inflammation and pd-l1 expression in pulmonary ... · tumor-associated inflammation and pd-l1...

https://doi.org/10.1530/ERC-17-0427http://erc.endocrinology-journals.org © 2018 Society for Endocrinology

Printed in Great BritainPublished by Bioscientifica Ltd.

25:3Endocrine-Related Cancer

A Kasajima et al. PD-L1 in lung neuroendocrine tumors

339–350

10.1530/ERC-17-0427

RESEARCH

Inflammation and PD-L1 expression in pulmonary neuroendocrine tumors

Atsuko Kasajima1,2, Yuichi Ishikawa3, Ayaka Iwata2, Katja Steiger1, Naomi Oka2,4, Hirotaka Ishida2, Akira Sakurada5, Hiroyoshi Suzuki4, Toru Kameya6, Björn Konukiewitz1, Günter Klöppel1, Yoshinori Okada5, Hironobu Sasano2 and Wilko Weichert1,7

1Department of Pathology, Technical University Munich, Munich, Germany2Department of Pathology, Tohoku University Graduate School of Medicine, Sendai, Japan3Pathology Department, The Cancer Institute Hospital of JFCR, Tokyo, Japan4National Hospital Organization, Sendai Medical Center, Sendai, Japan5Department of Thoracic Surgery, Institute of Development, Aging and Cancer, Tohoku University Graduate School of Medicine, Sendai, Japan6Division of Pathology, Shizuoka Cancer Center Hospital and Research Institute, Sizuoka, Japan7Member of the German Cancer Consortium (DKTK)

Correspondence should be addressed to A Kasajima: [email protected]

Abstract

In the light of novel cancer immune therapies, the status of antitumor inflammatory

response and its regulation has gained much attention in patients with lung cancer. Ample

datasets exist for non-small-cell lung cancer, but those for pulmonary neuroendocrine

tumors are scarce and controversial. Here, tumor-associated inflammation, CD8+ cell

infiltration and PD-L1 status were evaluated in a cohort of 57 resected carcinoids and

185 resected neuroendocrine carcinomas of the lung (58 large cell carcinomas and 127

small cell carcinomas). Data were correlated with clinicopathological factors and survival.

Moderate or high tumor-associated inflammation was detected in 4 carcinoids (7%) and

in 37 neuroendocrine carcinomas (20%). PD-L1 immunoreactivity was seen in immune

cells of 73 (39%) neuroendocrine carcinomas, while tumor cells were labeled in 21 (11%)

cases. Inflammatory cells and tumor cells in carcinoids lacked any PD-L1 expression. In

neuroendocrine carcinomas, PD-L1 positivity in immune cells, but not in tumor cells,

was associated with intratumoral CD8+ cell infiltration (P < 0.001), as well as with the

severity of tumor-associated inflammation (P < 0.001). In neuroendocrine carcinomas,

tumor-associated inflammation and PD-L1 positivity in immune cells correlated with

prolonged survival and the latter factor was also an independent prognosticator (P < 0.01,

hazard ratio 0.4 for overall survival, P < 0.001 hazard ratio 0.4 for disease-free survival).

Taken together, in neuroendocrine tumors, antitumor inflammatory response and PD-L1

expression are largely restricted to neuroendocrine carcinomas, and in this tumor entity,

PD-L1 expression in inflammatory cells is positively correlated to patient survival.

Introduction

Pulmonary neuroendocrine tumors, as defined in the 2015 classification of the World Health Organization (WHO), comprise three different types; carcinoid tumors,

large cell neuroendocrine carcinomas (LCNEC) and small-cell lung carcinomas (SCLC) (Travis et al. 2015). Carcinoid tumors are low-to-intermediate grade neoplasms that

Endocrine-Related Cancer (2018) 25, 339–350

3

Key Words

f neuroendocrine carcinoma

f pulmonary carcinoid

f PD-L1

f prognosis

f inflammation

25

Downloaded from Bioscientifica.com at 12/18/2020 04:18:26AMvia free access

https://doi.org/10.1530/ERC-17-0427

mailto:[email protected]

Printed in Great BritainPublished by Bioscientifica Ltd.https://doi.org/10.1530/ERC-17-0427

http://erc.endocrinology-journals.org © 2018 Society for Endocrinology

340A Kasajima et al. PD-L1 in lung neuroendocrine tumors


are subclassified into typical and atypical carcinoids on the basis of mitotic count and/or presence of necrosis. Although atypical carcinoids are more likely to metastasize and have a worse prognosis, these two neoplasms share the same phenotype and genotype (Simbolo et al. 2017). In contrast to carcinoids, SCLC and LCNEC are high-grade neuroendocrine carcinomas (NEC) (Travis et al. 2015). Patients with these malignancies share similar clinical features, e.g. higher incidence in males than in females, strong association with cigarette smoking, early metastasis and dismal prognosis (Asamura et al. 2006). Also, the genetic profiles of the two NEC types are largely overlapping (Rekhtman et al. 2016, Simbolo et al. 2017). Moreover, the few patients who are amenable to surgical treatment benefit from adjuvant chemotherapy (Kim et al. 2017).

Programmed death-ligand 1 (PD-L1) is a transmembrane protein that regulates immune cell activity through binding to its receptor, programmed death protein 1 (PD-1), expressed on inflammatory cells. The interaction of PD-L1 and PD-1 inhibits antitumor immune response through negative regulation of T-cell proliferation and cytokine production. Therefore, blockade of the PD-L/PD-1 pathway has become one of the most promising therapeutic options in various human malignancies including lung cancer (Brahmer et al. 2015, Herbst et al. 2016, Massard et al. 2016, Reck et al. 2016, Rittmeyer et al. 2017).

Pembrolizumab, a monoclonal antibody against PD-1, was reported to prolong survival in patients with metastasized non-small-cell lung cancer (NSCLC) (Reck et al. 2016). This study, as well as multiple additional clinical trials with a variety of other immune checkpoint inhibitors (Brahmer et al. 2015, Rittmeyer et al. 2017), has defined PD-L1 protein expression detected by immunohistochemistry as a valid biomarker for selecting patients who will benefit from immune checkpoint inhibitory agents. Although patient selection by PD-L1 expression is, due to various reasons, by no means perfect, this biomarker is the only currently approved method to predict response in NSCLC (Gridelli et al. 2017). Currently, PD-L1 staining is established as a companion diagnostic biomarker in first-line (>50% positivity in tumor cells) treatment of metastasized NSCLC patients prior to therapy with pembrolizumab and in second-line (>1% positivity in tumor cells) treatment of metastasized NSCLC, again prior to therapy with pembrolizumab.

