Evaluating the significance of density, localization, and PD-1/PD-L1 immunopositivity of mononuclear cells in the clinical course of lung adenocarcinoma patients with brain metastasis

Clinical characteristics of lung adenocarcinoma patients with brain metastasis included in this study

Characteristic	Patients (N = 208)
Characteristic	No.	%
Age, y
Mean (±SD)	59.28 (±8.96)
Range	25–79
Sex
Male	104	50
Female	104	50
Clinical stage at time of lung cancer diagnosis
IA	6	2.9
IB	24	11.5
IIA	12	5.8
IIB	8	3.8
IIIA	26	12.5
IIIB	15	7.2
IV (brain)	110	52.9
Unknown	7	3.4
Location of brain metastasis
Solitary	141	67.8
Supratentorial	112	79.4
Infratentorial	29	20.6
Multiple	67	32.2
Management of primary tumor
Surgical resection (±chemotherapy)	76	36.5
Chemotherapy	95	45.7
Best supportive care	37	17.8
Steroid treatment before brain surgery (≤1 mo)
Yes	137	65.9
No	52	25
Unknown	19	9.1
Chemotherapy before brain surgery
Yes	75	36.1
No	132	63.4
Unknown	1	0.5
Cerebral irradiation
Whole brain radiotherapy	82	57.7
Radiosurgery	25	17.6
Both	35	24.6

Characteristic	Patients (N = 208)
Characteristic	No.	%
Age, y
Mean (±SD)	59.28 (±8.96)
Range	25–79
Sex
Male	104	50
Female	104	50
Clinical stage at time of lung cancer diagnosis
IA	6	2.9
IB	24	11.5
IIA	12	5.8
IIB	8	3.8
IIIA	26	12.5
IIIB	15	7.2
IV (brain)	110	52.9
Unknown	7	3.4
Location of brain metastasis
Solitary	141	67.8
Supratentorial	112	79.4
Infratentorial	29	20.6
Multiple	67	32.2
Management of primary tumor
Surgical resection (±chemotherapy)	76	36.5
Chemotherapy	95	45.7
Best supportive care	37	17.8
Steroid treatment before brain surgery (≤1 mo)
Yes	137	65.9
No	52	25
Unknown	19	9.1
Chemotherapy before brain surgery
Yes	75	36.1
No	132	63.4
Unknown	1	0.5
Cerebral irradiation
Whole brain radiotherapy	82	57.7
Radiosurgery	25	17.6
Both	35	24.6

Table 1

Clinical characteristics of lung adenocarcinoma patients with brain metastasis included in this study

Characteristic	Patients (N = 208)
Characteristic	No.	%
Age, y
Mean (±SD)	59.28 (±8.96)
Range	25–79
Sex
Male	104	50
Female	104	50
Clinical stage at time of lung cancer diagnosis
IA	6	2.9
IB	24	11.5
IIA	12	5.8
IIB	8	3.8
IIIA	26	12.5
IIIB	15	7.2
IV (brain)	110	52.9
Unknown	7	3.4
Location of brain metastasis
Solitary	141	67.8
Supratentorial	112	79.4
Infratentorial	29	20.6
Multiple	67	32.2
Management of primary tumor
Surgical resection (±chemotherapy)	76	36.5
Chemotherapy	95	45.7
Best supportive care	37	17.8
Steroid treatment before brain surgery (≤1 mo)
Yes	137	65.9
No	52	25
Unknown	19	9.1
Chemotherapy before brain surgery
Yes	75	36.1
No	132	63.4
Unknown	1	0.5
Cerebral irradiation
Whole brain radiotherapy	82	57.7
Radiosurgery	25	17.6
Both	35	24.6

