Significance of programmed death‐ligand 1 expression in resected lung cancer and its relationship with EGFR mutation

Abstract

Background

Several studies have reported that the high expression of programmed death‐ligand 1 (PD‐L1) within tumor cells predicts a poor prognosis. However, the relationship between the PD‐L1 expression and lymph node metastasis or driver mutations in lung cancer remains poorly understood.

Methods

A total of 356 consecutive patients who underwent surgical resection for primary lung cancer were included in the study. There were 268 adenocarcinomas including 100 EGFR mutations, 67 squamous cell carcinomas (Sq), and 21 other histologies. The high expression of PD‐L1 was defined as a tumor proportion score (TPS) of ≥50. The relationship between the PD‐L1 expression and clinicopathological factors and recurrence‐free survival (RFS) was analyzed.

Results

The PD‐L1 expression was high in 75 patients. It was significantly related to smoking history, Sq histology, driver mutation negative, elevated serum carcinoembryonic antigen levels, and lymph node metastasis. Among patients with driver mutations, a high PD‐L1 TPS was found in patients with EGFR G719X mutation. A significant difference in RFS was observed in adenocarcinoma patients. A multivariate analysis of adenocarcinoma cases revealed that tumor size and lymph node metastasis were independent prognostic factors for poor RFS, while the PD‐L1 expression was not. A logistic regression analysis revealed that the absence of driver mutations, lymph node metastasis, and a history of smoking were significantly associated with the high expression of PD‐L1.

Conclusion

Lymph node metastasis was positively related with the high expression of PD‐L1, resulting in poor RFS. A high PD‐L1 TPS was observed in patients with the EGFR G719X mutation.

• The high expression of PD‐L1 was significantly related to smoking history, driver mutation negative, and lymph node metastasis.

• A significant difference in RFS was observed in adenocarcinoma patients and which was mainly resulted from the relationship between lymph node metastasis and the high expression of PD‐L1.

• A high PD‐L1 TPS was observed in patients with the EGFR G719X mutation.

INTRODUCTION

Lung cancer is the leading cause of death worldwide, ¹ and surgery remains the primary treatment option, despite advances in radiotherapy, chemotherapy and immunotherapy. Despite improvements in postoperative outcomes over the years, approximately half of patients experience recurrence, ² highlighting the urgent need for more optimal postoperative adjuvant therapy strategies.

Targeting programmed death‐ligand 1 (PD‐L1) is a crucial aspect of immunotherapy for lung cancer, ³ , ⁴ as most cases are characterized by the accumulation of genetic mutations caused by environmental exposure to carcinogens (e.g., tobacco‐related materials). ⁵ Neoantigens, present in cancer cells with accumulated genetic mutations, are recognized and targeted by cytotoxic T cells. ⁶ PD‐L1 in tumor cells enables the tumor to evade immunosurveillance. Therefore, PD‐L1 has been identified not only as a predictive factor for the effect of anti‐PD‐1 therapy ⁷ but also as a prognostic factor for survival. ⁸

In this study, we investigated the significance of the PD‐L1 expression in tumors, particularly in relation to driver mutations and lymph node metastasis.

METHODS

Patients

From January 2017 to June 2020, 356 consecutive patients with non‐small cell lung cancer (NSCLC) underwent surgery with curative intent. The pathological stage was defined based on the eighth edition of the American Joint Committee on Cancer lung cancer staging system and histological classification ⁹ was carried out based on the 2021 World Health Organization Classification of Tumors of the Lung, Pleura, Thymus and Heart. ¹⁰ Table S1 summarizes the patient background characteristics. A total of 60% of the patients were male, 69% were smokers, and 75% had adenocarcinoma. Driver mutations were found in 38% of patients who were screened. Five patients were finally diagnosed with pathological stage IV disease, all due to pleural dissemination or malignant effusion found after thoracotomy, and they were excluded from the survival analysis. The following clinicopathological features were analyzed: age at surgery, sex, smoking status, serum level of carcinoembryonic antigen (CEA) before surgery, histology, tumor size, and pathological tumor stage. This study was performed in line with the principles of the Declaration of Helsinki. This study was approved by the Institutional Review Board at Oita University Faculty of Medicine (IRB No. 698), and informed consent was obtained from each patient.

