The efficacy of bipolar and multipolar radiofrequency ablation of lung neoplasms — results of an ablate and resect study

Abstract

Objective: Radiofrequency ablation (RFA) has obtained increasing attention as an interventional approach for the local treatment of primary and secondary lung neoplasms. The local effect of the procedure is usually controlled by radiologic means. The objectives of this ‘ablate and resect’ study were to investigate the efficacy of bipolar and multipolar RFA by histologic evaluation and to compare the two techniques. Methods: In a total of 32 subjects with histologically proven non-small-cell lung cancer or pulmonary metastases from an extrathoracic primary tumor, bipolar, or multipolar RFA was performed during open thoracotomy. Curative resection (lobectomy or wedge resection including mediastinal lymph node dissection) was performed subsequently. The extent of cell death and early histologic findings following RFA were determined by histology and immunohistochemistry (nicotinamide adenine dinucleotide (NADH) and monoclonal anti-mitochondrial antibodies MAB 1273). Results: Intra-operative bipolar and multipolar RFA is a safe procedure, and there was no bleeding or thermal damage of the lung tissue. Routine histologic staining could not identify tumor cell death. However, immunohistochemistry was able to verify cell death in the ablated tumor tissue. Complete tumor cell necrosis was determined in 12 tumors (37.5%); and scattered vital tumor tissue was detected in 16 tumors (50%). Incomplete ablation with a ratio of >20% vital tumor tissue was found in four tumors (12.5%), particularly surrounding vascular structures within the tumor tissue or in marginal zones of the tumor tissue. The local efficacy of bipolar and multipolar RFA was comparable, and incomplete ablations were found only in adenocarcinoma. Conclusions: Bipolar and multipolar RFA in an open thoracotomy setting is a technically feasible and safe procedure. Early immunohistochemical findings after RFA showed complete tumor cell necrosis in 38% of cases. The high rate of viable tumor cells remaining after ablation casts doubt on RFA as a curative concept. This approach should be reserved for palliative indications. Patients fulfilling the criteria for curative resection should not be denied surgery.

Radiofrequency ablation, Local thermal therapy, Lung neoplasms, Histology

1 Introduction

Image-guided radiofrequency ablation (RFA) has garnered great interest as a minimally invasive approach in the treatment of primary and secondary lung neoplasms among patients, who are not candidates for surgical resection. Numerous reports on the application of RF current in lung neoplasms have been published up to now. At present, safety and local effectiveness are evaluated at the level of phase 1 trials [1–4]. Treatment response is commonly validated by radiological means. Subsequent to the RF procedure, the tumor diameter increases in size due to edematous and hemorrhagic alterations in the surrounding lung parenchyma [5,6]. For up to 3 months, the ablated area decreases in size; at 6 months, it is of the same size or smaller than before RFA. Later, it becomes stable or shrinks [5]. Conventional radiologic response evaluation criteria, therefore, cannot be applied. Animal investigations have indicated a high proportion of viable tumor tissue after RFA in lung tissue [7]. In human investigations, histology subsequent to RFA in lung neoplasms consistently detected vital tumor tissue at a frequency of up to 63.5% [8–10].

The objective of the present study was to evaluate the local efficiency of bipolar and multipolar RFA in a comprehensive series of different types of primary and secondary lung neoplasms using histological proof of tissue vitality subsequent to the ablation procedure. Bipolar and multipolar RFA procedures offer a further development of monopolar RFA, and histologic evaluation subsequent to multipolar RFA in human lung neoplasms has not been performed up to now. The physiology of the image-guided approach is maintained in an open thoracotomy setting, implying full ventilation and blood perfusion of the lung during the RFA procedure. The extent of cell death and early histologic findings following RFA were determined by histology and immunohistochemistry (nicotinamide adenine dinucleotide (NADH) supravital staining and monoclonal anti-mitochondrial antibodies MAB 1273).

2 Methods

A total of 32 subjects fulfilling the criteria for curative surgical resection for non-small cell lung cancer (NSCLC) or pulmonary metastases were enrolled in this study. The mean subject age was 62.6 years, with 17 female and 15 male subjects. The recruitment period lasted from November 2005 to December 2009. Prior to the intervention, histology was proven NSCLC by trans-bronchial biopsy in 14 subjects (adenocarcinoma: n = 10, and squamous cell carcinoma: n = 4). A further 18 subjects with at least two pulmonary metastases from an extrathoracic primary tumor (colorectal cancer n = 7, renal cancer n = 4, melanoma n = 2, and soft-tissue sarcoma n = 5) were also included in this series. The Ethics Committee of the University of Heidelberg approved this study (Study No. 503/2005). Each subject was informed comprehensively about the ablation and the surgical procedure; written informed consent was obtained prior to study participation.

