Introduction

Lung cancer is the most common cause of cancer-related death worldwide, with an estimated 1.8 million deaths each year [1]. Non-small cell lung cancer (NSCLC) is the major pathological type of lung cancer, consisting of approximately 85% of all lung cancer cases [2]. Among the various stages, Stage III NSCLC with N2 mediastinal nodal involvement, represents a heterogeneous patient population who require variable and individualized therapeutic approaches, and surgical resection is the primary treatment modality for these patients [3]. But there is still a notable risk of local–regional recurrence and distant metastasis in these patients who receive surgical resection alone [4].

For many years, postoperative radiotherapy (PORT) in completely resected pN2 NSCLC has been a subject of debate. Recently published PORT-C and Lung ART trials both concluded that PORT cannot be routinely recommended for pN2 NSCLC patients, posing a growing challenge to the acceptance of PORT in this context [5, 6]. Despite this, PORT has shown promising results in improving local control and overall survival (OS) rates in patients with high N2 mediastinal nodal burden disease [7,8,9,10].

The tumor-infiltrating lymphocytes (TILs), which is the major component of the tumor microenvironment (TME), was found to be significantly associated with enhanced antitumor activity and tumor destruction [11]. Several studies have demonstrated that the presence and density of TILs within the tumor tissue can serve as predictive and prognostic biomarkers of the treatment response and overall prognosis in NSCLC patients [12, 13]. Radiotherapy, as an important means for local treatment, can convert an immunologically silenced “cold” tumor to an active “hot” tumor by triggering the release of pro-inflammatory mediators and increasing tumor-infiltrating immune cells [14]. However, an important question remains unanswered: could TILs be used as a predictive biomarker to predict the efficacy of radiotherapy?

Hence, the present study aims to investigate the utility of TILs evaluated through H&E sections as biomarkers for predicting PORT efficacy in patients with completely resected stage III pN2 NSCLC and futher identify the benefit population for PORT from the perspective of immune infiltration.

Materials and methods

Patient cohort

This retrospective study included patients with stage III pN2 NSCLC who was histologically documented treated by surgery combined with chemotherapy ± radiotherapy in Shandong Cancer Hospital and Institute between 2014 and 2020. The inclusion criteria were (1) pathologic evidence of NSCLC and R0 resection was achieved; (2) pathologically documented N2 mediastinal nodal involvement at the time of surgery; (3) receiving 2–4 cycles of platinum-based doublet adjuvant chemotherapy following surgery, neoadjuvant chemotherapy was also allowed; (4) three-dimensional conformal radiation therapy or intensity-modulated radiation therapy was administered at a dose of 50 ~ 60 Gy; (5) available clinicopathologic data and H&E sections. The exclusion criteria were (1) pathological T4 and M1 status; (2) locoregional recurrence or distant metastasis during postoperative chemotherapy; (3) previous chest radiotherapy; (4) synchronous multiple primary cancers. Ultimately, a total of 244 patients were included in this study. The patient recruitment and selection process was shown in Fig. 1. The study was approved by the ethical review committee of our hospital (ethics approval number: SDTHEC2020004042).

Fig. 1
figure 1

Recruitment and selection process of patients

Full size image

Treatment response assessment

The follow-up of all patients was conducted every 3 months during the first 2 years after surgery. After that, the patients were followed up every 6 ~ 12 months. Treatment responses were assessed by CT imaging at each follow-up and compared to the images at baseline or from the last follow-up and were evaluated according to RECIST 1.1. Disease-free survival (DFS) was defined as the time from the surgery date to the date of first locoregional recurrence, distant metastasis or death of any cause. OS was defined as the time from surgery to the death of the patient or the last follow-up. All time-to-event data were censored at last follow-up if the corresponding event had not occurred.

