Stereotactic image‐guided microwave ablation of hepatocellular carcinoma using a computer‐assisted navigation system

Ablation plays an important role in the treatment of hepatocellular carcinoma. Because image‐guided navigation technology has recently entered the clinical setting, we aimed to analyse its safety, therapeutic and procedural efficiency.


| INTRODUC TI ON
Hepatocellular carcinoma (HCC) is the second most common cause of cancer-related death and the main cause of death in patients with cirrhosis. 1 Curative treatment options for patients with early stage HCC include surgical resection, orthotopic liver transplantation (OLT) and local ablation. 2 However, only a minority of patients qualify for curative treatments because of advanced liver disease, intrahepatic tumour distribution or comorbidities. 3 Even potentially "ideal" transplant candidates may not go on to have the surgery they would actually require, with many transplant centres rarely adopting OLT as a first-line option for very early or early HCC. 4 Furthermore, as resection or ablation do not preclude salvage OLT in case of treatment failure, the use of initially less invasive strategies is often favoured. 5,6 Indeed, with 5-year overall survival rates of up to 75%, ablation has been validated as an alternative to resection for very small tumours (Barcelona Clinic Liver Cancer classification [BCLC] stage 0) 7 and patients with limited disease non-resectable because of associated liver disease (BCLC stage A). 8 In addition, ablation has shown promising results within combination therapies for advanced disease 9,10 and is commonly used as bridging to OLT (as recommended by the current AASLD and EASL guidelines). [11][12][13] The major advantage of ablation lies in its minimalinvasive tissue-sparing yet locally destructive character, with short hospital stays and low morbidity. 14 Although radiofrequency ablation (RFA) is used mostly for HCC, microwave ablation (MWA) has recently entered the clinical field and is now increasingly used because of shorter ablation times, a potential advantage for larger tumours and reduced heat sink effect, which occurs when hepatic blood flow reduces the local heat distribution by its cooling effect. 15 Next to tumour burden and the extent of the underlying liver disease, initial complete response represents an independent predictive factor for recurrence-free and disease-specific survival in patients undergoing ablation for HCC. 16 Hence, the precision of ablation might represent a crucial factor for treatment success, relying directly on optimal visualization of intrahepatic lesions and accurate placement of ablation probes within the tumour targets.
To optimize guidance of the ablation probe and thus precision in tumour targeting and ablation, advanced stereotactic navigation technology has been introduced, and several navigation systems are now available for image-guided interventions. [17][18][19] First reports on stereotactic percutaneous computed tomography (CT)-guided ablation of liver tumours have been published, suggesting a safe and accurate treatment. [20][21][22] While a few works report oncological benefits when using stereotactic ablation for colorectal liver metastases, 23 no data exists on oncologic outcomes after stereotactic CT-guided microwave ablation in a large cohort of HCC patients. Overall, we hypothesize that stereotactic image-guided microwave ablation (SMWA) is a safe and precise ablative treatment for all HCCs including those lesions that are conventionally ineligible to ablation (difficult anatomical location/ not detectable on ultrasound or CT) or inoperable because of the underlying liver disease or patient comorbidities. In this first retrospective analysis, we report safety and therapeutic efficacy in terms of local tumour control and short-term survival, as well as procedural efficiency in terms of targeting accuracy and required time when using SMWA for the treatment of HCC.

| MATERIAL S AND ME THODS
All patients with HCC treated with SMWA at the Department of Visceral Surgery and Medicine and the Department of Radiology, University Hospital of Bern, Switzerland between January 2015 and December 2017 were retrospectively included in the analysis. The study protocol was approved by the Regional Ethical Review Board (KEK-No. 2017-01038), and registered at clinicaltrials.gov (NCT03630068).

| Patient population
All patients were discussed in the institutional multidisciplinary tumour board. General indications for SMWA included patients with a) single solitary lesions <2 cm (BCLC 0), b) single or up to 3 nodules ≤3 cm (BCLC A) not amenable to resection or OLT, c) multinodular disease and/or tumours >3 cm (BCLC B) whenever the tumour board agreed to perform a downstaging so that patients could be evaluated for liver transplantation. 24 Accordingly, SMWA was the method of choice as a bridge-to-transplant or as a downstaging procedure for patients awaiting OLT. SMWA is also the standard of care for all patients receiving an ablative treatment of HCC at our institution irrespective of the tumour location. Tumours had to be radiologically (using the LIRADS categorical system) 25 or histologically confirmed HCC requiring SMWA as indicated by the consensus tumour board decision. Written informed consent was obtained before the procedure in patients aged ≥18 years.

| Procedural technique
All procedures were performed under general anaesthesia in the interventional CT suite by an interdisciplinary team consisting of specifically trained surgeons and interventional radiologists.
Patients were placed on the CT table and High Frequency Jet ventilation was used to ensure minimal movement of the patient and

Key points
• Computer navigated minimal-invasive thermal ablation of hepatocellular carcinoma is a novel technique that allows a safe and efficient curative treatment in particular for inoperable, not detectable or conventionally unablatable liver lesions. the diaphragm during the procedure. The intervention always con-

| Data extraction and assessment
Demographic and clinical data were extracted from patient's medical records. Lesion-specific and radiological data were extracted from DICOM imaging files, and technical data acquired during the intervention was extracted from the navigation system log files.

