Материал: Advanced Imaging of the Abdomen - Jovitas Skucas

have at most three tumors measuring <3cm in diameter, and should have no portal extension. Transplantation rather than resection is considered for localized tumors in a cirrhotic liver where a sizable resection risks liver failure.

Is surgical resection or transplantation preferred in a patient with cirrhosis and a hepatocellular carcinoma? Data from six French medical universities reveal an overall similar 5-year survival rate for cirrhotic patients with hepatocellular carcinomas whether treated by liver resection or transplantation (114): 31% for the resection group and 32% for the transplantation group. Among these patients the 5-year survival rate without recurrence, however, was higher in the transplantation group (60%) than in the resection group (14%).

Chemoembolization

Intraarterial embolization techniques used to treat hepatocellular carcinomas consist of inert particle embolization, a variety of embolic material containing chemotoxic agents (chemoembolization), and injection of radioactive particles (radioembolization). Of these, chemoembolization has achieved the greatest acceptance.

As minimally invasive outpatient therapy, transarterial chemoembolization offers considerable palliation for patients with unresectable hepatocellular carcinomas. Undoubtedly in the future such therapy will expand. Treatment strategies adopted probably should depend on size and tumor number. Thus percutaneous ethanol injection appears superior for a single small tumor, but intraarterial chemoembolization is superior for more or larger tumors.

Chemoembolization consists of injecting a mixture of a chemotoxic agent and also an embolic material into a tumor feeding artery (a procedure often called transcatheter arterial chemoembolization, TACE). Ethiodol (Lipiodol) is included as one of the chemoembolization ingredients for several reasons: First, this oil acts as a chemotherapeutic agent carrier. Second, because of its relatively high viscosity, it acts as a temporary embolizing agent and thus prolongs chemotherapeutic agent contact with a tumor. In addition, Ethiodol remains within tumor neovascularity much longer than in adjacent normal liver parenchyma, at times for

ADVANCED IMAGING OF THE ABDOMEN

up to 7 days (Fig. 7.35). Later, after injecting a chemotherapeutic agent, if needed, embolization of the vessel in question is performed using particulate matter.

Although mostly supplied by hepatic artery branches, an occasional parasitic blood supply is detected with these tumors. Especially with a carcinoma located ventrally beneath the diaphragm, internal mammary arteries and inferior phrenic arteries serve as feeding arteries. If necessary, Lipiodol embolization of one of these arteries can be performed.

Temporary occlusion of a draining hepatic vein may aid arterial infusion chemotherapy. Such selective hepatic vein occlusion results in a dense hepatogram, but this issue is controversial, and some investigators believe that the occlusion causes unacceptable damage to normal liver parenchyma.

Transarterial chemoembolization can be repeated several times (called sequential chemoembolization). Some data suggest that such sequential therapy prolongs survival compared with no therapy in those with an unresectable carcinoma. At times a port system is implanted as an aid for future hepatic artery infusion chemotherapy. An infusion pump provides ready access for follow-up CT arteriography or conventional arteriography.

The most common indication for chemoembolization is palliation. At times, reduction in tumor size allows resection or additional

Figure 7.35. CT shows residual Ethiodol retained in a multifocal hepatocellular carcinoma. (Courtesy of Patrick Fultz, M.D., University of Rochester.)

LIVER

therapy, such as laser-induced thermotherapy or percutaneous ethanol injection. Initial results suggest that such a dual approach is more effective than chemoembolization alone, although survival depended considerably on the extent of underlying liver disease. With a resectable tumor, transcatheter arterial chemoembolization during the postoperative period appears to increase survival, but insufficient data are available on this subset of patients.

Portal vein thrombosis is considered a contraindication to transcatheter embolization, although embolization has been performed in patients with portal vein branch thrombosis. Arteriovenous shunting, common with hepatocellular carcinomas, is generally considered a contraindication to arterial chemoembolization. At times arteriovenous shunting, mostly arterioportal, develops during therapy. Other risk factors include coagulation abnormalities, severe portal or pulmonary hypertension, esophageal varices prone to bleeding, poor cardiovascular status, and advanced cirrhosis.

A severe postembolization syndrome and hepatic insufficiency are common complications; in general, predisposing factors—such as major portal vein obstruction, compromised hepatic reserve, biliary obstruction, excessive or nonselective iodized oil embolization— are evident in most patients with major complications. Other complications include those related to catheter placements, variceal bleeding, and severe hyperglycemia. Rare complications include carcinoma rupture after chemoembolization, adrenal hemorrhage, and acute adrenal insufficiency and ischemic cholecystitis. Procedure-related deaths have been reported. Tumor seeding along an implantable access port in the hepatic artery has developed.

Transcatheter arterial embolization affects arterial and portal perfusion hemodynamics. Arterial perfusion increases temporarily after embolization, presumably due to acute inflammation.