Characteristic	Patients (N = 208)
Characteristic	No.	%
Age, y
Mean (±SD)	59.28 (±8.96)
Range	25–79
Sex
Male	104	50
Female	104	50
Clinical stage at time of lung cancer diagnosis
IA	6	2.9
IB	24	11.5
IIA	12	5.8
IIB	8	3.8
IIIA	26	12.5
IIIB	15	7.2
IV (brain)	110	52.9
Unknown	7	3.4
Location of brain metastasis
Solitary	141	67.8
Supratentorial	112	79.4
Infratentorial	29	20.6
Multiple	67	32.2
Management of primary tumor
Surgical resection (±chemotherapy)	76	36.5
Chemotherapy	95	45.7
Best supportive care	37	17.8
Steroid treatment before brain surgery (≤1 mo)
Yes	137	65.9
No	52	25
Unknown	19	9.1
Chemotherapy before brain surgery
Yes	75	36.1
No	132	63.4
Unknown	1	0.5
Cerebral irradiation
Whole brain radiotherapy	82	57.7
Radiosurgery	25	17.6
Both	35	24.6

Intratumoral Stromal and Peritumoral Inflammatory Cells

Of the cases, 68.7% had <20% and 31.3% had ≥20% intratumoral stromal ICs; 55.7% of cases presented with peritumoral mononuclear ring and 44.3% without. Taken together, 38% of the cases had no mononuclear ring and displayed very few or no stromal ICs.

Correlation of the Amount of Mononuclear Cells and PD-L1/PD-1 Expression

Patients divided into different groups using different thresholds of protein expressions were analyzed for the amount of stromal IC and the presence of mononuclear ring (Table 2).

Table 2

Correlation between the amount of mononuclear cells associated with brain metastasis of lung adenocarcinoma and PD-L1 and PD-1 expression

	Patient groups Based on Percentage of Positive Cells	Stromal Immune Cells (<20% vs >20%)		Mononuclear Ring (not present vs present)
	Patient groups Based on Percentage of Positive Cells	P-value	Pearson’s R	P-value	Pearson’s R
PD-L1 tumor cells	<1% vs ≥1%	.021	0.180	ns	—
	<5% vs ≥5%	.028	0.172	ns	—
	<50% vs ≥50%	.009	0.203	ns	—
PD-L1 immune cells	<1% vs ≥1%	<.001	0.315	.011	0.212
	<5% vs ≥5%	<.001	0.341	<.001	0.323
	<10% vs ≥10%	.013	0.196	.006	0.231
PD-1 immune cells	<1% vs ≥1%	.004	0.226	.003	0.242
	<5% vs ≥5%	ns	—	ns	—
	<10% vs ≥10%	ns	—	.044	0.166
Groups used in the POPLAR study
PD-L1 tumor cells or immune cells	Non-expressors vs any-expressors	<.001	0.302	.026	0.185
	Lower-expressors vs high/medium-expressors	<.001	0.338	.001	0.275
	Lower-expressors vs high-expressors	<.001	0.361	<.001	0.292

	Patient groups Based on Percentage of Positive Cells	Stromal Immune Cells (<20% vs >20%)		Mononuclear Ring (not present vs present)
	Patient groups Based on Percentage of Positive Cells	P-value	Pearson’s R	P-value	Pearson’s R
PD-L1 tumor cells	<1% vs ≥1%	.021	0.180	ns	—
	<5% vs ≥5%	.028	0.172	ns	—
	<50% vs ≥50%	.009	0.203	ns	—
PD-L1 immune cells	<1% vs ≥1%	<.001	0.315	.011	0.212
	<5% vs ≥5%	<.001	0.341	<.001	0.323
	<10% vs ≥10%	.013	0.196	.006	0.231
PD-1 immune cells	<1% vs ≥1%	.004	0.226	.003	0.242
	<5% vs ≥5%	ns	—	ns	—
	<10% vs ≥10%	ns	—	.044	0.166
Groups used in the POPLAR study
PD-L1 tumor cells or immune cells	Non-expressors vs any-expressors	<.001	0.302	.026	0.185
	Lower-expressors vs high/medium-expressors	<.001	0.338	.001	0.275
	Lower-expressors vs high-expressors	<.001	0.361	<.001	0.292

