Stereotactic radiofrequency ablation of subcardiac hepatocellular carcinoma: a case-control study.

Abstract Purpose: The purpose of this study is to evaluate the safety and efficacy of stereotactic radiofrequency ablation (SRFA) for the treatment of subcardiac hepatocellular carcinoma (HCC). Material and methods: From 2003 to 2018, 79 consecutive patients underwent 104 multi-probe SRFA sessions for the treatment of 114 subcardiac HCC with a median size of 2.5 cm (0.5–9.5 cm). The results were compared with a randomly selected control group of 79 patients with 242 HCC with a median size of 2.0 cm (0.5–12 cm) following SRFA in other (non-subcardiac) locations with propensity score matching. Results: The 95.6% of the tumors were successfully treated by the first ablation session (primary technical efficacy rate) and 99.1% after the second session (secondary technical efficacy rate). Local tumor recurrence developed in 8 nodules (7.0%). Major complication and perioperative mortality rates were 7.7% (8/104) and 1% (1/104), respectively. The overall survival (OS) rates from the date of the first SRFA with single subcardiac HCCs were 92%, 77% and 65% at 1, 3 and 5 years, respectively, with a median OS of 90.6 months. The disease-free survival (DFS) after SRFA was 75, 34 and 34%, at 1, 3 and 5 years, respectively, with a median DFS of 19.1 months. There was no statistically significant difference with the control group in terms of local tumor control, safety, OS and DFS. Conclusion: SRFA of subcardiac tumors is as safe and efficacious as when treating tumors remote from the heart.


Introduction
Hepatocellular carcinoma (HCC) arises in the context of background cirrhosis in 90% of cases [1] and confers a poor prognosis even in liver-confined disease (18% 5-year survival) [2]. Whilst the healthy liver is resilient to insult, the cirrhotic liver has a poor regenerative capacity, which very often precludes patients from surgical resection (SR) [3]. Therefore, the management of HCC is a delicate balance between tumor and background liver destruction which has brought about an increasing demand for focal ablative treatments, including radiofrequency ablation (RFA) [4][5][6]. To achieve low rates of residual or recurrent disease, the lethal ablation zone must cover the tumor with a sufficient safety margin [7] which may be challenging where the tumor is located in the vicinity of critical structures such as the heart.
Since percutaneous treatment is more challenging in tumors which are located close to the liver capsule, early studies showed significantly higher local tumor recurrence compared to non-subcapsular tumors (65 vs. 19%) [8]. However, more recent studies suggest that whilst RFA for subcardiac locations is technically feasible and not significantly different to non-subcardiac locations [9,10], recurrence rates reported to be relatively high (19.1-20%). Three-dimensional (3D) navigation systems have been implemented to allow for a more sophisticated 3D planning of multiple overlapping ablation zones, precise probe placement and intraoperative assessment of the result by means of image fusion [11,12].
However, the role of this technique for the treatment of subcardiac tumors remains undefined. The aim of this study is to evaluate the safety and efficacy of stereotactic radiofrequency ablation (SRFA) for subcardiac tumors in terms of primary and secondary efficacy and local recurrence rate (LR). We also consider overall survival (OS) and disease-free survival (DFS) which, to the best of our knowledge have never previously been reported. Our hypothesis was that there is no difference in LR, DFS and OS for subcardiac vs. non-subcardiac HCC.

