/NA None/NA None/NA NA NA NA 19 NA
   Radical nephrectomy None NA NA NA NA NED (12 months)
   80 Montgomery E 2009 (1) 74 M NA NA None/NA None/NA None/NA NA NA NA 15
   NA Radical nephrectomy NA NA NA NA NA NED (36 months)
   81 Montgomery E 2009 (2) 75 F NA NA None/NA None/NA None/NA NA NA NA 20
   NA Radical nephrectomy NA NA NA NA NA NA
   82 Montgomery E 2009 (3) 49 M NA NA None/NA None/NA None/NA NA Renal
   sinus/hilus/medulla NA 13 NA Radical nephrectomy NA NA NA NA NA NED (12
   months)
   83 Omiyale AO 2015 64 M Incidentally None/NA None/NA Yes None/NA Left
   Renal sinus/hilus/medulla Unifocal 24 RCC Radical nephrectomy None + +
   NA NA NED (10 months)
   84 Pantelides NM 2012 57 F Incidentally Yes AAV None/NA None/NA Right
   Renal sinus/hilus/medulla Unifocal 27 NA Radical nephrectomy None NA NA
   NA NA NA
   85 Patel SR 2019 39 M Incidentally Yes SLE None/NA None/NA Bilateral
   Renal sinus/hilus/medulla Multifocal 15 AH Radical nephrectomy None NA
   NA NA NA NA
   86 Perdiki M 2017 (1) 64 F Back pain None/NA None/NA None/NA None/NA
   Right Renal sinus/hilus/medulla and cortex Unifocal 11 NA Partial
   nephrectomy NA + + NA NA NED (25 months)
   87 Perdiki M 2017 (2) 47 M Incidentally Yes SLE None/NA None/NA Left
   Renal sinus/hilus/medulla and cortex Multifocal 28 NA Radical
   nephrectomy NA + + NA NA NED (14 months)
   88 Rubio Fernández A 2015 41 M Incidentally Yes SLE None/NA None/NA
   Bilateral Renal sinus/hilus/medulla Multifocal 80 Malignant renal tumor
   Radical nephrectomy None + + NA NA NA
   89 Rupanshu 2023 35 F Incidentally Yes SLE None/NA None/NA Right NA
   Unifocal 17 RCC Radical nephrectomy None + + NA NA NED (NA)
   90 Sasaki Y 2022 65 M Incidentally None/NA None/NA None/NA None/NA
   Right Renal cortex Unifocal 16 AH Partial nephrectomy None + + + 0 NED
   (3 months)
   91 Silva MA 2017 53 M Incidentally None/NA None/NA None/NA None/NA Left
   Renal cortex Unifocal NA Complex renal cystic lesion Partial
   nephrectomy None NA NA NA NA NA
   92 Tahir M 2016 57 M Incidentally None/NA None/NA None/NA None/NA Left
   Renal sinus/hilus/medulla Unifocal 30 NA Radical nephrectomy None + + +
   NA NA
   93 Tao LL 2014 32 M Incidentally None/NA None/NA None/NA None/NA Left
   Renal sinus/hilus/medulla Unifocal 26 NA Radical nephrectomy None + +
   NA 0 NA
   94 Tran TA 2012 61 M Right-sided back pain with radiation to his right
   hip None/NA None/NA None/NA None/NA Right Renal sinus/hilus/medulla and
   cortex Unifocal 24 NA Radical nephrectomy None + NA + NA NA
   95 Veerwal A 2020 25 M Right fiank pain None/NA None/NA None/NA None/NA
   Right NA Unifocal NA RCC Radical nephrectomy None NA NA NA NA NA
   96 Wetherell DR 2013 74 M Incidentally None/NA None/NA None/NA None/NA
   Right Renal sinus/hilus/medulla Unifocal 50 RCC Radical nephrectomy
   None + + + NA Dead (6 weeks)
   97 Zhang W 2015 25 F Incidentally None/NA None/NA None/NA None/NA Right
   Renal sinus/hilus/medulla Unifocal 12 NA Partial nephrectomy None + +
   NA NA NED (16 months)
   98 Zhang X 2023 61 M Incidentally None/NA None/NA None/NA None/NA Left
   Renal sinus/hilus/medulla and cortex Unifocal 18 NA Radical nephrectomy
   None + + NA 0.1 NED (13 months)
   99 Zhao M 2013 48 M Incidentally None/NA None/NA None/NA None/NA Right
   Renal cortex Unifocal 25 RCC Partial nephrectomy None + + NA NA NED (12
   months)
   100 Our study 71 F Incidentally None/NA None/NA Yes Yes Left Renal
   sinus/hilus/medulla Unifocal 24 NA Partial nephrectomy NA + + NA 0.1
   NED (16 months)
   [142]Open in a new tab