Immunogenomic analysis

PD‐L1 immunohistochemistry was performed on a 4‐μm‐thick section of a formalin fixed, paraffin embedded tissue section of the primary tumor using 22C3 antibody (Agilent Technologies) and the tumor proportion score (TPS) was evaluated (SRL Inc.). The PD‐L1 expression was regarded as high when the TPS value was ≥50%. ¹¹

EGFR mutation and ALK rearrangement were examined when the tumor contained an adenocarcinoma component. EGFR mutation was examined using Cycleave‐PCR (SRL Inc.) from genomic DNA extracted from the primary tumor and ALK rearrangement was screened using Optiview ALK (D5F3, Roche Diagnostics) on serial sections from the primary tumor and was then confirmed by fluorescent in situ hybridization (SRL Inc.).

Statistical analysis

The statistical significance of differences between groups was determined using the chi‐squared test. p‐values of <0.05 were considered statistically significant. All patients were retrospectively analyzed for age and sex, smoking history, histology, preoperative serum CEA level, presence of driver mutations, pathological stage, size, lymph node metastasis, and recurrence‐free survival (RFS, between lung cancer resection and recurrence, second primary lung cancer or death caused by lung cancer). The probability of survival was analyzed by the Kaplan–Meier method using the date of lung cancer resection as the starting point. The log‐rank test was used to determine the significance of differences between subgroups. A Cox proportional hazards model was used to determine factors significantly associated with poor RFS. A logistic regression analysis was performed to determine factors associated with the high expression of PD‐L1. The Mann–Whitney U test with Bonferroni correction was performed to compare the TPS of PD‐L1 among factors. All statistical analyses were performed using EZR version 1.55. ¹²

RESULTS

PD‐L1 and clinicopathological factors

PD‐L1 TPS ranged from 0 to 100, with a median value of 5. The PD‐L1 TPS was high (≥50) in 21% of the patients. The PD‐L1 TPS was significantly higher in patients with a smoking history, nonadenocarcinoma histology, patients with a high serum CEA level, and wild‐type or uncertain driver mutations. The pathological stage was also related to the expression of PD‐L1. Patients with stage II–IV, or pN positivity had a higher PD‐L1 TPS, while the tumor size was not related to the PD‐L1 TPS (Table S2).

PD‐L1 and recurrence‐free survival

The relationship between the PD‐L1 expression and survival was analyzed in patients who received curative resection. We used RFS as an endpoint, because the follow‐up period was not long enough to analyze overall survival. Although the survival curve of patients with low expression of PD‐L1 was slightly better than that of their counterpart, the difference was not significant (Figure 1a). On the other hand, when we analyzed patients with adenocarcinoma, the difference in survival became significant (Figure 1b, p = 0.006). When we divided the PD‐L1 group into three categories (TPS 0, 1–49, 50‐), we did not find any difference between TPS 0 and 1–49 (see Figure S1). A multivariate analysis revealed that the tumor size (hazard ratio [HR] 0.29, 95% confidence interval [CI]: 0.14–0.60, p = 0.001) and lymph node (LN) metastasis (HR 4.77, 95% CI: 2.12–10.7, p = 0.0002) were independent risk factors for poor RFS in patients with adenocarcinoma, while the expression of PD‐L1 or other factors was not (Table 1).

FIGURE 1.

The PD‐L1 expression and recurrence‐free survival. Kaplan–Meier curves according to PD‐L1 TPS (≥50: black line or < 50: red line) for (a) all patients and for (b) adenocarcinoma patients. + indicates censored cases. PD‐L1, programmed death‐ligand 1; TPS, tumor proportion score.

TABLE 1.

The multivariate analysis of factors associated with survival in lung adenocarcinoma.

Factors	HR (95% CI)	p‐value
LN metastasis pN1‐2/pN0	4.77 (2.12–10.7)	0.0002
Size (cm) <3.0/≥3.0	0.29 (0.14–0.60)	0.001
PD‐L1 TPS < 50/≥50	0.67 (0.29–1.53)	0.342
Driver mutation pos/neg	0.78 (0.34–1.78)	0.554
CEA (ng/mL) < 4.0/≥4.0	0.75 (0.36–1.56)	0.433
Sex male/female	1.95 (0.79–4.80)	0.148
Age < 75/≥75	0.69 (0.35–1.38)	0.297
Smoking ever/never	0.60 (0.23–1.58)	0.300

Abbreviations: CI, confidence interval; CEA, carcinoembryonic antigen; HR, hazard ratio; LN, lymph node; TPS, tumor proportion score.

