DOI: https://doi.org/10.21203/rs.3.rs-23442/v1
Primary hepatic sarcomatous carcinoma (PHSC) and primary hepatic sarcoma (PHS) are rare malignancy with frequent overlap in the clinic and radiology. There has never been a comparative study of these tumors for the restricted cases. The purpose of our study was to analyze the clinical and imaging features of PHSCs and PHSs, with an emphasis on the particularities and similarities by comparing the two tumors.
We retrospectively analyzed the clinical and imaging features of 39 patients with pathologically proven PHSCs (n = 23) and PHSs (n = 16) from three university centres over a 9-year period from 2010–2019. Univariate analyses were performed to determine the consistent and distinctive features.
The background of chronic hepatitis or cirrhosis was observed with a high frequency in both of PHSCs (73.7%) and PHSs (62.5%). Tumors with a diameter more than 10 cm were significantly more common in PHSs than PHSCs (p = 0.043) and cystic masses were more detected in PHSs (P = 0.041). Both of PHSCs and PHSs mainly present hypovascularity (73.7% vs 82.6%). The ring hyper enhancement on the arterial phase (AP) and wash out were more frequently seen in PHSCs and the iso-hypo enhancement on the AP followed persistent or progressive enhancement were more commonly detected in PHSs (all, p < 0.05).
PHSC and PHS generally present as mass lesions with hypovascularity. The ring hyper enhancement on the AP and wash out favor the diagnosis of PHSC. The large size greater than 10 cm, cystic lesion, iso-hypo persistent or progressive enhancement pattern might suggest the possibility of PHSs.
Primary hepatic sarcomatous carcinoma (PHSC) and primary hepatic sarcoma (PHS), are rare malignancies accounting for only 0.2% (1) and 1% (2) of primary malignant liver tumors, respectively. These tumors present with overlapping features clinically, radiologically and morphologically, making it a challenging task to diagnose accurately.
Sarcomatous carcinoma is defined as a tumor containing an intimate mixture of carcinomatous (either hepatocellular or cholangiocellular) and sarcomatous elements. Sarcomatous change in hepatocellular carcinoma (HCC) or intrahepatic cholangiocarcinoma (ICC) is defined as ‘‘sarcomatous HCC (S-HCC)’’ or ‘‘sarcomatous ICC (S-ICC)’’ in the World Health Organization (WHO) classification (3). This entity is differentiated from a true hepatic sarcoma such as undifferentiated embryonal sarcoma (UES), leiomyosarcoma (LS), malignant solitary fibrous tumor (SFT), epithelioid sarcoma (ES) and other interstitial tumors deriving from the liver. It should be diagnosed as sarcomatous carcinoma when the sarcomatous component is predominantly composed of spindle cells, but the epithelial cells are still morphologically, immunohistochemically, and ultrastructurally identifiable (1). However, how to accurately identify the two tumors pathologically is still a challenge for pathologists (4–7). Clinically, these tumors are usually asymptomatic until they become significantly large at the time of diagnosis (5, 8). Moreover, these tumors have frequent overlap in the imaging appearances (8–10). Therefore, the preoperative diagnosis is challenging. However, some characteristic clinical and imaging features may exist that can suggest a specific diagnosis. To the best of our knowledge, the current literature about these tumors was limited to either case reports or small case series (11–15) and no reports were available on the comparison of these two tumors except a mention by Mani, H (16). Although the diagnosis is difficult, there still exist some characteristic clinical and imaging features that can suggest a specific diagnosis. Our research aims to explore the clinical and imaging features that can aid to differentiate PHSCs from PHSs.
We retrospectively reviewed the patients with PHSC and PHS proven pathologically according to the World Health Organization (WHO) definition in 2000, from three university centres between January 2011 and April 2019. For the inclusion of the subject,all the evaluated lesions did not have any prior treatment. In PHSC, one patient with preoperative intervention by transcatheter arterial chemoembolization (TACE) and two patients with liver metastasis from extrahepatic origin of SC were excluded. In the PHS group, sarcomas of vascular origin including epithelioid hemangioendothelioma (n = 7), angiosarcoma(n = 8), and Kaposi sarcoma(n = 1) were excluded for their relative specificity in the imaging or clinical characteristics. Our study included PHSCs (n = 23, 11 S-HCCs, 4 S-ICCs, 1 S-HCC–CC, 7 unclassified) and PHSs (n = 16, 1 UES, 2 SFTs, 2 ES, 3 LSs, 8 unclassified sarcomas). The clinical materials (demographic characteristics, laboratory data, clinical symptoms and prognosis), imaging findings and pathology results were reviewed. The request and approval for the study protocol was obtained from the Institutional Review Board of each hospital.
