Thermal ablation versus radiotherapy for inoperable stage III non-small cell lung cancer: a propensity score matching analysis

Abstract Objective To compare the survival benefits of thermal ablation (TA) and radiotherapy in inoperable patients with stage III non-small cell lung cancer (NSCLC). Method A retrospective analysis was conducted using the data from the Surveillance, Epidemiology, and End Results (SEER) program. Propensity score matching (PSM) was conducted to balance potential baseline confounding factors. Survival analyses were conducted using Kaplan–Meier and Cox regression methods. Results The present study included 33,393 inoperable patients with stage III NSCLC, including 106 patients treated with TA and 33,287 patients treated with radiotherapy. No statistical difference in overall survival (OS) (p = .065) or cancer-specific survival (CSS) (p = .996) was found between the patients treated with TA and those treated with radiotherapy. Using 1:3 matching, a matched cohort of 420 patients (105 patients treated with TA, 315 patients treated with radiotherapy) was identified. The differences in OS (p = .177) and CSS (p = .605) were still not significant between the radiotherapy and TA groups after PSM. According to subgroup analyses, TA showed comparable survival benefits in almost all subgroups compared to radiotherapy. Conclusion For inoperable stage III NSCLC, the survival benefit of TA was comparable to radiotherapy. TA may be a potential therapeutic modality for inoperable stage III NSCLC.


Introduction
Lung cancer is one of the most common cancers and a leading cause of cancer-related deaths worldwide [1].Surgical resection is the primary choice for early-stage lung cancer [2].However, many tumors are already unresectable at diagnosis because of their locally advanced or metastatic stages [3].In addition, a proportion of patients are not medically operable due to poor cardiac and pulmonary fitness, which results from advanced age and severe comorbidities [4].Thus, non-surgical therapeutic modalities for non-small cell lung cancer (NSCLC), including radiotherapy, targeted therapy, immunotherapy, and thermal ablation (TA) therapy, have recently been of increasing interest [5][6][7][8].
TA is a safe and effective treatment alternative to surgical resection for NSCLC [8].TA can cause thermal injuries to tumors using extremely high temperatures, which results in irreversible cellular damage and coagulation necrosis [9].According to the various energy sources used, TA can be classified as microwave ablation (MWA), radiofrequency ablation (RFA), or laser ablation, among others [10].Among these TA technologies, RFA is the most common technology applied in cancer therapy [11].Compared to RFA, MWA can more quickly heat tissue to higher temperatures, and the active heating zone of MWA is larger and more reproducible [10].Based on the current guidelines, some patients with stage I NSCLC are not amenable to surgical resection, and they are considered the best candidates for TA [12,13].However, there is no consensus on whether TA can benefit inoperable stage III NSCLC patients.
Stage III NSCLC is a highly heterogeneous disease with significant differences regarding clinicopathological characteristics and treatment modalities [14].According to the current guidelines, radiotherapy is the foundation of curative treatment for inoperable stage III NSCLC [15,16].Concurrent chemoradiotherapy is the standard treatment for inoperable stage III NSCLC.Patients for whom concurrent chemoradiotherapy is not suitable should be offered sequential chemoradiotherapy.Moreover, radiotherapy alone is still beneficial to inoperable patients with stage III NSCLC who cannot tolerate chemotherapy [17].Previous studies have indicated that radiotherapy and TA offered comparable survival benefits for stage I NSCLC [8,18].However, few studies have compared the efficacies of radiotherapy and TA in stage III NSCLC patients.Based on data from the Surveillance, Epidemiology, and End Results (SEER) program, this study aimed to compare the survival benefits of TA and radiotherapy in inoperable patients with stage III NSCLC.

Data source
The SEER program is a population-based program established by the National Cancer Institute.The SEER database is the largest publicly available cancer database, covering approximately one-third of the U.S. population [19].The SEER program collected information on patient demographics, tumor characteristics, treatment modalities, and survival data.Because the SEER database is publicly available and all records have been de-identified, no additional ethical approval or informed consent was required.3) patients who received TA or radiotherapy.The following exclusion criteria were applied: (1) patients treated with other surgical procedures in addition to TA, or a lack of information regarding surgery status and/or type; (2) patients treated with TA in combination with radiotherapy; (3) cases identified from autopsy or death certificates and (4) patients with missing data on the studied variables such as age, sex, laterality, T stage, N stage and marital status, among others.

