Radiofrequency ablation of stage IA non–small cell lung cancer in patients ineligible for surgery: results of a prospective multicenter phase II trial

Lobectomy and lymph node resection remains the first-line treatment and the best option for stage I non-small-cell lung cancer (NSCLC) [1]. Despite the development of sub-lobular resection [2] to limit the functional damage of lobectomy, approximately 20% of patients remain ineligible for surgery (mostly due to comorbidities). The 5-year overall survival (OS) without treatment ranges from 6 to 14% for these patients [3]. Until recently, these ‘non-surgical’ patients were treated with conventional radiotherapy [4]. However, current treatments such as stereotactic body radiotherapy (SBRT) [5‐8] and thermal ablation [9‐15] are increasingly offered as alternative therapies for non-surgical candidates. Radiofrequency ablation (RFA) has been the most commonly used and evaluated image-guided thermal ablation technique. RFA was shown to be feasible and safe in highly selected patients [9‐14]. One of the advantages of RFA is that it is a stand-alone therapy which can be repeated in case of local failure. RFA has the ability to eradicate a tumor with heat while causing minimal damage to the surrounding normal lung tissue. The objectives of this multicenter study were to prospectively analyze the efficacy and the tolerance of RFA in stage I NSCLC for non-surgical patients.

The study was approved by the ethics committee and was funded by the Programme Hospitalier de Recherche Clinique (PHRC, grant number A00812–53) and was conducted in six French institutions. This study followed Consolidated Standards of Reporting Trials guidelines.

Patient characteristics are given in Table 1. A total of 42 patients were included in the study out of which 36 patients were eligible and 32 patients were eligible and assessable for the main criteria (Fig. 1). Majority of patients were ECOG1 and the tumor characteristics (pathology, dimensions etc.) are presented Table 1. Histological analysis revealed an adenocarcinoma in 60% of the patients. T1a and T1b tumors were almost equally distributed. PET-CT at inclusion revealed that SUVmax ranged from 2.7 to 22.2 (mean = 6.9; median = 6.7) for all patients and from 2.7 to 14.5 (mean = 6.3; median = 5.4) for assessable patients. RFA treatments were performed under general anesthesia, except for 4 interventions where thoracic epidural anesthesia was administered. A 4 cm electrode was used in 60% of patients, even though half of the tumors were T1a. Moreover, overlapping ablations were frequent and were performed in more than half of the tumors. A 48 h hospitalization was systematically planned for every patient as part of the protocol for this old and fragile population with multiple comorbidities. Hospitalization duration ranged from 2 to 19 days (mean = 4 days; median = 3 days). An X-ray for pneumothorax was performed during ablation in 26 treatments (83.87%), with the need for a chest tube in 17. Chest tubes were always removed within 2 days. Other CT findings were pleural effusion (n = 4; 1 mild and 3 minimal), alveolar hemorrhage (n = 2; subsegmental), hemothorax (n = 1); however, no clinical symptoms were noted in relation to these findings. Two major adverse events were recorded in 2 patients (4.7% of the overall population) and 3 delayed minor adverse events were noted (Table 2).

This multicenter study is the second prospective study on RFA and non-surgical patients with stage I NSCLC. Dupuy et al. have published [20] the results of American College of Surgeons oncology group Z4033 trial, with a local recurrence-free rate of 68.9 and 59.8% at 1 and 2 years, respectively. In our study, the local control rate was 84.38% (95% CI, [67.21–95.72]) at 1 year and 81.25% (95% CI, [54.35–95.95]) at 3 years. In our study, local recurrence refers to a recurrence in the RFA zone (Fig. 4) whereas in the Z4033 trial it refers not only to a recurrence in the RFA site but also in the primary tumor lobe and the hilar lymph node. Our results are in accordance with our previously published retrospective study, with local control rates of 88.5 and 78.9% at 1 and 3 years, respectively [21].