Despite recent progress in the understanding of immune regulation in NSCLC, the role of antitumor immunity in neuroendocrine tumors of the lung is far

less clear. A clinical study has recently reported the efficacy and safety of pembrolizumab in patients with advanced SCLC (Ott et al. 2017). However, data regarding PD-L1 expression in pulmonary neuroendocrine tumors are sparse, only fragmented data are available especially in carcinoids and LCNEC (Fan et al. 2016, Tsuruoka et al. 2017). Recently, some studies reported on PD-L1 immunoreactivity in mostly small cohorts of SCLC (Ishii et al. 2015, Schultheis et al. 2015, Berghoff et al. 2016, Fan et al. 2016, Miao et al. 2016, Toyokawa et al. 2016, Tsuruoka et al. 2017, Yu et al. 2017). However, the reported incidence and localization of PD-L1 expression varied considerably, probably due to the fact that differing antibodies were used and/or variable cutoff values were applied. In addition, often expression was investigated in tissue microarrays (Schultheis et al. 2015, Berghoff et al. 2016, Fan et al. 2016, Miao et al. 2016, Takada et al. 2016, Tsuruoka et al. 2017). In this context, it is important to note that antibody/cutoff selection is crucial in the evaluation of PD-L1 positivity in NSCLC (Scheel et al. 2016, 2017).

The aim of this study is to clarify the status and clinical impact of inflammatory response and PD-L1 expression in a large series of surgically resected neuroendocrine tumors of the lung, employing a rigorously standardized methodology.

Materials and methods

Tissues and patients’ characteristics

Consecutive tissue samples from 242 patients with pulmonary neuroendocrine tumors (57 patients with carcinoids, and 185 patients with NEC), who underwent surgical resection at 10 institutions in Japan between 1996 and 2015 were included into the study (Supplementary Table 1, see section on supplementary data given at the end of this article). All tumors were primaries, recurrences were excluded. Detailed clinical information was obtained for 225 patients (93%; 168 patients with NEC and 57 patients with carcinoids) (Table 1). Clinical follow-up data was available for 224 patients (167 patients with NEC and 57 patients with carcinoids). Among 212 patients, for whom post-operative treatment information was available, 95 patients (45%) received adjuvant therapy (CDBCA/VP16 37%, CDDP/VP16 24%, CDDP/CPT11 13%, CDBCA/CPT11 6%, CDBCA/ETP 3%, others 19%). Histological diagnoses were confirmed after extensive reviewing of the specimens by five pathologists including two neuroendocrine tumor experts (AK and HS) and three lung pathology experts (TK, YI and WW), who based their



http://erc.endocrinology-journals.org/cgi/content/full/ERC-17-0427/DC1






assessment on the criteria of the 2015 WHO classification (Travis et al. 2015). The mean follow-up time was 42 months (72 and 32 months for carcinoids and NECs, respectively). Three patients (two with LCNEC and 1 with SCLC) died within 1 month of surgery and were excluded from the survival analyses. All tumors were reclassified according to the 8th edition of the TNM classification (Brierley et al. 2017). The study has been approved by the ethics committees of all participating institutions.

Immunohistochemistry and evaluation

Serial 3-μm tissue sections from paraffin blocks were prepared. For immunohistochemistry, a mouse monoclonal antibody directed against PD-L1 (clone 22C3, dilution 1:30, Dako, Glostrup, Denmark – the companion diagnostic antibody used in all Pembrolizumab studies – and a mouse monoclonal antibody directed against CD8 (clone C8/144B, dilution 1:100, Dako) as well as a mouse monoclonal antibody directed against CD68 (clone KP1, dilution 1:10,000, Dako) were used. Immunohistochemical stainings were performed using an automated staining module (BenchMark XT system, Roche/Ventana Medical Systems, Tuscon, AZ, USA). Briefly, for PD-L1

staining tissue sections were deparaffinized with EZ Prep (Roche/Ventana Medical Systems) at 75°C, heat pretreated in Cell Conditioning 1 (CC1) for antigen retrieval at 76°C–100°C and then incubated with the primary antibody for 32 min at 37°C after inactivation of the endogenous peroxidase using UV inhibitor for 4 min at 37°C. Slides were then incubated with a HRP Universal Multimer (Roche/Ventana Medical Systems) for 8 min. Antibody binding was detected using DAB as chromogen and slides were counterstained with hematoxylin. To ascertain staining quality, rigorous standards were applied, including internal and external positive/negative controls (Fig. 1H). The quality of staining was ascertained by participation in various round robin trials (Scheel et al. 2016, 2017).

Tumor-associated inflammation was assessed by two observers (AK and NO) as previously described (Kasajima et al. 2010). Briefly, a full-section hematoxylin-eosin (H&E) slide was screened for tumor-associated inflammation (including lymphocytes, macrophages and granulocytes), which was graded as absent (no inflammatory cells at the tumor margin), weak (mild and patchy inflammatory cells at the tumor margin), moderate (prominent band-like inflammatory reaction at the tumor margin) or

Table 1 Clinicopathological characteristics of 225 patients with neuroendocrine tumors of the lung.

N (%) Typical carcinoid N (%) Atypical carcinoid N (%) LCNEC N (%) SCLC N (%)

N (%) 225 (100) 39 (17) 18 (8) 53 (24) 115 (51)Age (year) Mean ± s.d. 68 ± 9 64 ± 11 65 ± 11 68 ± 10 70 ± 7Sex Male 170 (76) 15 (38) 9 (50) 48 (91) 98 (85) Female 55 (24) 24 (62) 9 (50) 5 (9) 17 (15)Smoking index Mean ± s.d. 842 ± 627 263 ± 381 425 ± 722 938 ± 437 1075 ± 597Size (cm) Mean ± s.d. 2.6 ± 1.5 2.1 ± 1.4 3.1 ± 2.4 3.1 ± 1.5 2.6 ± 1.3pTa

pT1a,b,c 122 (55) 29 (74) 9 (56) 18 (35) 66 (58) pT2a,b 71 (32) 9 (23) 4 (25) 25 (48) 33 (29) pT3 22 (10) 0 (0) 1 (6) 7 (13) 14 (12) pT4 5 (2) 1 (3) 2 (13) 2 (4) 0 (0)pNb

pN0 149 (72) 36 (97) 12 (75) 30 (61) 71 (67) pN1, 2, 3 59 (28) 1 (3) 4 (25) 19 (39) 35 (33)pMa

pM0 210 (95) 39 (100) 16 (100) 49 (94) 106 (94) pM1a,b 10 (5) 0 (0) 0 (0) 3 (6) 7 (6)pStagec

pStage I 131 (59) 35 (90) 10 (63) 24 (46) 62 (54) pStage II–IV 90 (41) 4 (10) 6 (38) 28 (54) 52 (46)

Clinical information was not obtainable for 17 patients of our cohort. Data not available in a5, b17, and c4 cases, respectively.LCNEC, large cell neuroendocrine carcinoma; SCLC, small-cell lung carcinoma; s.d., standard deviation.