Twenty-one patients with PHSC and all the patients with PHS were instructed to finish the examinations on the Computed tomographic (CT) scanner (LightSpeed VCT 64, GE Healthcare, Waukesha, Wisconsin, USA) with the following parameters as bellow: tube voltage, 120 kV; tube current, 189–200 mA; matrix, 512 × 512 mm; and section thickness 5 mm. All patients underwent dynamic three phases scanning including hepatic arterial phase (HAP) (25–40 s), portal venous phase (PVP) (45–90 s) and equilibrium phase (EP) (2–5 min) which were obtained after the bolus injection of contrast agent with lopromide (Ultravist 370, Bayer Schering Pharma, Berlin, Germany) at a dose of 1.5 mL/kg and rate of 3 mL/s.
Nine patients with PHSC and four patients with PHS were instructed to finish the examinations with 3.0T whole-body MRI systems (Trio, Siemens Healthineers, Erlangen, Germany) with an 8-channel phased array body coil. The parameters of T1-weighted fast low angle shot sequence were mentioned as below: TR/in phase: TE, 170/2.30; out-of phase TE, 3.67 ms; matrix size, 256 × 205; flip angle, 65°. The three-dimensional volumetric interpolated breath-hold examination (3D-VIBE) sequence was obtained before (pre-contrast) and after the injection of contrast agent (Gd-BOPTA, MultiHance, Bracco Pharma, Italy) at a rate of 2 ml/s. The serial dynamic contrast-enhanced scans including HAP, PVP and EP were collected at the time of 25–40 s, 45–90 s and 2–5 min.
All the images were retrospectively assessed by two abdominal radiologists with more than 7 years of experience in hepatic imaging. When there is disagreement in the assessment of the images, the two readers need to reassess them together.
For morphological lesion assessment, the following items were evaluated: 1. The location (right lobe, left lobe), 2.Size (> 10 cm, ≤ 10 mm), 3.Contour (round, lobulated or irregular), 4. Margin (sharp and indistinct), 5. Liver surface contour (retraction, smooth, bulging), 6.The presence of capsule appearance, hemorrhage, and perfusion alteration, 7.The cystic lesion (The cystic lesion was evaluated based on the predominant parts (75%) of the tumor with cystic changes without any enhancement), 8.The presence of vascular invasion, intrahepatic metastasis and extrahepatic metastasis.
AP enhancement was classified according to the categorizations provided by Rimola et al with modifications (17): 9.non-ring enhancement include the global enhancement that hyperenhancement involving > 75% of the lesion and the nodular enhancement that hyperenhancement involving 25–75% of the lesion; ring enhancement include the peripheral enhancement that hyperenhancement involving 25–75% of the lesion and rim enhancement that rimlike hyperenhancement involving < 25% of the lesion), and iso-hypointensity/density. 10.The vascularity of the whole tumor (lesions with heterogeneous enhancement were evaluated based on the predominant parts more than half of the entire tumor) (17), 11.Dynamic pattern of enhancement (washout, progressive or persistent enhancement).
The continuous variables including the age of patients and the diameter of tumors were expressed as mean ± SD, and the differences between PHSC group and PHS group were conducted using the independent t-test. The categorical variables were compared using -Fisher’s exact. P༜0.05 was considered to indicate a statistically significant difference. All statistical analyses were performed with the software SPSS® version 23.0 (IBM, Armonk, NY, USA).
