Epithelioid hemangioendothelioma (EHE) is a rare malignant vascular tumor, whose clinical behavior is between hemangioma and malignant angiosarcoma (1). EHE affects all ages with a peak incidence in the 4-5th decades of life and a slight female preponderance in visceral tumors (2). Histologically, conventional EHE is composed of strands and nests of epithelioid cells in a myxohyaline stroma, with the presence of intracytoplasmic vacuoles which may contain erythrocytes. In 2001, Mendlick et al. found recurrent chromosomal translocation, involving 1p36.3 and 3q25 in EHE, and then Errani et al. demonstrated this recurrent translocation result in the WWTR1::CAMTA1 fusion gene (3, 4). Hence, CAMTA1 immunohistochemistry (IHC) has been developed as a sensitive and specific tool for the diagnosis of EHE cases, and for the challenging cases, fluorescence in-situ hybridization (FISH) for the identification of WWTR1::CAMTA1 fusion could further confirm the diagnosis (5, 6).

While a subset of EHE were found to be CAMTA1-immunonegative and fusion-negative, with distinctive morphology, such as abundant eosinophilic cytoplasm, and well-defined vascular channels. These cases were proved to be TFE3-rearranged, harboring YAP1::TFE3 fusion, and TFE3 IHC serve as a useful screening tool for this subtype (7). Although most CAMTA1-negative cases are proved to be TFE3-subtype. In our clinical practice, we found there existed CAMTA1-immunonegative, but CAMTA1-rearranged EHE cases, and such peculiar cases were also reported in previous studies, with CAMTA1-rearrangemnt or other exceedingly rare fusion variant, including WWTR1::ACTL6A and WWTR1::TFE3 fusions (8–11).

Liver is one of the most common sites of EHE, in which the CAMTA1-subtype dominates (>90%), and the TFE3-subtype is extremely rare. Hence, we performed this study focusing CAMTA1-immunonegative EHEs involving the liver, to further elucidate the clinicopathological and molecular features of CAMTA1-immunonegative EHE lesions at our institution. To the best of our knowledge, this is the first study of these rare, peculiar hepatic CAMTA1-immunonegative EHE cases.

EHE is a rare malignant vascular neoplasm which was first described by Weiss and Enzinger in 1982 (13). The majority of EHEs (>90%) are characterized by WWTR1::CAMTA1 fusions, therefore, CAMTA1 IHC is commonly utilized as a diagnostic tool for EHE. Subsequently, a small subset EHE cases were found to harbor YAP1::TFE3 fusion gene and negative for CAMTA1 IHC but showed positivity to TFE3. Most CAMTA1-immunonegative cases are composed of the TFE3 subtype, however, in our clinical practice, CAMTA1-immunonegative but CAMTA1 gene rearrangements cases were identified, and further literature review found more such cases (8, 9). Moreover, with the development of NGS, a few EHE cases were found to be negative for CAMTA1 rearrangement, but with variant WWTR1 fusion, including WWTR1::MAML2 (n=2), WWTR1::ACTL6A (n=2), WWTR1::TFE3 (n=1) ( Table 2 ) (8, 10, 11). Furthermore, there existed EHE cases with WWTR1 rearrangement, but NGS failed to identify fusion gene candidate (11). So far, no studies have systematically reported these peculiar cases. Hence, we focused on liver, one of the most common sites of EHE, to further explore the clinicopathological and molecular features of these CAMTA1-immunonegative tumors.

In our series, 4 CATMA1-immunonegative hepatic cases were proved to be TFE3-subtype EHE, including 2 males and 2 females with a median age of 60 years. There were 17 TFE3-subtype hepatic EHE cases that had been reported in English literatures ( Table 3 ) (14–23). Sixteen (including our study) historical, hepatic TFE3-subtype EHEs had clinical information, including 8 males and 8 females. Histologically, 2 TFE3-rearranged cases (case1 and 2) in our cohort displayed classic TFE3-subtype morphologic change, and case 3 also showed TFE3-subtype morphology with a certain degree variation, harboring a lower degree of atypia, smaller tumor cells and moderate cytoplasm. Furthermore, case 4 only displayed mild atypia that could potentially be misdiagnosed as CAMTA1-subtype EHE morphologically, harboring cells with intracytoplasmic vacuoles and lumina formation, presenting a signet-ring like appearance. In the hepatic TFE3-subtype cases we reviewed, morphological manifestations were available in 7 cases, of which 4 showed classical TFE3-type morphology and only 1 case had similar morphology to case 4, with morphologic change as CAMTA1 subtype EHE (23). However, unlike our case, this case was not genetically confirmed. Besides, the morphology of the other 2 historical TFE3-subtype cases were uncommon, with dual growth pattern. One case exhibited pseudo-alveolar architecture and cords, nests, and single neoplastic cells in myxoid area simultaneously (14). Another case had classic TFE3-subytpe area and focal area resembling nodular hyperplasia (17). Both in our series and historical cases, the hepatic TFE3-subtype EHE harbored unusual morphologic change, and great care with ancillary analysis should be taken to avoid misdiagnosis.

