Drs. Solbiati , Tonolini, and Cova are with the Department of Radiology, General Hospital, Busto Arsizio, Italy.

Currently, percutaneous radiofrequency (RF) ablation is widely employed for minimally invasive treatment of primary and metastatic liver tumors. Its safety and effectiveness have been demonstrated in several large studies and its advantages include lower morbidity and mortality compared with surgical resection, the possibility of repeated treatment sessions if local recurrence or new lesions develop, and markedly reduced treatment costs and hospital stays. ^1-7 The results of RF ablation for both hepatocellular carcinoma and liver metastases compare favorably with those reported in recent surgical series, although patients are not usually candidates for resection. ^8-10

Recent technical improvements including internally cooled electrodes, ¹¹ cluster electrodes, ¹² pulsed current application, ¹³ and peritumoral saline injection prior to energy deposition ¹⁴ were developed in order to increase the extent of induced coagulation and, therefore, to allow the treatment of larger tumors.

Thermal ablative therapies destroy neoplastic lesions. Coagulation necrosis is achieved through a mechanism of protein denaturation and irreversible inactivation of cellular and mitochondrial enzymes and of nucleic acid-histone complexes. The treatment of hepatocellular carcinoma (HCC) in a cirrhotic liver using RF ablation benefits from the "oven effect": the surrounding, densely fibrotic, and poorly vascularized liver tissue and the fibrous capsule hinder thermal conduction away from the target lesion, and thus maintain optimal heat diffusion in the softer, usually well-circumscribed tumoral nodule. ⁹ Conversely, liver metastases are not encapsulated and tend to infiltrate the surrounding, well-vascularized liver that can act as a heat sink to limit tissue heating. ^6,7 Therefore RF ablation should aim to necrotize not only the metastatic nodule, but also a 0.5 to 1.0 cm thick rim (ie, a surgical "safety margin") of surrounding liver tissue in order to kill infiltrating tumor undetectable by currently available imaging methods and reduce risk of recurrence.

Diagnostic imaging is of paramount importance in all of the four steps of tumor ablation: 1) detection of lesions and selection of patients for treatment; 2) targeting of lesion(s) with optimal positioning of the energy applicator; 3) immediate assessment of therapeutic result; and 4) long-term follow-up. In our experience, the use of contrast-enhanced ultrasound (CEUS) represents a significant improvement in each of these steps and may optimize patients' management and treatment results.

Herewith we report our experience with the use of a second-generation ultrasound (US) contrast agent (SonoVue , sulphur hexafluoride, Bracco, Milan, Italy) and low mechanical index (MI) (or low acoustic pressure) continuous-mode ultrasound, employing a new imaging technique (Contrast Pulse Sequencing [CPS], Acuson/Siemens, Mountain View, CA). Since it has been demonstrated that the detectability of contrast agents at low MIs can be improved by processing received nonlinear energy that exists in the same frequency band as the transmitted signal, CPS processes signals from multiple transmit pulses to extract nonlinearly generated signals in the fundamental frequency. This can improve sensitivity and specificity to contrast agents compared to second harmonic imaging techniques. ¹⁵

Detection of Lesions and Selection of Patients

Early detection and accurate assessment of the extent of neoplastic liver disease at the time of diagnosis or during the course of treatment is crucial to optimal patient management and may result in prolonged survival and improved chance for cure. Patients with chronic liver disease/cirrhosis may be treated with RF ablation for up to 4 to 5 HCC and/or dysplastic lesions, in absence of portal thrombosis and liver function decompensation. Patients with previously treated colorectal or other primary tumors may undergo RF ablation of 1 to 4 metachronous liver metastases, each smaller than 4 cm.

Multiphasic contrast-enhanced helical computed tomography (CT) and dynamic gadolinium-enhanced magnetic resonance imaging (MRI) provide convenient staging of hepatic and extrahepatic neoplastic involvement. Conventional ultrasound, which represents the most widely available low-cost imaging modality for the screening of liver disease, is less accurate than CT and MRI for the detection of focal lesions, particularly of smaller ones (<1 cm).

In our preliminary experience, CEUS with SonoVue and CPS further improved the results achieved with a first-generation contrast agent and color-power Doppler ¹⁶ and with low-MI harmonic imaging. ¹⁷ In the first 3 patients, conventional doses of 2.4 or 4.8 mL of SonoVue were administered in rapid bolus injection, followed by a 3- to 5-mL saline flush. For the remaining 6 patients, progressively decreasing doses of SonoVue (up to 0.6 to 1.2 mL) were given and no significant impairment of image quality and sensitivity was found. The whole vascular phase was studied in all patients, consisting of arterial (15 to 25 sec following the injection, with some delay in cirrhotic patients), early portal (45 to 60 sec) and full portal phase (90 to 240 sec).