Study population
Demographic data extracted from the SEER database included age at diagnosis, sex, race, and marital status.Tumor characteristics included primary tumor site, laterality, pathological type, histological grade, TNM substage, T stage, N stage, and tumor size.Treatment modalities included surgery, radiotherapy, and chemotherapy.The TNM stage in the SEER database was based on the sixth edition of the TNM classification.The patients were divided into a TA group and a radiotherapy group according to the treatment modalities they received.The primary endpoints were cancer-specific survival (CSS) and overall survival (OS).CSS was defined as the time interval between diagnosis and death resultant from NSCLC.

Statistical analysis
Analyses were performed with SPSS version 26.0 (IBM Corporation, Armonk, NY, USA) and R version 4.0.0 (R Foundation for Statistical Computing, Vienna, Austria).The propensity score matching (PSM) method was used to balance the patient characteristics and reduce the influence of potential confounders.The matching approach was 1:3 nearest neighbor with a caliper of 0.05.Baseline characteristics were compared between groups using the chi-squared test or Fisher's exact test.Survival analysis was performed using the Kaplan-Meier method and log-rank test.Cox proportional hazards regression was used for univariable and multivariable analysis.Variables with p < .05 in the univariable analyses were entered into multivariable models.Differences were considered statistically significant with a p < .05.

Baseline characteristics
In total, 33,393 patients with stage III NSCLC, of whom 33,287 patients (99.68%) underwent radiotherapy and 106 patients (0.32%) underwent TA therapy, were included in this study.As shown in Table 1, the baseline characteristics of patients were significantly different between the radiotherapy and TA groups.Specifically, the patients receiving TA had an older age at diagnosis (p ¼ .003),an advanced TNM substage (p < .001),and a smaller tumor size (p < .001)compared to the radiotherapy group.The TA group had a higher proportion of stage T4 (74.53%, p < .001),whereas the radiotherapy group had a higher proportion of patients with lymph node metastases (p < .001).Moreover, the patients receiving radiotherapy tended to receive chemotherapy (75.32%, p < .001)more often than those receiving TA.Therefore, a 1:3 PSM analysis was performed to reduce the effects of differences in baseline characteristics.After PSM, a matched cohort of 420 patients (105 patients treated with TA, 315 patients treated with radiotherapy) was identified.The baseline characteristics were well balanced in these two groups.There were no statistically significant differences in any of the baseline characteristics between the groups after PSM (p > .05).In the matched cohort, more than half were males, 87.94-89.52%were white, 65.71-66.67%were right laterality, and over half of tumors were squamous cell carcinoma.The majority of patients had stage T4 disease, and more than half of each group were stage N0.In addition, approximately half of the tumors were smaller than 3 cm, and more than 80.00% of either group did not receive chemotherapy.

Survival analysis in the full cohort
According to the Kaplan-Meier analyses, no statistical difference was found in OS between the radiotherapy and TA groups (p ¼ 0.065, Figure 1(A)).The one-and three-year OS of the radiotherapy and TA groups was 54.44% versus 46.23% and 22.39% versus 20.64%, respectively.The median OS was 14 months (95% confidence interval [CI]: 13.779-14.221months) and 10 months (95% CI: 4.235-15.765months) for the radiotherapy and TA groups, respectively.Meanwhile, the CSS of patients receiving TA was also comparable to those receiving radiotherapy (p ¼ .996, Figure 1(B)).The one-and three-year CSS of the radiotherapy and TA groups was 58.36% versus 57.61% and 27.12% versus 30.24%, respectively.The median CSS was 16 months (95% CI: 15.726-16.274months) and 21 months (95% CI: 13.484-28.506months) for the radiotherapy and TA groups, respectively.In addition, the Cox regression analyses showed that TA was not a risk factor for OS (hazard ratio [HR] ¼ 1.20, 95% CI: 0.99-1.47,p ¼ .069)and CSS (HR ¼ 1.00, 95% CI: 0.79-1.26,p ¼ .989)compared with radiotherapy (Table 2).Other variables, including age, sex, race, primary tumor site, pathological type, T stage, N stage, tumor size, chemotherapy, and marital status, were independent prognostic factors for OS and CSS.