Figure 1Histological and immunohistochemical images of inflammation in neuroendocrine carcinomas of the lung. (A) H&E of a small cell carcinoma. Numerous inflammatory cells infiltrating the tumor area can be seen. (B) Abundant infiltration of CD8+ T-cells observed in the stroma of a small cell carcinoma. (C) Tightly gathered PD-L1-expressing immune cells, closely attached to cancer cell nests. (D) A case with abundant PD-L1-positive macrophages, while lymphocytes (arrow) and tumor cells (arrow head) were mostly negative. CD8-positive T-cells (E) and CD68-positive macrophages (F) in consecutive sections to D. The distribution of CD68+ cells (F) showed a strong overlap to PD-L1+ IC cells (D). CD8+ cells (E) were found in the same area, but their density was slightly lower than that of CD68+ cells. (G) Focal PD-L1 expression observed in tumor cells in a case of small cell carcinoma. (H) For PD-L1 staining, quality control human tonsil was used. PD-L1 is strongly positive in crypt epithelium and weakly positive in dendritic cells.







high (florid cup-like inflammation at the invasive edges) without further differentiation of the different cell types (Mohammed et al. 2012). Cases with absent or weak infiltrates were defined as tumor-associated inflammation negative, while cases with moderate or high infiltrate were defined as positive. CD8+ cells were counted by two observers (AK and AI) in the area of highest immune cell density by using a 40× objective lens. For the assessment of PD-L1, membranous staining in immune cells (IC) and tumor cells (TC) were independently evaluated by three observers (AK, AI and WW). Cases with disagreement were discussed using a multiheaded microscope. The proportion of the tumor area occupied by PD-L1-positive cells was scored regardless of PD-L1 staining intensity. A score of <1% was considered as PD-L1 negative, a score of ≥1% was considered as PD-L1 positive (Herbst et al. 2016, Ott et al. 2017).

Statistical analyses

JMP Pro, version 13.1.0 software (SAS Institute, Inc., Cary, NC, USA) was used for all statistical analyses. The sample distribution between groups was compared using Pearson’s chi-square test. The Wilcoxon test was applied for the comparison of continuous values or scores between groups. The probability of differences in overall survival (OS) and disease-free survival (DFS) was determined using the Kaplan–Meier method, with a log-rank test to probe for significance. Multivariate analysis was undertaken with the Cox model of proportional hazards. A P value of <0.05 was considered statistically significant.

Results

Patients’ characteristics and clinical relevance of standard pathological factors

Thirty-nine patients (17%) had a typical carcinoid, 18 patients (7%) an atypical carcinoid, 58 patients (24%) a LCNEC and 127 patients (52%) a SCLC (Table 1). The associations of clinicopathological parameters with survival of patients with carcinoids and NECs are shown in Supplementary Tables 2 and 3, respectively. Univariate survival analysis of patients with NEC revealed ≥pN1 and ≥pStage II as adverse prognosticators for OS (P < 0.0001 for pN, P = 0.004 for pStage) as well as for DFS (P < 0.0001 for pN, P = 0.002 for pStage). Significant prognostic factors obtained from univariate analyses were included into multivariate analyses (see the ‘Correlation of tumor-associated inflammation and PD-L1 immunoreactivity with survival’ section below).

Tumor-associated inflammation and CD8+ cell infiltrate

Moderate/high inflammation was observed in only 4 (7%) carcinoid tumors (one typical and 3 atypical), but in 37 NECs (20%, P < 0.0001; 12 LCNEC, 25 SCLC) (Fig. 1 and Table 2). The degree of CD8+ cell infiltration (Fig. 1) was positively correlated with tumor-associated inflammation (mean number 76 in the negative group, 156 in the positive group, P < 0.0001, Fig. 2). The number of CD8+ cells was significantly higher in NEC (mean 92)

Table 2 Histomorphological and immunohistochemical characteristics of 242 patients with neuroendocrine tumors of the lung.

N (%)

Typical carcinoid N (%)

Atypical carcinoid N (%)

P value

LCNEC N (%)

SCLC N (%)

P value

N (%) 242 (100) 39 (16) 18 (7) 58 (24) 127 (52)Tumor-associated inflammation Negative 201 (83) 38 (97) 15 (83)

NS46 (79) 102(80)

NS Positive 41 (17) 1 (3) 3 (17) 12 (21) 25 (20)CD8+ cells Mean ± s.d. 73 ± 82 7 ± 10 19 ± 21 0.01 83 ± 70 96 ± 91 NSPD-L1 in ICa

Mean ± s.d. (%) 5 ± 12 0 ± 0 0.01 ± 0.03 0.03 7 ± 12 7 ± 15 NS Negative 167 (70) 39 (100) 18 (100)

–32 (56) 78 (62)

NS Positive 73 (30) 0 (0) 0 (0) 25 (44) 48 (38)PD-L1 in TCa

Mean ± s.d. (%) 0.5 ± 2.7 0 ± 0 0 ± 0 – 0.7 ± 4 0.7 ± 3 NS Negative 219 (91) 39 (100) 18 (100)

–52 (91) 110 (87)

NS Positive 21 (9) 0 (0) 0 (0) 5 (9) 16 (13)

Data not available in a2 cases.IC, immune cells; LCNEC, large-cell neuroendocrine carcinoma; NS, not significant; SCLC, small-cell lung carcinoma; s.d., standard deviation; TC, tumor cells.









than in carcinoids (mean 12, P < 0.0001). The number of CD8+ cells was higher in atypical carcinoids (mean 19) than in typical carcinoids (mean 7, P = 0.05), no difference was found between LCNEC and SCLC (mean 83 vs mean 97) (Table 2).