For local staging of both primary and metastatic liver cancer, CEUS with CPS demonstrated sensitivity rates equal to that of helical CT in 7 (77.8 %) of 9 cases and in the remaining 2 (22.2%) cases, it was superior to helical CT, particularly for infracentimetric lesions (Figure 1).

During pretreatment CEUS, images and/or movie clips were stored digitally in order to "map" the lesions to be targeted during the operating session and to compare pretreatment findings with posttreatment enhanced images.

RF Procedure and Targeting of Lesions

Although ultrasound represents the modality of choice for guiding ablative procedures, inherent limitations of conventional US (related to small size, poor conspicuity, and inhomogeneity of liver parenchyma) may hinder the targeting of focal tumors. In these occurrences, CEUS with CPS is repeated in the interventional room after inducing anesthesia and the electrode is inserted during the vascular phase in which the maximum lesion conspicuity is achieved, eg, in the arterial phase for hypervascular lesions such as HCC and in the portal phase for hypovascular metastases. In our experience with CPS, 11 malignancies in 6 patients were punctured during the vascular phases, with 100% targeting accuracy.

All the 17 malignancies were treated using RF ablation: either single or clustered 17G cool-tip ablation electrodes connected to an RF generator (CC1, Radionics, Burlington, MA) were inserted into the targets percutaneously, under general anesthesia for 12 to 24 minutes.

Assessment of Therapeutic Response

Whereas most RF ablation series report a high rate of apparently complete tumor necrosis on initial postablation evaluation, a moderate rate of local recurrences probably resulting from a lack of radicality may occur. ¹⁸ Achieving only partial necrosis implies the need to perform retreatments with increased costs, patient discomfort, greater technical difficulties, and a higher rate of failure. ⁸ When the first RF treatment has not eradicated the tumor effectively, it is extremely difficult to differentiate active tumor from coagulation necrosis, and therefore to target residual tumor foci. ⁸

Ultrasound provides real-time guidance of ablation treatments, but sonographic and color/power Doppler findings observed during the ablation procedure provide only a gross estimate of the extent of induced coagulation necrosis and, therefore, are not useful to reliably assess the completeness of the treatment. Furthermore, additional repositioning of the RF electrode may become obscured because during the application of thermal energy a progressively increasing hyperechogenic "cloud" due to gas microbubble formation and tissue vaporization appears for 5 to 8 minutes around the distal probe. The most important imaging finding that suggests complete treatment of a focal liver tumor is the disappearance of any previously visualized vascular enhancement on contrast-enhanced images. ¹⁹ This assessment is usually achieved using either biphasic helical CT or dynamic gadolinium-enhanced MRI. Radiologic-pathologic correlation studies demonstrated that both modalities can predict the extent of coagulation area to within 2 to 3 mm. ³

The lack of neoplastic or parenchymal contrast enhancement in any vascular phase throughout the entire lesion is the hallmark of complete ablation. This evaluation is easily feasible for hypervascular HCCs, whereas for hypovascular metastases the confident assessment of complete ablation can only be inferred based upon the necrosis volume exceeding the original lesion and a 0.5 to 1 cm "safety margin" in every diameter. Unfortunately, neither CT nor MRI can be performed during the ablative procedures under general anesthesia, unless treatment is performed under CT or MR guidance.

First-generation contrast agents with color Doppler and power Doppler ^16,20,21 and, more recently, with pulse inversion sonography ^22,23,24 have been used successfully to evaluate the therapeutic response to ablative treatments, demonstrating the capability to detect persistently perfused tissue consistent with residual viable tumor, with a reported accuracy for detection of incompletely ablated areas ranging from 75% to 82%.

Our preliminary experience with CPS and SonoVue was even more encouraging. In 9 patients, 17 liver malignancies (9 metastases from colorectal cancer and 8 primary liver tumors) ranging in size from 1.0 to 3.9 cm, preliminarly studied with CPS and SonoVue, underwent immediate postablation evaluation using the same modality, 5 to 10 minutes after the assumed completion of the RF session, with the patient still under general anesthesia. Comparison of immediate postablation images with stored pre-ablation scans was performed in all cases.

For 15 of the 17 lesions, no residual enhancement was demonstrated either inside the or at the periphery of the tumors (Figures 1 and 2). These lesions did not undergo further treatment and CT scans performed the following day confirmed the complete treatment in all 15 tumors. No immediate CEUS-guided targeted retreatment was needed. In the remaining 2 malignancies, no residual enhancement was found with CEUS, but the following CT scan depicted few tiny intralesional enhancing areas suggesting local regrowth.

Conclusion

Based on our experience, CEUS with CPS and a second-generation US contrast agent is extremely accurate for the detection, local staging, CEUS-guided ablation, and immediate assessment of the therapeutic results. This simplifies patient management and reduces costs by decreasing the number of RF procedures and follow-up examinations. *