Prospective evaluation of the safety and efficacy of microwave ablation for hepatocellular carcinoma in Korean patients

Compliance with Ethical Standards

This study was approved by the institutional review board of Samsung Medical Center (IRB File No. 2017-09-086), and written informed consent was obtained from all participants. The study was conducted in accordance with the ethical principles of the Declaration of Helsinki.

Study Design

This prospective, single-arm clinical trial evaluated the 3-year safety and efficacy of MWA using the NEUWAVE system in Korean patients with early-stage HCC. As the first clinical study of the NEUWAVE system in Korea, it was designed as a preliminary feasibility investigation. Based on clinical judgment and practical considerations, the investigators determined that a full analysis set (FAS) of 30 patients would be appropriate to assess procedural safety and preliminary efficacy over a 3-year follow-up period. To ensure this target was achieved, 33 patients were enrolled to allow for potential dropouts or protocol deviations.

Participants were enrolled at Samsung Medical Center (Seoul, Korea) between December 2017 and May 2020. The study followed a structured schedule of 11 visits over a 3-year period, including screening, ablation, and follow-up assessments, with flexibility for additional unscheduled visits as needed. The protocol adhered to institutional standards for local ablation therapy of liver tumors, ensuring consistency in treatment procedures and follow-up evaluations.

Patients

Eligibility criteria included adults aged 19 years or older with a confirmed diagnosis of HCC, characterized as a single tumor measuring 2–5 cm in diameter. All tumors were classified as Barcelona Clinic Liver Cancer stage A based on imaging and biopsy findings. Patients were either treatment-naive or had recurrent HCC after prior local ablation or surgical resection. Additional requirements included a scheduled MWA procedure, an Eastern Cooperative Oncology Group performance status of 0–2, Child-Pugh class A or B hepatic function, and an American Society of Anesthesiologists Physical Status Classification score below 3.

Exclusion criteria included active bacterial or fungal infection, recent systemic steroid use, or chemotherapy or radiation therapy within 30 days before the procedure. Patients with implantable electronic devices, scheduled liver surgery, or coagulopathy (platelet count <50,000/mm³ or international normalized ratio >1.5) were also excluded. Additional exclusions were renal failure requiring dialysis, pregnancy or breastfeeding, and participation in other clinical studies within the past 3 months. Inability to attend follow-up visits or conditions that could impede protocol adherence were further grounds for exclusion.

MWA Procedure

Percutaneous MWA was performed under monitored anesthesia care supervised by anesthesiologists. Four interventional radiologists (H.R., M.W.L., T.W.K., and K.D.S.) with at least 5 years of experience in RFA for liver tumors performed the procedures. Each radiologist received pre-enrollment training specific to MWA, including supervised observation of procedures. The antenna was placed in the target lesion under the guidance of fusion imaging (LOGIQ E9, GE Healthcare, Chicago, IL, USA), combining real-time ultrasound with pre-acquired computed tomography (CT) or magnetic resonance imaging (MRI). The number and type of antennas were determined by tumor size and morphology, as well as operator preference.

The antenna type—PR (perfusion-reducing) or LK (large-zone ablation kit)—was selected by the operating radiologist in accordance with institutional standards, based on tumor size, location, and morphology. In general, the PR antenna was used for smaller or superficially located lesions requiring precise ablation with limited propagation beyond the tip, whereas the LK antenna, which delivers higher power and creates larger ablation zones, was preferred for deeper or larger tumors necessitating broader treatment coverage.

The NEUWAVE Certus 140 Microwave Ablation System (NEUWAVE Medical, Madison, WI, USA) was used. The system operates at a frequency of 2.45 GHz and includes three independent amplifiers, each capable of delivering up to 140 W of power. Up to three antennas could be activated simultaneously, with power settings adjusted according to the number and type of antennas used, as well as tumor characteristics. Most ablation sessions used settings of 60–100 W per antenna. The system incorporates a CO₂-based cooling mechanism and a Tissu-Loc function to stabilize antenna placement and improve targeting accuracy. Continuous reflected power monitoring enhances safety by enabling immediate cessation of energy delivery in response to impedance changes, thereby minimizing unintended tissue injury. For each procedure, the number of antennas used, and the frequency of repositioning were recorded to assess technical efficiency and lesion coverage. Artificial fluid was introduced when necessary to protect adjacent organs or improve the sonographic window. A 5% solution of dextrose in water was instilled via a percutaneous catheter under ultrasound guidance, and the total instilled volume and the indication for use were recorded for each case.

Contrast-enhanced (CE)–MRI or CT was performed after the procedure to confirm complete tumor ablation and assess for procedure-related complications.

Efficacy Endpoints

The primary efficacy endpoint was technical success, defined as complete coverage of the tumor by the ablation zone as confirmed by CE-MRI or CT immediately after the procedure [19]. Secondary efficacy endpoints included technique efficacy—defined as complete tumor ablation confirmed on CE-CT or MRI at the 1-month follow-up—as well as LTP, progression-free survival (PFS), and overall survival (OS) [19,20].