Subgroup analysis stratified by age
Subgroup analyses stratified by age were conducted using the Kaplan-Meier method.As shown in Figure 3, the patients treated with radiotherapy had better OS than those treated with TA when age at diagnosis was <65 years (p ¼ .002, Figure 3(A)) and 75-84 years (p ¼ .021, Figure 3(C)).However, in patients aged !85 years, the patients treated with TA had longer OS than those treated with radiotherapy (p ¼ .049, Figure 3(D)).In patients aged 65-74 years, the OS was comparable between the TA and radiotherapy groups (p ¼ .731, Figure 3(B)).Figure S2 shows the Kaplan-Meier curves of CSS for patients with different ages at diagnosis.The patients receiving radiotherapy had better CSS than patients receiving TA when age at diagnosis was <65 years (p ¼ .039, Figure S2A).However, in patients aged !85 years, the patients treated with TA had longer CSS than those treated with radiotherapy (p ¼ .032, Figure S2D).There was no statistical difference in CSS between the radiotherapy and TA groups when age was 65-74 years (p ¼ .283, Figure S2B) or 75-84 years (p ¼ .304, Figure S2C).After PSM, no statistical difference was identified in OS between the TA and radiotherapy groups when age at diagnosis was <65 years (p ¼ .104, Figure 4(A)), 65-74 years (p ¼ .555, Figure 4(B)), or !85 years (p ¼ .327, Figure 4(D)).The patients treated with radiotherapy had longer OS than those treated with TA when age at diagnosis was 75-84 years (p ¼ .004, Figure 4(C)).As shown in Figure S3, there was no statistical difference in CSS between patients receiving radiotherapy and those receiving TA in any age subgroup after PSM (p > .05).

Subgroup analysis stratified by TNM substage
In patients with stage IIIA NSCLC, the OS (p ¼ .001, Figure 5(A)) and CSS (p ¼ .012, Figure 5(B)) of patients receiving radiotherapy were longer than those receiving TA.The threeyear OS and CSS rates of the radiotherapy and TA groups was 25.70% versus 9.09% and 31.15%versus 15.73%, respectively.However, the OS (p ¼ .882, Figure 5(C)) and CSS (p ¼ .145, Figure 5(D)) were not statistically different between the patients receiving TA and those receiving radiotherapy in the stage IIIB subgroup.The three-year OS and CSS rates of the radiotherapy and TA groups was 20.12% versus 23.65% and 24.33% versus 33.75%, respectively.
Figure S4 shows the subgroup analyses stratified by TNM substage after PSM.As shown in Figure S4A, the OS of patients who received radiotherapy remained higher than those receiving TA in the stage IIIA subgroup (p ¼ .033).However, the CSS of patients receiving radiotherapy was comparable to those receiving TA in the stage IIIA subgroup after PSM (p ¼ .060, Figure S4B).In the stage IIIB subgroup, the OS (p ¼ .600, Figure S4C) and CSS (p ¼ .761, Figure S4D) were still comparable between the patients receiving TA and those receiving radiotherapy after PSM.

Subgroup analysis after PSM
The Cox analyses were performed to explore whether TA was associated with better OS in different population subgroups (Figure 6).With the exception of 75-84 years, stage IIIA, and stage N2, all subgroups had comparable effects on OS from TA therapy and radiotherapy.Moreover, the subgroup analyses of CSS showed no difference between TA and radiotherapy in any subgroup (Figure S5).