Table 2

Correlation between the amount of mononuclear cells associated with brain metastasis of lung adenocarcinoma and PD-L1 and PD-1 expression

	Patient groups Based on Percentage of Positive Cells	Stromal Immune Cells (<20% vs >20%)		Mononuclear Ring (not present vs present)
	Patient groups Based on Percentage of Positive Cells	P-value	Pearson’s R	P-value	Pearson’s R
PD-L1 tumor cells	<1% vs ≥1%	.021	0.180	ns	—
	<5% vs ≥5%	.028	0.172	ns	—
	<50% vs ≥50%	.009	0.203	ns	—
PD-L1 immune cells	<1% vs ≥1%	<.001	0.315	.011	0.212
	<5% vs ≥5%	<.001	0.341	<.001	0.323
	<10% vs ≥10%	.013	0.196	.006	0.231
PD-1 immune cells	<1% vs ≥1%	.004	0.226	.003	0.242
	<5% vs ≥5%	ns	—	ns	—
	<10% vs ≥10%	ns	—	.044	0.166
Groups used in the POPLAR study
PD-L1 tumor cells or immune cells	Non-expressors vs any-expressors	<.001	0.302	.026	0.185
	Lower-expressors vs high/medium-expressors	<.001	0.338	.001	0.275
	Lower-expressors vs high-expressors	<.001	0.361	<.001	0.292

	Patient groups Based on Percentage of Positive Cells	Stromal Immune Cells (<20% vs >20%)		Mononuclear Ring (not present vs present)
	Patient groups Based on Percentage of Positive Cells	P-value	Pearson’s R	P-value	Pearson’s R
PD-L1 tumor cells	<1% vs ≥1%	.021	0.180	ns	—
	<5% vs ≥5%	.028	0.172	ns	—
	<50% vs ≥50%	.009	0.203	ns	—
PD-L1 immune cells	<1% vs ≥1%	<.001	0.315	.011	0.212
	<5% vs ≥5%	<.001	0.341	<.001	0.323
	<10% vs ≥10%	.013	0.196	.006	0.231
PD-1 immune cells	<1% vs ≥1%	.004	0.226	.003	0.242
	<5% vs ≥5%	ns	—	ns	—
	<10% vs ≥10%	ns	—	.044	0.166
Groups used in the POPLAR study
PD-L1 tumor cells or immune cells	Non-expressors vs any-expressors	<.001	0.302	.026	0.185
	Lower-expressors vs high/medium-expressors	<.001	0.338	.001	0.275
	Lower-expressors vs high-expressors	<.001	0.361	<.001	0.292

There was a positive correlation between the amount of stromal ICs and the presence of mononuclear ring (P = .001; chi-square, Pearson’s R = 0.324).

Sparse stromal ICs in a brain metastasis sample correlated with low levels of PD-L1 positivity of TCs at any cutoff level (P = .021, .028 and .009 and Pearson’s R = 0.180, 0.172, and 0.203 at 1%, 5%, and 50% cutoff, respectively) (Table 2).

Regarding PD-L1 positivity of ICs, ≥20% stromal ICs correlated with higher levels of PD-L1 positive ICs for every cutoff level (P < .001, P < .001, and P = .013 with Pearson’s R = 0.315, 0.341, and 0.196 at 1%, 5%, and 10% cutoff, respectively). In addition, the presence of mononuclear ring correlated with more PD-L1+ ICs at every cutoff level as well (P = .011, P < .001, and P = .006 with Pearson’s R = 0.212, 0.323, and 0.231, respectively) (Table 2).

More than 1% PD-1+ ICs showed significant correlation with ≥20% of stromal ICs (P = .004, Pearson’s R = 0.226) and the presence of mononuclear ring (P = .003, Pearson’s R = 0.242) (Table 2).