Immunohistochemical PD-L1 expression

In carcinoid tumors, PD-L1 immunoreactivity was not detected in either IC or TC (IC: 0%, TC: 0%). In NEC, PD-L1 immunoreactivity ≥1% was seen in IC in

Figure 2Correlation of the number of CD8+ cells with tumor-associated inflammation and PD-L1 expression. (A) The number of CD8+ cell was positively correlated with tumor-associated inflammation assessed on H&E slides (P < 0.0001). (B) The number of CD8+ cell was significantly higher in cases with PD-L1 expression in immune cells (P < 0.0001). (C) No difference was detected in the number of CD8+ cell between cases with and without PD-L1 expression in tumor cells. IC, immune cells; NS, not significant; TC, tumor cells.







73 tumors (39%) with a relatively low percentage of covered tumor area even in the positive group (median percentage of tumor area covered by PD-L1 positive IC: 10, range 1–60). High PD-L1 expression in ICs (≥50%) was observed in 5 of 126 SCLC (3%), but was not detected in LCNEC (Supplementary Table 3). Most of the PD-L1-positive ICs were directly associated with tumor cell nests (Fig. 1). The overall location and distribution of PD-L1-positive cells was equivalent to those of CD68-expressing macrophages on consecutive sections (Fig. 1). In contrast, CD8+ cells were only sparsely distributed in the corresponding areas of the tumors (Fig. 1). IC cell characteristics were determinable by morphology in 62 IC PD-L1-positive NEC; 39 NEC (76%) expressed PD-L1 only in macrophages, 23 NEC (24%) expressed PD-L1 both in macrophages and lymphocytes. No significant difference was detected for PD-L1 expression in IC between LCNEC and SCLC (Table 2). PD-L1 positivity in IC was significantly correlated with the number of CD8+ cells (Fig. 2).

PD-L1 immunoreactivity in TC was scarcely detected. Only 21 NECs (12%) were found to be positive (≥1%) and even these cases usually showed a low percentage of positive cells. None of the tumors expressed PD-L1 in ≥50% of TC, PD-L1 positivity in TC ranged from 0 to 30%. No differences were detected in tumor-associated inflammation and PD-L1 status in TC between LCNEC

and SCLC (Table 2). The number of CD8+ cells did not differ between tumors with and without PD-L1 immunoreactivity in TC (Fig. 2).

Correlation of tumor-associated inflammation and PD-L1 immunoreactivity with survival

In patients with carcinoids, no prognostic impact was detected for tumor-associated inflammation. Survival impact of PD-L1 expression in carcinoids could not be assessed due to the complete absence of PD-L1 expression in both TC and IC in our cohort. NEC patients with moderate/high tumor-associated inflammation showed a significantly prolonged OS (P = 0.03) and DFS (P = 0.03, Fig. 3 and Table 3). The same was true for NEC with high PD-L1 expression in IC (P = 0.03 for OS and P = 0.003 for DFS, Fig. 3 and Table 3). Since PD-L1 positivity in IC was usually seen in macrophages, we assessed whether the positive survival impact of PD-L1-expressing IC was a mere consequence of stronger macrophage infiltration in these tumors. However, macrophage infiltration in NEC by itself had no survival impact and when stratified for both parameters combined (PD-L1 positivity/macrophage infiltration) only those patients with PD-L1-positive macrophages had a survival advantage compared to patients without macrophages (with or without other

Figure 3Kaplan–Meier survival curves for tumor-associated inflammation (A for overall survival, B for disease-free survival) and PD-L1 expression in immune cells (C for overall survival, D for disease-free survival). Patients with moderate/high tumor-associated inflammation demonstrated a longer overall survival (A, P = 0.03) and a longer disease-free survival (B, P = 0.03). Patients with PD-L1 expression detected in immune cells demonstrated a longer overall survival (C, P = 0.03) and disease-free survival (D, P = 0.003) compared to the patients with no PD-L1 expression in immune cells (IC, immune cells).








Tab

le 3

U

niv

aria

te s

urv

ival

an

alys

is in

pat

ien

ts w

ith

neu

roen

do

crin

e ca

rcin

om

as o

f th

e lu

ng

.

N

Overa

ll s

urv

ival

Dis

ease

-fre

e s

urv

ival

Even

tM

edia

n s

urv

ival

(m

on

ths)

Log

-ran

k te

st (

P-va

lue)

Even

tM

edia

n s

urv

ival

(m

on

ths)

Log

-ran

k te

st (

P-va

lue)

Ag

e

<70

≥

7083 81

35 3974 47

NS

47 5219 15

NS

Sex

M

ale

Fe

mal

e14

2 2264 10

55 55N

S86 13

16 14N

S

Smo

kin

g in

dex

a

<

1000

≥

1000

83 6940 29

47 76N

S55 38

12 21N

S

Size

(m

m)b

<

25

≥25

77 7928 41

80 32N

S43 50

24 14N

S

pTc

p

T1a,

b,c

p

T2a,

b

pT3

p

T4

83 55 21 2

29 32 11 1

80 47 22 –

NS

43 38 16 1

24 14 10 –

NS

pN

d

p

N0

p

N1,

2, 3

99 5236 32

80 20<

0.00

0150 39

44 11<

0.00

01

pM

c

p

M0

p

M1a

,b15

1 1069 3

55 24N

S91 6

19 17N

S

pSt

agee

p

Stag

e I

p

Stag

e II–

IV84 78

30 4380 22

0.00

443 55

36 110.

002

WH

O20

15

LCN

EC

SCLC

51 113

28 4647 59

NS

34 6512 22

NS

Tum

or-

asso

ciat

ed in

flam

mat

ion

N

egat

ive

Po

siti

ve13

0 3461 13

37 800.

0382 17

13 360.

03

PD-L

1 in

ICe

N

egat

ive

Po

siti

ve94 68

50 2237 10

80.

003

67 3013 74

0.00

03

PD-L

1 in

TC

e

N

egat

ive

Po

siti

ve14

1 2165 7

47 87N

S87 10

16 68N

S

Thre

e p

atie

nts

wh

o d

ied

wit

hin

on

e m

on

th a

fter

su

rger

y w

ere

excl

ud

ed f

rom

su

rviv

al a

nal

ysis

. Dat

a n

ot

avai

lab

le in

a12

, b8,

c 3, d

13, a

nd

e2

case

s, r

esp

ecti

vely

.IC

, im

mu

ne

cells

; LC

NEC

, lar

ge

cell

neu

roen

do

crin

e ca

rcin

om

a; N

S, n

ot

sig

nifi

can

t; S

CLC

, sm

all-

cell

lun

g c

arci

no

ma;

TC

, tu

mo

r ce

lls.