Histological evaluation

Permanent full-face H&E sections from surgical specimens from each case were retrieved from the pathology archives and evaluated for TILs. TILs were assessed by two experienced pathologists who were blinded to each other, and the evaluation of TILs was primarily based on the guidelines from International Immuno-Oncology Biomarkers Working Group [15]. We assessed TILs level within the areas including invasive margin and the stromal compartment within the tumor border. Only mononuclear cells (lymphocytes and plasma cells) were included in the assessment. Areas with necrosis, crush artifacts, and fibrosis were excluded for TILs evaluation. The level of TILs was scored as a continuous variable according to the percentage of stroma area occupied by stromal TILs and was evaluated as a percentage in 5% increments. The percentage assessments for multiple slides from each case were averaged. And then patients were subdivided into TILlow group (≤50%) and TILhigh group (>50%) according to arbitrary TILs level of 50%.

Statistical analysis

Patients’ clinicopathological characteristics were described for the whole cohort and compared using the Chi-squared or Fisher exact test, as appropriate. The Mann–Whitney test was used for the comparison of TILs level between patients with and without PORT. OS and DFS were described by the Kaplan–Meier (KM) curves, and the log-rank test was used to compare whether there was a survival difference between two groups. Median follow-up was estimated with the reverse KM method [16]. Cox proportional hazard model was used for univariate and multivariate analyses. All statistical analyses were conducted using SPSS version 25.0 software (IBM, USA). The optimal cutoff value of TILs was determined by the extension module Survival Module of ClinicoPath in Jamovi software (version 2.3.21). All graphs were made using GraphPad software (version 9.5.1). The factors with p < 0.10 on univariate analysis were included in the multivariate analysis, and p < 0.05 was considered to be statistically significant. All tests were two-tailed.

Results

Baseline characteristics

We retrospectively performed analyses on 244 stage III pN2 patients who met the inclusion criteria in the present study, and patients’ clinicopathological characteristics were shown in Table 1. The distribution of TILs was illustrated in Fig. 2. TILs level in patients with PORT was significantly higher than that in patients without PORT (Fig. 2d). For further analyses the cohort was divided into two groups based on TILs density, of whom 190 patients (77.9%) were assigned to TILlow group and 54 patients (22.1%) were TILhigh group. Representative images of two groups were provided in Fig. 3.

Table 1 Patients clinicopathological characteristics of the whole cohort
Full size table
Fig. 2
figure 2

Distribution of tumor-infiltrating lymphocytes (TILs) within the (a) whole population, (b) non-PORT population and (c) PORT population and (d) the differences of TILs between patients with and without PORT . The dashed black line indicates cut-off value

Full size image
Fig. 3
figure 3

Example of stromal tumor-infiltrating lymphocytes (TILs) assessment with hematoxylin and eosin (H&E) sections. (a) Low and (b) high level of TILs in NSCLC tissue

Full size image

Associations between TILs and prognosis

At final data cut-off, the median follow-up time was 44.6 months. During the follow-up period, there were 159 DFS events (65.2%) and 65 OS events (26.6%) in the entire cohort. Median DFS and OS were 26.5 and 95.5 months, respectively.

We found that TILhigh group was associated with an improved prognosis for the entire chort. The TILhigh group had longer OS than the TILlow group [hazard ratio (HR) 0.334, 95% CI 0.199−0.561, p = 0.001] (Fig. 4a). The median OS was 64.7 months in the TILlow group but not reached in the TILhigh group. Similar results were observed for DFS (HR 0.304, 95% CI 0.217−0.423, p < 0.001) (Fig. 4b). The median DFS time in the TILlow group was 20.5 months, while the median DFS time in the TILhigh group was also not reached. In multivariable analysis, high TILs level remained a good predictive factor for both OS (HR 0.589, 95% CI 0.356−0.975, p = 0.007) (Table 2) and DFS (HR 0.290, 95% CI 0.174−0.483, p < 0.001) (Table 3) after correcting for factors which were significant in univariate analysis. Notworthy, the significant positive correlation between high TILs infiltration and improved survival outcomes was observed in both PORT chort (Fig. 4c and d) and non-PORT chort (Fig. 4e and f).

