Prognostic impact of primary cancer adjoining emphysematous bullae in non-small cell lung cancer patients treated with immune checkpoint inhibitors

Abstract

Background

Immune checkpoint inhibitors (ICIs) have become a standard therapy in non-small cell lung cancer (NSCLC). Although lung cancer adjoining emphysematous bullae (Ca-ADJ) were reported to express higher programmed cell death-ligand 1 (PD-L1), the predictive impact of Ca-ADJ on the response to ICIs is unknown.

Methods

Two hundred and fifty-seven advanced or recurrent NSCLC patients treated with ICI monotherapy at Kyushu University Hospital and National Hospital Organization Kyushu Cancer Center were analyzed. To minimize the bias arising from the patients’ background, adjusted Kaplan–Meier survival curves and Cox proportional hazards regression analyses using inverse probability of treatment weights (IPTW) were performed.

Results

Of the 257 patients, 55 had Ca-ADJ. Patients with Ca-ADJ were significantly associated with younger age (P = 0.0343), male sex (P = 0.0070), and smoking (P = 0.0080). The objective response rate of cases with Ca-ADJ was significantly higher than that of those without Ca-ADJ (36.4% vs. 20.8%, respectively; P = 0.0167). The disease control rate of cases with Ca-ADJ was also significantly higher than tumors without Ca-ADJ (63.6% vs. 47.5%, respectively; P = 0.0341). The IPTW-adjusted Kaplan–Meier curves showed that patients with Ca-ADJ had significantly longer progression-free survival (PFS) and overall survival (OS) than those without Ca-ADJ (P = 0.0407 and P = 0.0126, respectively). On IPTW-adjusted Cox analysis, Ca-ADJ was an independent predictor of PFS and OS (P < 0.0001 and P < 0.0001, respectively).

Conclusions

Patients with Ca-ADJ may be good candidates for ICIs. These findings should be validated prospectively.

Electronic supplementary material

The online version of this article (10.1007/s00262-020-02783-6) contains supplementary material, which is available to authorized users.

Introduction

Lung cancer is one of the most fatal malignancies worldwide. Non-small cell lung cancer (NSCLC) accounts for 85% of lung cancer [1, 2]. Recently, immune checkpoint inhibitors (ICIs) targeting programmed cell death-1 or programmed death-ligand 1 (PD-L1) have received attention as a novel pharmacotherapy for NSCLC patients [3–5]. Numerous studies investigating predictors for the response to ICIs have shown that PD-L1 expression on tumor cells and blood-based tumor mutation burden (TMB) are promising predictors of clinical benefit in NSCLC patients treated with anti-PD-L1 therapy [6–9]. However, these predictors are not definitive, and additional biomarkers for predicting the response to ICIs are currently being investigated.

The relationship between lung cancer and the presence of emphysematous bullae (EB) in smokers has been described in previous studies [10–14]. Iwama et al. [13] reported that most of the lung cancers in smokers with EB presented as cancer adjoining EB (Ca-ADJ). Our previous studies have shown that in lung adenocarcinoma patients, Ca-ADJ expressed higher levels of PD-L1 than those without EB [15, 16]. However, whether the primary Ca-ADJ influences the efficacy of ICIs in NSCLC patients is unknown. Therefore, in the current retrospective study, we investigated the clinical impact of Ca-ADJ on the efficacy of ICIs in patients with NSCLC.

Materials and methods

Patients

We retrospectively identified 257 patients with advanced or recurrent NSCLC who were treated with ICI monotherapy (nivolumab, pembrolizumab or atezolizumab) between January 2016 and August 2018 at Kyushu University Hospital (n = 113) and National Hospital Organization Kyushu Cancer Center (n = 144). The patients received nivolumab, pembrolizumab, and atezolizumab intravenously at a dose of 3 mg/kg every 2 weeks, at a fixed dose of 200 mg every 3 weeks, and at a fixed dose of 1200 mg every 3 weeks. Patients’ clinicopathological features, including age at the time of treatment initiation, sex, Eastern Cooperative Oncology Group performance status (ECOG PS), line of treatment, smoking history, history of radiation therapy, clinical stage (8th edition of the TNM Classification of Malignant Tumors) [17], driver oncogene mutation status (epidermal growth factor receptor [EGFR]), histology, and Ca-ADJ were examined. Bullae were defined as an airspace with a diameter greater than 1 cm that was sharply demarcated by a smooth wall with a thickness of 1 mm or less [18]. Ca-ADJ was defined as tumor arising from the EB. A representative image of Ca-ADJ is shown in Fig. 1. In recurrent cases, the primary NSCLCs, not the recurrent lesions, were investigated. Tumor response was assessed by computed tomography every 6–8 weeks according to the Response Evaluation Criteria in Solid Tumors, version 1.1 [19]. Disease-control rate was defined as summing partial response rate and stable disease rate. Clinical information and follow-up data were obtained from patients’ medical records. The end of the follow-up period was September 30, 2019. This study was approved by our institutional review boards (Kyushu University, IRB no. 2019-195 and National Hospital Organization Kyushu Cancer Center, IRB no. 2019-45).

