First described in 1931 , SFT has since appeared in the form of hemangiopericytoma, giant cell angiofibroma, fat-forming variant, lipomatous hemangiopericytoma, and, rarely, as a mesenchymal neoplasm that commonly originates from pleura. In the 2013 WHO classification of tumors of soft tissue and bone, extrapleural SFT was considered a fibroblastic/myofibroblastic neoplasm with intermediate, rarely metastasizing biological behavior . Subsequently, in the updated WHO classification of the digestive system, SFT is considered a benign tumor with the potential for malignant transformation . Because the hepatic SFT is extremely rare, whether the SFT should be classified as a benign or malignant tumor remains controversial regarding this tumor originating from the liver.
Only 85 patients with SFTs of the liver, including ours, have been reported in the English-language literature. According to the review by Chen et al. , the average age of patients with this affliction is 57.1 (range 16–87) years. It appears to occur more frequently in women than in men (1.4:1.0), and the histological incidence of malignant features is 19.0% (16/84 patients). The clinical presentation of this disease is generally non-specific. Extrapleural SFTs present as relatively slow-growing masses that are often asymptomatic. When symptoms do occur, they are caused by the pressure the mass exerts on adjacent structures . Similarly, our patient complained of steadily increasing abdominal bloating at presentation, which proved to be due to the huge tumor compressing other organs. In previous reports, hepatic SFTs were found incidentally during routine examinations.
Among radiological studies, ultrasonography often reveals a heterogeneous mass that may be either hypoechogenic or hyperechogenic (or both) with or without calcification. Contrast-enhanced CT shows early arterial enhancement with delayed venous washout. MRI of our patient showed that the tumor mass was heterogeneous with slight hyperintensity on T2-weighted images and slight hypointensity on T1-weighted images. Findings on MRI are similar to those seen on CT scans. DWI revealed greater signal intensity than that of normal liver . These findings often mimic those of high-grade HCCs (including scirrhous and sarcomatous HCC) or leiomyomas. Our radiological results were consistent with these patterns.
In our case, a percutaneous biopsy was not performed to obtain a tissue diagnosis. Although a fine-needle biopsy can distinguish the SFT from malignant tumors such as HCC or sarcoma, the procedure could lead to rupture or seeding of malignant cells. Some reports suggested that, although liver biopsy could be performed safely, the biopsy of the SFT was misdiagnosed as HCC or metastasis from adenocarcinoma of the pancreas or a gastrointestinal tract lesion [16, 17]. In our patient, the possibility of a malignant tumor could not be ruled out preoperatively because we decided to resect this huge tumor without a definitive diagnosis due to the patient’s symptoms and the risk of hemorrhage.
A definitive diagnosis of SFT of the liver requires histopathological and immunohistochemical studies. Microscopically, the tumor is composed of ovoid spindle-shaped cells with its characteristic architecture in a storiform pattern or a haphazard, “patternless pattern.” These cells are separated from thick bands of keloid-like collagen bundles and display branching of staghorn vessels in a hemangiopericytoma-like pattern. Myxoid changes are also commonly observed. Mitoses and necrotic changes—characteristically suggesting malignancy—are rare for this tumor . Features identified by WHO as being associated with malignancy include hypercellularity, cytological atypia, tumor necrosis, infiltrative margins, and high mitotic activity (≥ 4/10 HPF) .
Immunohistochemically, the staining of CD34, vimentin, and Bcl-2 is useful for distinguishing SFTs from other liver tumors. However, CD34 staining is imprecise in this case because 5–10% of typical SFTs are nonreactive to CD34 . To make up for this imprecision, the NAB2–STAT6 fusion gene has recently been identified as the genetic hallmark of SFT. This aberration drives the nuclear relocation of STAT6 . Immunohistochemical detection of STAT6 nuclear expression and the NAB2–STAT6 fusion gene identified by RT-PCR assay offers a strong surrogate diagnostic technique for distinguishing SFTs from histological mimics. Our patient, in fact, was immunohistochemically negative for CD34, which made the definitive diagnosis difficult. However, the spindle cells were immunohistochemically positive for STAT6, and the NAB2–STAT6 fusion gene was detected by RT-PCR and direct sequencing, thereby allowing a definitive SFT diagnosis.
Chen et al.  previously reported that 16 of 84 SFTs were malignant, which was similar to the intrapleural SFT recurrence rate of 20–67% among malignant tumors. These malignant cases were diagnosed by histological examination, which showed a high incidence of mitotic changes (> 4/10 HPF) and local recurrence or distant metastasis in 17.9%. Furthermore, 26.7% of patients with these malignant SFTs had a local recurrence within 9 months to 6 years, and 53% had distant metastasis within 1 month to 6 years. England et al.  established the criteria for malignant SFT: mitotic changes (> 4/10 HPFs), tumor necrosis and hemorrhage, nuclear pleomorphism, and metastasis were the major criteria, and large tumor size (> 10 cm) and cellular atypia were the minor criteria. Wilky et al.  classified the SFTs into “benign” with no England’s criteria, “borderline” with 1 or more England’s criteria but final classification as benign, and “malignant.” This report described that “borderline” SFTs with any of England’s criteria had been related to high risk of recurrence. Our patient described herein met two of the six criteria (necrosis/hemorrhage and tumor size), suggesting a possibility of malignancy.
SFTs of the head, neck, and intracranial meninges are also rare and display benign behavior without metastasis . We found no descriptions in the English-language literature that resembled that of our patient, who presented with suspected localized liver metastasis from a head SFT. Du et al. , however, reported a rare case of non-malignant SFT that appeared 5 years after initial liver resection. The tumor’s appearance had no marked variances from other non-malignant SFTs.
To clarify whether this second tumor was a recurrence from a cranial SFT in our patient, we attempted to examine specimens from her previous, cranial SFT. Unfortunately, no tissue was available because the maintenance term of the specimen had reached its limit, and it had been discarded. Assessment of the radiological findings, however, showed that there had been no abnormality in other organs, and no important mitotic changes or low levels of the Ki67 labeling index. Thus, it was more likely that our patient currently had a primary tumor in the liver rather than a recurrence from her original cranial SFT. Nevertheless, we are committed to performing carefully follow-up every 3 months in the first 2 years, twice each year up to 5 years after surgery, and then once a year after the fifth year. The best method for follow-up is not well established, but whole-body CT is suggested in our patient.