Treatment responses, including LTP, PFS, and OS, were evaluated at multiple time points (3, 6, 9, 12, 18, 24, 30, and 36 months) after ablation (Supplementary Table 1). Immediate CE-CT and MRI were performed in all patients within 4 days after ablation (visit 2B/2C) to confirm complete tumor coverage and assess procedure-related complications. At the 1-month follow-up (visit 3), both CE-CT and MRI were repeated to evaluate technique efficacy. Thereafter, CE-CT was performed at each scheduled follow-up visit, while MRI could be obtained instead of CT once per year during the 3-year follow-up, as appropriate, based on the investigator’s clinical judgment and the patient’s condition. LTP was defined as the development of new tumor foci at the ablative margin following initial tumor eradication with MWA [19]. PFS was measured as the time from ablation to the first occurrence of any recurrence (local, intrahepatic, or distant) or death, whichever occurred first [20].

All CT and MRI examinations obtained throughout the study were independently reviewed by an experienced radiologist (S.H.K.) to ensure objective assessment of image-based endpoints, including technical success, technique efficacy, and LTP.

Safety Assessments

Safety was assessed independently of efficacy endpoints. Adverse events (AEs) were monitored from the day of the procedure through the 3-year follow-up period. AEs were defined as any untoward medical occurrence, including newly arising events or worsening of pre-existing conditions, regardless of any direct association with the MWA procedure.

All reported AEs were reviewed and classified by the treating physician according to their relationship to the MWA procedure and/or device. Severity was evaluated using the Society of Interventional Radiology (SIR) classification, with AEs categorized as minor or major. Minor AEs were defined as transient events requiring minimal or no intervention, whereas major AEs were those that required significant medical intervention, prolonged hospitalization, or resulted in lasting disability or death. The cumulative incidence of AEs was calculated and summarized at 3 and 36 months post-ablation, including a detailed breakdown of AE types and frequencies, severity, and potential association with the MWA procedure or device.

Data Analysis

Two distinct analysis sets were established for statistical evaluation: the FAS and the safety analysis set (SAF). The FAS included all patients (n=30) who underwent MWA according to the study protocol and had sufficient data available to assess technical success and technique efficacy; this set formed the basis for evaluating baseline characteristics and efficacy endpoints. The SAF comprised all patients (n=33) who underwent MWA, regardless of protocol adherence or completeness of efficacy data, and was used exclusively for safety endpoints, including the frequency, severity, and classification of AEs.

Summary statistics were calculated for all endpoints, tailored to continuous or categorical data as appropriate. Technical success and technique efficacy rates were summarized as the proportion of treated tumors meeting the predefined criteria. Cumulative LTP rates were estimated at 1-, 2-, and 3-year intervals using the Kaplan-Meier method. Kaplan-Meier analysis was also used to calculate cumulative PFS and OS rates over the 36-month follow-up period.

AEs were summarized descriptively, including frequency, severity, and classification. The proportion of patients experiencing at least one procedure- or device-related AE was reported, and individual events were categorized as minor or major using the SIR classification.

Datasets

The distribution of patients across the enrollment (ENR), FAS, and SAF datasets is shown in Fig. 1. A total of 33 patients provided informed consent and underwent MWA, comprising the ENR and SAF datasets. Three patients were excluded from the FAS due to protocol deviations or incomplete data for evaluating technical success and technique efficacy. Thus, the FAS included 30 patients who met all protocol criteria and completed the procedure as intended.

Fig. 1.

Study flowchart and dataset classification.

^a)The enrollment set (ENR) included all patients (n=33) who met the inclusion/exclusion criteria and provided informed consent. Percentages are calculated based on the total number of enrolled patients. ^b)The safety analysis set (SAF) included all patients who underwent microwave ablation (MWA). ^c)The full analysis set (FAS) consisted of patients treated with MWA per protocol who had complete technical success and technique efficacy data. Five patients were censored during follow-up due to death (n=3) or withdrawal (n=2).

During the study period, five patients (15.2%, 5/33) discontinued early, including three who died (9.1%) and two who withdrew from the study (6.1%). Consequently, 28 of the 33 patients (84.8%) completed the 3-year follow-up, providing comprehensive intermediate-term data for analysis.

Patient Characteristics

The FAS dataset included 30 Korean patients. Twenty-five were men (83.3%, 25/30) and five were women (16.7%, 5/30), aged 45 to 79 years (Table 1). Liver function was predominantly well preserved, as reflected by the Child-Pugh classification, with 93.3% (28/30) of patients categorized as Child-Pugh class A and only 6.7% (2/30) as Child-Pugh class B. The most common etiology of liver disease was hepatitis B, accounting for 50.0% (15/30) of cases, followed by hepatitis C (16.7%, 5/30), alcohol-related liver disease (16.7%, 5/30), and other causes (16.7%, 5/30).

Table 1.