   ESRD, end stage renal disease; M, male; F, female; RCC, renal cell
   carcinoma; NED, no evidence of disease; DWOD, died without disease; AH,
   anastomosing hemangioma; ACKD, acquired cystic kidney disease; AAV,
   ANCA-associated vasculitis; SLE, systemic lupus erythematosus; NA, not
   available.

3.3.2. Comparison of RAHs with ESRD and RAHs without ESRD

   In the comparison of RAHs with ESRD and RAHs without ESRD, as shown in
   [143]Table 3 and [144]Figures 14 , [145]15 , age, tumor site, tumor
   focality, largest tumor diameter and surgical approach were
   significantly different (p<0.05). Patients with ESRD and RAHs without
   RSRD were diagnosed at a younger age than those with RAHs without RSRD.
   Moreover, smaller bilateral and multifocal tumors were more prevalent
   among RAHs with ESRD.

Table 3.

   Comparison of features of renal Anastomosing Hemangiomas between
   patients with end-stage renal disease (ESRD) and patients without ESRD.
   All patients (n=100) With ESRD (n=46) Without ESRD (n=54) p
   Age (Mean ± SD) 50.32 ± 16.96 44.34 ± 14.73 55.41 ± 17.19 0.001
   Gender (n, %)
    Female 32 (32.0%) 16 (34.8%) 16 (29.6%) 0.582
    Male 68 (68.0%) 30 (65.2%) 38 (70.5%)
   Clinical Manifestation (n, %)
    Absent 83 (83.0%) 40 (87.0%) 43 (79.6%) 0.363
    Present 17 (17.0%) 6 (13.0%) 11 (20.4%)
   Site (n, %)
    Left 40 (40.0%) 13 (28.3%) 27 (50.0%) <0.001
    Right 31 (31.0%) 11 (23.9%) 20 (37.0%)
    Bilateral 11 (11.0%) 11 (23.9%) 0 (0.0%)
    Unknown 18 (18.0%) 11 (23.9%) 7 (13.0%)
   Location (n, %)
    Renal cortex 5 (5.0%) 2 (4.3%) 3 (5.6%) 0.943
    Renal sinus/hilus/medulla 35 (35.0%) 15 (32.6%) 20 (37.0%)
    Renal sinus/hilus/medulla and cortex 11 (11.0%) 5 (10.9%) 6 (11.1%)
    Unknown 49 (49.0%) 24 (52.2%) 25 (46.3%)
   Focality (n, %)
    Unifocal 61 (61.0%) 19 (41.3%) 42(77.8%) <0.001
    Multifocal 25 (25.0%) 23 (50.0%) 2 (3.7%)
    Unknown 14 (14.0%) 4 (8.7%) 10 (18.5%)
   Largest Tumor Diameter (mm) (n=81) 24.72 ± 22.31 16.79 ± 13.62 33.26 ±
   26.51 0.001
   Preoperative misdiagnosis (n=21) 21 (100.0%) 5 (23.8%) 16 (76.2%)
   Surgery approach (n, %) (n=91)
    Partial nephrectomy 20 (22.0%) 2 (4.5%) 18 (38.3%) <0.001
    Radical nephrectomy 71 (78.0%) 42 (95.5%) 29 (61.7%)
   Cooccurrence/History of other renal tumors (n=24) 24 (100.0%) 18
   (75.0%) 6 (25.0%)
   [146]Open in a new tab

   Significant values are in bold.

Figure 14.

   [147]Figure 14
   [148]Open in a new tab

   (A) The boxplots of age in the patients with end-stage renal disease
   (ESRD) and patients without ESRD. (B) The column charts of patients
   with ESRD and patients without ESRD in the female and male patients.
   (C) The column charts of patients with ESRD and patients without ESRD
   in the patients absent clinical manifestation and patients present
   clinical manifestation. (D) The column charts of patients with ESRD and
   patients without ESRD in the female with tumor in the left kidney,
   right kidney and bilateral kidney.