Regarding the groups used in the POPLAR study, “higher PD-L1 expressors” were associated with the presence of mononuclear ring and ≥20% stromal ICs of brain metastases compared with “lower PD-L1 expressors” (or “non-expressors”) (Table 2).

Interestingly, positive correlations were found between the amount of PD-1+ ICs and PD-L1+ TCs and ICs using any cutoff levels, except for 10% cutoff for PD-1+ ICs and 1% cutoff for PD-L1+ ICs (Table 3). Moreover, “higher PD-L1 expressors” according to the POPLAR study were associated with more PD-1+ ICs at any cutoff level (Table 3).

Table 3

Correlation between PD-L1 and PD-1 expression in brain metastasis of lung adenocarcinoma

	Percentage of Positive Cells	PD-1 Immune Cells
		<1% vs ≥1%		<5% vs ≥5%		<10% vs ≥10%
		P-value	Pearson’s R	P-value	Pearson’s R	P-value	Pearson’s R
PD-L1 tumor cells	<1% vs ≥1%	.024	0.179	<.001	0.300	<.001	0.276
	<5% vs ≥5%	.040	0.163	<.001	0.314	<.001	0.282
	<50% vs ≥50%	.023	0.179	<.001	0.291	<.001	0.311
PD-L1 immune cells	<1% vs ≥1%	<.001	0.367	.005	0.224	ns	—
	<5% vs ≥5%	<.001	0.398	.001	0.258	.037	0.167
	<10% vs ≥10%	<.001	0.308	<.001	0.285	.011	0.203
Groups used in the POPLAR study
PD-L1 tumor cells or immune cells	Non-expressors vs any-expressors	<.001	0.343	.002	0.248	.008	0.210
	Lower-expressors vs high/ medium- expressors	<.001	0.404	<.001	0.285	<.001	0.279
	Lower-expressors vs high-expressors	<.001	0.355	.006	0.218	.001	0.269

	Percentage of Positive Cells	PD-1 Immune Cells
		<1% vs ≥1%		<5% vs ≥5%		<10% vs ≥10%
		P-value	Pearson’s R	P-value	Pearson’s R	P-value	Pearson’s R
PD-L1 tumor cells	<1% vs ≥1%	.024	0.179	<.001	0.300	<.001	0.276
	<5% vs ≥5%	.040	0.163	<.001	0.314	<.001	0.282
	<50% vs ≥50%	.023	0.179	<.001	0.291	<.001	0.311
PD-L1 immune cells	<1% vs ≥1%	<.001	0.367	.005	0.224	ns	—
	<5% vs ≥5%	<.001	0.398	.001	0.258	.037	0.167
	<10% vs ≥10%	<.001	0.308	<.001	0.285	.011	0.203
Groups used in the POPLAR study
PD-L1 tumor cells or immune cells	Non-expressors vs any-expressors	<.001	0.343	.002	0.248	.008	0.210
	Lower-expressors vs high/ medium- expressors	<.001	0.404	<.001	0.285	<.001	0.279
	Lower-expressors vs high-expressors	<.001	0.355	.006	0.218	.001	0.269

Table 3

Correlation between PD-L1 and PD-1 expression in brain metastasis of lung adenocarcinoma

	Percentage of Positive Cells	PD-1 Immune Cells
		<1% vs ≥1%		<5% vs ≥5%		<10% vs ≥10%
		P-value	Pearson’s R	P-value	Pearson’s R	P-value	Pearson’s R
PD-L1 tumor cells	<1% vs ≥1%	.024	0.179	<.001	0.300	<.001	0.276
	<5% vs ≥5%	.040	0.163	<.001	0.314	<.001	0.282
	<50% vs ≥50%	.023	0.179	<.001	0.291	<.001	0.311
PD-L1 immune cells	<1% vs ≥1%	<.001	0.367	.005	0.224	ns	—
	<5% vs ≥5%	<.001	0.398	.001	0.258	.037	0.167
	<10% vs ≥10%	<.001	0.308	<.001	0.285	.011	0.203
Groups used in the POPLAR study
PD-L1 tumor cells or immune cells	Non-expressors vs any-expressors	<.001	0.343	.002	0.248	.008	0.210
	Lower-expressors vs high/ medium- expressors	<.001	0.404	<.001	0.285	<.001	0.279
	Lower-expressors vs high-expressors	<.001	0.355	.006	0.218	.001	0.269