Biliary complications can be evaluated with Tc-99m–HIDA. Postembolization scans reveal that gallbladder filling time and contractility is worse shortly after embolization. Also, curiously, most survivors have gallbladder nonvisualization.

Transarterial chemoembolization is most effective when a tumor is limited in size and

liver function is preserved. The response to chemoembolization is poor with extensive parenchymal invasion. Postprocedure survival depends on underlying liver disease (Child’s class), size and number of tumors, and patient age. Patients with a predominantly hypervascular tumor respond better than those with a hypovascular tumor (115).

Published survival results vary, and a comparison of different studies using different selection criteria is difficult. Types of agents employed together with their doses and administration rates also differ and presumably influence results. In a typical study, among patients responding to therapy, survival was 90%, 67%, and 36% at 1, 2, and 3 years, respectively, while nonresponders (mostly with hypovascular tumors) achieved 70%, 17%, and 10% survival, respectively (115). Survival is better in patients with smaller tumors. Tumor decrease in size after arterial chemoembolization is of prognostic significance. Nevertheless, not all results are positive. Thus a multicenter randomized trial of patients with unresectable hepatocellular carcinoma but without severe liver disease or portal vein occlusion compared Lipiodol chemoembolization (Lipiodol, cisplatin, lecithin, and gelatin sponge injected into the hepatic artery) plus tamoxifen versus tamoxifen alone (116); although an objective response was more frequent in the Lipiodol group (24%) than in the tamoxifen group (5.5%), overall the two groups showed no difference in survival at 1 year. In a study comparing hepatic artery infusion chemotherapy versus transcatheter arterial Lipiodol chemoembolization in patients with advanced hepatocellular carcinomas, tumor response rates were significantly higher in the former group and these patients tended to have longer survival rates (117).

Some tumors are difficult to evaluate with CT after embolization due to retained iodized oil artifacts. Nevertheless, that portion of tumor retaining iodized oil after chemoembolization is typically necrotic.

Both contrast-enhanced wide-band harmonic gray-scale US and power Doppler US are used to detect tumor recurrence after chemoembolization, achieving about 90% sensitivity and specificity. Tumor vascularity detected by contrast enhanced harmonic wideband gray-scale US after arterial chemoembolization usually implies residual tumor (118).

Postgadolinium T1-weighted MRI is reliable in evaluating chemoembolization outcome and is more accurate than precontrast MRI. Loss of enhancement postgadolinium T1-weighted images correlated with necrosis. On the other hand, focal enhancement generally signifies viable tumor.

Tc-99m-MAA SPECT angiography (hepatic arterial flow study) provides an estimate of liver vascularity and is another method of gauging the success of arterial chemoembolization.

Radioembolization

Intrahepatic artery Tc-99m-MAA injection estimates the relative tumor-to–normal parenchyma uptake of various size microspheres, including those containing therapeutic radiopharmaceutical agents. Theoretically, intraarterial infusion of microspheres containing the beta-emitter yttrium 90 selectively delivers a high radiation dose to a malignant hepatic tumor while mostly sparing adjacent normal parenchyma. One problem is that hepatocellular carcinomas have a wide tumor- to–normal parenchymal uptake ratio. Tumor vascularity, as assessed by hepatic angiography, does not predict the relative tumor-to– normal parenchyma uptake for most of these tumors.

Hepatocellular carcinomas have been embolized with iodine-131 iodized oil (radioiodinated Lipiodol) and gelatin sponges through a superselectively placed arterial catheter. Such embolization has provided long-term palliation without complications, but few conclusions can be drawn from the limited number of patients studied.

Yttrium-90 glass microspheres have also been injected percutaneously into hepatocellular carcinomas, but this technique is controversial.

Percutaneous Techniques

Hepatocellular carcinomas can be treated using a direct percutaneous approach. Such therapy can be divided into two broad categories; first, injection of chemical agents such as ethanol, acetic acid, and hot saline induce tumor coagulation necrosis. Second, thermally mediated techniques such as radiofrequency (RF) ablation, laser photocoagulation, microwave

ADVANCED IMAGING OF THE ABDOMEN

therapy, and cryotherapy result in similar tumor cell death.

Imaging provides needle and catheter insertion guidance. The introduction of open-bore MR units and nonferromagnetic instrumentation should expand interventional MR techniques.

Although comparisons of various percutaneous tumor ablation techniques are being published, they should be viewed cautiously. Type of tumor, tumor size and specific techniques used influence results. Thus a conclusion that, say, radiofrequency ablation achieves necrosis in 90% of tumors versus 80% with percutaneous ethanol injection may not be valid at other institutions.

Ethanol Injection: Percutaneous ethanol injection into unresectable hepatocellular carcinomas nodules has had wide international application, especially in patients with underlying cirrhosis, yet has been little used in the United States. Results are mostly palliative and in most patients tumors either recur or new tumors develop. Therapy is repeated with tumor recurrence, with the interval between successive treatments decreasing. A solitary nodule, pretreatment serum a-fetoprotein level <20ng/ mL, and limited cirrhosis favor longer survival. Severe complications of percutaneous ethanol injection are uncommon.