Patient cohort and inclusion criteria
This study was approved by our Institutional Review Board and written informend consent was obtained from all patients. In every patient, the treatment plan was established by a multidisciplinary tumor board consisting of hepatologists, oncologists, transplant surgeons and interventional radiologists. Decisions for therapy were based on tumor characteristics, Child-Pugh classification, anatomical considerations and performance status. Due to the multiprobe approach there were no limitations in the size or number of tumors. Thus, BCLC B and BCLC C patients were not excluded from SRFA despite BCLC guidelines recommending such patients are treated with transarterial chemoembolization or medical therapy. Patient selection was carried out retrospectively as part of a single-center, single-arm study.
435 consecutive patients with HCC were treated with SRFA between 2003 and 2018 ( Figure 1). Arterial phase CT was reviewed by a board-certified radiologist with 10 years' experience (PES), who assigned patients to the subcardiac group, defined as tumor location 1 cm from the pericardium in any direction. 79 patients met the inclusion criteria and 79 patients with non-subcardiac tumors were randomly selected using nearest neighbor propensity score matching by the R package 'MatchIt' with sex, age, tumor number and size and liver biochemistry as matching variables. The baseline characteristics of the two groups are shown in Table 1.
Exclusion criteria for SRFA were a platelet count of <50,000/mm 3 and prothrombin activity <50%.
Diagnosis of HCC was based on the identification of typical imaging hallmarks of HCC with detection of contrast hyperenhancement in the arterial phase (wash-in) and 'wash-out' in the portal venous and/or delayed phases on CT, MRI or CEUS for nodules 1 cm. In addition, thoracic CT was performed in all patients to exclude the presence of pulmonary metastases. All HCCs were pathologically confirmed by needle biopsy during the SRFA procedure.

Multi-probe SRFA
Our method of SRFA has been described in detail previously [13,14]. In brief, the procedure is carried out in the interventional CT-suite with anesthetized patients fixated on the CT-table by a single (Bluebag, Medical Intelligence Schwabm€ unchen, Germany) or double vacuum fixation technique (BodyFix, Medical Intelligence Schwabm€ unchen, Germany) using full muscle paralysis. For image-to-patient registration, 10-15 registration markers, (Beekley Spots, Beekley Corporation, Bristol, CT, USA) were broadly attached to the skin of the thorax and upper abdomen.
Subsequently, a contrast-enhanced planning CT (Siemens SOMATOM Sensation Open, sliding gantry with 82 cm diameter, Siemens AG, Erlangen, Germany) with 3 mm slice thickness was performed in the arterial and venous phase. The obtained CT-data were transferred to an optical based navigation system (Stealth Station Treon plus, Medtronic Inc., Louisville, KY, USA). One or multiple trajectories were planned using the navigation system with multiplanar and 3D reconstructed images. To eliminate respiratory motion, temporary disconnections of the endotracheal tube (ETT) were performed during the planning CT, during each stereotactic needle placement and the final CT. Multiple RF electrode positions were planned in order to cover the entire tumor volume and a sufficient safety margin with overlapping necrosis. Therefore, multiple coaxial needles (1-15, median 3) were placed in the entire tumor including the most peripheral parts with a maximum inter-probe distance of 2 cm. For all subcardiac lesions the planned distance between the tip of the probe and the diaphragm was >1 cm.
After registration, an accuracy check and sterile draping, the ATLAS aiming device (Elekta PSC Medical Intelligence Inc., Schwabmuenchen, Germany) was used for navigated trajectory alignment. 15 G/17.2 cm coaxial needles (Bard Inc., Covington, GA, USA) were subsequently placed through the aiming device during temporary ETT disconnection without the need for real-time imaging, using needle trajectories as defined by the coordinates on the planning CT. The depth from the aiming device to the target was automatically calculated by navigation software. The coaxial needles served as guides for the radiofrequency electrodes. For verification of correct needle placement, an unenhanced CT was performed following ETT disconnection and fused with the planning CT using the navigation system's image 3D registration algorithm.
A 16 G biopsy sample was obtained via one of the co-axial needles in all cases. Thereafter, RF-electrodes (Cool-tip, Medtronic, Mansfield, MA, USA) were introduced through the coaxial needles for serial tumor ablation. A post-ablation completion contrast-enhanced CT was finally carried out in the arterial and portal venous phase during ETT disconnection and fused with the planning CT for verification of ablation zone coverage and for assessment of complications.
RFA was performed using a unipolar Cool-tip_RFsystem (Cool-tip, Medtronic, Mansfield, MA, USA), a Cool-tip_RF switching controller and a 17 G Cool-tip electrode with a length of 25 cm and an exposure of 3 cm. The standard ablation time for three electrodes (switching control) was 16 min. Needle track ablation was performed during repositioning and during final removal of the RFA electrodes to reduce the risk of bleeding and tumor seeding.
Example images from subcardiac multiprobe SRFA are shown in Figures 2 and 3.

