Discussion
For stage pIII-N2 NSCLC after pneumonectomy followed by adjuvant chemotherapy, the results showed good compliance and safety of PORT without any severe radiation pneumonitis and esophagitis. Compared with the control, PORT significantly improved OS, DFS, LRFS and DMFS, although there were more patients with R1/R2 resection in PORT group. To our knowledge, this is the first study focusing on the safety and efficacy of PORT for stage pIII-N2 NSCLC after pneumonectomy followed by chemotherapy.
Safety of radiotherapy after pneumonectomy is the most important concern in clinical practice. Pneumonectomy leads to great change on cardiopulmonary function. The removal of an entire lung decreases the tolerance to radiotherapy and amplifies potential risks, leading to few application of radiotherapy. In our study, 32 patients (26.9%) received PORT. In 104 patients with R0 resection, 24 cases (23.1%) were in the PORT group, which was consistent with previous reports (less than 25%) [
5,
10,
11], and was significantly less than patients receiving PORT after lobectomy (43.4%) [
12]. The postoperative residual tumor in gross or positive margin in microscopy was an important indication of PORT. In this study, although more patients (8/15, 53.3%) with non-R0 resection received PORT comparing with those with R0 resection (23.1%), there were still nearly half of the patients not receiving PORT, which revealed the concern on safety of PORT in both physicians and patients.
Contrary to concerns on the risks of PORT after pneumonectomy, this study showed that PORT was well tolerant with high compliance (completion rate of 93.8%). Only one patient (3.1%) suffered grade 2 radiation pneumonitis and one patient (3.1%) suffered grade 3 radiation pneumonitis, while no severe radiation pneumonitis (≥ grade 4) occurred. The incidence of ≥ grade 2 radiation pneumonitis in this study was significantly lower compared with the toxicities in previous studies of radical chemoradiotherapy (10–35%) [
13‐
16], and previous studies of radiotherapy or chemoradiotherapy after lobectomy [
17,
18]. Bradly [
17] reported ≥grade 3 radiation pneumonitis with the incidence rate of 6% in 88 patients receiving adjuvant chemotherapy and concurrent radiotherapy after surgery, while pneumonectomy accounted for 14%. Besides, Zhao [
18] demonstrated that the incidence of ≥grade 2 radiation pneumonitis was 13% after lobectomy, and 0% after pneumonectomy, which verified the safety of radiotherapy after pneumonectomy. Zhao believed that strict lung dose constraint and experienced beam arrangement (outside pulmonary parenchyma) were all protective and beneficial factors, which might be the reason for no serious radiation pneumonitis after pneumonectomy. In addition, in our research, low prescription dose (median dose 54Gy) and modern techniques could effectively reduce the volume and dose of the contralateral lung (median V20 4.75%), which was also an important reason for the low occurrence of radiation pneumonitis. Radiation esophagitis was another common side effect of radiotherapy. However, large-scaled analysis showed that the rate of ≥grade 3 radiation esophagitis was only 4% in NSCLC patients after postoperative chemotherapy and radiotherapy [
19]. Although there were nearly 1/3 of patients suffering grade 2 radiation esophagitis in our study, no severe radiation esophagitis was noted, which may be related to the precise techniques and careful protection of normal tissue in radiotherapy after pneumonectomy. Therefore, with careful assessment of pulmonary function, reasonable prescription dose, contralateral lung dose limitation and beam arrangement, radiotherapy after pneumonectomy was safe and feasible for stage pIII-N2 patients.
In recent years, a number of retrospective studies have shown survival benefits brought by PORT in stage pIII-N2 NSCLC patients after surgery. However, most patients in these studies received lobectomy. As reported, only a few patients received pneumonectomy with the rate of 16% in Corso’s study [
1], 8.3% in Robinson’s study [
4], 10.8% in Hui’s study [
12], 27.4% in Shen’s study [
20], and highest in the ANITA study, which was only 36.9% [
21]. Therefore, the benefit of PORT to survival in patients receiving pneumonectomy is not clear and still needs to be evaluated. Our study showed that PORT significantly improved OS, DFS, LRFS and DMFS in stage pIII-N2 NSCLC patients receiving pneumonectomy and adjuvant chemotherapy, although there were more patients with R1/R2 resection in PORT group. In subgroup analysis of patients with R0 resection, PORT was significantly related to better OS and LRFS, and DFS and DMFS in the PORT group trended to be improved. Even in patients with R1/R2 resection, there was a positive correlation between OS and PORT.