PD‐L1, LN metastasis, and driver mutations

Finally, a logistic regression analysis was carried out to identify factors contributing to the expression of PD‐L1 in the tumor (Table 2). As predicted, LN metastasis was most related to the expression of PD‐L1 with an odds ratio of 2.38 (1.19–4.76, p = 0.014). Smoking history was also related to the expression of PD‐L1 (odds ratio [OR] 2.86, 95% CI: 1.18–6.67, p = 0.020), while other factors (including driver mutation, histology, tumor size, serum CEA level, and age) were not. This observation became even more significant when we limited the analysis to patients with adenocarcinoma (Table S3). The PD‐L1 TPS was significantly lower in patients without LN metastasis than in those with pN1 or pN2 (p = 0.002 and p = 0.005, respectively) (Figure 2a). As for driver mutations, the PD‐L1 TPS was in patients with EGFR exon 19 deletion (Del19, median TPS = 1, range 0–70, p = 0.002) and L858R (median TPS = 1, range 0–65, p = 0.01) was significantly lower than that in EGFR wild‐type (median TPS = 5, range 0–100), but it was not lower in patients with ALK rearrangement (median TPS = 62.5, range 0–100) or in patients with uncommon EGFR mutations (median TPS = 15, range 0–95) (Figure 2b). Among uncommon EGFR mutations, six patients harbored G719X, four of whom had a PD‐L1 TPS of ≥50 (Table 3).

TABLE 2.

Factors contributing to the tumor expression of PD‐L1.

Factors	OR (95% CI)	p‐value
LN metastasis pN1‐2/pN0	2.38 (1.19–4.76)	0.014
Smoking ever/never	2.86 (1.18–6.67)	0.020
Size (cm) ≥ 3.0/<3.0	1.24 (0.66–2.33)	0.510
Histology Ad/non‐Ad	0.98 (0.52–1.84)	0.954
Driver mutation neg/pos	1.59 (0.76–3.34)	0.218
CEA (ng/mL) ≥ 4.0/<4.0	1.27 (0.74–2.21)	0.386
Sex (female/male)	1.22 (0.59–2.54)	0.148
Age ≥ 75/<75	1.47 (0.84–2.57)	0.173

Abbreviations: Ad, adenocarcinoma; CEA, carcinoembryonic antigen; CI, confidence interval; LN, lymph node; OR, odds ratio.

FIGURE 2.

Low PD‐L1 TPS in patients with pN0 or with common EGFR mutations (Del19 and L858R). Box and whisker plots according to (a) the pN status and (b) according to driver alterations. PD‐L1, programmed death‐ligand 1; TPS, tumor proportion score; Del19, exon 19 deletion.

TABLE 3.

PD‐L1 TPS in cases with uncommon EGFR mutations.

Case	EGFR mutation type	PD‐L1 TPS	Pack‐year index
69M	G719X	95	30
70M	G719X, S768I	80	48
79M	G719X	50	33
85F	G719X	50	0
69M	Exon 20 insertion	20	80
72M	L861Q	10	48
63M	G719X	1	49.5
70M	L861Q	1	20
80F	L861Q	1	0
74F	G719X, S768I	0	0

Abbreviation: TPS, tumor proportion score.

DISCUSSION

The relationship between the expression of PD‐L1 and LN metastasis has also been described in another study. ¹³ However, the mechanism through which tumor metastasis to the LN regulates the expression of PD‐L1 is unknown. Although the tumor draining LNs are places where tumor infiltrating lymphocytes are activated in response to neoantigens presented to dendritic cells, ¹⁴ this only explains why the expression of PD‐L1 in the tumor is elevated in the presence of neoantigens. However, it does not explain why tumor cells are able to migrate to LNs. Recently, several studies have reported that PD‐L1 expression could promote LN metastasis by reducing tumor immunity in the lymphatic system. ¹⁵ , ¹⁶ How LN metastasis occurs remains unclear. However, it has been well demonstrated that antigen recognition enhances stromal cytokine gradients, resulting in elevated lymph flow to the LNs. ¹⁷ This overflow possibly allows tumor cells to migrate to the draining LNs. Taken together, the expression of PD‐L1 in the tumor is a response to the presentation of neoantigens in the draining LNs, and also enables the tumor to migrate to the draining LNs. Therefore, it is also suggested that tumors in patients with LN metastasis should be sensitive to anti‐PD‐1 therapy.