PD-L1-positive inflammatory cells) and patients with macrophages, which do not express PD-L1 (Supplementary Fig. 1). PD-L1 expression in TC of NEC had no prognostic implication.

A multivariate survival analysis including tumor size, local tumor extent (pT), absence or presence of lymph node metastases (pN), tumor-associated inflammation and PD-L1 immunoreactivity in IC showed an independent poor prognostic impact for the presence of lymph node metastasis (≥pN1, HR 2.6 (95% CI 1.6–4.4) P = 0.001 for OS, ≥pN1, HR 2.5 (95% CI 1.6–3.9) P < 0.0001 for DFS) and an independent favorable prognostic impact for PD-L1 immunoreactivity in IC (HR 0.4 (95% CI 0.2–0.7) P = 0.001 for OS, HR 0.4 (95% CI 0.3–0.7) P = 0.001) (Table 4).

Discussion

In this study investigating the so far largest pulmonary neuroendocrine tumor cohort, we demonstrate an association between severity of antitumor inflammation as well as of PD-L1 expression in IC and favorable prognosis in pulmonary NEC. Our results indicate that T-cell-mediated antitumor immunity seem to play a role in high-grade neuroendocrine tumors of the lung, but not in carcinoids, suggesting that the latter group of pulmonary neoplasms is ‘immune deserted’.

The status of PD-L1 expression in lung neuroendocrine tumor has previously not been fully established; most series were small in number, employed tissue microarrays and/or included a limited number of histological types (Supplementary Table 4) (Ishii et al. 2015, Schultheis et al. 2015, Berghoff et al. 2016, Fan et al. 2016, Miao et al. 2016, Takada et al. 2016, Toyokawa et al. 2016, Tsuruoka et al. 2017, Yu et al. 2017). In our study, a large number of surgically resected SCLC could be included, mainly due to the frequent usage of lung cancer surveillance in Japan, which results in the detection of a higher rate

of potentially resectable SCLC at the time of diagnosis. This uniqueness in our cohort allowed for a more precise evaluation of the inflammatory status in NEC, compared to what can be achieved in cohorts of biopsy specimen or tissue microarrays.

Interestingly, we found PD-L1 expression in intratumoral IC to be prognostic in NEC. An association of PD-L1 expression in IC, but not in TC with a favorable prognosis in SCLC patients has also been reported in one of the previous studies that addressed this issue (Berghoff et al. 2016) and in extrapulmonary human malignancies (Bellmunt et al. 2015, Darb-Esfahani et al. 2016, Kim et al. 2016). PD-L1 expression in IC has also been reported to be correlated with high-histologic grade and solid subtype in NSCLC (Driver et al. 2017). The detailed mechanisms of PD-L1 overexpression in IC and the reason for its association with a better prognosis are largely unknown. A compensatory upregulation of PD-L1 expression by pro-inflammatory cytokines that are generated in the tumor microenvironment has been discussed as one potential option (Darb-Esfahani et al. 2016).

Our observation that most PD-L1-expressing ICs in NEC were macrophages indicated that this cell type might be centrally involved into the antitumor inflammatory reaction in this tumor entity. It has been reported that inflammation-associated signals by TNF-alpha induce inflammatory monocytes/macrophages expressing PD-L1 in the tumor microenvironment (Hartley et al. 2017). Moreover, an inhibitory function of PD-L1 expressing intratumoral macrophages on T-cell proliferation has been observed in tumor tissues (Wang et al. 2017). Thus, our results, in the light of these former findings, further support the hypothesis that PD-L1 expressed in intratumoral macrophages might be a tumor-induced protective mechanism against antitumor inflammation in pulmonary NEC. This tumor protective function was likely dependent on the ability of macrophages to express

Table 4 Multivariate survival analysis in patients with neuroendocrine carcinomas of the lung.

Overall survival Disease-free survival

Hazard ratio 95% CI P value Hazard ratio 95% CI P value

pN pN0 pN1, 2, 3

12.6 1.6–4.4 0.001 1

2.5 1.6–3.9 <0.0001

Tumor-associated inflammation Negative Positive

10.7 0.3–1.2 NS 1

0.7 0.4–1.2 NS

PD-L1 in IC Negative Positive

10.4 0.2–0.7 0.001 1

0.4 0.3–0.7 0.001

CI, confidence interval; IC, immune cells; NS, not significant.










PD-L1, since the density of macrophages in itself had no prognostic impact whatsoever in our cohort.

Multiple clinical trials are currently evaluating the efficacy and safety of immune checkpoint inhibitors in SCLC (Karachaliou et al. 2017). Some preliminarily results demonstrated a favorable response rate and manageable toxicity (Antonia et al. 2016). Recently, the first large dataset on immune checkpoint therapy in pulmonary neuroendocrine tumors was fully published; in this study, pembrolizumab proved to be effective in a subset of patients with these tumors (Ott et al. 2017). From these studies, somewhat controversial results emerged with respect to the predictive value of PD-L1 immunoreactivity in tumor cells, with the majority of investigations showing only minor effects. Regrettably, none of the studies reported on the predictive effect of immune cell positivity for PD-L1 in these scenarios.

Multiple PD-L1 antibodies are commercially available, of which especially four (22C3, 28-8, SP142, SP263) have been used as clinical trial assays in the context of currently available immune checkpoint inhibitor treatment (Scheel et al. 2016, 2017). Some previous studies have investigated PD-L1 expression in neuroendocrine tumors of the lung applying a wide range of different antibodies and staining/evaluation methods (Supplementary Table 4), which may account for the heterogeneous results reported. Ours is the first study in pulmonary neuroendocrine tumors that analyzed PD-L1 expression employing the 22C3 clone in the context of a rigorous quality controlled assay setup. Since we have recently shown that PD-L1 staining results for three of the trial antibodies (SP263, 22C3, 28-8) are well comparable, when used in a quality controlled way (Scheel et al. 2016, 2017). We believe that our data obtained with the 22C3 antibody can be seen as representative and are transferable to other trial antibody scenarios.

