Articial Pneumothorax Improves Radiofrequency Ablation of Pulmonary Metastases close to Mediastinum

To investigate the feasibility, safety and ecacy of percutaneous radiofrequency ablation (RFA) of pulmonary metastases from hepatocellular carcinoma (HCC) contiguous with the mediastinum using the articial pneumothorax technique. Method: A total of 40 lesions in 32 patients with pulmonary metastases from HCC contiguous with the mediastinum accepted RFA treatment from August 2014 to May 2018 via the articial pneumothorax technique. After ablation, clinical outcomes were followed up by contrast enhanced CT. Technical success, local tumor progression (LTP), intrapulmonary distant recurrence (IDR), and adverse events were evaluated. Overall survival (OS) and local tumor progression free survival (LTPFS) were recorded for each patient.


Background
Image-guided percutaneous local ablation has been widely used for the treatment of pulmonary metastasis, based on the high capability on accurate orientation of needles and real-time visualization of targets [1][2][3][4][5] . However, the minimal invasive thermal technique remains challenging for treatment of pulmonary metastasis abutting the mediastinum, as a result of the proximity to the heart or aorta. In Subsequently, studies on the e cacious improvement of RFA for the treatment of lung tumors contiguous with mediastinum were rarely reported because of the insurmountable technique risk. Therefore, a safe and effective adjuvant method should be developed to reduce the risk of puncture and ablation for improvement of RFA e ciency of lung tumor contiguous with mediastinum.
Arti cial pneumothorax is a standard technique that serves as an effective adjuvant for the clinical ablation of peripheral lung tumors 7,8 , which could be used for separation between the punctured needle path and normal tissues after gas administration into the pleural cavity 9 . It is suggested that the pulmonary tumors could be accurately separated from mediastinal structures during an arti cial pneumothorax procedure. In that case, thermal damage to heart, aorta, or main nerves would be controlled, and pain caused by pleura heat could be effectively alleviated. In addition, the heat-sink effect of RFA could be reduced as well due to the arti cial-induced space between tumors and vessels 10 . Therefore, this strategy could not only signi cantly reduce the risks of puncture and ablation, but also enhance RFA e ciency for complete ablation. Compared with the air administration to form arti cial pneumothorax 9,11 , CO 2 is less common but displays good properties including high diffusion coe cient, inert reaction to RFA, and pure gas injection without unidenti ed substance 12 . Therefore, we investigate the feasibility, safety and e cacy of using arti cial CO 2 pneumothorax in this study for mediating RFA of pulmonary metastases contiguous with the mediastinum.

