Abstract
Keywords
Introduction
Percutaneous ablation therapy is a common used method for minimal invasive treatment of liver tumors. Microwave ablation (MWA) is widely accepted as an effective treatment option for hepatic malignancies, especially hepatic metastasis and hepatocellular carcinoma (HCC). MWA uses electromagnetic waves and leads to coagulation necrosis of the surrounding tissue and a complete loss of microcirculation [14]. Several clinical trials proofed the comparable efficacy of thermal ablation and surgical resection [11, 21].
The characteristic feature of all HCC in all contrast-enhanced imaging modalities is the early arterial enhancement combined with portal-venous wash-out [4, 20]. The most important criteria for complete ablation in post-treatment imaging evaluation of HCC is the loss of hypervascular tumor tissue [16]. Frequent findings after ablation of esp. larger tumors are reactive arterial rim enhancement which in most cases disappears over time and small arterio-portal-venous shunts [5, 8].
These findings often make it difficult to differentiate residual hypervascular tumor tissue from reactive changes immediate after ablation using contrast-enhanced CT or MRI due to the limited temporal resolution. Contrast-enhanced ultrasound (CEUS) which is nowadays a common imaging modality to detect and characterize liver lesions may be of advantage in the immediate post-interventional imaging because of its ability to depict the micro- and macro-circulation in real-time [2, 23].
Aim of this study was to compare the diagnostic value of dynamic CEUS and 3-T MRI with liver-specific contrast agent to evaluate ablation success and differentiate residual tumor tissue from reactive changes one day after ablation with long-term follow-up as the gold standard.
Materials and methods
Patients
Between 01.03.2014 and 01.04.2016 30 consecutive patients (5 women, 25 men, mean age 64 years, age range 54 to 73 years) with 30 HCC larger than or equal to 3 cm identified by characteristic imaging features or histopathology underwent CT-guided MWA. 1 male patient was excluded of this study because of an artificial pacemaker. The enrolment criteria for this study were: (1) untreated HCC ≥3 cm (2) no contraindication for MRI or CEUS examination, especially no anamnestic allergy to either of the contrast agents. On day 1 after tumor ablation contrast enhanced ultrasound examination and 3-T MRI with liver specific contrast media Gd-EOB-DTPA (Primovist®, Bayer Schering Pharma, Berlin) were performed in all patients to assess the ablation success. Baseline characteristics of patients and treated lesions are shown in Table 1.
CEUS acquisition
CEUS was performed 1 day after MWA by an experienced radiologist with national DEGUM stage III with a multi-frequency probe (1–5 MHz, LOGIQ E9, GE). The complete liver was examined using conventional B-mode as well as color-coded duplex sonography and native power Doppler. Dynamic CEUS of the ablation zone was conducted with a bolus injection of 2.4 ml highly echogenic sulfur hexafluoride microbubbles (SonoVue®, Bracco, Milan, Italy) followed by 10 ml of a 0.9% saline bolus applying contrast harmonic imaging. Scanning was accomplished in real-time for 3 minutes at minimum starting directly after the injection. Irregular enhancement in the periphery of the ablation zone during early arterial phase, ideally combined with a washout starting during portal venous phase was seen as a characteristic criterion for residual HCC tissue. A uniform peripheral rim enhancement without washout was considered as physiologic response to thermal injury [16]. Wedge-shaped, homogenous arterial enhancement peripheral to the ablation zone with enhancing portal-venous branching structures and without washout was defined as an arterio-portal-venous shunt [9].
MRI acquisition
MRI with liver-specific contrast agent Gd-EOB-DTPA (Primovist®, Bayer Schering Pharma, Berlin) was performed 1 day after MWA using a 3-T MRI (Magnetom Skyra, Siemens Healthcare, Erlangen, Germany) with a combination of body and spine array coil elements (18-channel body matrix coil, 24-channel spine matrix coil) for signal reception. After acquisition of a localizer the following native sequences were acquired: coronary T2-weighted HASTE-sequences (half-fourier acquisition single shot turbo spin-echo) transversal T1 FLASH in- and opposed-phase in breathhold technique transversal T1-weighted volume interpolated breathhold examination (VIBE) 3D-gradient echo sequences with fat saturation transversal T2 BLADE with fat saturation transversal diffusion weighted imaging with ADC-map.
After the intravenous administration of 0.025 mmol per kilogram body weight of Gd-EOB-DTPA with a flow rate of 1 ml/s further imaging was performed using T1-VIBE sequences with fat suppression in the early arterial, arterial, portal-venous and late phase (20 minutes after contrast agent application). Evaluation of all MRI examinations regarding the presence of residual tumor tissue was retrospectively performed by an experienced abdominal radiologist according to the same criteria as for CEUS.
Follow up
All patients underwent follow-up using CEUS and 3-T MRI 6 weeks after percutaneous MWA. If unequivocal residual tumor tissue was seen after 6 weeks, ablation was considered incomplete and patients underwent re-ablation. Otherwise further follow-up examinations with CEUS and MRI were performed every 3 months.
Statistics
The analysis was performed using R (version 3.2.3, R Foundation for Statistical Computing, Vienna, Austria). Descriptive statistics, with a calculation of sensitivities and specificities for CEUS and MRI, were performed following standard methodology [17]. To test for differences in sensitivity and specificity we employed an exact binomial test [24]. To test for differences in complete ablation depending on tumor size we used Fishers exact test. A p value of <0.05 was considered the cut-off point of statistical significance.