2.1 Surgery and RFA

RFA was performed in an open thoracotomy setting. The standard surgical incision was an anterolateral thoracotomy in the fourth intercostal space. In primary neoplasms, histologic proof was obtained before surgery via endoscopic biopsy. In secondary pulmonary neoplasms, histologic proof was obtained by wedge resection of one tumor, followed by immediate intra-operative frozen sectioning.

The RFA electrodes were placed through the thoracotomy incision into the pulmonary lesion (Fig. 1 ). The correct placement of the electrode in the pulmonary tumor was verified by manual palpation. RFA was performed by the bipolar CelonPro Surge^® RFA system. Depending on the tumor size, the conductive electrode length and the number of electrodes were chosen. Bipolar RFA was performed in tumors with a maximum diameter of 2.5 cm, and multipolar RFA was performed in tumors with a maximum diameter of 3.5 cm (Fig. 5). In bipolar RFA, the electric current flow is located between the electrodes at the tip of the RFA probe. In contrast to monopolar RFA, no grounding pads are required. In addition, there is no current flow through the patient’s body. In multipolar RFA, the current flow switches between the different RFA probes placed into the tumor to enhance thermal efficiency. A total of 18 bipolar ablations and 14 multipolar ablations were included in this series.

Fig. 1

Intra-operative approach – RFA at an open thoracotomy setting with two RFA probes placed in the lung tumor.

The RF current was applied according to the manufacturer’s protocol for the treatment of lung tumors. The energy was provided by a power control unit (CelonLab Power^®), which offers the opportunity of connecting a number of applicators simultaneously for multipolar application and monitoring targeted energy and tissue impedance during the procedure.

The lung was ventilated with a continuous positive airway pressure (CPAP) of 10 mmHg. The central vascular structures were surgically dissected prior to the RF procedure, but blood perfusion of the lung was perpetuated during the ablation procedure.

After completion of the RFA procedure, curative surgical resection was performed. In primary lung neoplasms, lobectomy was performed; in secondary lung neoplasms, the ablated and non-ablated tumors were resected by wedge resection. Radical mediastinal and hilar lymphadenectomy was realized concurrently with all procedures, including four compartments in right-sided (paratracheal, infracarinal, inferior mediastinal, and hilar) and left-sided (aortic, infracarinal, inferior mediastinal, and hilar) thoracotomies [12].

2.2 Pathologic work-up

The tumors were bisected through the largest diameter at a right angle with reference to the RFA electrode direction. In the cross section, the inlet ducts of the RFA probes were investigated and correct positioning inside the tumor tissue verified (Fig. 2 ). The specimens for routine diagnostics and immunohistochemistry were fixed in 4% buffered formalin solution. The specimens were examined by light microscopy. Routine staining was carried out with hematoxylin–eosin (H&E), and analysis of cellular vitality was performed with a mouse anti-human mitochondria monoclonal antibody (MAB 1273; Millipore UK). MAB 1273 recognizes a 65-kDa protein of human mitochondria by immunoprecipitation that gives a ‘spaghetti-like’ staining pattern in the cytoplasm. The histopathologic criterion for cell death in MAB 1273 immunostaining was the lack of staining in the RFA-treated tissue. MAB 1273 staining was performed in 18 of the total 32 RFA procedures.

Fig. 2

Postoperative gross sectioning of an RF ablated tumor – the RF probes were re-induced into the ducts of the electrodes to demonstrate the correct placement of the probes.

NADH staining relies on the reduction of nitroblue tetrazolium chloride by cells expressing NADH. Non-viable cells lack staining, and positive staining thus implies cellular viability. For NADH staining, the tumors were bisected after resection as described above; one section of the ablated tumor was snap-frozen in liquid nitrogen and stored immediately at −80 °C. NADH staining was performed in 22 of the total 32 RFA procedures. The effect of RFA was evaluated by two experienced pathologists and was categorized into three groups: (1) complete ablation (i.e., no vital tumor cells were found after RFA in immunohistochemistry with MAB 1273 and/or NADH); (2) incomplete ablation with the minimal finding of vital tumor cells (i.e., scattered tumor cells were found within the ablated tumor tissue either surrounding vascular structures or at the outer region of the tumor); and (3) incomplete ablation with the finding of >20% vital tumor tissue after RFA.