Endpoints and follow-up
Primary and secondary efficacy, as well as the local recurrence rate (LR) were the primary endpoints. The secondary endpoints included DFS and OS. Follow-up contrast-enhanced CT/MR scans were performed first at 1-month and then at 3-month intervals after the SRFA. The images were evaluated by two abdominal radiologists with more than 10 years of experience by consensus (PES and BR). Technical success was defined as accurate electrode placement according to protocol with immediate, complete coverage by the ablation zone. Primary technical efficacy rate was evaluated for every tumor as the absence of residual tumor on 1-month follow-up CT. Secondary technical efficacy rate included tumors with repeated ablation after identification of residual tumor. Evidence of newly detected tumors within or immediately adjacent to the ablation zone or initial tumor was deemed as LR. Complete ablation was defined as a circumscribed non-enhancing zone within and/or extending beyond the initial tumor borders with a well-defined margin. Newly detected nodules distant to the ablation zone and or to initial tumor location were defined as distant tumor recurrence. Complications were defined according to the Society of Interventional Radiology (SIR) Standards of Practice Committee classification [15]. In patients with a single subcardiac HCC vs. those with a single HCC elsewhere in the control cohort, survival was calculated from the date of first stereotactic RF ablation to the date of death attributable to malignancy (HCC) or other causes ('event') vs. the most recent follow-up visit (i.e. censoring).   (1:1 matching with the nearest neighbor) was used for the propensity matching process to select patients from the control group.

Patient characteristics
Baseline characteristics of patients undergoing RFA are shown in Table 1 At initial SRFA therapy, 37 patients had a solitary liver nodule, 20 patients had two nodules, 11 had three nodules and 11 patients had more than three nodules (defined as multiple nodules), including nodules remote from the heart. Of the patients in the subcardiac group with >1 HCC at initial SRFA, 1 patient had 3 subcardiac HCCs, 4 patients 2 subcardiac HCCs and 37 patients 1 subcardiac HCC. 11 patients had SR (2 local recurrences), 24 patients underwent TACE (21 residual or recurrent tumors) and 5 patients (5 residual or recurrent tumors) conventional RF ablation prior to SRFA. Due to untreatable tumor progression with diffuse/disseminated HCC infiltration as detected in follow-up imaging 4 patients received TACE and 9 patients received systemic chemotherapy subsequently. Liver transplantation could be performed in 19 patients.
From the control group, 79 patients, 11 females and 68 males, with a median age of 65 years (41-85) underwent SRFA for treatment of 242 HCCs in 125 ablation sessions. The median tumor size was 2.0 cm (0.5-12 cm).

Perioperative complications
Perioperative complications are shown in Table 2. One death occurred following ablation due to therapeutic induced liver failure (mortality rate of 1.0% (1/104)). The total major complicate rate was 7.7% (8 of 104). Three of these complications were clearly related to simultaneous thermal ablation of nodules in other locations, leading to a major complication rate of 4.8% (5 of 104) in subcardiac tumors. Thermal injury to the pericardium in one patient resulted in recurrent pericardial effusion and a prolonged intensive care admission with death 51 days after SRFA in one patient with a Child B cirrhosis. Two patients developed postoperative liver abscesses. Two patients were transferred to intensive care for temporary postoperative organ support and two patients developed perihepatic hemorrhage managed by angiographic coil embolization in the same general anesthetic session. Despite subcardiac location we did not observe any modifications in heart pace and/or hypo/ hypertension.
Fever (>37 C, measured peripherally) developed in all patients but subsided in 1-2 days with symptomatic treatment. The median hospital stay was 5 days .