On the other hand, patients with EGFR mutation or ALK rearrangement are not likely to respond to anti‐PD‐1 therapy. Many clinical trials using immune checkpoint inhibitors excluded patients with EGFR mutation or ALK rearrangement. Even if they included those patients, a subgroup analysis of patients with EGFR mutation or ALK rearrangements exhibited unfavorable results. ¹⁸ Unlike tumors with PD‐L1 expression, tumors with driver mutations are characterized by a lower tumor mutation burden and neoantigens. ¹⁹ Our study also revealed that the PD‐L1 expression level was significantly lower in tumors with EGFR common mutations, as previously reported. ²⁰ Meanwhile, tumors with ALK rearrangement have been reported to show high expression of PD‐L1 through an intrinsic mechanism where the JAK‐STAT3 pathway, which is the downstream of ALK, promotes PD‐L1 transcription. ²¹ In tumors with uncommon EGFR mutations, considerably high expression of PD‐L1 was observed (although it should be noted that the population was small). Most cases harbored EGFR G719X mutations, which are sensitive to EGFR‐tyrosine kinase inhibitors. Because we could not find activation in the JAK‐STAT3 pathway or the expression of PD‐L1 in a patient‐derived cell line with EGFR G719S mutation that we previously established ²² (data not shown), there could be an extrinsic mechanism in the expression of PD‐L1; the secretion of some neoantigens or cytokines took place in cells with EGFR G719X. An examination of the copy number variation revealed that the patient‐derived cell line had IL‐6 amplification as well as EGFR amplification. ²² Different from common EGFR mutations, the incidence of EGFR G719X is not related to a smoking history; on the contrary, it increases with the smoking pack‐years value. ⁵ Further studies are warranted to clarify the relationship between the expression of PD‐L1 and EGFR G719X.

The expression of PD‐L1 has been the most important predictive factor for anti‐PD‐1 therapy across different cancer types. ²³ Furthermore, in advanced NSCLC, anti‐PD‐1 therapy showed a greater effect when it was used in a first‐line setting in comparison to a second‐line or later settings. ²⁴ It has been explained that the therapeutic effects from immune checkpoint inhibitors are stronger in cases with a lower tumor burden and earlier disease stage, supporting the immunological importance of regional LNs in anti‐PD‐1 therapy, not of circulating immune cells to target metastasized tumors or the exhausted tumor microenvironment after cytotoxic therapy. ¹⁴ , ²⁵ Several ongoing studies are testing anti‐PD‐1 antibody therapy in an adjuvant setting with promising results. ²⁶ , ²⁷ , ²⁸ , ²⁹ Further results and subsequent biomarker findings are awaited.

In conclusion, the expression of PD‐L1 was related to LN metastasis, which resulted in poor survival, in patients with resected NSCLC. PD‐L1 expression was significantly lower in tumors with common EGFR mutations, but not in tumors with uncommon EGFR mutations or with ALK rearrangement.

AUTHOR CONTRIBUTIONS

Atsushi Osoegawa: Conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, software, validation, visualization, roles/writing—original draft, writing—review and editing.

Miyuki Abe: Conceptualization, data curation, formal analysis, funding acquisition, methodology, project administration, resources, validation.

Yohei Takumi: Data curation, resources.

Takafumi Hashimoto: Data curation, resources.

Takashi Karashima: Data curation, resources, investigation, project administration.

Michiyo Miyawaki: Data curation, investigation.

Kenji Sugio: Funding acquisition, supervision, validation, writing—review and editing.

All authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

CONFLICT OF INTEREST STATEMENT

Miyuki Abe, Yohei Takumi, Takafumi Hashimoto, Takashi Karashima, and Michiyo Miyawaki declare no financial interests. Atsushi Osoegawa has received speaker and consultant honoraria from AstraZeneca (Cambridge, UK), speaker honoraria from MSD Japan (Tokyo, Japan), Chugai Pharmaceutical (Tokyo, Japan), Bristol Myers Squibb (New York, NY), and Ono Pharmaceutical (London, UK). Kenji Sugio has received speaker and consultant honoraria from AstraZeneca (Cambridge, UK) and Chugai Pharmaceutical (Tokyo, Japan), speaker honoraria from MSD Japan (Tokyo, Japan), Bristol Myers Squibb (New York, NY), and Ono Pharmaceutical (London, UK), and grants from AstraZeneca (Cambridge, UK), Chugai Pharmaceutical (Tokyo, Japan), MSD Japan (Tokyo, Japan), Bristol Myers Squibb (Tokyo, Japan), and Ono Pharmaceutical (Tokyo, Japan).

CONSENT TO PARTICIPATE

Informed consent was obtained from all individual participants included in this study.

Supporting information

Table S1. Patient characteristics.

Table S2. PD‐L1 and clinicopathological factors.

Table S3. Factors contributing to the tumor expression of PD‐L1.

Osoegawa A, Abe M, Takumi Y, Hashimoto T, Karashima T, Miyawaki M, et al. Significance of programmed death‐ligand 1 expression in resected lung cancer and its relationship with EGFR mutation. Thorac Cancer. 2023;14(24):2467–2472. 10.1111/1759-7714.15031

DATA AVAILABILITY STATEMENT

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.