3 Results

A total of 32 subjects were enrolled in this study. The intra-operative placement of the electrode into the tumors under manual control in the non-ventilated lung was successful, and the positioning of the electrode into the tumor tissue remained stable under CPAP ventilation in all procedures. We observed no intra-operative complications during the RFA procedures, and there was no evidence of intra-operative bleeding or thermal injury along the coagulation electrode tract. No sign of prolonged air-leak or bronchopleural fistula subsequent to surgical resection was evident. The amount of applied energy during the RFA procedure ranged from 3.25 to 50.4 kJ, depending on the diameter of the tumors (range: from 0.7 to 3.5 cm). Curative surgical resections were performed as described above, subsequent to the ablation procedure in all patients.

Immediately after surgical resection, the ablated lung specimens were transferred for pathologic examination. In gross examinations of the RFA-treated lesions, the correct locations of the duct of the RFA electrodes inside the tumor tissue were confirmed (Fig. 2). In routine H&E staining, the cytoplasm of RFA-treated tissues presented with increased eosinophilia and streaks of chromatin. In addition, changes including cytoplasmic dissolution, nuclear elongation and blood vessel dilation inconsistent with cautery artifacts were noted. In all subjects, the tissue architecture of the tumor cells was preserved; non-viable cells subsequent to RFA could not be differentiated from viable cells using standard H&E staining. Performance of the previous RFA procedure did not hamper histological diagnosis of the specimens.

Immunostaining was performed using an anti-human mitochondria monoclonal antibody (MAB 1273) and NADH–diaphorase staining. Histopathologic proof of tumor cell liveliness revealed complete ablation in 12 subjects (37.5%). In 16 subjects (50%), scattered tumor cells were found within the ablated tumor tissue either surrounding vascular structures (Fig. 4(b)) or at the outer regions of the tumors (Figs. 3 and 4(a)). In four subjects (12.5%), ablation was incomplete and accompanied by the finding of >20% vital tumor tissue after RFA. In NADH staining, artifacts subsequent to the application of heat (in the procedure) were seen along the duct of the RFA electrodes (Fig. 4(c) ). Small areas with a lack of staining attributable to thermal injury could also be found in NADH-stained specimens in tumor-adjacent lung tissue.

Fig. 3

Histologic proof of lively tumor cells in marginal zones of an ablated tumor in MAB 1273 staining (viable cells are colored in red). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

Fig. 4

Histologic proof of lively tumor cells in an ablated tumor in NADH staining (vital cells are colored in black). (a) Vital tumor tissue in the marginal zones of the tumor; (b) vital tumor tissue encircling vascular structures crossing through the tumor; (c) cautery induced heat-artifacts along the duct of the RF probes; no vital tumor tissue can be seen. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article.)

There was a good linear correlation between the applied energy and the diameter of the lesions; tumors showing incomplete ablation (<80%) had received an amount of energy equivalent to that received by completely ablated tumors of the same diameter (Fig. 5 ). The efficacy of the bipolar and multipolar approaches was comparable. Two incomplete ablations were found after the bipolar approach (2/18, 11%) and multipolar approach (2/14, 14%).

Fig. 5

Efficacy of the RFA procedure depending on diameter of the tumors, the amount of applicated energy and the RFA approach. Linear correlation between applicated energy and diameter of the lung neoplasm’s in both bipolar and multipolar RFA. ○ bipolar RFA, complete ablation or minimal vital tumor tissue; ● bipolar RFA, incomplete ablation; ▵ multipolar RFA, complete ablation or minimal vital tumor tissue; ▴ multipolar RFA, incomplete ablation.

The grade of ablation with regard to histology is shown in Table 1 . No difference in the local efficiency was noticed between primary neoplasms and the different histologic subtypes of secondary lung neoplasms. However, all incomplete (<80%) ablations (total n = 4) were found in adenocarcinomas, either primary lung cancer (2/10, 20%) or secondary neoplasms (metastatic colorectal cancer (2/7, 29%)).

Table 1

Tumor characteristics and grade of ablation.

In one patient with multiple pulmonary metastases of malignant melanoma, a pulmonary artery pseudo-aneurysm in the same lobe was detected 6 weeks after RFA and surgical resection. This has to be regarded as a late sequela of the RFA procedure (Fig. 6 ). In further follow-up observation, this patient remained asymptomatic; pulmonary metastases have also been resected on the opposite side.

Fig. 6

Intrapulmonary pseudo-aneurysm in the same lobe occurring six weeks later after RFA.