The following limitations apply to our study. First, we cannot exclude a certain selection bias due to the inclusion of only surgically resected patients in earlier stages in our cohort. Second, since this is a multi-institutional cohort, potential inter-institutional variability in adjuvant treatment and post-operative surveillance methods might have a confounding influence in our dataset. A platinum-based chemotherapy was performed in all NEC patients, but the chemotherapeutic partner used was variable, as was the number and timing of chemotherapy cycles applied. A phase III clinical study conducted by the Japan Clinical Oncology Group (JCOG1205/1206, HGNEC-EP/IP-P3) to determine the optimal adjuvant chemotherapy in resected NEC patients is currently ongoing. To date, however, a standardized adjuvant

therapy or post-operative monitoring protocol for patients with SCLC/LCNECs is lacking. Third, some clinical data were missing in our cohort such as e.g. WHO performance status. And although our mean follow-up for carcinoid tumor patients was more than 5 years, even longer follow-up times could be potentially useful for this group of quite indolent tumors (Cao et al. 2011). All mentioned limitations are very hard to avoid when studying tissue cohorts of pulmonary neuroendocrine tumors, which are rare and in case of SCLC are only very occasionally resected in most countries. To assemble an extremely large cohort like ours, one has to go for a logistically demanding assembly of a multi-institutional tumor collection, which – outside a clinical trial – always comes with some data heterogeneity. We nevertheless believe that despite all mentioned limitations, much can be learned from our data set.

In conclusion, we demonstrated antitumor inflammation to be prognostic in NEC of the lung, while it was virtually not present and possibly played no role in carcinoids. The quantity of PD-L1-positive IC, which were mainly macrophages, was closely linked to tumor-associated inflammation and cytotoxic T-cell infiltration and was also indicative of a better prognosis. In the future, it would be interesting to explore the predictive value of inflammation in general and of PD-L1-expressing inflammatory cells in the context of immune checkpoint inhibition.

Supplementary dataThis is linked to the online version of the paper at https://doi.org/10.1530/ERC-17-0427.

Declaration of interestThe authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.

FundingThis work is supported by funding provided by the Alexander von Humboldt foundation for AK.

Author contribution statementA K, N O and H I have contributed to specimen preparation, evaluation and statistical analysis. A K, A I, K S, N O, H I and W W were responsible for immunohistochemical staining and case evaluation. A K, K S, B K, G K, H S and W W contributed to writing, editing and finalizing the manuscript. A K, T K, Y I, G K and W W were responsible for the determination of the histological diagnosis. A S and Y O provided clinical data.










AcknowledgementsThe authors would like to acknowledge Dr Nobuyuki Sato and Dr Hidekachi Kurotaki (Aomori Prefectural Central Hospital, Aomori, Japan), Dr Satoshi Suzuki, Dr Shinsaku Ueda, Dr Yamato Suzuki, Dr Yuko Itakura and Dr Tohru Takahashi (Ishinomaki Red Cross Hospital, Miyagi, Japan), Dr Kazuyuki Ishida (Iwate Medical University, Iwate, Japan), Dr Hiroyuki Oura and Dr Tsutomu Sakuma (Iwate Prefectural Central Hospital, Iwate, Japan), Dr Jotaro Shibuya (Iwate Prefectural Isawa Hospital, Iwate, Japan), Dr Jiro Abe, Dr Satomi Takahashi, Dr Ikuro Sato (Miyagi Cancer Center, Miyagi, Japan), Dr Akira Miyamoto (Miyagi Cardiovascular and Respiratory Center, Miyagi, Japan), Dr Kazuhiro Sakamoto and Dr Kazuyoshi Shimada (Osaki Citizen Hospital, Miyagi, Japan), Dr Hasumi Tohru, Dr Junko Sakurada (Sendai Medical Center, Miyagi, Japan), Dr Nobuaki Tamahashi (Surgical Pathology Japan Inc., Miyagi, Japan), Dr Toshiharu Tabata, Dr Naoya Ishibashi, Dr Kazuhiro Murakami (Tohoku Medical and Pharmaceutical University Hospital, Miyagi, Japan), Dr Kazuma Kobayashi (Tohoku University Hospital, Miyagi, Japan) for providing clinical data and tissue samples. They also thank Dr Samaneh Yazdani, Dr Shuko Hata, Kazue Ise, Erina Iwabuchi, Katsuhiko Ono, Tsuyoshi Miura, Yasuko Furukawa (Tohoku University, Miyagi, Japan), Maki Takahashi (Sendai Medical Center, Miyagi, Japan), Marion Mielke, Olga Seelbach, and Ulrike Mühlthaler (Technical University Munich, Munich, Germany) for their excellent technical support.

ReferencesAntonia SJ, Lopez-Martin JA & Bendell J 2016 Nivolumab alone and

nivolumab plus ipilimumab in recurrent small-cell lung cancer (CheckMate 032): a multicentre, open-label, phase 1/2 trial. Lancet Oncology 17 883–895. (https://doi.org/10.1016/S1470-2045(16)30098-5)

Asamura H, Kameya T, Matsuno Y, Noguchi M, Tada H, Ishikawa Y, Yokose T, Jiang SX, Inoue T, Nakagawa K, et al. 2006 Neuroendocrine neoplasms of the lung: a prognostic spectrum. Journal of Clinical Oncology 24 70–76. (https://doi.org/10.1200/JCO.2005.04.1202)

Bellmunt J, Mullane SA, Werner L, Fay AP, Callea M, Leow JJ, Taplin ME, Choueiri TK, Hodi FS, Freeman GJ, et al. 2015 Association of PD-L1 expression on tumor-infiltrating mononuclear cells and overall survival in patients with urothelial carcinoma. Annals of Oncology 26 812–817. (https://doi.org/10.1093/annonc/mdv009)

Berghoff AS, Ricken G, Wilhelm D, Rajky O, Widhalm G, Dieckmann K, Birner P, Bartsch R & Preusser M 2016 Tumor infiltrating lymphocytes and PD-L1 expression in brain metastases of small cell lung cancer (SCLC). Journal of Neuro-Oncology 130 19–29. (https://doi.org/10.1007/s11060-016-2216-8)

Brahmer J, Reckamp KL, Baas P, Crino L, Eberhardt WE, Poddubskaya E, Antonia S, Pluzanski A, Vokes EE, Holgado E, et al. 2015 Nivolumab versus docetaxel in advanced squamous-cell non-small-cell lung cancer. New England Journal of Medicine 373 123–135. (https://doi.org/10.1056/NEJMoa1504627)

Brierley JD, Gospodarowicz MK & Wittekind C 2017 Classification of Malignant Tumours, 8th ed. Hoboken, NJ, USA: John Wiley & Sons.