Fig. 4
figure 4

Kaplan-Meier estimates of DFS and OS for patients with high and low tumor-infiltrating lymphocytes (TILs) level within the whole population, PORT population and non-PORT population

Full size image
Table 2 Univariate and multivariate analyses of the OS in all patients
Full size table
Table 3 Univariate and multivariate analyses of the DFS in all patients
Full size table

Association between TILs and PORT Benefit

Cox regression analysis showed that PORT was not an independent prognostic factor in patients with pN2 NSCLC (Tables 2 and 3). We further explored whether TIL infiltrations could predict PORT efficacy and better identify the population that will benefit from PORT. In the low TILs density group, patients treated with PORT had similar DFS (HR 0.821, 95% CI 0.587–1.148, p = 0.244) and OS (HR 0.798, 95% CI 0.470–1.354, p = 0.404) compared to patients without PORT (Fig. 5a and b). Similar results were observed in the high TILs density group, there was no difference in DFS (HR 2.000, 95% CI 0.830–4.817, p = 0.167) and OS (HR 1.038, 95% CI 0.262–4.117, p = 0.958) related to treatment (Fig. 5c and d).

Fig. 5
figure 5

Kaplan–Meier curves comparing PORT versus non-PORT for two groups of patients with high and low tumor-infiltrating lymphocytes (TILs) level

Full size image

Discussion

In this study, we aimed to determine whether the TILs evaluated with H&E sections could predict the response to PORT in patients with completely resected pN2 NSCLC. Our results demonstrated that baseline TILs, an indicator of immunogenicity, could serve as a potent prognostic biomarker in pN2 patients, regardless of the treatment modality applied. The present study is the first study to investigate whether TILs can predict the efficacy of PORT. However, similar prognosis was achieved for patients treated with and without PORT in different TILs infiltration subgroups, and our study did not confirm the predictive value of TILs for the efficacy of PORT.

Immune cells play a critical role in the development and progression of cancer. Considering the important role of host immune infiltration in controlling tumor progression, immune infiltration could be served as a new prognostic factor in various tumor types [17]. It has been shown that higher level of TILs in the primary tumor site is associated with good prognosis with completely resected stage III-N2 NSCLC patients [13, 18]. Our results reached a consistent conclusion and coincided with those of previous reports. However, this study differed from previous studies in the following two ways. First, the sample sizes of patients treated with and without PORT enrolled in this study were similar. Second, our study further validated our results in different treatment subgroups.

The use of PORT in routine treatment settings for pN2 NSCLC had been a controversial issue since a meta-analysis in 1998 [19]. Recently, two large randomized clinical trials have both reported that PORT cannot be recommended in all stage III-N2 patients, but should identify the appropriate patients who will optimally benefit from PORT [5, 6]. Previous studies have shown that select patients could benefit from PORT from the perspective of N2 mediastinal nodal burden, such as a high lymph node ratio and multiple N2 stations [7,8,9]. Departing from previous research perspectives, we expect to identify the benefit population for PORT by utilizing TILs infiltration status, as a study has reported that TILs can be used as predictive markers to predict the efficacy of radiotherapy in breast cancer [20].

However, the present study failed to find the predictive value for the efficacy of PORT in completely resected pN2 NSCLC, as DFS and OS were not statistically different for patients receiving PORT and adjuvant chemotherapy in the subgroup analysis based on the TILs infiltration level. The results may be explained by the high cut-off value. The cut-off value of 50% in the present study was determined in the prognostic analysis, which is consistent with previous studies [21]. However, this cutoff value may not reflect the predictive value of the efficacy of radiotherapy. It was reported that stratification of TILs with a cutoff value of 10% could predict the risk for ipsilateral tumor recurrence in breat cancer research [20]. High TILs infiltration was a protective factor, reducing the risk of regional recurrence and distant recurrence. Thus, high cutoff value might underestimate the benefits of PORT and a lower cutoff value should be futher explored. Although the results were negative, our study fills a research gap in this field.

Our study still has some limitations. First, it is a retrospective study with a relatively small sample size, and the results might be flawed because of residual confounding factors. Second, it is based on single-institution data, and exploration in larger datasets is required to understand any differential efficacy of PORT based on TILs infiltration in patients with completely resected pN2 disease.

Conclusions

Histologic assessment of TILs with H&E section had a strong prognostic value in patients with completely resected pN2 NSCLC. However, TILs had limited clinical utility in the setting of identifing the benefit population for PORT and futher studies shuold be should be conducted to investigate the predictive value of TILs for radiotherapy.