Patients enrollment
This study was approved by the institutional ethics committee of our hospital, the patients diagnosed with pulmonary metastases contiguous with the mediastinum who underwent RFA were retrospectively investigated at our hospital from August 2014 to May 2018. The target lesions of all cases were con rmed clinically or pathologically as hepatocellular carcinoma (HCC). "Contiguous with mediastinum" was de ned as the distance between metastatic lesion and mediastinum was less than 3 mm on CT or MR images.
The inclusion criteria for this study were: I) a single metastatic lesion contiguous with mediastinum in the unilateral lobe of lung; II) lesion size less than 3 cm in diameter; III) intrahepatic lesion shown to be stable as indicated by enhance imaging after surgical and (or) interventional therapy; IV) no increase in numbers of pulmonary lesions after 1-2 months of observation; V) patients refuse to accept surgical resection; and VI) Karnofsky performance status (KPS) greater than 70%.
Arti cial pneumothorax adjuvant RFA procedure All patients underwent contrast enhanced CT (CECT) of the chest prior to RFA in order to evaluate the anatomic relationship between lesions and peripheral cardiovascular structures. CECT was performed using the SOMATOM Force CT system (Siemens, Munich, Germany) with intravenous administration of Omnipaque (GE Medical, USA) as a contrast agent. Before the procedure, patients in a supine position received continuous electrocardiogram (ECG) monitoring and were hypodermically injected with 1% lidocaine. Then, the puncture point and needle track were determined by CT scan (MIYABI, Siemens, Munich, Germany). Subsequently, the biopsy needle was penetrated into the target in a direction away from the heart and vessels. When the procedure was accomplished, the biopsy needle was removed.
Thereafter, the arti cial pneumothorax technique was performed using a 22-G puncture needle (NPAS-100, COOK, USA) with a blunt tip. The needle was inserted along the well-designed puncture proposal, which speci es insertion route, depth and angle. When the needle tip reached the edge of the pleura, the needle core was pulled out, 1-2 ml of saline was injected locally to form a water capsule using Tstopcock and tubes. Then, saline in the tube owed into the cavity, and the water capsule disappeared gradually as the needle tip continued to be inserted and reached into the pleura cavity. Then, 100-500 ml of CO 2 gas was frequently administered to separate the lung parenchyma. CT scanning was performed individually after injection of 100, 200, 300, 400, and 500 ml CO 2 until the tumor was separated from the mediastinum and a feasible puncture path was established.
After the RF electrode was successfully inserted into the tumor based on the CT scanning assessment, percutaneous RFA (Model 1500, RITA Medical System, Mountain View, CA, USA) was performed at 60-70 W for 5-12 min at a rating temperature of 90℃. During the process, an 18-G unipolar electrode (Uniblate) was chosen (the length of the needle was 15 cm and the maximum ablation range of a single electrode was 3 × 3 cm 2 ). The ablation procedure was terminated when the ablation zone completely overlapped the target tumor and an ablative margin of 5-10 mm beyond the tumor boarder was achieved. Then, the RF electrode was withdrawn for coagulation of the needle track to avoid needle implantation and bleeding. During the procedure, 5-10 mg of morphine was intravenously injected for analgesia and sedation based on pain degree of patients. The ablation zone was de ned as the pulmonary texture around the tumor that showed a circular exudation shadow with ground-glass appearance on CT imaging. After RFA, CO 2 in the pleura cavity was aspirated with a 50-ml syringe and expelled through the T-stopcock until no additional gas could be aspirated.

Follow up
CECT was performed to evaluate the therapeutic response of tumors at 1, 3, 6, and 12 months post ablation and at 6-month intervals thereafter. If incomplete tumor ablation was detected by CECT, secondary RFA was performed. According to the reporting criteria of image-guided tumor ablation 13,14 , technical success was de ned as a tumor that was treated according to the initial protocol and was covered completely by the ablation zone. Local tumor progression (LTP) was de ned as the appearance of new tumor foci at the ablative margin. Intrapulmonary distant recurrence (IDR) was de ned as any occurrence of a new tumor foci in the lung.
During follow-up, adverse events (AEs) after ablation, local tumor progression free survival (LTPFS), and overall survival (OS) were recorded for each patient. AEs were classi ed in this study according to AE classi cation of the Society of Interventional Radiology 15 . The de nition of major complication included moderate AE, severe AE, life-threatening or disabling event, patient death or unexpected pregnancy abortion. Mild AE was de ned as minor complication.

Statistical analysis
Statistical analysis was performed using SPSS software (version 22.0; SPSS Inc, Chicago, IL, USA). Categorical variables were described as numbers (percentages). Continuous variables were described as the mean ± standard deviation (SD) or medians (range) according to the normality results using the Kolmogorov-Smirnov test. Survival was calculated by Kaplan-Meier survival analysis.

Patients characteristics
The graphical diagram of this study is shown in Fig. 1. A total of 32 patients with 40 pulmonary metastatic tumors contiguous with the mediastinum were con rmed from HCC using biopsy and then accept RFA treatment during this study. As shown in Table 1, 37.5% (12/32) of patients were males, and 62.5% (20/32) of patients were females. The median age was 61 years old, ranging from 44 to 72 years old. The mean tumor diameter was 1.4 ± 0.6 cm with a median distance of 0.1 cm from the mediastinum, ranging from 0 to 0.3 cm. 87.5% (28/32) patients had cirrhosis, mainly due to viral hepatitis B or C (96.4%). In total, 56.2% (18/32) of patients were characterized as Child-Pugh classi cation A, and 43.7% (14/32) of patients were characterized as Child-Pugh classi cation B.  Fig. 2A-L.

