Abstract
Aim: To clarify the clinical and radiological features of isolated tumor thrombi in the inferior vena cava (IVC)/right atrium in patients with hepatocellular carcinoma (HCC) without hepatic vein invasion. Patients and Methods: In this retrospective study, from January 2007 to December 2019, a total of 35,163 chemoembolization sessions were performed in 7,704 patients with HCC. Among them, 10 (0.13%) patients had tumor thrombi in the IVC/right atrium without definitive hepatic vein invasion. Computed tomographic (CT) scans, digital subtraction angiograms, and cone-beam CT images were retrospectively reviewed and interpreted. Results: The tumor thrombi were supplied by the right inferior phrenic artery (n=8) or the right internal mammary artery (n=2). Follow-up CT scans in eight patients showed linear accumulation of iodized oil along the diaphragm, which was presumed to be a thrombosis of the phrenic vein. Retrospective review of formal radiological reports of pre-procedural CT scans revealed that a correct diagnosis of tumor thrombi of the IVC/right atrium was made in only three cases. Conclusion: HCC invading the phrenic vein may have tumor thrombi in the IVC/right atrium without hepatic vein invasion.
- Hepatocellular carcinoma
- chemoembolization
- right atrium thrombus
- inferior vena cava thrombus
- inferior phrenic vein
Hepatocellular carcinoma (HCC) is the most common primary malignant liver tumor and often invades the portal vein (1). The hepatic vein is infrequently involved, and tumor thrombi are seldom detected in the inferior vena cava (IVC) or right atrium (2). Although patients with HCC with IVC/right atrial tumor thrombi have bleak prognoses and systemic therapy is a standard treatment option, local treatments, such as surgery, external radiotherapy, and chemoembolization, have been increasingly performed in these patients (3-6).
Most tumor thrombi in the IVC and right atrium develop as continuous extensions of tumor thrombi in the hepatic veins (2) but there are occasional reports of isolated right atrium metastasis without hepatic vein invasion (7, 8). A recent report presented cases in which the tumor thrombi of the IVC/right atrium extended to the phrenic vein, as confirmed by surgical resection (9). We have infrequently encountered patients with HCC with IVC/right atrial thrombi without hepatic venous thrombi who were treated with chemoembolization. Our hypothesis is that HCC can directly invade the diaphragm and the phrenic vein, resulting in tumor thrombi in the IVC and/or right atrium. The purpose of this study was to clarify the clinical and radiological features of isolated tumor thrombi in the IVC/right atrium in patients with HCC without hepatic vein invasion.
Patients and Methods
Patients. The Institutional Review Board approved this retrospective study (no, 2104-112-1212), and the requirement for informed patient consent was waived. From January 2007 to December 2019, a total of 35,163 chemoembolization sessions were performed in 7,704 patients with HCC. Among them, 10 (0.13%) patients had tumor thrombi in the IVC/right atrium without definite hepatic vein invasion. This study population consisted of eight men and two women (mean age=68.6 years, range=55-80 years) (Table I).
All 10 patients had a previous history of treatment for HCC, including chemoembolization (n=9), sorafenib (n=6), radiofrequency ablation (n=6), external radiation therapy (n=6), percutaneous alcohol injection (n=3), and surgical resection (n=3). Of the 10 patients, nine had previous chemoembolization (median=8 sessions; range=1-22 sessions), and seven patients had been previously treated with chemoembolization in the right inferior phrenic artery. The time interval between the initial HCC diagnosis and chemoembolization of the tumor thrombi of the IVC/right atrium ranged from 2.8 to 12.2 years.
Chemoembolization procedure. For all patients, enhanced dynamic computed tomographic (CT) scans (pre-contrast, hepatic arterial, portal venous, and equilibrium phase) of the liver were obtained before chemoembolization. Detailed protocols for CT imaging were described previously (10). All chemoembolization procedures were performed by, or under the supervision of, two experienced interventional radiologists (H.C.K. and J.W.C., with 14 and 28 years of experience in interventional oncology, respectively). A 5.0 Fr catheter was used for celiac arteriography, and various 1.7 Fr–2.0 Fr microcatheters (Progreat - Terumo, Tokyo, Japan; Carnelian - Tokai Medical Products, Kasugai, Japan; Veloute - Asahi intecc, Aichi, Japan; and Radiostar - Taewoong Medial, Gimpo, Republic of Korea) were used for selective arteriography.