Fig. 1.

Representative image of lung cancer adjoining emphysematous bullae

PD-L1 tumor expression and EGFR mutation analysis

Immunohistochemical staining for PD-L1 was performed with the pharmDx antibody (clone 22C3, Dako North America, Inc., Agilent/Dako, Carpinteria, CA, USA) in accordance with the manufacturer’s recommended methods [20]. All PD-L1 data were reviewed by two experienced observers who were blinded to the patients’ clinical status. EGFR status was determined in tumor samples using the peptide nucleic acid-locked nucleic acid polymerase chain reaction clamp method (Mitsubishi Chemical Medience, Tokyo, Japan) [21]. All data regarding PD-L1 expression and EGFR mutation status were extracted from medical records. The final assessments were achieved by consensus, and membrane PD-L1 tumor expression was grouped by tumor proportion score levels.

Statistical analysis

With the aim of minimizing the bias arising from the patients’ background, an inverse probability of treatment weights (IPTW) model was conducted [22]. Possible variables associated with Ca-ADJ, including age, sex, and smoking history, were selected. Associations between Ca-ADJ and clinical factors before and after weighting were analyzed using Student’s t test, Pearson’s χ² test, and Fisher’s exact test where appropriate. Progression-free survival (PFS) was defined as the period from initial treatment to clinical or radiographic progression or death, and overall survival (OS) was defined as the period from initial treatment to the date of last follow-up or death. Adjusted Kaplan–Meier survival curves using IPTW and weighted log-rank test were generated [22]. Cox proportional hazards regression analysis was used to estimate hazard ratios for risk factors with the backward elimination method. P < 0.05 was considered significant. All statistical analyses were performed using JMP® 14.0 and SAS 9.4 (SAS Institute Inc., Cary, NC, USA).

Results

Patient characteristics

Patient characteristics are shown in Table 1. Of the 257 patients, 55 (21.4%) had Ca-ADJ. Driver oncogene mutation (EGFR) status data were available for 231 (89.9%), and PD-L1 tumor expression data were available for 174 (67.7%). The characteristics of patients with Ca-ADJ treated with ICIs before and after weighting are shown in Table 2. Before weighting, patients with Ca-ADJ were significantly associated with younger age (P = 0.0343), male sex (P = 0.0070), and smoking (P = 0.0080). In the current overall cohort, 42 of 55 patients with bullae formation (76.4%) had emphysema and 22 of 202 patients without bullae formation (10.9%) had emphysema.

Table 1.

Clinicopathological characteristics of all patients (N = 257)

Factors	Value or no. of patients
Age (years)
Median	66
Range	31–88
Sex
Female	52 (20.2%)
Male	205 (79.8%)
ECOG PS
0	104 (40.4%)
1	130 (50.6%)
2	20 (7.8%)
3	3 (1.2%)
Line of treatment
First	47 (18.3%)
Second	79 (30.7%)
Third or higher	131 (51.0%)
Smoking history
Never-smoker	45 (17.5%)
Ex-smoker	119 (46.3%)
Current smoker	93 (36.2%)
History of radiation therapy
No	131 (60.4%)
Yes	86 (39.6%)
Clinical stage
IIIB	37 (14.4%)
IV	178 (69.3%)
Recurrent	42 (16.3%)
Mutation status (EGFR)
Wild-type	204 (79.4%)
Mutant-type	27 (10.5%)
Unknown	26 (10.1%)
Histology
Adenocarcinoma	167 (65.0%)
Squamous cell carcinoma	68 (26.5%)
Others or unknowna	22 (8.5%)
Immune checkpoint inhibitors
Nivolumab	133 (51.8%)
Pembrolizumab	95 (37.0%)
Atezolizumab	29 (11.2%)
PD-L1 (22C3) TPS
< 1%	42 (16.3%)
≥ 1% and < 50%	56 (21.8%)
≥ 50%	76 (29.6%)
Unknown	83 (32.3%)
Primary cancer adjoining emphysematous bullae
Yes	55 (21.4%)
No	202 (78.6%)