	Percentage of Positive Cells	PD-1 Immune Cells
		<1% vs ≥1%		<5% vs ≥5%		<10% vs ≥10%
		P-value	Pearson’s R	P-value	Pearson’s R	P-value	Pearson’s R
PD-L1 tumor cells	<1% vs ≥1%	.024	0.179	<.001	0.300	<.001	0.276
	<5% vs ≥5%	.040	0.163	<.001	0.314	<.001	0.282
	<50% vs ≥50%	.023	0.179	<.001	0.291	<.001	0.311
PD-L1 immune cells	<1% vs ≥1%	<.001	0.367	.005	0.224	ns	—
	<5% vs ≥5%	<.001	0.398	.001	0.258	.037	0.167
	<10% vs ≥10%	<.001	0.308	<.001	0.285	.011	0.203
Groups used in the POPLAR study
PD-L1 tumor cells or immune cells	Non-expressors vs any-expressors	<.001	0.343	.002	0.248	.008	0.210
	Lower-expressors vs high/ medium- expressors	<.001	0.404	<.001	0.285	<.001	0.279
	Lower-expressors vs high-expressors	<.001	0.355	.006	0.218	.001	0.269

Correlation of Mononuclear Ring or Stromal ICs and Clinical Characteristics

Importantly, the proportion of patients with abundant stromal ICs or mononuclear ring was similar with or without receiving chemotherapy or steroid before brain metastasis surgery (P = .997 and P = .452; P = .186 and P = .879, respectively; chi-square).

Correlation of the Expression of PD-L1 and PD-1 and Clinical Characteristics

Steroid administration significantly correlated with less PD-L1 expression of ICs at 10% cutoff value compared with patients without steroid use (P = .021; chi-square, Pearson’s R = 0.208); however, no correlation was revealed with PD-L1 expression of TCs and PD-1 expression. No difference was detected in PD-L1 or PD-1 expression comparing patients having received chemotherapy or not before brain metastasis surgery (data not shown).

Correlation of Clinical Characteristics with Survival

Kaplan–Meier survival curves of different variable values were tested for significant differences with log-rank tests. Following the general steps of model selection described in Supplementary File 1, the final set of variables included in the Cox proportional hazards model of the whole cohort for OS and survival from brain metastasis surgery prediction are listed in Table 4. Both results of univariate analyses and the final multivariate models are presented below.

Table 4

Multivariate proportional hazards Cox models for OS and survival from brain surgery in the investigated patient groups

Variable	P	HR	CI
Whole cohort, OS
Later stages	.282	1.675	0.655–4.284
Surgery	.004	0.318	0.145–0.701
Lack of chemotherapy	<.001	3.139	1.715–5.745
Age <60 y	.032	0.554	0.323–0.949
Solitary metastasis	.174	0.796	0.573–1.106
Lack of MR	.089	1.635	0.928–2.881
No expressor (POPLAR)	.546	0.816	0.421–1.581
PD-1 immune cells >1%	.840	0.930	0.456–1.893
Supratentorial localization	.440	1.267	0.695–2.313
Whole cohort, survival from brain surgery
Later stages	.161	1.595	0.831–3.062
Lack of chemotherapy	.224	1.430	0.804–2.543
Solitary metastasis	.010	0.656	0.476–0.905
Lack of MR	.050	1.731	0.578–2.994
Medium- and high-expressor (POPLAR)	.102	1.604	0.910–2.825
PD-1 immune cells >5%	.840	0.941	0.524–1.691
PD-L1 tumor cells >50%	.090	0.562	0.288–1.095
Supratentorial localization	.357	0.739	0.388–1.407
Subgroup 1, OS
Later stages	.375	1.634	0.552–4.837
Lack of chemotherapy	.999	2.717	0.829–8.901
Lack of MR	.027	3.054	1.137–8.202
No expressor (POPLAR)	.731	0.829	0.285–2.414
PD-1 immune cells >1%	.526	0.722	0.264–1.974