Figure 15.

   [149]Figure 15
   [150]Open in a new tab

   (A) The column charts of patients with ESRD and patients without ESRD
   in the female with tumor arising from renal cortex, renal
   sinus/hilus/medulla and renal sinus/hilus/medulla and cortex. (B) The
   column charts of patients with ESRD and patients without ESRD in
   patients with unifocal tumor and multifocal tumors. (C) The boxplots of
   largest tumor diameter in the patients with end-stage renal disease
   (ESRD) and patients without ESRD. (D) The pie chart of patients with
   ESRD and patients without ESRD in the patients with preoperative
   misdiagnosis. (E) The column charts of patients with ESRD and patients
   without ESRD in patients underwent partial nephrectomy and radical
   nephrectomy. (F) The pie chart of patients with ESRD and patients
   without ESRD in the patients with cooccurrence/History of other renal
   tumors.

3.3.3. Comparison of RAHs and Renal Hemangiosarcoma

   In the comparison of clinical and tumor features between 100 RAH
   patients and 47 renal hemangiosarcoma patients, as shown in
   [151]Table 4 , RAH patients were diagnosed earlier than renal
   hemangiosarcoma patients were (p=0.001). The RAH tended to be bilateral
   at presentation (p<0.001) and to be associated with a smaller tumor
   size (p=0.001). A total of 47 RAH patients and 54 renal hemangiosarcoma
   patients with complete follow-up data were included in the survival
   analysis. As shown in [152]Figure 16 , the RAH patients exhibited
   significantly better OS and CSS than the renal hemangiosarcoma patients
   did (p<0.001).

Table 4.

   Comparison of clinical and tumor features between renal Anastomosing
   Hemangioma and Hemangiosarcoma.
   Anastomosing Hemangioma (n=100) Hemangiosarcoma (n=53) p
   Age (Mean ± SD) 50.32 ± 16.96 63.00 ± 17.72 0.001
   Gender (n, %)
    Female 32 (32.0%) 12 (22.6%) 0.224
    Male 68 (68.0%) 41 (77.4%)
   Site (n, %)
    Left 40 (40.0%) 32 (60.4%) <0.001
    Right 31 (31.0%) 20 (37.7%)
    Bilateral 11 (11.0%) 0 (0.0%)
    Unknown 18 (18.0%) 1 (1.9%)
   Tumor size (mm, Mean ± SD) 24.72 ± 22.31 116.08 ± 53.12 0.001
   [153]Open in a new tab

   Significant values are in bold.

Figure 16.

   [154]Figure 16
   [155]Open in a new tab

   Survival analysis of overall survival (A) and cancer-specific survival
   (B) between renal anastomosing hemangioma and renal hemangiosarcoma
   patients.

4. Discussion

   RAH is a rare benign subtype of benign hemangioma whose pathogenesis
   has not yet been fully elucidated. In our study, we present the case of
   a female patient with left RAH who underwent robotic-assisted
   laparoscopic partial left nephrectomy and remained in good condition
   with no signs of recurrence or metastasis after 27 months of regular
   follow-up. To clarify the genetic alterations and preliminarily explore
   the potential mechanism of the pathogenesis of RAH, we first applied
   WES to RAH and conducted a population-based study to comprehensively
   understand the clinicopathological characteristics of RAH.

   There are two main accepted hypotheses for the pathogenesis of RAH. One
   is the ESRD-related mechanism, supported by RAH arising in ESRD
   patients as reported in publications from different institutions.
   Moreover, the greater percentage of bilateral (p=0.021) and multifocal
   (p<0.001) tumors in ESRD patients in our population-based study also
   implies a potential association between ESRD and RAH. ESRD is a fertile
   ground for malignant renal epithelial tumor development through the
   following proposed factors: the cellular and humoral immune deficiency
   caused by renal failure results in a decrease in antioxidant defenses,
   subsequent accumulation of reactive oxygen species, enhanced chronic
   infections and inflammation in association with phagocytic activity and
   increased free radical release, resulting in DNA damage, mutations and
   cancer progression. The accumulation of carcinogenic compounds,
   including carcinogenic heterocyclic amines, in ESRD and
   immunosuppressive medications can also impair methylation-dependent
   repair, which can inhibit the repair of DNA mutations and promote
   malignant cell transformation ([156]27, [157]28). While some
   researchers hold different views, Kryvenko ON et al. suggested that the
   observed association between ESRD and AH was due to the bias that the
   surveillance imaging of ESRD patients promoted the detection of RAH
   ([158]12). However, while this incidental association between ESRD and
   RAH has been proposed as an alternative hypothesis, the exact
   underlying mechanisms remain unknown and still need in-depth
   exploration.