Immunohistochemically, nuclear TFE3 was uniformly expressed, while CAMTA1 was negative in the 4 cases. In terms of genetics, all 4 cases harbored YAP1::TFE3 fusion and had gene fusions between YAP1 exon 1 and TFE3 exon 4. Of the 17 historical cases, 11 had genetic information, harboring YAP1::TFE3 fusions gene or TFE3 rearrangement. Three of the 11 cases had exon information, involving YAP1 exon 1 and TFE3 exon 4, like our report. The above data was also similar to the previous largest TFE3-subtype EHEs study, among which the majority (14/16, 88%) cases had YAP1 exon 1 fused to TFE3 exon 4 (21). These results suggested the TFE3 immunohistochemistry could aid in the EHE diagnosis and further genetic analysis could help to establish the final diagnosis.

In our study, another 3 hepatic CAMTA1-immunonegative cases were proved to be positive for WWTR1::CAMTA1 gene fusion. Four historical cases harbored similar immunohistochemical and genetic change, with CAMTA1 rearrangement and negative CAMTA1 immunoreactivity (8, 9). However, the 4 cases lacked the description of their morphology. In our study, 2 of the 3 cases exhibited classic histology of CAMTA1-subtype, showing nests of epithelioid tumor cells with small, oval nuclei and intracytoplasmic vacuoles. It needs to be pointed out, another 1 case (case 7), unlike characteristic EHE, showing the morphology analogous to TFE3-subtype, with medium-sized epithelioid cells having moderate amounts of eosinophilic cytoplasm and vascular lumen, containing erythrocytes. More cases and further studies and needed to clarify the pathological feature of these peculiar cases. More importantly, when meeting such peculiar and challenging cases with unusual morphologic and immunohistochemical results, molecular studies such as FISH and furthermore, NGS are needed to establish the diagnosis and reveal the reason for these changes.

In this study, we identified a novel fusion between WWTR1 exon 2 and CAMTA1 exon 9 in one case (case 7) through NGS. This breakpoint is located at the 1663 nucleotide of CAMTA1 gene (NM_015215.4), a previously unreported site, and the breakpoint of WWTR1 gene is identical to previous studies. The current CAMTA1 antibody only recognizes 84 amino acids encoded by nucleotides 1428 to 1682 of the CAMTA1 gene, which is located downstream of all previously reported CAMTA1 breakpoints (4, 6, 24–27). However, the antibody recognition sequence is located upstream of the breakpoint of case 7, with only 20bp nucleotide overlapped and corresponding to 6 amino acids ( Figure 5G ). This result suggested that the uncommon fusion point might result in the negative IHC result. This finding underscores the importance of understanding the precise location of fusion points, as breakpoint could impact the gene translation and antibody recognition and may lead to unexpected results. Additionally, the CAMTA1 immuno-negativity may also occurred in TFE3-rearranged EHE cases and more rarely, in cases with variant WWTR1 gene rearrangements. Hence, when meeting such challenging cases with uncommon IHC results, the FISH and NGS is valuable to assist the establishment of correct diagnosis.

The histopathologic morphology of hepatic EHE is diverse and heterogeneous, particularly in small biopsy samples, and it is usually too difficult to confirm endothelial differentiation and to exclude histologic mimics. EHE includes a wide range of differential diagnoses, including poorly differentiated adenocarcinoma, perivascular epithelioid cell tumor (PEComa), epithelioid angiosarcoma and so on.

Poorly differentiated adenocarcinoma can also have distinct signet-ring cell components and morphologically overlapped with EHE. The typical features of adenocarcinoma are a large number of cells with intracytoplasmic vacuoles containing mucus, readily confirmed by epithelial mucinous histochemical staining such as PAS (28). Immunohistochemically, adenocarcinoma expresses epithelial markers such as CK and EMA, but negative for CD31, CD34 and especially CAMTA1 and TFE3. Genetically, CAMTA1 or TFE3-rearrangement is present in EHE, but not in poorly differentiated adenocarcinomas.

PEComa also has spindle or epithelioid cells arranged in fascicular or nested distribution, like EHE, but the former lacks evident vascular differentiation. PEComa exhibit expression of melanocyte markers such as HMB45 and Melan-A, as well as myogenic markers including SMA and desmin, while they do not express the endothelial marker CD34 (29). Generally, specific immunohistochemical markers can be utilized for distinguishing between these two entities. Since both have TFE3 gene break, they can be identified by gene sequencing if necessary to prevent misdiagnosis or missed diagnosis. Unlike YAP1::TFE3 EHEs, PEComa is frequently accompanied by SFPQ::TFE3 fusion.