Cao C, Yan TD, Kennedy C, Hendel N, Bannon PG & McCaughan BC 2011 Bronchopulmonary carcinoid tumors: long-term outcomes after resection. Annals of Thoracic Surgery 91 339–343. (https://doi.org/10.1016/j.athoracsur.2010.08.062)

Darb-Esfahani S, Kunze CA, Kulbe H, Sehouli J, Wienert S, Lindner J, Budczies J, Bockmayr M, Dietel M, Denkert C, et al. 2016 Prognostic impact of programmed cell death-1 (PD-1) and PD-ligand 1 (PD-L1) expression in cancer cells and tumor-infiltrating lymphocytes in ovarian high grade serous carcinoma. Oncotarget 7 1486–1499. (https://doi.org/10.18632/oncotarget.6429)

Driver BR, Miller RA, Miller T, Deavers M, Gorman B, Mody D, Ge Y, Barrios R, Bernicker E, Kim M, et al. 2017 Programmed death ligand-1 (PD-L1) expression in either tumor cells or tumor-

infiltrating immune cells correlates with solid and high-grade lung adenocarcinomas. Archives of Pathology and Laboratory Medicine 141 1529–1532. (https://doi.org/10.5858/arpa.2017-0028-OA)

Fan Y, Ma K, Wang C, Ning J, Hu Y, Dong D, Dong X, Geng Q, Li E & Wu Y 2016 Prognostic value of PD-L1 and PD-1 expression in pulmonary neuroendocrine tumors. OncoTargets and Therapy 9 6075–6082. (https://doi.org/10.2147/OTT.S115054)

Gridelli C, Ardizzoni A, Barberis M, Cappuzzo F, Casaluce F, Danesi R, Troncone G & De Marinis F 2017 Predictive biomarkers of immunotherapy for non-small cell lung cancer: results from an Experts Panel Meeting of the Italian Association of Thoracic Oncology. Translational Lung Cancer Research 6 373–386. (https://doi.org/10.21037/tlcr.2017.05.09)

Hartley G, Regan D, Guth A & Dow S 2017 Regulation of PD-L1 expression on murine tumor-associated monocytes and macrophages by locally produced TNF-alpha. Cancer Immunology, Immunotherapy 66 523–535. (https://doi.org/10.1007/s00262-017-1955-5)

Herbst RS, Baas P, Kim DW, Felip E, Perez-Gracia JL, Han JY, Molina J, Kim JH, Arvis CD, Ahn MJ, et al. 2016 Pembrolizumab versus docetaxel for previously treated, PD-L1-positive, advanced non-small-cell lung cancer (KEYNOTE-010): a randomised controlled trial. Lancet 387 1540–1550. (https://doi.org/10.1016/S0140-6736(15)01281-7)

Ishii H, Azuma K, Kawahara A, Yamada K, Imamura Y, Tokito T, Kinoshita T, Kage M & Hoshino T 2015 Significance of programmed cell death-ligand 1 expression and its association with survival in patients with small cell lung cancer. Journal of Thoracic Oncology 10 426–430. (https://doi.org/10.1097/JTO.0000000000000414)

Karachaliou N, Sosa EA & Rosell R 2017 Immunotherapy for small-cell lung cancer: rationale and clinical evidence. Journal of Xiangya Medicine 2 1–4. (https://doi.org/10.21037/jxym.2017.01.03)

Kasajima A, Sers C, Sasano H, Johrens K, Stenzinger A, Noske A, Buckendahl AC, Darb-Esfahani S, Muller BM, Budczies J, et al. 2010 Down-regulation of the antigen processing machinery is linked to a loss of inflammatory response in colorectal cancer. Human Pathology 41 1758–1769. (https://doi.org/10.1016/j.humpath.2010.05.014)

Kim HR, Ha SJ, Hong MH, Heo SJ, Koh YW, Choi EC, Kim EK, Pyo KH, Jung I, Seo D, et al. 2016 PD-L1 expression on immune cells, but not on tumor cells, is a favorable prognostic factor for head and neck cancer patients. Scientific Reports 6 36956. (https://doi.org/10.1038/srep36956)

Kim KW, Kim HK, Kim J, Shim YM, Ahn MJ & Choi YL 2017 Outcomes of curative-intent surgery and adjuvant treatment for pulmonary large cell neuroendocrine carcinoma. World Journal of Surgery 41 1820–1827. (https://doi.org/10.1007/s00268-017-3908-8)

Massard C, Gordon MS, Sharma S, Rafii S, Wainberg ZA, Luke J, Curiel TJ, Colon-Otero G, Hamid O, Sanborn RE, et al. 2016 Safety and efficacy of durvalumab (MEDI4736), an anti-programmed cell death ligand-1 immune checkpoint inhibitor, in patients with advanced urothelial bladder cancer. Journal of Clinical Oncology 34 3119–3125. (https://doi.org/10.1200/JCO.2016.67.9761)

Miao L, Lu Y, Xu Y, Zhang G, Huang Z, Gong L & Fan Y 2016 PD-L1 and c-MET expression and survival in patients with small cell lung cancer. Oncotarget 8 53978–53988. (https://doi.org/10.18632/oncotarget.9765)

Mohammed ZM, Going JJ, Edwards J, Elsberger B, Doughty JC & McMillan DC 2012 The relationship between components of tumour inflammatory cell infiltrate and clinicopathological factors and survival in patients with primary operable invasive ductal breast cancer. British Journal of Cancer 107 864–873. (https://doi.org/10.1038/bjc.2012.347)

Ott PA, Elez E, Hiret S, Kim DW, Morosky A, Saraf S, Piperdi B & Mehnert JM 2017 Pembrolizumab in patients with extensive-stage small-cell lung cancer: results from the phase Ib KEYNOTE-028 study. Journal of Clinical Oncology 35 3823–3829. (https://doi.org/10.1200/JCO.2017.72.5069)