Technical success
SRFA was completed according to plan in all 114 nodules (technical success rate 100%). Tumor enhancement disappeared after the initial RF session in 109/114 cases resulting in a primary technical efficacy rate of 95.6%. Thereafter, four nodules were successfully retreated resulting in a secondary technical efficacy rate of 99.1%. One nodule was not retreated due to untreatable tumor progression involving multiple newly developed tumors (Table 3).

Local recurrence rate and distant recurrence
Local tumor recurrence developed in 8 of 114 nodules (7.0%). In five nodules, pretreatment, including TACE,

OS and DFS
OS and DFS rates are depicted in Figures 4 and 5. The OS rates at 1, 3 and 5 years from the date of the first SRFA were 92%, 77% and 65% with a median OS of 90.6 months in patients with exclusively single subcardiac HCC (n ¼ 37). The corresponding DFS at 1, 3 and 5 years was 75, 34 and 34% with a median DFS of 19.1 months.
The OS rates of the control group at 1, 3 and 5 years were 87, 65 and 53% with a median OS of 61.9 months for the patient control group with single HCC in nonsubcardiac location, respectively. The DFS rates at 1, 3 and  There was no significant difference in OS or DFS of the subcardiac vs. control group (p ¼ 0.28 and p ¼ 0.89, respectively).