The distinction between EHE and epithelioid angiosarcoma can be challenging due to their similar histological change with large epithelioid cells, eosinophilic cytoplasm, distinct nucleoli, and occasional intracytoplasmic vacuoles containing erythrocytes. However, the latter exhibits higher atypia and marked mitosis. Furthermore, CAMTA1 or TFE3 immunohistochemistry may provide valuable insights. Moreover, confirming the presence of WWTR1::CAMTA1 or YAP1::TFE3 gene fusion in the tumor would offer more convincing evidence.

In clinical practice, the diagnosis of EHE is mainly based on the histological evaluation and IHC results, and for the challenging cases with ambiguous morphology or unexpected IHC profile, the FISH for CAMTA1 and TFE3 should be performed, and furthermore, NGS is recommended for cases which is highly suspected for EHE to make the final diagnosis. We recommend the diagnostic algorithm shown in the figure to minimize misdiagnosis ( Figure 6 ).

Owing to the rarity and heterologous clinical course of hepatic EHE, there is no well-established treatment strategy for such entity. In our series, 4 patients were treated with surgical resection only, 1 patient received surgical resection and chemotherapy and 2 patients underwent chemotherapy. According to previous large series studies of hepatic EHE, the common treatment modalities include hepatic resection, liver transplantation, ablations, chemotherapy, and radiotherapy. Hepatic resection is recommended for resectable intrahepatic lesions, and liver transplantation is considered to be the first choice for the treatment of tumor patients when there is intrahepatic metastasis or the tumor is too large to be resected (30, 31). In addition, adjuvant chemoradiotherapy is recommended when metastasis occurs.

In cases of EHE, the discovery of the WWTR1::CAMTA1 and YAP1::TFE3 fusion genes has provided a new perspective for the treatment of EHE. WWTR1 and YAP1 are key components of the Hippo signaling pathway, playing a central role in a variety of physiological and pathological processes. The WWTR1::CAMTA1 fusion gene can induce endothelial cells to transform into vascular tumors with EHE characteristics. This process involves WWTR1 (TAZ)::CAMTA1 as a continuously activated form of TAZ, which is mainly located in the cell nucleus and activates its pro-tumor transcription program. Similarly, the carcinogenic potential of the YAP1::TFE3 fusion gene also requires interaction with TEAD. The fusion protein utilizes the transcriptional activation domain and nuclear localization sequence of TFE3, binding to DNA through the TEAD binding site of YAP1, forming a continuously activated chimeric transcription factor (32). Additionally, trametinib, an MEK inhibitor, has demonstrated certain therapeutic effects in EHE patients with the WWTR1::CAMTA1 fusion (33). In the future, it is necessary to further collect and analyze cases to gain a deeper understanding of the molecular basis of EHE, in order to provide more precise treatment strategies for patients.

In our series, 4 patients (4/6; 66.7%) were alive with no evidence of disease, and 2 patients (2/6;33.3%) were alive with disease. Previous studies found the 5-year overall survival rate of hepatic EHE ranging from 57% to 67% (18, 34). Pathologic parameters including tumor size >3 cm, mitotic activity >3 mitoses per 50 high-powered fields and histologic grade were associated with worse outcome of EHE patients (18, 35). Moreover, a recent large series study discovered the CAMTA1-subtype EHE correlated with a worse prognosis compared to TFE3 subtype, with the 5-year overall survival (OS) as 59% versus 86%, respectively (18). Hence, we carefully reviewed the historical hepatic EHE cases and found 74 CAMTA1 subtype and 13 TFE3 subtype EHE cases with available follow-up information (including our cases)) (9, 16, 26, 31, 36–45). The survival results showed 10 of 74 (13.5%) CAMTA1 subtype cases died of disease and only 1 of 13 (7.7%) TFE3-subtype EHEs succumbed to disease. Combined with this study and retrospective data analysis, most patients with hepatic EHE have a good prognosis, and TFE3 subtype cases seems to harbor a favorable outcome. While the prognosis difference between the TFE3-subtype and CAMTA1-subtype hepatic EHE still needs to be further explored in a larger study cohort.

In summary, our study reported a series of 7 peculiar, CAMTA1-immunonegative hepatic EHE cases and most of them were TFE3-rearranged EHEs. Moreover, our study showed that there existed the CAMTA1-immunonegative but CAMTA1-rearranged EHE cases. Therefore, the diagnosis of EHE should be based on histological morphology, combined with the dual detection of CAMTA1 and TFE3 IHC, and if necessary, supplemented by comprehensive analysis, including FISH and NGS for the genetic aberration, to ensure correct diagnosis.