The PHSC cohort included 23 patients (20 men, 3 women; median age, 56 years; range, 32–77 years), and the PHS cohort consisted of 16 patients (11 men, 5 women; median age, 58 years; range, 22–75 years). The diagnosis proved pathologically was got after surgical resection (15 PHSCs, 6 PHSs) or biopsy (8 PHSCs, 10 PHSs). The clinical data of the patients with PHSC and PHS were summarized in Table 1. There were not significant difference in the tumor markers including AFP, CEA and CA1-99. The majority of patients in both groups were middle-aged men (87.0% vs 68.8%) with the background of liver cirrhosis (73.7.0% vs 62.5%). The most common complaints in PHS and PHSC were abdominal discomfort (43.8% vs 30.4%) and in PHSCs, 34.8% of these patients were detected incidentally in their routine check up for the hepatitis or some other disease. Vascular invasion (56.5% vs 25.0%), intrahepatic metastasis (40.9% vs 25.0%), and extrahepatic metastasis (57.1 vs 37.5%) tended to be more commonly seen in PHSCs than PHSs. In the PHSC group, 13 patients underwent surgery, 5 of them were combined with TACE or RFA (radiofrequency ablation),4 patients received interventional therapy and one patient underwent liver transplantation. 64.7% (11 of 17) of PHSC patients progressed or died between 1–16 months. In the PHS group, 7 patients underwent surgery and 7 received interventional therapy. 76.9% (10 of 13) of PHS patients progressed or died between 1–26 months.
| Variable | PHSC (n = 11) | PHS (n = 11) | P value |
|---|---|---|---|
| Age (y)* | 56 (32–77) | 58 (22–75) | 0.276 |
| Male: female ratio | 20: 3 | 11: 5 | 0.235 |
| Liver hepatitis cirrhosis | 14 (73.7) | 10 (62.5) | 0.716 |
| Clinical manifestions | 11 (52.4) | 11 (73.3) | 0.500 |
| Tumor markers | |||
| AFP | 9 (47.4) | 5 (33.0) | 0.495 |
| Ca19-9 | 5 (35.7) | 2 (14.3) | 0.385 |
| CEA | 1 (6.7) | 0 | 0.999 |
| Vascular invasion | 13 (56.5) | 4 (25.0) | 0.099 |
| Intrahepatic metastasis | 9 (40.9) | 4 (25.0) | 0.490 |
| Extrahepatic metastasis | 12 (57.1) | 6 (37.5) | 0.325 |
| Note. Unless otherwise specified, data are numbers of patients, with percentages in parentheses, AFP a-fetoprotein, CA 19 − 9 carbohydrate antigen 19 − 9, CEA carcinoembryonic antigen. | |||
| *Data are medians, with ranges in parentheses. | |||
Table 2 summarizes the morphologic characteristics and accompanying findings of PHSCs and PHSs. The PHSCs and PHSs occurred more frequently from the subcapsular area in the right hepatic lobe (69.6% vs 75.0%) with a sharp margin. The tumor greater than 10 mm was more commonly seen in PHSs than PHSCs (P = 0.043). The PHSCs and PHSs mainly showed smooth (47.8% vs 37.5%) or bulging surface (43.5% vs 56.3%) and retraction of the capsule was rare. The capsule appearance occurred in around half of theses patients and 54.5% (6 of 11) of capsule in PHSCs was not complete. Although there was no significant difference, the hemorrhage was more common in PHSs than PHSCs (50.0% vs 26.1%)(Fig. 1)
| Variable | PHSC (n = 23) | PHS (n = 16) | P value |
|---|---|---|---|
| Right lobe | 16 (69.6) | 12(75.0) | 0.999 |
| Tumor diameter | 72.6 ± 38.9 | 94.8 ± 44.4 | 0.823 |
| ≥10 cm | 5 (21.7) | 9 (56.3) | 0.043 |
| ༜10 cm | 18 (78.3) | 7 (43.8) | |
| contour | 0.999 | ||
| round | 16 (69.6) | 10 (62.5) | |
| Lobulated | 2 (8.7) | 2 (12.5) | |
| irregular | 5 (21.7) | 4 (25) | |
| margin | 0.444 | ||
| sharp | 19 (82.6) | 11 (68.8) | |
| indistinct | 4 (17.4) | 5 (31.3) | |
| Liver surface contour | 0.786 | ||
| retraction | 2 (8.7) | 1 (6.3) | |
| smooth | 11 (47.8) | 6 (37.5) | |
| bulging | 10 (43.5) | 9 (56.3) | |
| Capsule appearance | 11 (47.8) | 6 (37.5) | 0.743 |
| Hemorrhage | 6 (26.1) | 8 (50.0) | 0.179 |
| Perfusion alteration | 11 (47.8) | 9 (52.2) | 0.748 |
| Cystic mass | 4 (17.4) | 8 (50.0) | 0.041 |
| Note. Data are numbers of lesions, with percentages in parentheses. | |||
Table 3 summarizes the enhancement characteristics of PHSC and PHS. There was significant difference in the AP enhancement (P = 0.027). Ring hyper-enhancement was more commonly seen in PHSCs than PHSs (60.1% vs 18.8%). Of the 14 patients in PHSCs with ring hyper enhancement on AP, 7 patients demonstrated rim enhancement and the others showed peripheral enhancement (Fig. 2, 3). The iso-hypo enhancement was more frequently detected in PHSs (Fig. 4) than PHSCs (56.3% vs 30.4%). Pattern of persistent and progressive enhancement was observed in both of PHSCs and PHSs, especially in the PHSs, but wash out was more commonly seen in PHSCs (p = 0.017) (Fig. 5). For the vascularity of the whole tumor, PHSCs and PHSs were predominantly hypovascularity (78.3% vs 87.5%) and the cystic mass was more commonly seen in the PHSs (Fig. 1) than the PHSCs (50.0% vs 17.4%,p = 0.041). Notably, no or minimal enhancement with a nearly complete cystic appearance existed in PHSs in our study (Fig. 1).