Discussion
In this study, a total of 33,393 patients with stage III NSCLC, including 33,287 cases who received radiotherapy and 106 cases who received TA, were retrieved from the SEER program.The findings suggested that the survival benefit of TA was comparable to radiotherapy in inoperable patients with stage III NSCLC.
Stage III NSCLC is highly heterogeneous in terms of clinicopathologic characteristics and treatment modalities [14].According to the current guidelines, radiotherapy is the basis for the treatment of inoperable stage III NSCLC [15,16].Concurrent chemoradiotherapy or sequential chemoradiotherapy is recommended for inoperable stage III NSCLC.For those who cannot tolerate chemotherapy, radiotherapy alone is still considered effective [17].In the current study, the median OS of inoperable stage III NSCLC patients receiving radiotherapy was 14 months, which was significantly better than the previously reported median OS of patients without treatment [20].These results revealed that radiotherapy was an effective therapy for stage III NSCLC.TA has been proven as an effective and minimally invasive treatment for NSCLC [8,12].TA has several advantages over surgical resection, including fewer complications, shorter hospital stays, and faster recovery times [10].Several studies have shown that TA could benefit patients with both locally advanced and metastatic NSCLC.Findings from a recent retrospective cohort study reported that TA could benefit patients with stage II-III NSCLC [21].Both OS and CSS for patients treated with TA were significantly longer than those for patients who did not undergo TA.Another retrospective study by Yang et al. showed that patients receiving chemotherapy with stage IV NSCLC could gain survival benefits from TA [22].However, few studies have examined whether TA is sufficiently effective for inoperable stage III NSCLC compared to radiotherapy.
In the present study, significant differences in many baseline characteristics were observed between the radiotherapy and TA groups.Compared to the patients who received radiotherapy, the patients receiving TA had an older age at diagnosis (p ¼ .003),an advanced TNM substage (p < .001),and a smaller tumor size (p < .001).These findings were similar to a previous study based on the SEER Database by Liang et al. [23].They compared the survival outcomes of TA and stereotactic body radiotherapy in inoperable stage I NSCLC patients.They found that the socioeconomic factors and clinicopathological characteristics were significantly different between the TA and radiotherapy groups.In addition, compared to the patients who received TA, the patients receiving radiotherapy had a greater tendency to receive chemotherapy in the present study (p < .001).This phenomenon may be related to the younger median age at diagnosis in the radiotherapy group.As a result of their relatively better physical condition, younger patients generally received more aggressive treatment than elderly patients [24].Overall, numerous confounding factors may influence the results of direct comparisons.As a result, PSM was performed to compare the survival outcomes of radiotherapy and TA in inoperable stage III NSCLC patients.
In the full cohort, no statistical difference in OS and CSS was found between the radiotherapy group and TA group.After 1:3 matching, the OS and CSS of patients treated with TA were still comparable to those who received radiotherapy in the matched cohort.These results showed that TA was an effective therapy for inoperable stage III NSCLC.In addition, the multivariable Cox regression analysis revealed that age, sex, race, primary tumor site, pathological type, T stage, N stage, tumor size, chemotherapy, and marital status were independent prognostic factors for OS and CSS.These prognostic factors for NSCLC were consistent with the previous studies [21,23].
The age-stratified analysis revealed that the OS and CSS for patients treated with TA were significantly longer than those treated with radiotherapy when age at diagnosis was !85 years in the full cohort.However, when age at diagnosis <65 years, OS and CSS were significantly longer in the radiotherapy group than in the TA group.After matching, the OS and CSS for patients treated with TA were comparable to those treated with radiotherapy in almost all age subgroups, except for patients aged 75-84 years.A previous study by Zeng et al. showed that the survival benefit of TA in stage I NSCLC was associated with age [25].In patients aged >75 years, the OS and CSS did not differ statistically between the wedge resection and TA groups (p > .05).However, in patients aged 75 years, the OS and CSS of patients receiving TA were significantly shorter than those treated with wedge resection (p < .05).Another study on stage IV NSCLC also showed that the survival benefits of TA were different in different age subgroups [22].When age at diagnosis was !70 years, TA improved the prognosis of patients treated with chemotherapy.However, the survival benefit of TA was not statistically significant when age at diagnosis was <70 years.Numerous studies have shown that the survival benefits to NSCLC from various therapies, including chemotherapy, radiotherapy, and immunotherapy, were affected by age [24,26].However, the impacts of age on TA efficacy in NSCLC patients remain unclear.Additional studies are needed to explore the relationship between TA efficacy and age in NSCLC.
The OS and CSS, both before and after PSM, did not statistically differ between the patients receiving TA and those receiving radiotherapy in the stage IIIB subgroup.However, in the stage IIIA subgroup, the patients treated with radiotherapy had longer OS and CSS than those treated with TA before PSM.After PSM, the OS of patients receiving radiotherapy was still better than those receiving TA, whereas the CSS did not differ between these two groups.These results also suggested that TA was an effective therapy for inoperable patients with stage III NSCLC compared to radiotherapy.However, for inoperable patients with stage IIIA NSCLC, radiotherapy was still recommended as the preferred treatment.
The survival benefits from TA in patients with stage III NSCLC may be attributable to several factors.First, TA can lead to thermal damage to tumors using extremely high temperatures, which results in irreversible cellular damage and coagulation necrosis [9].Second, TA can increase the cell membranes permeability, which could facilitate the delivery of anti-cancer drugs to the tumor site [22].The efficacy of anti-cancer drugs, including chemotherapeutic drugs, targeted therapeutic drugs, and immunotherapeutic drugs, can be increased after TA therapy [22,27,28].Finally, recent studies revealed that TA could lead to the release of immunogenic intracellular substances, which can activate anti-tumor immune responses [29].TA can promote dendritic cell maturation and T cell activation, which plays an essential role in anti-tumor immunity [30].
The present study had several limitations.First, the information on TNM stages in the SEER database was based on the sixth edition of the TNM classification.The latest eighth edition of the TNM staging system for NSCLC has several changes compared to the sixth edition of staging system [31].Stage III NSCLC was subdivided into IIIA and IIIB in the sixth edition of staging system.However, stage III was subdivided into IIIA, IIIB, and IIIC in the latest eighth edition of staging system, which might affect the results.Second, the TA modalities, such as RFA, MWA, and cryoablation, were not distinguished in the SEER database.We could not extract detailed information on the TA modalities used for each patient, which might have an effect on the results.Third, the information regarding treatment in the SEER program was limited.The SEER database did not provide specific details regarding radiotherapy, including total dose, dose fractionation, and overall treatment time, among others.The detailed information on chemotherapy, such as regimen, dose, and duration, were also unavailable.In addition, the SEER database lacks data on immunotherapy and targeted therapy.Thus, these factors could not be investigated in this study.Finally, the present study was a retrospective study.The credibility is weaker than a prospective or intervention study.The findings need to be confirmed in prospective, randomized, controlled trials.