Ethanol therapy is achieved by cell membrane lysis and protein denaturation and through necrosis of vascular endothelium leading to surrounding tissue ischemia and necrosis. Absolute alcohol has a very low CT attenuation and injected alcohol can be identified on CT images. No blood flow is identified in necrotic tumors, and they are avascular at color Doppler US.

Using color Doppler US guidance, instead of injecting directly into a tumor, ethanol can be injected into a vessel supplying the tumor; limited data are available for such a technique.

Positron emission tomography performed by injecting carbon-11 ethanol via a percutaneous needle reveals high C-11 uptake in these carcinomas; no significant elimination of C-11 ethanol is evident from these tumor.

Accurate tumor localization is, of course, important. Some can be located with US. Others require CT arteriography or arterial injection of iodized oil and real-time CT fluoroscopy

LIVER

during needle placement. Computed tomography fluoroscopy permits rapid assessment of needle position. Also, real-time CT fluoroscopy reconstruction evaluates ethanol distribution during injection, and the injected dose can then be modified accordingly.

Study technical variables, such as accuracy of needle placement, amount injected, and number of sessions, influence survival and make comparison of published studies difficult. Also, inhomogeneous drug distribution and dilution limit therapeutic response. Survival results are modified by tumor size and underlying liver disease. In some studies of patients with a single lesion <3cm in diameter, the results of percutaneous ethanol injection approach those reported for resection.

Computed tomography performed shortly after ethanol injection therapy reveals an unchanged or even increased tumor diameter with a well-defined margin; necrotic tissue tends to be hypodense, while residual tumors enhance during the arterial phase, gradually becoming hypodense on later phases. These changes are less pronounced after multiple therapy sessions.

After ethanol injection, arterial phase con- trast-enhanced MRI reveals increased contrast enhancement adjacent to treated tumors, presumably due to increased blood flow to surrounding tissues; similarly to US, early tumor contrast enhancement implies residual tumor.

Acetic Acid Injection: Percutaneous acetic acid injection using acid concentrations of 15% to 50% has been used for hepatocellular carcinoma nodule therapy. The number of treatment sessions needed is less with the higher concentrations, although even a 15% concentration appears adequate to successfully treat tumors. A single percutaneous injection into nodules <3.0cm in diameter has resulted in no local recurrence of most nodules (119).

The 1-, 2-, and 3-year survival rates for patients with hypervascular hepatocellular carcinomas <3cm in diameter treated with percutaneous acetic acid injection were 100%, 94%, and 83% (120); corresponding survival rates for those treated with transcatheter arterial embolization were 72%, 65%, and 39%, respectively. Local recurrence, as gauged by enlargement of original tumor, occurred in 3% of

tumors treated with acetic acid injection and 50% of those treated by transcatheter arterial embolization.

Other Injection Techniques: Less often used is hot saline injection. Usually a larger volume of hot saline is injected than ethanol; thus fewer treatment sessions are required. Currently insufficient data exist to draw meaningful conclusions about such therapy efficacy.

Computer tomography–guided percutaneous intratumoral injection of a cisplatin/epinephrine gel (part of a clinical phase II study) in eight patients with hepatocellular carcinomas led to a local control rate of about 80% (121); a similar technique in eight patients with metastases achieved a control rate of only 38%.

Radiofrequency Coagulation: Radiofrequency ablation, laser-induced photocoagulation, and microwave therapy induce tumor thermal coagulation necrosis (thermotherapy). Radiofrequency energy is applied via shielded needle electrodes inserted into a tumor to deliver sufficient energy to induce tumor necrosis. Percutaneous needles are inserted under CT, US, or MR control. Alternate approaches are laparoscopic or laparotomic, with each one having advantages and disadvantages. Laparotomy is especially useful for tumors close to other critical organs, such as the diaphragm. The extent of ablation depends on needle tip size and energy delivered. The greatest tissue heating occurs closest to the needle tip, a disadvantage overcome by the use of continuous needle tip cooling, which results in tissue heating away from the needle.

Typically several sessions are needed for tumor ablation. Published nodule necrosis has ranged from 50% to 85% of tumors; the most sensitive sign of necrosis is lack of nodule contrast enhancement during follow-up arte- rial-phase CT (Fig. 7.36), with contrastenhanced pulse inversion harmonic US and contrast-enhanced power Doppler US being less sensitive (122), although some studies suggest that contrast-enhanced, phase-inversion harmonic US is almost as accurate as CT in detecting tumor necrosis (123). Doppler US reveals a rich peripheral vascularity persisting even after successful therapy.

At times radiofrequency ablation is combined with intraarterial chemoembolization. The