It was reported that approximately 50–70% of Asian and 30–40% of non-Asian NSCLC patients harbor gene mutations [24]. Of these, EGFR mutation occurs in 55% of Asian patients and 15% of non-Asian patients with lung adenocarcinoma, followed by ALK rearrangement occurs in 5–8%, and other mutations are limited to 5% of non-squamous NSCLC [25, 26]. According to the latest cancer statistics in the USA, the median overall survival of lung cancer increased to 13 months and the three-year relative survival rate was up to 38% [3]. Molecular targeted therapy prompts this progression, which is mainly against gene mutations and is recommended as the standard treatment [6]. Percutaneous CNB was one of the primary methods to obtain specimens for pathological diagnosis and genomic testing, and was especially suitable for peripheral or unresectable lung lesions, with a diagnostic accuracy rate of 90% [6, 20, 27]. It should be noticed that the patients were at a high risk of occurring pulmonary hemorrhage or pneumothorax when CNB was performed for pulmonary lesions that with hypervascularity or were adjacent to vessels or bronchi, which may influence the precise biopsy.

Thermal ablation was recommended as a treatment option for stage I NSCLC patients who have contraindications to surgery or radiotherapy, or be considered as a salvage treatment for patients who developed progression on EGFR or ALK therapy [6]. Thermal ablation can not only conduct the coagulative necrosis of tumor tissues but also cause the collapse of small or medium-sized blood vessels depending on the hyperthermia directly, to some extent, has hemostatic effects [28]. In general, the biopsy was supposed to be performed before ablation to obtain accurate diagnoses. Nevertheless, a high risk of occurring hemorrhage existed in pre-ablation CNB, especially for GGO nodules and lesions with small diameters or adjacent to vessels, which may disturb the subsequent ablation [11]. In 2012, a retrospective study analyzed 33 lung neoplasm patients treated with simultaneous CNB and RFA, and found a local tumor control rate of 77% in a median follow-up of one year [29]. Then, Wang et al. [30] attempted simultaneously coaxial MWA and biopsy in suspicious malignant lung lesions and found this procedure has lesser AEs but similar efficacy when compared with separate procedures, which could achieve the diagnosis and treatment concomitantly and was recommended by SIR standards for the lesions with a high risk of hemorrhage that may interfere with the ablation [12]. Nevertheless, a high risk of occurring hemorrhage still existed in the biopsy immediately before ablation despite a synchronous procedure, which may lead to indeterminate tumor positioning and increase the risk of incomplete ablation [11]. Therefore, several authors attempted to perform CNB immediately after ablation in highly suspicious malignant lung lesions, and indicate the accuracy and safety of this procedure [13,14,15,16,17,18].

It was reported that the pathological diagnosis rate of CNB immediately after thermal ablation ranged from 70 to 100% [13,14,15,16,17,18], with the potential mechanisms of apoptosis progressing in tumor cells subjected to hyperthermia gradually, and cell morphology remaining in the tumor within the first month after ablation [14, 31,32,33]. In 2016, Hasegawa et al. [14] performed a biopsy immediately after RFA for three patients with lung malignancy, including two metastases and one adenocarcinoma, all of whom achieved the precise pathological diagnosis. Then, a study attempted coaxial biopsy immediately after RFA, and found histological subtype can be distinguished in 70% of patients despite most of the tumors being lung metastases [16]. Wei et al. [18] performed CNB immediately after MWA in 69 confirmed NSCLC patients, and found the pathological diagnosis can be distinguished in 85.3% of patients and 69.1% of patients have identical histological subtypes when compared with previous results, which indicated that the accuracy of post-ablation CNB for determining the tumor subtypes. Another study conducted by Hasegawa et al. [15] enrolled 13 solid pulmonary lesions and six GGO nodules that had undergone CNB immediately after RFA, with the overall pathological diagnosis rate reaching 79% while that was only 50% for GGO nodules. In a study of 74 patients with GGO nodules, the pathological diagnosis rates of pre- and post-MWA CNB were 85.1% and 74.3%, respectively, and the histological subtypes could also be distinguished, which indicated the comparability of pre- and post-ablation CNB [17]. Compared to the above studies, all of the patients enrolled in our study were NSCLC with solid lesions, and pathological diagnoses between pre- and post-ablation CNB were compared directly, with a high concordance rate of 93.9%. Two patients (6.1%) presented with the absence of atypical cells in post-ablation CNB and the potential interpretation was the overlong ablation time between twice CNB that lead to the carbonization of specimens. Moreover, the attenuation of immunohistochemistry staining in post-ablation CNB was also found, which is per the results from a previous study [15].

Two studies have investigated the accuracy of genomic testing in CNB immediately after thermal ablation, with the technical success rate ranging from 74 to 84% [13, 15]. In 2018, a study reported that EGFR and KRAS mutations can be detected in 74% of the specimens obtained from post-ablation CNB although GGO lesions were included and the percentage of NSCLC was less than 50% [15]. Then, Chi et al. [13] presented a success rate of 100% and 84% could be achieved for pre- and post-MWA CNB in GGO nodules, respectively, with no significant difference being found despite only EGFR mutation being detected. Of these, the MWA between twice CNB was at a power of 20 W and this procedure could decrease the incidence of AEs. However, the quantitative analysis of DNA/RNA extracted from specimens is scarce previously. Our study verified the applicability of these results in solid NSCLC and found a concordance rate of 90.9% between pre- and post-ablation CNB, which was higher than that in previous studies [13, 15]. Besides, the scope of gene mutations was extended, with the verification of accuracy in MET, ROS1, and BRAF. Adequate amounts of DNA/RNA extracted from specimens were critical for the quantitative analyses of genomic testing. In this study, the mean concentration of DNA/RNA in post-ablation CNB was significantly lower than that in pre-ablation CNB, which demonstrated that the post-ablation biopsy influences the DNA/RNA level, in other words, has an impact on quantitative analyses for genomic testing. The potential mechanism was that the heat delivered from the ablation antenna damaged the DNA in tumor cells and induced apoptosis, and the RNA structure was unstable and was also prone to be damaged by hyperthermia [34]. In theory, the more heat absorbed by the tumor cells, the more severe damage is brought to DNA/RNA. Therefore, we attempted to investigate the potential cut-off values of ablation-related parameters in MWA between twice CNB, which could not only achieve the qualitative analysis of genomic testing but also had no significant impacts on DNA/RNA levels. This study indicated that the quantitative analyses of DNA/RNA may not be influenced significantly when the MWA between pre- and post-ablation CNB was performed at a power of 30 or 40 W and ablation time within five minutes.

Several limitations in this study should be presented. First, the selection bias may exist due to the retrospective nature of this study. Second, the patients are from single-center and the sample size was still limited. Third, although both DNA- and RNA-related gene mutations were detected, the results are still needed to be verified beyond the scope of gene mutations in this study. Finally, the precise evaluation of heat distribution in the tumor was complicated and vulnerable to being affected by multiple factors, including the tumor volume, margins, density, intratumoral vascularity, blood supply, ablation power, duration, and so on, and further investigation was warranted to assess the potential impacts of these factors on genomic testing precisely.