| Vascularity*1 | 0.999 | ||
|---|---|---|---|
| Hypervascularity | 5 (21.7) | 3 (17.6) | |
| Hypovascularity | 18 (78.3) | 14 (82.3) | |
| AP enhancement | 0.027 | ||
| Ring hyperintensity | 14 (60.1) | 3 (18.8) | |
| Non-ring hypertensity | 2 (8.7) | 4 (25.0) | |
| Iso- or hypointensity | 7 (30.4) | 9 (56.3) | |
| Dynamic pattern | 0.017 | ||
| Wash out | 12 (52.2) | 2 (12.5.0) | |
| Persistent or progressive | 11 (47.8) | 14 (87.5) | |
| Note. Data are numbers of lesions, with percentages in parentheses. | |||
| *1 The whole tumor was evaluated according the predominant parts more than 50%. | |||
In our study, we did not find any significant difference between the two tumors in the background of liver cirrhosis and the tumor markers such as AFP, CEA and CA1-99. Being different from previous reports that patients with PHSs had no evidence of hepatitis or cirrhosis (8), Ten (62.5%) of 16 PHSs were positive for hepatitis or cirrhosis (18). Fourteen (73.7%) of 19 PHSCs had the background of liver cirrhosis, similar to previous reports that liver hepatitis virus infection might have relationship with the occurrence of PHSCs (5, 19). For PHSs, most of the laboratory data were not helpful in diagnosis (11, 15). Less than half of the patients with PHSCs were positive for AFP, which is lower than that in the ordinary HCC.
Similar to previous study (3, 12, 13, 20), the PHSCs and PHSs demonstrate hypovascularity probably for hemorrhage, necrosis, fibrous tissue or myxoid degeneration (21–24). However, the AP enhancement and dynamic enhancement pattern were significantly different. The current study concluded that PHSCs mainly showed ring hyper-enhancement on the AP, followed a washout on the later phase. It was reported that the diverse tissue compositions of PHSC determine its enhancement pattern (6). The PHSCs, especially the S-HCCs, were characterized by the peripheral viable cancerous tissue (viable cells, higher microvascular densities and relatively less fibrous tissue) and central necrosis. The sarcomatous component consists of poorly differentiated cells that grow rapidly with the neovasculature unable to adequately supply the fast-growing malignant cells, resulting in necrosis. The PHSs generally present iso-or hypo enhancement on the AP and persistent or progressive enhancement on the later phase, similar to previous studies (12, 25). The myxoid degeneration and the loose arrangement of the cells in PHSs could expand the extracellular space and the contrast agent in the extracellular space were accumulated gradually and expurgated slowly, leading to hypo-iso continuous or progressive enhancement.