Conclusions
In conclusion, for inoperable stage III NSCLC, the survival benefit of TA was comparable to radiotherapy.TA may be a potential therapeutic modality for inoperable stage III NSCLC.
Information on patients with NSCLC diagnosed in 2004-2015 were retrieved from the 'Incidence-SEER Research Plus Data, 18 Registries, Nov. 2020 Sub (2000-2018)' dataset, based on the 'Site Recode ICD-O-3/WHO 2008' and 'ICD-O-3 His/Behav, malignant'.The data were retrieved using SEER Ã Stat software (version 8.4.0).A flow chart is shown in Figure S1.Inclusion criteria were as follows: (1) age at diagnosis greater than 18 years; (2) patients diagnosed with stage III NSCLC and (

Figure 1 .
Figure 1.Kaplan-Meier survival curves for OS (A) and CSS (B) in inoperable patients with stage III NSCLC.OS: overall survival; CSS: cancer-specific survival; NSCLC: non-small cell lung cancer.

Figure 2 .
Figure 2. Kaplan-Meier survival curves for OS (A) and CSS (B) in inoperable patients with stage III NSCLC after PSM.OS: overall survival; CSS: cancer-specific survival; NSCLC: non-small cell lung cancer; PSM: propensity score matching.

Figure 5 .
Figure 5. Kaplan-Meier survival curves between the thermal ablation and radiotherapy groups when TNM substage was IIIA (A, B) or IIIB (C, D) before PSM.PSM: propensity score matching.

Figure 6 .
Figure 6.Forest plot depicting subgroup analysis of OS between the TA and radiotherapy groups after PSM.NOS: not otherwise specified; SCC: squamous cell carcinoma; ADC: adenocarcinoma; LCC: large cell carcinoma; Inf: infinity; OS: overall survival; TA: thermal ablation; PSM: propensity score matching.

Table 1 .
Characteristics of patients with stage III NSCLC before and after PSM, n (%).
NSCLC: non-small cell lung cancer; PSM: propensity score matching; NOS: not otherwise specified.

Table 2 .
Univariable and multivariable analyses of OS and CSS in stage III NSCLC patients receiving radiotherapy or thermal ablation before PSM, n (%).