   A potential driving mechanism is that constitutive GTPase activity and
   the mitogen-activated protein (MAP) kinase signaling pathway are
   regulated by the recurrent mutant G protein alpha subunit GNAQ and its
   paralogs GNA11 and GNA14, which are also expressed in other
   hemangiomas. GNAQ and GNA11 also participate in vascular endothelial
   growth factor (VEGF) signaling and angiogenesis ([159]13, [160]29,
   [161]30). Our study also detected mutant GNAQ mutations via WGS while
   didn’t confirm them via Sanger sequencing.

   We identified mutations in the predisposing genes CNTNAP2, NCOA2, FAT1,
   MET, TJP2, MAML2, SRGAP3, and CSMD3, driver gene HIP1 and validated
   them by Sanger sequencing. We also demonstrated that FAT, MET and MAML2
   mutations may be involved in the progression of RAH. FAT1 (FAT atypical
   cadherin 1), which is a member of the FAT cadherin family in
   vertebrates, is a highly mutated gene in human cancers that encodes a
   protocadherin ([162]31, [163]32). FAT1 participated in multiple
   signaling pathways, including the Wnt/β-catenin signaling pathway, the
   Hippo signaling pathway, and the MAPK/ERK signaling pathway, and it
   mediates epithelial–mesenchymal transition to promote cell
   proliferation, migration and invasion ([164]33–[165]37). Owing to the
   importance of FAT1 in organisms, mutations in FAT1 may cause diverse
   malignant biological behaviors and have been detected in multiple
   diseases, such as acute lymphoblastic leukemia ([166]38),
   hepatocellular carcinoma ([167]39), pituitary spindle cell tumors
   ([168]40), and head and neck squamous cell carcinoma ([169]41,
   [170]42). Moreover, Zhou L et al. reported a mutation in FAT1 in an
   acquired cystic disease-associated renal cell carcinoma that may have
   participated in its pathogenesis ([171]43). The FAT1 mutation we
   confirmed suggested a potential role for this mutation in the
   progression of RAH. MET is a proto-oncogene encoding the hepatocyte
   growth factor (HGF) tyrosine kinase receptor and it participates in
   regulating embryogenesis, angiogenesis, wound healing and liver
   regeneration ([172]44). Alterations in MET can drive tumorigenesis
   through various molecular mechanisms in several types of cancer.
   Mutations in the tyrosine kinase domain (TKD) can lead to receptor
   phosphorylation and signaling unrelated to ligands. Mutations in this
   domain have been confirmed in type 1 papillary renal cell carcinomas
   and even cause resistance to MET tyrosine kinase inhibitors in lung
   cancers and neuroblastoma ([173]44–[174]47). Fang Y et al. reported
   that HGF/Met pathway deficiency caused by MET mutation impedes late
   thyroid expansion and subsequently leads to thyroid dysgenesis, and the
   MET mutation identified in the RAH in this study may be involved in the
   pathogenesis of innate developmental disorders ([175]48).
   Mastermind-like transcriptional coactivator 2 (MAML2) is a member of
   the Mastermind-like family and is a transcriptional regulator of Notch
   signaling ([176]49). The rearrangement of MAML2 has been confirmed to
   be associated with a variety of diseases, such as atypical
   intraparenchymal meningioma ([177]50) and metaplastic thymoma ([178]51,
   [179]52). Linos K and Dermawan JK et al. also detected MAML2
   rearrangement in composite hemangioendotheliomas, which supports the
   function of MAML2 in tumors originating from the vasculature ([180]53,
   [181]54). The confirmed mutated genes and subsequent explorations may
   also contribute to the diagnosis of RAH and the development of targeted
   therapies in the future.