Our study demonstrated the cystic mass was commonly seen in PHSs (P < 0.05). In current study, there were PHSs displaying as nearly complete cyst-like masses nearly without any enhancement simulating benign tumors, which were never seen in PHSCs. There had been an emphasis on cystic-like appearance in PHSs, which was mainly attributed to the varying degrees of myxomatous change (11, 12, 14, 26, 27). Hemorrhage also played a role in the cystic appearance, which was reported more frequently seen in PHSs than some other rare liver malignant tumors and attributed by rupture of the tumor for the serpiginous vessels (10, 11). In previous studies, there was often extensive hemorrhage in PHS creating a huge cyst mass so that the underlying tumor was obscured and misdiagnosed as a hematoma, abscess or cystic tumor (28), which also occurred in our study. Although there was not statistical difference in the current study, hemorrhage was more commonly seen in PHSs than PHSCs (50.0% vs 26.1%). Besides, the tumor larger than 10 cm in PHSs was significantly more detected than PHSCs (P < 0.05). It was reported that solid or cystic manifestations were different stages of PHSs and as the tumor grew, necrosis increased,tending to result in a cystic appearance. In summary, cystic lesions occurred more often in the PHSs and it might help us distinguish PHS from PHSCs.
The capsule invasion (19), vascular invasion or thrombosis, intrahepatic metastasis and lymph node metastasis were more prevalent in PHSCs and in our study the vascular invasion in PHSCs was close to significantly more common than the PHS (P = 0.099). The PHSCs were highly aggressive, and the presence of SC were considered to be closely related to the more invasive tumor biology, more common metastasis, low resectability and frequent postoperative recurrences (19, 29, 30). By contrast, the PHSs usually involved the adjacent anatomic structures, and vascular invasion, metastases and lymph node involvement were less common (9, 28).
These tumors should also be differentiated from other liver masses (18). The ring hyper-enhancement of PHSCs may mimic those of ICCs (31). The elevated CA19-9 levels, bile duct dilation around the lesion and capsule retraction may be helpful for the differentiation of these lesions (32). The global avid enhancement with washout and elevated AFP levels help us to differentiate the HCC from PHS and PHSC (18, 33, 34). When the PHSs displaying as nearly complete cyst-like mass as seen in our study, they should be distinguished from other cyst-like lesions such as hydatid cyst, abscess, biliary cyst and adenoma. It was reported in studies that a cyst-like PHS could be frequently misdiagnosed as a hepatic cyst (8, 14, 26). However, the presence of feeding vessels, the findings of hemorrhage and the abrupt increase in its size should alert us to the diagnosis of PHS (11, 28).
We should acknowledge several limitations to our study. First, for the retrospective study, it was technically unworkable to make the imaging-pathology match slice-by-slice. Second, the relatively small sample size had its internal disadvantages, but it was inevitable for the rare incidence of the tumors. Third, there wasn’t a recognized international standard for the evaluation of the cystic tumors.
In conclusion, the PHSC and PHS generally presented as a large subcapsular hypovascular mass. The ring hyperenhancement, wash out and more common vascular invasion favored the diagnosis of PHSC. The large mass with a diameter more than 10cm,iso-hypo intensity/density on AP and pattern of persistent or progressive enhancement might alert us to the possibility of PHS. In spite of the presence of these meaningful diagnostic features, there were not specific for the diagnosis and differential diagnosis of PHSC and PHS. However, the absence of characteristic imaging manifestations of primary hepatic tumors should remind us of the possibility of these tumors.
Primary hepatic sarcomatous carcinoma
Sarcomatous intrahepatic cholangiocarcinoma
Sarcomatous hepatocellular carcinoma
Sarcomatous combined hepatocellular and cholangiocarcinoma
primary hepatic sarcoma
undifferentiated embryonal sarcoma
leiomyosarcoma;
solitary fibrous tumor
Hepatic arterial phase
Portal venous phase
Equilibrium phase
Alpha-fetoprotein
Carbohydrate antigen 19–9
carcinoembryonic antigen
Computed tomography
Magnetic resonance imaging
Original data and material can be available from the corresponding author if necessary.
This retrospective study was conducted with the approval of the Ethics Committee of the Capital medical University Affiliated Youan hospital，Zhengzhou University Affiliated Henan Cancer Hospital. Southeast University Affiliated Xuzhou Central Hospital. Written informed consent was obtained from the patients or their parents before MRI or CT.
Not applicable
The authors declare that they have no competing interest.
No Funding
DlS: data acquisition and analysis, manuscript preparation and writing. LM:collection of clinical data, manuscript editing. JS: data acquisition and analysis. JC: pathological data analysis .The others: collected and review the data. All authors read and approved the final manuscript.
We thank Dr Yibing Shi and jingrong Qu for their precious help in data collection.