4 Discussion

In this study, we translated the most commonly used image-guided RFA procedure into an intra-operative approach and were able to characterize pathomorphology subsequent to RFA. Ablation was complete (with a proportion of 100% necrotic tumor tissue in histology) in 37.5% of the procedures. Unequivocal incomplete ablation (with a portion of >20% vital tumor tissue in histology) was found in 12.5% of cases, and scattered vital tumor cell agglomerates subsequent to the RFA procedures were seen in 50% of the procedures. Our findings are in line with two other protocols of intra-operative monopolar RFA on early-stage NSCLC [8,9]. One series performing surgical resection and histology 3 days subsequent to image-guided RFA reported a maximum local effectiveness of 90.1% [11]. Vital tumor cell aggregates in our study were located either in the marginal zones of the tumor or encircling vascular structures crossing through the tumor tissue. The incomplete ablation in the marginal tumor zones may be explained in part by the so-called heat sink effect (removal of heat by blood flow in vascular structures) in the case of vascular structures coursing through the tumor tissue (Fig. 4(b)). Interestingly, the energy applied in the incompletely ablated tumors was not lower than that applied in completely ablated tumors of the same diameter (Fig. 5). Furthermore, the heat sink effect due to small vessels intersecting the tumor was not predictable. We have to conclude that the local effect of RFA cannot be anticipated merely by the amount of energy applied; additionally, there was no warning during the RFA procedure that incomplete ablation would occur.

The histologic proof of the efficacy of bipolar RFA in lung metastases has already been reported [10]; but the local efficacy of multipolar RFA in human lung neoplasms was investigated for the first time in this study. When we consider our findings of the bipolar and multipolar approaches in our series as well as findings from the literature on monopolar RFA, the local efficacy of the three RFA techniques appears to be equivalent. A dependence of the success of RFA on the histology of the ablated lesions has not been reported so far. In our series, complete and incomplete ablations were seen among both primary pulmonary neoplasms and various metastatic subtypes. All incomplete ablations (4/32), however, occurred in lesions with adenocarcinoma histology in both the primary neoplasm and secondary neoplasm (metastatic colorectal cancer) groups, resulting in incomplete ablation rates of 20% and 29%, respectively. Although this is an interesting finding, the results must be regarded with caution due to the small number of cases included in this study.

The early histopathologic detection of lethal cell injury is a limitation of the ‘ablate and resect’ approach. Routine H&E staining fails to identify irreversible cellular injury within 1 week after RFA. In the literature, methods using NADH staining on fresh frozen tissue, polyclonal rabbit anti-single-strand DNA (ssDNA), terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate (dUTP)-biotin nick end labeling (TUNEL), mouse anti-human mitochondria monoclonal antibody MAB 1273, and anti-mitochondrial antibody 113-1 were applied. Particular features of the different techniques have been discussed previously [10]. H&E staining showed increased cytoplasmic eosinophilia and streaks of chromatin subsequent to local heat application; NADH and the human mitochondria monoclonal antibody MAB 1273 were able to detect undamaged tumor tissue. Of course, the fate of small vital tumor cell aggregates remains uncertain. These cells could die off within a few days because of cellular injury or local hypoxemia, thus endowing upon the procedure a potential success rate of 100%. On the other hand, single cells may survive and originate further local tumor recurrence. Basically, the process of decay seems to proceed. Roughly 9.1% of cases exhibited incomplete ablation in a series of surgical resections 3 days subsequent to image-guided RFA [11]. Subsuming completely ablated tumors and those with a minimal finding of scattered vital tumor cell agglomerates in our series actually results in an incomplete ablation rate of 12.5%. A similar ratio of ‘incomplete ablation’ has been reported in other ‘ablate and resect’ studies applying monopolar RFA [8,9,11].

The manual placement of the RFA electrodes may be another limitation of this study. In our series, we were able to control the correct placement of the RFA probes by performing an immediate postoperative cross section through the ablated tumor tissue and sampling the tumor tissue for histologic evaluation (Fig. 1). Intra-operative ultrasound may be helpful for RFA probe placement in atelectasis; during the ablation procedure in the ventilated lung, however, it is unfortunately not applicable.

An intrapulmonary pseudo-aneurysm in the same lobe occurring 6 weeks after RFA must be regarded as a late complication that ultimately may require surgical resection or trans-catheter coil embolization [13].

5 Conclusion

RFA appears to be a safe and feasible technique. The assumed efficiency rates of bipolar and multipolar RFA in our series were equivalent to and in conformity with other authors, and these rates are likely to be inferior to the results of surgical resection [8,9,11]. Our findings demonstrate that RFA, even when properly targeted, still can result in an incomplete local ablation. The rate of incomplete ablation seems to be higher in lesions of the adenocarcinoma subtype. Therefore, we conclude that this approach should be reserved only for patients considered ‘medically inoperable’ in primary and secondary lung neoplasms. Surgical resection, when oncologically and medically feasible, remains the preferred approach.

☆

Presented at the 18th European Conference on General Thoracic Surgery, Valladolid, Spain, May 30–June 2, 2010.

☆☆

The RFA probes and the bipolar CelonPro Surge^® radiofrequency ablation system were provided by CELON^®.