   Another striking finding was that mutated genes may influence disease
   development by regulating the infiltration of immune cells, including
   NK/T cells, mast cells, M1 macrophages, M2 macrophages and plasmacytes,
   in the tumor microenvironment through the PPI analysis, the higher
   distribution of the tumor-associated macrophages (CD163) also validate
   the role in the disease. The regulation of macrophages in hemangioma
   has been thoroughly described in previous studies ([182]55). M2 and M1
   macrophages play almost opposite roles in infantile hemangioma, with M2
   predominantly present in large numbers in the early stages of IH
   proliferation, promoting microvessel formation and lymphatic growth,
   whereas M1 macrophages exert antiproliferative and antitumor effects in
   the later stages of proliferation. Furthermore, the regulatory
   mechanisms of angiogenesis by which M1 macrophages contribute to
   regulating the proliferation and differentiation of hemangioma-derived
   stem cells and M2 macrophages facilitate the endothelial
   differentiation of hemangioma-derived stem cells, which also
   demonstrates the crucial role of immune cells in the molecular
   mechanisms and pathological processes of hemangioma ([183]56, [184]57).
   Moreover, the identified combination of CSMD3 and MET with PD-1 and two
   reported cases of hepatic cavernous hemangioma after treatment with
   PD-1 inhibitors suggest that PD-1 may regulate angiogenesis by
   modulating certain mechanisms ([185]58).

   However, because the clinical and radiographic features between RAH,
   RCC, and AS are highly similar and accurate preoperative diagnosis of
   RAH currently remain challenging, only a small minority of patients who
   undergo preoperative puncture biopsy can receive a definitive diagnosis
   on the basis of histology and immunochemistry. The routine radiographic
   examinations for renal masses include ultrasonography, CT and magnetic
   resonance imaging (MRI), but these methods still cannot distinguish RAH
   from RCC or angiosarcoma. These entities all present echogenicity on
   ultrasonography and a solid, boundary-cleared lesion with heterogeneous
   avid enhancement on contrast-enhanced CT ([186]59). Nevertheless, RAH
   has some specific features on CT and MRI. On dynamic CT and MRI, RAH
   usually shows contrast enhancement from the periphery toward the
   center, with marked T2 hyperintensity similar to that of a cyst and
   persistent enhancement in the venous phase ([187]7, [188]60). Although
   the RAH patient presented by Chua WM et al. underwent CT and
   ^68Ga-DOTATATE PET/CT examinations, it was still misdiagnosed
   preoperatively as a neuroendocrine tumor ([189]61). The differentiated
   diagnosis of RAH from AH in angiosarcoma patients with extensive
   hemorrhage described by Heo SH et al. was also confirmed by
   immunohistochemistry of operative samples rather than preoperative CT
   and MRI ([190]62). Although the imaging findings are nonspecific and
   cannot further confirm the diagnosis of RAH, the above manifestations
   on imaging examinations, including CT, MRI and PET-CT, can also provide
   clues for radiologists and urologists to perform percutaneous renal
   biopsy to confirm the diagnosis to avoid nephrectomy. Notably, when
   invasive procedures, including percutaneous renal biopsy and
   ureteroscopy, are considered in patients with a suspected diagnosis of
   RAH, caution should be taken to prevent hemorrhage and spontaneous
   rupture of the RAH. The precise preoperative diagnosis of the RAH and
   its ability to differentiate it from other malignant renal tumors, such
   as renal cell carcinoma and angiosarcoma, could help patients choose
   more suitable treatment approaches, such as selective embolization and
   partial nephrectomy, rather than radical nephrectomy, and deep learning
   of radiographic images may serve useful purposes.

   There are also several limitations of our study. Given the rare
   occurrence of RAH, more RAH samples are needed to verify the
   reliability and reproducibility of genetic alterations screened by WGS,
   and cell and animal experiments are also needed to validate the role of
   genetic alterations in RAH pathogenesis. Moreover, WGS of data from RAH
   patients with ESRD may contribute to further understanding of the
   mechanisms associated with ESRD. In our population-based study, we
   excluded some patients with RAH who lacked detailed information, which
   may have led to bias and inaccuracies. Moreover, owing to the limited
   number of RAH cases, the impact of disorders other than ESRD, including
   hypertension, type 2 diabetes, and pyelonephritis, on RAH has not been
   explored, and larger cohorts may contribute to clarification.
   Furthermore, we would also like to conduct deep learning using CT or MR
   images to distinguish the RAH from RCC and AS in the future.