Local-regional recurrence is common in stage pIII-N2 NSCLC patients receiving surgery alone [
22]. PORT could increase local-regional control. However, the important premise of transforming improved local-regional control into better OS was the effective prevention of distant metastasis by adjuvant chemotherapy. Thus, we enrolled patients with adjuvant chemotherapy. A meta-analysis showed that some prospective studies failed to show efficacy of PORT in improving overall survival without adjuvant chemotherapy [
23]. On the other hand, a large sample retrospective analysis from national cancer database in the United States showed that the implementation of PORT on the basis of adjuvant chemotherapy, could prolong patients’ median OS by 5 months, and could increase the absolute 5-year survival rate by 5% (39.3% vs 34.8%,
p = 0.014, 4]. Additionally, less side effects of chemotherapy and radiotherapy in patients with both R0 resection and non-R0 resection had positive impact on survival. Previous randomized trials showed that ≥grade 3 adverse events occurred in 91% of patients with postoperative concurrent chemoradiotherapy [
24], which was significantly higher than that of patients treated with chemotherapy and radiotherapy. A recent retrospective study revealed that the OS in patients, with R0, R1 or R2 resection, receiving chemotherapy and radiotherapy was higher than those receiving concurrent chemoradiotherapy, although the survival benefits were not statistically significant in patients with R1/R2 resection [
19]. The researchers believed that toxicities of concurrent chemoradiotherapy counteract the benefits. Therefore, pneumonectomy followed by chemotherapy and radiotherapy was a suitable mode of treatment for stage pIII-N2 NSCLC patients in order to ensure safety of surgery and adjuvant treatment.
As for failure mode, this study showed that distant metastasis was still the main reason (48.7%) of failure after pneumonectomy, followed by local-regional recurrence (32.8%). The result was consistent with the failure patterns in previously published studies, which showed that distant metastasis being the main cause of failure in stage pIII-N2 NSCLC after surgery (about 37–51%), and local-regional recurrence rate accounting for 10–35% [
21,
25‐
27]. In several case-controlled studies enrolling stage pIII-N2 patients, PORT significantly reduced local and regional recurrence rates, compared with non-radiotherapy groups. In the ANITA study [
21], local-regional recurrence occurred in 25.7% of patients, which was significantly reduced to 14.6% after combined radiotherapy. In the previous study in our center [
27], 30.2% of locoregional recurrence occurred in the PORT group while 39.2% in patients without PORT (
p = 0.025). In this study, for patients receiving pneumonectomy, PORT also achieved a significant reduction in local-regional failure compared with patients received adjuvant chemotherapy alone (15.6% vs. 39.1%,
p = 0.016). Moreover, PORT could effectively reduce local-regional recurrence in patients after R0 resection or non-R0 resection, although the difference was not statistically significant (57.1%vs. 37.5%,
p = 0.405) in those with non-R0 resection as small sample size (15 cases). In addition, local-regional recurrence in PORT group mostly occurred out of the radiation field (in-filed 1/5 vs. out-of filed 4/5), which also suggested the effective local-regional control of PORT. In summary, distant control still needed to be highlighted for stage pIII-N2 patients as NSCLC is a highly malignant tumor. On condition of effective prevention of chemotherapy in distant metastasis, radiotherapy could reduce local-regional recurrence and further improve survival.
Lobectomy combining with adjuvant chemotherapy and radiotherapy had showed median OS of 20–48 months, and 5-year OS of 25–61.3% for stage pIII-N2 NSCLC [
3,
21,
26‐
29]. In the INT0139 study, stage pIII-N2 patients receiving chemotherapy and radiotherapy after pneumonectomy approached the median OS of 18.9 months, and the 5- year OS of 21.9% [
25], while 5- year OS was usually lower than 20% in other studies [
11,
27,
30,
31]. In this study, the median OS was 46 months and 5- year OS was 46.3%, which was similar to the prognosis of patients after lobectomy and significantly better than patients after pneumonectomy. The reasons for better prognosis in this study were as follows. Firstly, patients who died during peri-operative period were excluded. In previously reported studies, the mortality rate was high after pneumonectomy, which was about 6–20% within 30 days, and up to 25% within 6 months in some studies [
5,
9]. In the INT0139 study, 14/54 died of postoperative complications after pneumonectomy [
25]. Secondly, only patients receiving adjuvant chemotherapy after pneumonectomy were enrolled in this study. Thirdly, in this study, KPS scores in all patients were more than 80, and the preoperative pulmonary function approached a median FEV1 of 2.36 L. Assessment of the cardiopulmonary function was necessary before pneumonectomy. Some studies have demonstrated that patients with at least 2 L of preoperative FEV1 could tolerate pneumonectomy [
7].
There were some limitations in this study. Firstly, as a retrospective, single-center based study, there might be selective bias in enrollment. Secondly, radiation induced heart injury was evaluated by heart related symptoms instead of routine ECG and echocardiography, which might lead to underestimating side effects of the heart. Thirdly, in this article, only patients who could tolerate pneumonectomy and some cycles of chemotherapy were included, which might underestimate the potential harm of pneumonectomy. However, we aimed to focus to this group of people with high KPS score and survived chemotherapy. Finally, although the clinical characteristics between PORT group and Control group were comparable except that non-R0 resection rate was higher in the former, the number of patients in the PORT group was much lower than the control group (32 patients vs. 87 patients), which might greatly weaken our findings and conclusions. Despite these limitations, this is the largest population based study exploring safety and efficacy of PORT in stage pIII-N2 NSCLC patients after pneumonectomy and adjuvant chemotherapy, which can provide high reference and guidance for PORT of stage pIII-N2 NSCLC after pneumonectomy.