In this study, expandable multitined electrodes were used with an electrode array diameter of the at least 10 mm larger than the tumor, such oversizing the ablation volume was previously reported to increase the success rate of the treatment [22]. Moreover, a pathological study following RFA on NSCLC demonstrated the benefit of oversizing the ablation zone by encompassing the tumor; < 5 mm margins were observed in patients with incomplete ablation whilst a mean margin of 8 mm was noted in complete ablations [23]. This emphasizes the need for an extensive RFA to take into consideration the microscopic extension of NSLCC. Beland et al. [10] recommended an ablation zone “that includes the primary tumor plus at least an additional 8–10 mm of ablation beyond the visible tumor margin in all directions”. In our series, all the patients apart from three were treated with a 3.5 or a 4 cm electrode for a mean tumor size of 20.7 mm, probably explaining the good high rate of local control we report herein. Other ablation techniques such as microwaves may also improve the local control rate by providing a larger ablation volume compared to RFA, as reported in an animal study [24]. Cryotherapy [25], offers the possibility of using multiple probes to obtain a larger ablation volume.

Ever since the use of thermal ablation on lung tumors, different studies [10, 23, 26] have shown that repeated imaging is important to follow the variations in size of the ablation zone, to identify the different imaging patterns of lung tumors after RFA and finally, to distinguish total ablation from local tumor progression. The strength of this study is the strict prospective continuous imaging follow-up during 3 years, for all patients. Recurrence was determined using CT and PET-CT, as well as the possible use of biopsy. In one case, there was discordance between CT and PET-CT at 1 year, a local recurrence was confirmed by a biopsy. A new RFA was proposed for 2 out of 5 local failures. Moreover, with the possibility of re-treatment, loco-regional progression did not seem to negatively impact the OS [27]. The 1- and 3-year OS rates were of 91.67 and 58.33%, respectively which are similar to the American prospective trial results [20] with an OS rate of 86.3 and 69.8% (at 1 and 2 years, respectively). Furthermore, our patient population was old, frail and contraindicated for surgery. We note that RFA results are difficult to compare with those of sub-lobular resection and Stereotactic Body Radiation Therapy (SBRT). Patients for thermal ablation are often the frailest and the oldest [28], in this series all patients were medically inoperable and at 3 years, half of the patients died of causes other than cancer. Nevertheless, SBRT resulted in similar survival rates (86 and 45% at 1 and 3 years, respectively) in an elderly population [6]. The local control rate at 3 years was 89% which is comparable to ours. However, in this series at 12 months of follow-up, imaging (CT scans) was available for 88% of patients, and PET-CT was obtained only when there was suspicion of disease recurrence. Moreover, histological confirmation of malignancy was available only for 39% of patients vs 100% in our series. The absence of pathological confirmation of cancer in a significant number of patients may raise the concern of whether those lung lesions may be benign. This is a weakness of numerous SBRT studies. Comparison with surgery is difficult as patients treated with ablation are not surgical candidates. Nevertheless, using propensity scores to match patient subgroups, Kwan et al. recently demonstrated comparable overall and lung-cancer specific survival between RFA and sub-lobular resection in patients ≥65 years old with stage IA or IB NSCLC [29]. The lack of pathological mediastinal lymph node information, which may impair patient survival compared to surgery, remains a limitation of RFA and SBRT for early stage NSCLC.

RFA seems to have been well tolerated. No delayed pneumothorax occurred and all pneumothoraces cleared rapidly. In a large series of 1000 RFA sessions [30], pneumothorax requiring pleural sclerosis was rare (1.6%). Besides pneumothorax, we observed two major adverse events. An ECOG 2 patient with a FEV of 800 mL died a few weeks later from respiratory decompensation that has been attributed to RFA. Exacerbation of interstitial pneumonia has been mentioned as a major cause of death [30] and has also been reported after lung surgery and radiotherapy. Consequently, the indication of a thermal ablation must be carefully weighted in patients with lung insufficiency, even though it is difficult to propose a clear lower threshold for respiratory function for lung RFA [23]. The stable lung function and the positive impact on QoL in our series also highlight the good tolerance of the RFA. QLQ C 30 tests propose large scale measurements and are not focused on lung disorders. A majority of patients considered that RFA did not modify either their life status perception or their main vital function, including lung function. A significant deterioration of cognitive function and fatigue was observed.

This prospective study confirms that RFA is an effective, safe and well-tolerated option in the treatment of patients presenting stage IA NSCLC who are ineligible for surgery. The 3-year local control rate is encouraging and similar to SBRT results. The survival is comparable to another recent prospective study confirming the place of thermal ablation in this fragile non-surgical population. A prospective trial comparing RFA to other techniques such SBRT is required.