An emulsion of 2-10 ml iodized oil (Lipiodol; Andre Guerbet, Aulnay-sous-Bois, France) and 10-50 mg doxorubicin hydrochloride (Adriamycin RDF; Ildong Pharmaceutical, Seoul, Republic of Korea) was infused into the tumor-feeding branch as selectively as possible. Then additional embolization was performed with absorbable gelatin sponge particles (Gelfoam - Upjohn, Kalamazoo, MI, USA; Cutanplast - Mascia Brunelli, Milan, Italy; or EG gel - Engain, Seongnam, Republic of Korea) to achieve near or complete stasis of the target vessels. For the target vessels involved in tumor thrombosis, 50-100 mg cisplatin was additionally infused at a rate of 2-5 mg/min (11).
Cone-beam CT images were obtained at the hepatic artery and extrahepatic collateral arteries using a flat-panel detector angiography unit (Axiom Artis dTA/VB30, Artis zee or Artis Q; Siemens, Erlangen, Germany). Detailed protocols for cone beam CT were described previously (12).
Analysis. CT scans, digital subtraction angiograms, and cone-beam CT images were retrospectively reviewed and interpreted by consensus between the two authors (M.K, H.C.K), and the patients’ medical records were reviewed by one author (M.K.). Formal radiological reports of CT scans were also reviewed to check whether tumor thrombi in the IVC or right atrium were mentioned.
Because contrast media and unopacified blood were mixed in the IVC and heart, it was not possible to evaluate the size and viability of the tumor thrombi on the hepatic arterial or portal venous phase images. The size and viability of the tumor thrombi were assessed based on the equilibrium phase images. Because conventional chemoembolization was performed, the areas where Lipiodol did not accumulate were considered as viable tumor thrombi. Using the enhanced CT scan of the first follow-up, the response of tumor thrombi in the IVC and right atrium was evaluated and categorized as complete response, partial response, stable disease, or progressive disease. A complete response was observed when Lipiodol accumulated in the whole tumor thrombus in the IVC/right atrium. A partial response was observed when the viable portion (area where Lipiodol did not accumulate) decreased by at least 30% in diameter. Progressive disease was observed when the viable portion increased by at least 20% in diameter. Stable disease was defined as cases that did not exhibit either partial response or progressive disease.
Results
The maximum diameter of the tumor thrombi of the IVC/right atrium ranged from 1.0 to 3.1 cm (median=1.95 cm). Tumor thrombi extended to the IVC (n=5) (Figure 1A) or right atrium (n=5) (Figure 2B and Figure 3A). Tumor thrombi in the IVC joined the IVC above the hepatic vein confluence in nine and below it in one (Figure 1A). Multifocal intrahepatic viable tumors were observed in nine patients, and one patient (patient number 5) had only diaphragmatic thickening without a viable intrahepatic tumor. Diffuse thickening of the diaphragm, suggesting diaphragmatic metastasis, was noted in two patients (numbers 3 and 5) who had radiofrequency ablation for tumors abutting the diaphragm. The feeding artery to the tumor thrombi of the IVC/right atrium was the right inferior phrenic artery in eight cases (Figure 1B and Figure 2C) and right internal mammary artery in two (Figure 3C).
The responses of the tumor thrombi in the IVC/right atrium observed in the first follow-up CT scan were complete response in four (Figure 2D), partial response in five (Figure 1D), and stable disease in one. Linear accumulation of iodized oil along the diaphragm was observed in eight patients (Figure 3E). No patients experienced pulmonary embolism related to thrombi of the IVC/right atrium or right heart failure.