ECOG Eastern Cooperative Oncology Group, EGFR, epidermal growth factor receptor, PD-L1 programmed cell death-ligand 1, PS performance status, TPS tumor proportion score

^aSix patients with sarcomatoid carcinoma and 12 patients with NOS (not-otherwise specified)

Table 2.

Characteristics of the patients with non-small cell lung cancer adjoining emphysematous bullae (Ca-ADJ) treated with immune checkpoint inhibitors, before and after weighting

Factors	Ca-ADJ	Unweighted, N (%)			Weighted, N%
		Yes (n = 55)	No (n = 202)	P value	Yes	No	P value
Age (years)	Mean	61.8	65.0	0.0343a	64.4	65.0	0.6310a
Sex	Male	51 (92.7)	154 (76.2)	0.0070b	155.7	154.0	0.4143 b
	Female	4 (7.3)	48 (23.8)		39.8	48.0
ECOG PS	0	53 (96.4)	181 (89.6)	0.1807c	186.1	181.0	0.0377b
	≥ 2	2 (3.6)	21 (10.4)		9.5	21.0
Line of treatment	First or second	33 (38.8)	59 (44.7)	0.3928b	103.0	97.0	0.3531b
	Third or higher	52 (61.2)	73 (55.3)		92.5	105.0
Smoking history	Neversmoker	3 (5.5)	42 (20.8)	0.0080b	30.2	42.0	0.1657b
	Smoker	52 (94.5)	160 (79.2)		165.4	160.0
Clinical stage	IIIB/IV	46 (83.6)	167 (82.7)	0.8665b	158.0	167.0	0.6281b
	Recurrent	9 (16.4)	35 (17.3)		37.6	35.0
Mutation status (EGFR)	Wild type	47 (94.7)	157 (86.7)	0.1573b	155.9	157.0	0.8499b
	Mutant	3 (6.0)	24 (13.3)		22.5	24.0
Histology	Sq	10 (18.2)	58 (28.7)	0.1165b	162.1	144.0	0.0060b
	Non-Sq	45 (81.8)	144 (71.3)		33.5	58.0
PD-L1	≥ 50%	14 (40.0)	62 (44.6)	0.6235b	72.8	77.0	0.6650b
	< 50%	21 (60.0)	77 (55.4)		52.6	62.0

Ca-ADJ cancer adjoining emphysematous bullae, ECOG Eastern Cooperative Oncology Group, EGFR, epidermal growth factor receptor, PD-L1 programmed cell death-ligand 1, PS performance status, Sq squamous cell carcinoma

a Student’s t test

b χ² test

c Fisher’s exact test

Associations between Ca-ADJ and treatment outcomes in NSCLC patients treated with ICIs

There were 62 (24.1%) cases with partial response, 70 (27.2%) cases with stable disease, and 111 (43.2%) cases with progressive disease. Figure 2 shows the objective response rate and disease control rate according to Ca-ADJ. The objective response rate of cases with Ca-ADJ was significantly higher than that of those without Ca-ADJ (36.4% vs. 20.8%, respectively; P = 0.0167). The disease control rate of cases with Ca-ADJ was also significantly higher than tumors without Ca-ADJ (63.6% vs. 48.0%, respectively; P = 0.0399). Figure 3a, b show the Kaplan–Meier curves of NSCLC patients treated with ICIs according to Ca-ADJ. Patients with Ca-ADJ had significantly longer PFS and OS compared with those without Ca-ADJ (log-rank test; P = 0.0156 and P = 0.0194, respectively). The IPTW-adjusted Kaplan–Meier curves are shown in Fig. 3c, d. Patients with Ca-ADJ had significantly longer PFS and OS than those without Ca-ADJ (log-rank test; P = 0.0407 and P = 0.0126, respectively).