Complications
No major complications were observed during and following RFA procedure, such as severe pneumothorax, aeroembolism, hemothorax, lung abscess, alveolar bleeding, and tumor seeding. No AEs were relevant to the proximity of tumors to the mediastinum, such as vagus/recurrent laryngeal/phrenic nerve injury, vessel or oesophageal injury, pericardial effusion, arrhythmia, and cardiac infraction. In the 32 patients with 40 tumors, the occurrence of minor complication included slight pain of visual analogue scale (VAS) score 1-3 (n = 32), asymptomatic pneumothorax (n = 10), asymptomatic pleural effusion (n = 4), transient postablation syndrome of low-grade fever (n = 12) and general malaise (n = 16), as shown in Table 2. All complications resolved after 1-2 days postablation without any treatment.

Discussion
Arti cial pneumothorax was rst reported by Francini a century ago to treat pulmonary tuberculosis 16 . By injecting nitrogen between the parietal and visceral pleura, tuberculosis progression was suppressed by the reduction of blood ow and lymphatic re ux after pulmonary compression. During the past decades and enlightened by this report, arti cial pneumothorax has been successfully performed for protection of chest wall 17 , relief of chest pain 18 , and biopsy of pulmonary or mediastinal tumors 19 .
In this study, the technique of arti cial pneumothorax was rst validated as a feasible, safe, and e cient adjuvant method for RFA upon treatment of pulmonary metastases contiguous with the mediastinum. As shown in the results, technique success of RFA was 100% (32/32) without major complications under arti cial pneumothorax intervention, and the 1, 2, 3-year LTPFS rates were 90.6% and 81.2%, and 71.8%, respectively, resulting in the valid treatment e ciency and effective protection of proximate mediastinal structures. Compared to 43% only of the primary technical effectiveness reported by Iguchi et al, these results were comparable to the overall technique e cacy of standard RFA of lung intraparenchymal tumors away from mediastinum 20,21 . In general, our ndings indicated that the use of arti cial pneumothorax had safe and e cient local control for CT-guided RFA of lung metastasis contiguous with the mediastinum.
Given that 18-G needles are commonly used for gas administration in thorax, a potential complication for that procedure is mainly iatrogenic pneumothorax 8,9,11,22 . To avoid damage to the visceral pleura or the pulmonary parenchyma, we used a 22-G blunt needle with negative pressure during the procedure and withdrew CO 2 gas from the plural space after ablation. As a result, persistent pneumothorax or pleural hemorrhage with a chest tube drainage did not occurred.
In comparison with the air used before, the administration of CO 2 to produce arti cial pneumothorax in this study could be adsorbed more quickly by blood and reduce the risks of air embolism and injuries to normal tissues 23 . As a result, the blood oxygen saturations in all of the 32 patients were not in uenced based on the ECG monitoring.
However, excessive CO 2 in the pleural cavity can cause several adverse effects, including a) decrease in returned blood volume, increase in central venous pressure, and decrease in blood pressure; and b) impaired respiratory function and results in dyspnea, especially in the patients with poor pulmonary functions 24 . Furthermore, when excessive compression of lung parenchyma changes the electrical conductivity and heat conduction of RF, the ablation e ciency will be in uenced, resulting in the damage enlargement of lung tissues 9 . Therefore, it is preferable to minimize the amount of CO 2 that is injected to form an arti cial pneumothorax. During the procedure, 100-500 ml of CO 2 was slowly injected to establish a safe puncture path between the tumor and mediastinum. Compared to the previously reported injection volume of 400-1400 ml 22 , the amount of injected CO 2 in this study was more e cient.
Of note, the technique is non-indicated in some cases, including patients suffering from pleural adhesion or severe respiratory insu ciency. Limitations on this study was related to the fact that only 32 patients with 40 tumors were subjected to this method, owning to the focus is on the metastatic tumors from HCC. Further studies should explore the arti cial pneumothorax procedure as an adjuvant for RFA to understand the full extent of clinical bene ts.

Conclusion
CT-guided RFA after arti cially pneumothorax is a safe and effective method for the treatment of pulmonary metastases contiguous with the mediastinum. More samples should be examined to con rm the present results.   Overall survival (OS) and local tumor progression-free survival (LTPFS) curves during patient follow-up.