Retrospective review of formal radiological reports of pre-procedural CT scans revealed that a correct diagnosis of tumor thrombi of the IVC/right atrium had been made for only three patients. Myxoma was mentioned in the report for one patient, and tumor thrombi were completely missed in six patients.
Discussion
While the left inferior phrenic vein drains into the IVC, left suprarenal vein, left renal vein, or left hepatic vein, according to an anatomical study of 300 cadavers (13), the right inferior phrenic vein drains into the IVC in most cases. A surgical study of 77 patients revealed that the right inferior phrenic vein ran into the IVC just cranially to the right hepatic vein caval confluence (14). Thus, if the right inferior phrenic vein is involved by tumor thrombi, isolated IVC tumor thrombi without hepatic vein thrombosis can develop in patients with HCC.
Lee et al. reported that IVC tumor thrombi were frequently supplied by extrahepatic collateral arteries such as the right inferior phrenic, adrenal, or internal mammary arteries (3). The right inferior phrenic artery supplies most of the right hemidiaphragm. Because the right inferior phrenic artery has a branch supplying the IVC (15), it is to be expected that IVC tumor thrombi are fed by the right inferior phrenic artery. If the right inferior phrenic artery is obliterated by previous chemoembolization, adjacent collateral arteries, such as the internal mammary artery, may take over blood supply to the IVC tumor thrombi (3). In this study, the right inferior phrenic artery (n=8) and internal mammary artery (n=2) supplied the IVC/right atrial tumor thrombi.
Based on the formal radiological reports, this study found that tumor thrombi in the IVC/right atrium were missed in 6 out of 10 patients. In hepatic artery and portal venous phase images, contrast media in the heart and unopacified blood coming from the hepatic vein and IVC mingle, resulting in heterogeneous enhancement of the suprahepatic IVC. IVC/right atrial tumor thrombi can mimic this feature, which can be easily overlooked by abdominal radiologists. In addition, without hepatic venous thrombosis, abdominal radiologists can easily miss the presence of the tumor thrombi of the IVC or right atrium.
Standard treatment for HCC with vascular invasion is systemic therapy, such as sorafenib or atezolizumab plus bevacizumab (16). However, in actual practice, some patients with advanced-stage HCC, including those with extrahepatic metastasis, are treated using locoregional treatments, such as chemoembolization, because most patients die of intrahepatic tumor progression (17). In this study, six patients had already received sorafenib due to vascular invasion or extrahepatic metastasis. Thus, chemoembolization may play a role in controlling intrahepatic HCC refractory to systemic therapy. In this study, four patients had dense, complete accumulation of Lipiodol in the tumor thrombi. An objective response of the tumor thrombi was observed in nine patients. Thus, chemoembolization may be effective in controlling tumor thrombi in the IVC or right atrium.
This study has some limitations. Firstly, even though linear accumulation of iodized oil along the diaphragm, which is presumed to be a phrenic vein thrombosis, was observed in eight patients, there was no surgical or pathological confirmation of the right inferior phrenic vein thrombosis. Secondly, because of the small study population, statistical analysis of any clinical and radiological findings was not performed.
In conclusion, when HCC invades the diaphragm or phrenic vein, it may manifest as isolated tumor thrombi in the IVC or right atrium without hepatic vein thrombosis. Tumor thrombi can be supplied by the right inferior phrenic artery or internal mammary artery.
Footnotes
Authors’ Contributions
Guarantor of integrity of the entire study: Hyo-Cheol Kim. Study concepts and design: Hyo-Cheol Kim. Literature research: Hyo-Cheol Kim, Minuk Kim. Clinical studies: Hyo-Cheol Kim, Minuk Kim, Jin Wook Chung. Data analysis: Hyo-Cheol Kim, Minuk Kim. Manuscript preparation: Minuk Kim. Article editing: Hyo-Cheol Kim, Jin Wook Chung.
Conflicts of Interest
The Authors have no conflicts of interest to disclose in relation to this study.
- Received July 11, 2021.
- Revision received August 9, 2021.
- Accepted August 25, 2021.
- Copyright © 2021 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.