Fig. 2.

a The objective response rate and b disease control rate according to cancer adjoining emphysematous bullae in non-small-cell lung cancer patients treated with immune checkpoint inhibitors are shown

Fig. 3.

The Kaplan–Meier curves of a progression-free survival (PFS) and b overall survival (OS) in non-small cell lung cancer (NSCLC) patients treated with immune checkpoint inhibitors (ICIs) according to cancer adjoining emphysematous bullae (Ca-ADJ) are shown. The inverse probability of treatment weights-adjusted Kaplan–Meier curves of c PFS and d OS in NSCLC patients treated with ICIs according to Ca-ADJ are shown

Relationship between PFS and OS with clinical factors in patients with NSCLC treated with ICIs

We next analyzed the associations between survival and clinical factors in NSCLC patients treated with ICIs. The median follow-up time was 413 days (range 3–1316). Supplementary Table 1 shows the results of univariate and multivariable survival analyses. Multivariable analyses revealed that ECOG PS, driver gene mutation, PD-L1 expression, and Ca-ADJ were independent prognostic factors for PFS (P = 0.0188, P = 0.0071, P < 0.0001, and P = 0.0296, respectively) and C-ADJ was an independent prognostic factor for OS (P = 0.0207). As shown in Table 3, after IPTW adjusted Cox analysis, Ca-ADJ was an independent prognostic factor for PFS and OS (P < 0.0001 and P < 0.0001, respectively). The forest plot analyses assessing the PFS and OS of each subgroup are shown in Figs. 4 and 5, respectively.

Table 3.

Inverse probability of treatment weight adjusted univariate and multivariable analyses of PFS and OS in non-small cell lung cancer patients treated with immune checkpoint inhibitors

Factors	PFS				OS
	Univariate analysis		Multivariable analysis		Univariate analysis		Multivariable analysis
	HR (95% CI)	P value	HR (95% CI)	P value	HR (95% CI)	P value	HR (95% CI)	P value
Age (years)
Continuous variable	1.00 (0.99–1.01)	0.9063			1.01 (0.99–1.02)	0.8908
Sex
Female/Male	1.56 (1.20–2.03)	0.0010			1.62 (1.16–2.27)	0.0048	1.91 (1.18–3.09)	0.0085
ECOG PS
2 or 3/0 or 1	2.00 (1.34–2.98)	0.0007	2.45 (1.25–4.79)	0.0090	2.50 (1.66–3.75)	< 0.0001
Line of treatment
Third or higher/First or second	1.63 (1.30–2.05)	< 0.0001			1.31 (1.01–1.71)	0.0452
Smoking history
Never smoker/Smoker	1.32 (0.99–1.75)	0.0525			1.63 (1.14–2.34)	0.0079
Clinical stage
Recurrent/IIIB or IV	1.26 (0.95–1.67)	0.1089			1.70 (1.24–2.33)	0.0010
Mutation status (EGFR)
Mutant/Wild-type	2.24 (1.62–3.11)	< 0.0001	2.35 (1.58–3.51)	< 0.0001	1.53 (0.99–2.37)	0.0552
Histology
Sq/Non-Sq	1.20 (0.92–1.57)	0.1743			1.11 (0.81–1.52)	0.5129
PD-L1
< 50%/ ≥ 50%	2.19 (1.61–3.00)	< 0.0001	2.57 (1.81–3.66)	< 0.0001	1.32 (0.90–1.92)	0.1527
Ca-ADJ
No/Yes	1.49 (1.19–1.87)	0.0006	1.98 (1.43–2.72)	< 0.0001	1.90 (1.46–2.48)	< 0.0001	3.71 (2.42–5.68)	< 0.0001

Ca-ADJ cancer adjoining emphysematous bullae, CI confidence interval, ECOG Eastern Cooperative Oncology Group, EGFR, epidermal growth factor receptor, HR hazard ratio, OS overall survival, PD-L1 programmed cell death-ligand 1, PFS progression-free survival, PS performance status, Sq squamous cell carcinoma

Fig. 4.

Summary of progression-free survival hazard ratios according to subgroups

Fig. 5.