https://doi.org/10.1016/S1470-2045(16)30098-5

https://doi.org/10.1016/S1470-2045(16)30098-5

https://doi.org/10.1200/JCO.2005.04.1202

https://doi.org/10.1093/annonc/mdv009

https://doi.org/10.1007/s11060-016-2216-8

https://doi.org/10.1007/s11060-016-2216-8

https://doi.org/10.1056/NEJMoa1504627


https://doi.org/10.1016/j.athoracsur.2010.08.062

https://doi.org/10.1016/j.athoracsur.2010.08.062

https://doi.org/10.18632/oncotarget.6429

https://doi.org/10.5858/arpa.2017-0028-OA

https://doi.org/10.2147/OTT.S115054

https://doi.org/10.21037/tlcr.2017.05.09

https://doi.org/10.21037/tlcr.2017.05.09

https://doi.org/10.1007/s00262-017-1955-5

https://doi.org/10.1016/S0140-6736(15)01281-7

https://doi.org/10.1016/S0140-6736(15)01281-7

https://doi.org/10.1097/JTO.0000000000000414

https://doi.org/10.21037/jxym.2017.01.03

https://doi.org/10.1016/j.humpath.2010.05.014

https://doi.org/10.1038/srep36956

https://doi.org/10.1038/srep36956

https://doi.org/10.1007/s00268-017-3908-8

https://doi.org/10.1200/JCO.2016.67.9761



https://doi.org/10.1038/bjc.2012.347

https://doi.org/10.1038/bjc.2012.347

https://doi.org/10.1200/JCO.2017.72.5069

https://doi.org/10.1200/JCO.2017.72.5069





Reck M, Rodriguez-Abreu D, Robinson AG, Hui R, Csoszi T, Fulop A, Gottfried M, Peled N, Tafreshi A, Cuffe S, et al. 2016 Pembrolizumab versus chemotherapy for PD-L1-positive non-small-cell lung cancer. New England Journal of Medicine 375 1823–1833. (https://doi.org/10.1056/NEJMoa1606774)

Rekhtman N, Pietanza MC, Hellmann MD, Naidoo J, Arora A, Won H, Halpenny DF, Wang H, Tian SK, Litvak AM, et al. 2016 Next-generation sequencing of pulmonary large cell neuroendocrine carcinoma reveals small cell carcinoma-like and non-small cell carcinoma-like subsets. Clinical Cancer Research 22 3618–3629. (https://doi.org/10.1158/1078-0432.CCR-15-2946)

Rittmeyer A, Barlesi F, Waterkamp D, Park K, Ciardiello F, von Pawel J, Gadgeel SM, Hida T, Kowalski DM, Dols MC, et al. 2017 Atezolizumab versus docetaxel in patients with previously treated non-small-cell lung cancer (OAK): a phase 3, open-label, multicentre randomised controlled trial. Lancet 389 255–265. (https://doi.org/10.1016/S0140-6736(16)32517-X)

Scheel AH, Dietel M, Heukamp LC, Johrens K, Kirchner T, Reu S, Ruschoff J, Schildhaus HU, Schirmacher P, Tiemann M, et al. 2016 Harmonized PD-L1 immunohistochemistry for pulmonary squamous-cell and adenocarcinomas. Modern Pathology 29 1165–1172. (https://doi.org/10.1038/modpathol.2016.117)

Scheel AH, Baenfer G, Baretton G, Dietel M, Diezko R, Henkel T, Heukamp LC, Jasani B, Johrens K, Kirchner T, et al. 2017 Interlaboratory concordance of PD-L1 immunohistochemistry for non-small-cell lung cancer. Histopathology 72 449–459. (https://doi.org/10.1111/his.13375)

Schultheis AM, Scheel AH, Ozretic L, George J, Thomas RK, Hagemann T, Zander T, Wolf J & Buettner R 2015 PD-L1 expression in small cell neuroendocrine carcinomas. European Journal of Cancer 51 421–426. (https://doi.org/10.1016/j.ejca.2014.12.006)

Simbolo M, Mafficini A, Sikora KO, Fassan M, Barbi S, Corbo V, Mastracci L, Rusev B, Grillo F, Vicentini C, et al. 2017 Lung

neuroendocrine tumours: deep sequencing of the four World Health Organization histotypes reveals chromatin-remodelling genes as major players and a prognostic role for TERT, RB1, MEN1 and KMT2D. Journal of Pathology 241 488–500. (https://doi.org/10.1002/path.4853)

Takada K, Toyokawa G, Okamoto T, Akamine T, Takamori S, Katsura M, Fujishita T, Shoji F, Oda Y & Maehara Y 2016 An immunohistochemical analysis of PD-L1 protein expression in surgically resected small cell lung cancer using different antibodies and criteria. Anticancer Research 36 3409–3412.

Toyokawa G, Takada K, Haratake N, Takamori S, Akamine T, Katsura M, Fujishita T, Shoji F, Okamoto T, Oda Y, et al. 2016 Favorable disease-free survival associated with programmed death ligand 1 expression in patients with surgically resected small-cell lung cancer. Anticancer Research 36 4329–4336.

Travis WD, Brambilla E, Burke AP, Marx A & Nicholson AG 2015 WHO Classification of Tumours of the Lung, Pleura, Thymus and Heart. Lyon, France: IARC Press.

Tsuruoka K, Horinouchi H, Goto Y, Kanda S, Fujiwara Y, Nokihara H, Yamamoto N, Asakura K, Nakagawa K, Sakurai H, et al. 2017 PD-L1 expression in neuroendocrine tumors of the lung. Lung Cancer 108 115–120. (https://doi.org/10.1016/j.lungcan.2017.03.006)

Wang X, Ni S, Chen Q, Ma L, Jiao Z, Wang C & Jia G 2017 Bladder cancer cells induce immunosuppression of T cells by supporting PD-L1 expression in tumour macrophages partially through interleukin 10. Cell Biology International 41 177–186. (https://doi.org/10.1002/cbin.10716)

Yu H, Batenchuk C, Badzio A, Boyle TA, Czapiewski P, Chan DC, Lu X, Gao D, Ellison K, Kowalewski AA, et al. 2017 PD-L1 expression by two complementary diagnostic assays and mRNA in situ hybridization in small cell lung cancer. Journal of Thoracic Oncology 12 110–120. (https://doi.org/10.1016/j.jtho.2016.09.002)

Received in final form 18 December 2017Accepted 11 January 2018Accepted Preprint published online 11 January 2018





https://doi.org/10.1158/1078-0432.CCR-15-2946

https://doi.org/10.1016/S0140-6736(16)32517-X

https://doi.org/10.1016/S0140-6736(16)32517-X

https://doi.org/10.1038/modpathol.2016.117

https://doi.org/10.1111/his.13375

https://doi.org/10.1111/his.13375

https://doi.org/10.1016/j.ejca.2014.12.006

https://doi.org/10.1002/path.4853

https://doi.org/10.1002/path.4853

https://doi.org/10.1016/j.lungcan.2017.03.006

https://doi.org/10.1002/cbin.10716

https://doi.org/10.1002/cbin.10716

https://doi.org/10.1016/j.jtho.2016.09.002

https://doi.org/10.1016/j.jtho.2016.09.002

inflammation and pd-l1 expression in pulmonary ... · tumor-associated inflammation and pd-l1...

Documents