5. Conclusion

   RAH is a rare benign tumor, and the positive staining of markers,
   including ERG, FLI-1, CD31 and CD34, can contribute to its differential
   diagnosis. Pathologically, the expression content of tumor-associated
   macrophages and fibroblasts differs between cancerous and precancerous
   tissues. In the genomics field, mutations in predisposing genes,
   including CNTNAP2, NCOA2, FAT1, MET, TJP2, MAML2, SRGAP3, and CSMD3 and
   driver gene HIP1, which were confirmed by WGS and verified by Sanger
   sequencing, may participate in the pathogenesis of RAH. However,
   further molecular biology experiments are still needed to explore the
   mechanism driving the development of RAH.

Funding Statement

   The author(s) declare that financial support was received for the
   research and/or publication of this article. The present study was
   supported by the Joint Fund Project of Liaoning Provincial Science and
   Technology Programe (2023-MSLH-021), the Scientific Research Project of
   Ministry of Education of Liaoning Province (LJKZZ20220100,
   LJ212410161046, JYTZD2023046), the Interdisciplinary Research
   Cooperation Project Team Funding of Dalian Medical University Planning
   and research category (focusing on planning for recreation)
   (JCHZ2023001), the Interdisciplinary Research Cooperation Project Team
   Funding of Dalian Medical University Youth-specific category of free
   exploration(JCHZ2023020), the United Foundation for Medico-engineering
   Cooperation from Dalian Neusoft University of Information and the
   Second Hospital of Dalian Medical University (LH-JSRZ-202201),
   ”1+X”Program for Clinical Competency Enhancemen–interdisciplinary
   Innovation Project, the Second Hospital of Dalian Medical University
   (2022JCXKYB15), the United Foundation for Dalian Institute of Chemical
   Physics Chinese Academy of Sciences and the Second Hospital of Dalian
   Medical University (DMU-2-DICP UN202304, YJ20240002), Chinese Medicine
   Scientific Research Program Project of Dalian Municipal Health
   Commission (23Z12002) and Industry-University Cooperation Collaborative
   Education Program of Ministry of Education (231005073090218), Huilan
   Charity Foundation (HX20240096) and Weikang Charity Foundation
   (HX20250027) and “1+X” Program for Upgrading Clinical Skills of Young
   Physicians and Nurses, the Second Affiliated Hospital of Dalian Medical
   University’s (2024LCJSYL06).

Data availability statement

   The datasets presented in this study can be found in online
   repositories. The names of the repository/repositories and accession
   number(s) can be found below: [191]https://www.ncbi.nlm.nih.gov/,
   SRR28364768 [192]https://www.ncbi.nlm.nih.gov/, SRR28364767

Ethics statement

   The studies involving humans were approved by the Ethics Committee of
   the Second Hospital of Dalian Medical University (approval number:
   KY2024-391-01). The studies were conducted in accordance with the local
   legislation and institutional requirements. The participants provided
   their written informed consent to participate in this study. Written
   informed consent was obtained from the individual(s) for the
   publication of any potentially identifiable images or data included in
   this article.

Author contributions

   LZ: Software, Visualization, Writing – original draft. HL: Validation,
   Writing – original draft. HT: Methodology, Writing – original draft.
   HC: Data curation, Software, Writing – original draft. HG: Formal
   analysis, Supervision, Writing – original draft. SW: Data curation,
   Methodology, Writing – original draft. ZS: Formal analysis,
   Methodology, Writing – original draft. JC: Data curation, Formal
   analysis, Writing – original draft. SX: Formal analysis, Investigation,
   Writing – original draft. ZL: Funding acquisition, Project
   administration, Writing – review & editing. BF: Funding acquisition,
   Project administration, Writing – review & editing. LW: Funding
   acquisition, Project administration, Writing – review & editing.

Conflict of interest

   The authors declare that the research was conducted in the absence of
   any commercial or financial relationships that could be construed as a
   potential conflict of interest.

Generative AI statement

   The author(s) declare that no Generative AI was used in the creation of
   this manuscript.

Publisher’s note

   All claims expressed in this article are solely those of the authors
   and do not necessarily represent those of their affiliated
   organizations, or those of the publisher, the editors and the
   reviewers. Any product that may be evaluated in this article, or claim
   that may be made by its manufacturer, is not guaranteed or endorsed by
   the publisher.

Supplementary material

   The Supplementary Material for this article can be found online at:
   [193]https://www.frontiersin.org/articles/10.3389/fimmu.2025.1554203/fu
   ll#supplementary-material
   [194]DataSheet1.docx^ (224.6KB, docx)

References