Summary of overall survival hazard ratios according to subgroups

Discussion

In this retrospective study, we demonstrated that patients with Ca-ADJ had significantly better treatment outcomes and longer PFS and OS compared with those without Ca-ADJ. With regard to independent predictive factors of PFS, ECOG PS, EGFR mutation status, PD-L1, and Ca-ADJ were selected. ECOG PS and EGFR mutation status were reported to be significant predictors of the efficacy to ICIs in NSCLC patients [23–25]. Given that our results were in line with these previous reports, our findings may be applicable to a general NSCLC population treated with ICIs. In the current study, IPTW-adjusted analyses were performed to reduce the retrospective bias. Age, sex, and smoking history were selected as possible variables and the results remained significant. The forest plot analyses of each subgroup indicated that patients with Ca-ADJ generally had favorable PFS independent of each subgroup. Thus, the effect of Ca-ADJ was suggested to be independent of other clinical factors.

Patients with Ca-ADJ were significantly associated with younger age, male sex, and smoking. These results were in line with several previous reports [13–16]. EBs are reported to be closely related to cigarette smoking and have the potential to develop lung cancer [26, 27]. The accumulation of multiple genetic alterations due to smoking is believed to contribute to lung carcinogenesis [28], and such tumors would be associated with a higher TMB. Given that TMB is a promising predictive biomarker of the response to ICIs [8], the accumulation of genetic alterations in Ca-ADJ may explain the better response to ICIs in patients with Ca-ADJ compared with those without Ca-ADJ. Although our study only showed a slight tendency for a low frequency of EGFR mutation in patients with Ca-ADJ, several previous studies showed the frequency of EGFR-mutant NSCLC to be significantly lower in the patients with EBs than in those without EBs [13, 15, 16]. These results also support the notion that Ca-ADJ is associated with smoking-induced DNA damage. Regarding PD-L1 expression of Ca-ADJ, Toyokawa et al. [15] reported that EB-associated lung adenocarcinomas expressed PD-L1 protein more frequently than those without EB. In our study, no difference in PD-L1 expression between Ca-ADJ and the other tumors was observed. This may be because we investigated a smaller sample size and/or because biopsy specimens were used to examine PD-L1 positivity (tumor heterogeneity) [29].

In a previous report investigating 109 advanced lung cancer patients who were smokers and treated with chemotherapy, OS was slightly worse in patients with EBs than in those without EBs [13]. In contrast, our study showed that patients with Ca-ADJ had a significantly longer OS compared with those without Ca-ADJ. Given the discrepancy, our findings should be validated prospectively in advanced NSCLC patients treated with a combination therapy of chemotherapy and ICI, a strategy that has been approved by the Food and Drug Administration [30, 31]. Notably, however, Ca-ADJ was the only independent prognostic factor of OS in our study, suggesting that patients with Ca-ADJ may be good candidates for ICIs rather than chemotherapy.

There are some limitations to our study. First, this was a retrospective study investigating a relatively small sample size. The results observed in the current study should be validated prospectively, and the biological effect of Ca-ADJ on the response to ICIs should be elucidated in further studies. Second, this cohort included recurrent cases as well as advanced NSCLC patients. Our cohort also includes several histological types. This heterogeneity of patients makes it difficult to draw definitive conclusions.

In conclusion, the present study showed the favorable clinical impact of Ca-ADJ in NSCLC patients treated with ICIs. NSCLC patients with Ca-ADJ would be good candidates for ICIs. These findings need to be validated prospectively with a larger sample size.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

Ca-ADJ

Cancer adjoining emphysematous bullae

EB

Emphysematous bullae

ECOG PS

Eastern Cooperative Oncology Group performance status

EGFR

Epidermal growth factor receptor

ICI

Immune checkpoint inhibitor

IPTW

Inverse probability of treatment weights

NSCLC

Non-small cell lung cancer

OS

Overall survival

PD-L1

Programmed death-ligand 1

PFS

Progression-free survival

TMB

Tumor mutation burden

Funding

There are no sources of funding to report.

Availability of data and materials

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Compliance with ethical standards

Conflict of interest

The authors declare no conflicts of interest in association with this study.

Ethical approval

This study was approved by our institutional review boards (Kyushu University, IRB no. 2019–195 and National Hospital Organization Kyushu Cancer Center, IRB no. 2019-45).