Baseline characteristics of patients

Target Lesion Characteristics and Procedural Aspects

Tumor characteristics and procedural details are summarized in Table 2. A total of 30 HCCs were treated with MWA, corresponding to one lesion per patient. The median tumor size was 2.3 cm (range, 2.0 to 4.4 cm). Tumors were most frequently located in liver segment 8 (30.0%, 9/30), followed by segment 4 (20.0%, 6/30), segment 7 (16.7%, 5/30), and segments 5 and 6 (each 13.3%, 4/30). Segment 3 had the fewest tumors (6.7%, 2/30), and no tumors were identified in segments 1 or 2.

Table 2.

Tumor characteristics and details of the MWA procedure

Approximately 46.7% (14/30) of tumors were adjacent to critical structures such as blood vessels or bile ducts. Additionally, 16.7% (5/30) were subphrenic, while more than half (53.3%, 16/30) were in a non-subphrenic subcapsular location.

The median duration of the MWA procedure was 19.0 minutes, with a median lesion ablation time of 7.8 minutes. The median post-ablation hospital stay was 2 days. The PR antenna was used in 96.7% (29/30) of procedures, while the LK antenna was utilized in 3.3% (1/30).

Treatment Efficacy

Technical success and technique efficacy were achieved for all treated tumors (100.0%, 30/30) after a single ablation session. Two antennas were used in most cases (93.3%, 28/30). Antenna repositioning was performed in 11 procedures: once in six cases (20.0%) and twice in five cases (16.7%). These details are summarized in Table 2. Artificial ascites was employed in 15 patients (50.0%)—to prevent collateral thermal injury in 10, to improve the sonographic window in four, and for both purposes in one. The median instilled volume was 850 mL (range, 200–1,000 mL) of 5% dextrose in water. No cases required artificial pleural effusion.

During the follow-up period (median, 36.2 months; range, 17.8 to 39.1 months), two patients experienced LTP: at 23.8 months and at 29.0 months post-ablation. Both cases were successfully managed, the first with RFA and the second with transarterial chemoembolization. Local tumor control was sustained for the remainder of the study. The cumulative LTP rates at 1, 2, and 3 years were 0.0%, 3.6%, and 7.3%, respectively.

Intrahepatic distant recurrence occurred in 15 patients (50.0%, 15/30) over the 3-year follow-up period. Of these, two patients died of disease progression, while another patient died of a stroke unrelated to tumor recurrence. The cumulative PFS rates at 1, 2, and 3 years were 86.7%, 60.0%, and 38.5%, respectively; the corresponding OS rates were 100.0%, 96.7%, and 90.0%.

Adverse Events

A total of 53 AEs were reported during the study, with detailed classifications shown in Table 3. Nearly half (49.1%, 26/53) were related to the MWA procedure, while 5.7% (3/53) were attributed to the MWA device. Device-related AEs comprised two thermal burns and one hepatic infarction; all three were also considered probably procedure-related due to uncertain causality. These events were categorized as device-related because a potential link to energy delivery or antenna performance could not be excluded. In contrast, procedure-related AEs—such as post-ablation pain, fever, or duodenal thermal injury—were associated with the ablation process itself or patient factors rather than device function. The remaining events were considered unrelated. Among the 33 patients in the SAF dataset, 22 (66.7%) experienced at least one AE. Procedure-related AEs occurred in 18 patients (54.5%) and device-related AEs in two patients (6.1%). The most frequently reported AE was procedural pain (45.5%, 15/33), followed by fever (12.1%, 4/33) and nausea (9.1%, 3/33). These events were classified as minor and resolved without long-term effects.

Table 3.

Procedure- and device-related AEs

According to the SIR classification, two patients experienced major procedure-related AEs: one case of pyrexia (3.0%, 1/33) and one case of duodenal thermal injury (3.0%, 1/33). Both required medical intervention and extended hospitalization but were managed successfully, with no lasting complications (Table 4, Fig. 2).

Table 4.

Categorization of AEs based on SIR classification

Fig. 2.

Pre- and post–microwave ablation (MWA) imaging for a 2.3-cm hepatocellular carcinoma in a 59-year-old man with alcohol-related liver disease who experienced a major procedure-related adverse event.

Serum alpha-fetoprotein and protein induced by vitamin K absence-II levels were elevated to 17.7 ng/mL and 58.0 mAU/mL, respectively. A. Hepatobiliary phase magnetic resonance (MR) image obtained 3 weeks before ablation shows a well-defined mass (arrow) in segment 4 of the liver. B. B-mode ultrasound image obtained during MWA reveals a slightly hypoechoic mass (arrow). C. Two NEUWAVE perfusion-reducing antennas were positioned in parallel, and ablation was performed simultaneously at a power of 65 W. D. After 4 minutes of ablation at 65 W, the tumor was entirely covered by dense echogenic clouds. E. Arterial-phase MR obtained immediately after the procedure demonstrates complete tumor eradication with an adequate ablative margin; however, adjacent duodenal wall thickening is visible (arrowheads). F. Endoscopy performed 5 days after MWA shows an approximately 5-cm ulcerative lesion with surrounding hyperemia. G. One-month follow-up computed tomography images demonstrate resolution of the wall thickening, with no evidence of perforation.