Results
Gold standard
The 6-week follow-up, which was considered the gold standard, was available for the ablation success analysis. Complete ablation was achieved in 23 of 29 treated lesions (79%). The number of complete ablations for lesions ≥3 cm but <4 cm and for lesions ≥4 cm are represented in Table 2 with the difference being significant (p = 0.034).
CEUS and MRI imaging results
The results regarding detection of complete ablation for CEUS and MRI are represented in Table 3.
The sensitivity for the diagnosis of remaining tumor tissue was higher for CEUS when compared to MRI without the difference being significant (p = 0.25). The specificity was also higher for CEUS, also without a statistically significant difference (p = 1.0). The exact values for sensitivity, specificity, positive predictive (PPV) and negative predictive values (NPV) are shown in Table 4.
False classifications
All ablations classified incorrect either in CEUS or MRI are listed in Table 5. The main reason for false positive classifications in MRI (incomplete ablations reported as complete) was bad image quality either due to breathing artefacts or ascites. Typical reactive changes like rim enhancement faced a problem for MRI imaging on day 1 after ablation resulting in three complete ablations reported as incomplete (false negatives). One patient showed residual tumor tissue directly adjacent to the diaphragm in segment VIII which was not detected by CEUS (false positive).
Discussion
Percutaneous microwave ablation is a special form of thermal ablation and a minimally invasive treatment modality for the management of hepatic malignancies such as colorectal liver metastasis and HCC. It has been accepted as one of the best treatment options for HCC lesions which are not suitable for transplantation. Multiple randomized trials have proven that survival rates after thermal ablation of HCC are comparable to surgery [1, 12].
Novel techniques such as microwave ablation can be applied to treat large hepatic lesions. MWA uses electromagnetic waves to create coagulation necrosis following a rapid heating of the surrounding tissue. The higher thermal efficiency compared to radiofrequency ablation allows ablation of tumors up to 7 cm in size in a decent time [22] with studies reporting complete ablation rates of 74 to 88% [10, 15] which is similar to our results (79% complete ablations).
Thermal ablation often results in inflammation of the surrounding tissue which appears as arterial periablation rim enhancement in contrast-enhanced imaging [7, 19]. Another common post-ablative perfusion abnormality is the formation of small peripheral arterio-portal-venous shunts as a consequence of thermal damage (Fig. 1). They can be seen as peripheral wedge-shaped areas of arterial enhancement adjacent to the ablation zone [19].
Both facts can complicate evaluation of residual tumor tissue in immediate post-ablative contrast-enhanced CT and MRI imaging especially in large tumors. This is why follow-up imaging to evaluate local effectiveness is performed not less than 4 weeks after the ablation in most centers. On the contrary early detection of residual tumor would be desirable to facilitate retreatment at an early stage [8].
CEUS might be of advantage to distinguish reactive changes in the periablation zone from incomplete ablation because of its ability to visualize micro- and macro-vascularization in real-time [3]. Studies have shown that CEUS can be a valuable aid for immediate assessment of therapeutic success during percutaneous ablation [12]. Other advantages of CEUS compared to MRI and CT are the lower cost of the examination [13] and the good tolerability profile of the contrast agent, esp. for patients with impaired renal function [18].
In our study we have included only patients with HCC lesions larger than or equal to 3 cm because they pose a major challenge in the immediate post-interventional imaging to depict tumor tissue from reactive changes. This is due to the fact that large ablation zones tend to show marked reactive changes in the surrounding area and have a higher rate of incomplete ablations [10, 22].
Our study showed sensitivities and specificities of 100% and 83% for CEUS and 87% and 67% for MRI for the detection of residual tumor tissue on day 1 after percutaneous MWA of large HCC lesions (≥3 cm). Although the differences are not statistically significant, CEUS might be superior in the immediate post-interventional imaging control because of its ability to visualize the perfusion of the periablation zone in real-time. This allows CEUS to identify the characteristic micro- and macro-circulation features of reactive changes like arterial rim enhancement and arterio-portal-venous shunts and differentiate it from residual tumor tissue with characteristic early arterial enhancement combined with portal-venous wash-out.
In our study both imaging modalities, CEUS and MRI, have limitations in the detection of residual tumor tissue but show comparable specificities. In MRI this was due to poor image quality, esp. breathing artefacts and ascites (Fig. 2). CEUS on the other hand was not able to detect residual tumor tissue in only one case where the HCC was located directly next to the diaphragm. Different from the post-ablation MRI examinations CEUS showed no false negative cases. This is hugely important for a potential planning of a re-ablation to prevent tumor progress.
This study is limited by the relatively low number of patients, its retrospective design, the single-center setup and the lack of histological verification of residual tumor tissue. Follow-up examination after 6 weeks with MRI and CEUS was performed in all patients.
Conclusion
In summary CEUS allows a reliable immediate assessment of therapeutic success of percutaneous ablation of large HCC lesions. Its ability to visualize dynamic perfusion changes might be of advantage in the depiction of residual tumor tissue when compared to MRI imaging with its limited time resolution.