| Statistics
Descriptive statistics were used for presentation of patient characteristics and outcome data. Continuous data are shown as mean and standard deviation or median and range where appropriate.
The student T-Test was used to evaluate possible differences between groups. The Kaplan-Meier method, log rank test, logistic regression and cox regressions were applied to analyse the association of variables with local tumour progression, progression-free and overall survival. The threshold for statistical significance was set to P ≤ .05. Descriptive statistics and graphs were analysed using spss version 25.

| Clinical Data
Over a period of 36 months, 88, predominantly male patients were treated for 174 lesions in a total of 119 SMWA interventions at our institution ( Table 1). Half of the patients (n = 44) had previously been treated for HCC by other means (Table 1). Sixteen (36.4%) of these SMWA treated lesions were local tumour progressions after previous treatment failure, 28 (63.6%) were new tumours located elsewhere in the liver. In 33 patients, SMWA was performed as a bridging procedure prior to liver transplantation.

| HCC-specific data
At the time of intervention the majority of patients were BCLC stage A (n = 74, 62.2%; Table 1). Treated lesions were located in all liver segments, including segment I ( Figure 2). The majority of tumours were located in segment VIII (32.2%), followed by a relatively even distribution of lesions in all other segments. While more than half of the treated lesions (59.2%) were localized subcapsular (within 10 mm), 29.9% were in close proximity (within 5 mm) to major blood vessels (>3 mm). Among these, 32 lesions were located close to a portal vein branch, 11 close to a liver vein, one close to the vena cave and three close to the hepatic artery. Forty-five (25.9%) lesions were within 10 mm of the diaphragm or heart. Eleven lesions were in close proximity to other organs including the gallbladder (n = 4), kidney (n = 2), duodenum (n = 1), stomach (n = 2) and colon (n = 2; Figure 2).

| Intervention-specific data
The majority of patients had one (n = 77, 64%) or two (n = 31, 26.1%) tumours treated during one intervention. Nine patients had three (7.6%) and two patients had four (1.7%) lesions ablated during one intervention. All interventions were performed with microwave energy between 60 and 140 Watt for a median of five (1:30-18) minutes per tumour (Table 2). Thirteen lesions were treated with two needles and four lesions with three parallel needles to achieve overlapping ablation zones. We treated tumours >3 cm significantly more often with multiple needles (66.7%) than smaller lesions (33.3%), (P > .005, Chi square test). For further technical details, see Table 2.
The control of the ablation zone revealed an insufficient margin for 26 lesions immediately post-ablation, 19 of which could be immediately re-ablated. Three lesions could not be re-treated because of close proximity of the region of a potential insufficient margin to a major blood vessel (n = 1), gallbladder (n = 1), or lung (n = 1). One lesion was ablated to prepare a contralateral hepatectomy and three lesions were planned for an early follow-up. For 13 interventions, post-interventional imaging showed small intrahepatic or subcapsular hematomas and after one procedure a small pneumothorax was detected. None of these radiological findings were clinically symptomatic and they did not require any further treatment or follow-up.
These incidental findings because of the detailed post-ablation scan were therefore not considered complications.

| Post-interventional data
Median length of hospital stay (LOS) after the intervention was two (range 1-4) days. The overall 90-day complication rate was 5.9% (7/119), with six grade I-IIIa and one grade IIIb complication (Table 3). While we did not observe significant changes in the haemoglobin and creatinine levels before and after the intervention, the alanine transaminase increased significantly from a median 38 U/l to 90 U/l (P = .000) without any clinical correlation. This increase was not considered a pathological finding as it reflects the destruction of liver tissue through ablation.
successfully re-ablated, one patient was transplanted before a reablation could be performed and one patient showed additional new intrahepatic lesions requiring a change to systemic therapy.
Close proximity to a major blood vessel (P = .014) was the only factor significantly correlated with local tumour progression in the logistic regression analysis (Table 4). Tumour size ≥3 cm showed a trend to an inverse correlation with recurrence.
In the time to local tumour progression analysis ( Figure 3A (Table S1).

| SMWA as bridging therapy
Thirty-three (37.5%) patients awaiting transplantation underwent   In both reports, manual repositioning of the needle was performed and led to final positioning errors of 1.9 ± 1.7 mm resp. 1.6 ± 1.3 mm.

| D ISCUSS I ON
A randomized study on computer assisted electromagnetic navigation shows a median initial targeting accuracy of 4.1 mm for 60 needle placements. 40 In this study too, further needle corrections where performed manually after the initial navigated needle placement. Comparing these results, we see that our results of 3. trigger a re-ablation in the same session when insufficient margins are observed. A further aspect in standardizing ablation treatment is the use of JET-ventilation which guarantees minimal breathing movements that are required to achieve the precise placement of probes. Overall, our highly standardized protocol with reproducible workflows probably contributes to our low intra-and post-interventional complication rates. Although many argue that using general anaesthesia and computer navigation for liver ablation leads to pro- also being put forward for oncological treatments with curative intent. This patient population particularly profits from locally effective, but low risk, tissue-sparing interventions, where repeat therapy sessions are well-tolerated even in case of local tumour progression 10,11 .
In conclusion, SMWA using a computer-assisted navigation system is a safe and locally effective procedure for patients with HCC. While it improves the eligibility to a potentially curative and minimally invasive treatment option for patients conventionally unablatable it also represents a promising treatment alternative for lesions unresectable because of patient and liver comorbidity. Highly standardized treatment protocols as well as a specifically dedicated team, put together of hepatobiliary surgeons as well as interventional radiologists, ensure procedural safety while maximizing oncological outcome.