Discussion
The results of this study show that SRFA of subcardiac tumors is as safe and efficacious as when treating tumors remote from the heart. Specifically, we found no statistically significant difference between cases (i.e. patients with subcardiac tumors) vs. controls (those without subcardiac tumors) in terms of major complications (p ¼ 0.76), primary and secondary efficacy (p ¼ 0.93 and 0.23, respectively), local recurrence (p ¼ 0.25), OS (p ¼ 0.28, in terms of singular HCC) and DFS (p ¼ 0.89, in terms of singular HCC). These results are important because they show that patients with subcardiac tumors should still be offered the procedure, that the consent process should be similar for tumors remote from the heart.
Historically, the literature suggested that percutaneous RFA of liver tumors in difficult locations results in more complications, insufficient ablation and poor tumor control [16,17]. More specifically, subcapsular tumors were associated with increased risk of complications [8,16,18,19] although more recent studies have been contradictory, reporting similar complication rates for subcardiac vs. non-subcardiac tumors (1-4%) [7,9]. Since the major complication rate in our study was 7.7% (8/104) and not significantly different from the control group (p ¼ 0.56), our results are in keeping with these more recent reports. However, the higher complication rate found in our study may be explained by larger tumor size and the treatment of multiple tumors in one session. Indeed, three patients with major complications in our study were attributed to the treatment of other tumors, meaning the major complication rate specific to subcardiac tumors was 4.8% (5/104). However, by comparison, the surgical literature reports higher levels of postoperative morbidity ranging from 5 to 45% [20].
Due to the stereotactic multiprobe approach, our treatment strategy deliberately challenges BCLC guidelines, whereby >60% of cases were BCLC B and C. We strongly believe that SRFA is preferable to TACE ± sorafenib in these patients, since ablation offers homogenous, lethal cell kill with the potential for curative treatment.
In terms of technical success, we were able to treat tumors with complete disappearance of tumor enhancement in the first session in 95.6% of cases and 99.1% after retreatment with a local recurrence rate of 7% and no significant difference compared with the control group. We recently published a study which used the same radiological methods to assess ablation zone coverage [21] and found no viable tumor in 97.3% of cases as part of a histopathological validation study with whole explanted livers as the reference standard, meaning our data are likely to be histologically accurate. Whilst other groups have used iatrogenic pleural effusions [22] and pneumothoraces [23] to protect the thoracic structures, our reported high levels of complete treatment suggest such approaches may be unnecessary, whereby increased procedure time, radiation exposure and reduction of lung capacity in patients with co-existing respiratory disease are all considerations.
Early investigations [8,24] reported higher local recurrence rates for subcapsular tumors, which was thought to be due to undertreatment for fear of sustaining thermal damage to adjacent structures or causing liver capsular pain. However, more recent investigations which specifically focused on subcardiac tumors [9,10] reported no statistical difference in complete ablation (83-96%) or local recurrence rates between subcardiac and non-subcardiac tumors, although reported relatively high recurrence rates of 15-22%. Cha et al. compared the therapeutic outcomes of conventional US-guided radiofrequency (RF) ablation for subcardiac and non-subcardiac HCC in 73 patients with subcardiac HCC and the same number of patients with non-subcardiac HCC. Similar to our study there was no significant difference observed between both groups in terms of technical success rates and cumulative LR rates. However, despite their inclusion of BCLC A and B patients only, with a mean tumor size of 1.6 cm, their LR rates were > 15% in the subcardiac group. By contrast, we included tumors with a median size of 2.5 cm (maximum 9.5 cm) and found LR was 4-7%, which is similar to the surgical literature [25]. We attribute the good local tumor control after SRFA to consistent achievement of a sufficient ablation margin of at least 5 mm. SRFA offers three-dimensional planning and very precise targeting to create multiple overlapping coagulation volumes. In case of poor tumor visibility, SRFA planning may include fusion with previously acquired MR images. Immediate post ablation contrast-enhanced CT fusion with the planning CT allows for rapid, reliable judgment of the ablation results with the option for re-ablation.
We have previously reported the OS for SRFA in intrahepatic cholangiocarcinomas [25] and hepatic colorectal [14], breast cancer [26] and melanoma metastases [27] and found a comparable survival advantage to liver resection. In the present study, we report OS rates of 92, 77 and 65% at 1, 3 and 5 years, respectively, with a median OS of 90.6 months in patients with exclusively single HCC. Another study of HCC ablation for tumors abutting the diaphragm [28] showed OS rates of 91.7, 56.7 and 36.7% at 1, 3 and 5 years with a median OS of 44 months. SRFA achieved a higher survival rate despite unfavorable subcardiac location. We also found DFS of 75, 34 and 34%, at 1, 3 and 5 years respectively, with a median DFS of 19.1 months, whereas Kim et al. [29] reported rates of 66.5, 20.4 and 17% at 1, 3 and 5 years, respectively, for HCC in all locations.
In comparison with the surgical literature, a meta-analysis showed OS of 72.3 and 55.5% and DFS of 35 and 25.7% at 3 and 5 years for non-anatomical liver resections for HCC, meaning our survival statistics are very similar [30], even for large tumors as were ablated in the current study.
The relatively high rates of distant tumor recurrence (48.1% of patients) can be explained by the inclusion of BCLC B and BCLC C patients and the relatively long follow-up time (mean: 25.2 months and 0-162 months) in patients' cirrhotic livers with a propensity to develop further HCCs.
The limitations of our study include its retrospective design which precluded a prospective power calculation. In addition, use of additional treatments such as liver transplantation (24 vs. 19%), TACE (5 vs. 6%) and chemotherapy (11 vs. 3%) differed slightly between each of the groups and can impact on OS. Whilst our results are not highly generalizable due to limited use of navigation systems as part of ablation procedures, their use is growing and we have previously demonstrated our technique is reproducible [31]. In this way, we hope that the findings of this study will help generate further interest in our technique with a view to a prospective powered study, ideally in a multicenter setting. The lower complication rates and similar survival statistics vs. surgical alternatives suggests a comparison in the form of a randomized trial would be welcome.
In conclusion, our results suggest that stereotactic RFA is a safe and efficacious treatment strategy for subcardiac HCC with no statistically significant difference vs. tumors remote from the heart in terms of major complication, technical success and local recurrence rates and overall and DFS. These results are important because they show that patients with subcardiac tumors should still be offered stereotactic ablation and that the consent process should be the same as for tumors remote from the heart.

Disclosure statement
No potential conflict of interest was reported by the authors. RB is a consultant for CAScination and i-Sys Medizintechnik.