Variable	P	HR	CI
Whole cohort, OS
Later stages	.282	1.675	0.655–4.284
Surgery	.004	0.318	0.145–0.701
Lack of chemotherapy	<.001	3.139	1.715–5.745
Age <60 y	.032	0.554	0.323–0.949
Solitary metastasis	.174	0.796	0.573–1.106
Lack of MR	.089	1.635	0.928–2.881
No expressor (POPLAR)	.546	0.816	0.421–1.581
PD-1 immune cells >1%	.840	0.930	0.456–1.893
Supratentorial localization	.440	1.267	0.695–2.313
Whole cohort, survival from brain surgery
Later stages	.161	1.595	0.831–3.062
Lack of chemotherapy	.224	1.430	0.804–2.543
Solitary metastasis	.010	0.656	0.476–0.905
Lack of MR	.050	1.731	0.578–2.994
Medium- and high-expressor (POPLAR)	.102	1.604	0.910–2.825
PD-1 immune cells >5%	.840	0.941	0.524–1.691
PD-L1 tumor cells >50%	.090	0.562	0.288–1.095
Supratentorial localization	.357	0.739	0.388–1.407
Subgroup 1, OS
Later stages	.375	1.634	0.552–4.837
Lack of chemotherapy	.999	2.717	0.829–8.901
Lack of MR	.027	3.054	1.137–8.202
No expressor (POPLAR)	.731	0.829	0.285–2.414
PD-1 immune cells >1%	.526	0.722	0.264–1.974

Table 4

Open in new tab Download slide

Multivariate proportional hazards Cox models for OS and survival from brain surgery in the investigated patient groups

Variable	P	HR	CI
Whole cohort, OS
Later stages	.282	1.675	0.655–4.284
Surgery	.004	0.318	0.145–0.701
Lack of chemotherapy	<.001	3.139	1.715–5.745
Age <60 y	.032	0.554	0.323–0.949
Solitary metastasis	.174	0.796	0.573–1.106
Lack of MR	.089	1.635	0.928–2.881
No expressor (POPLAR)	.546	0.816	0.421–1.581
PD-1 immune cells >1%	.840	0.930	0.456–1.893
Supratentorial localization	.440	1.267	0.695–2.313
Whole cohort, survival from brain surgery
Later stages	.161	1.595	0.831–3.062
Lack of chemotherapy	.224	1.430	0.804–2.543
Solitary metastasis	.010	0.656	0.476–0.905
Lack of MR	.050	1.731	0.578–2.994
Medium- and high-expressor (POPLAR)	.102	1.604	0.910–2.825
PD-1 immune cells >5%	.840	0.941	0.524–1.691
PD-L1 tumor cells >50%	.090	0.562	0.288–1.095
Supratentorial localization	.357	0.739	0.388–1.407
Subgroup 1, OS
Later stages	.375	1.634	0.552–4.837
Lack of chemotherapy	.999	2.717	0.829–8.901
Lack of MR	.027	3.054	1.137–8.202
No expressor (POPLAR)	.731	0.829	0.285–2.414
PD-1 immune cells >1%	.526	0.722	0.264–1.974

Variable	P	HR	CI
Whole cohort, OS
Later stages	.282	1.675	0.655–4.284
Surgery	.004	0.318	0.145–0.701
Lack of chemotherapy	<.001	3.139	1.715–5.745
Age <60 y	.032	0.554	0.323–0.949
Solitary metastasis	.174	0.796	0.573–1.106
Lack of MR	.089	1.635	0.928–2.881
No expressor (POPLAR)	.546	0.816	0.421–1.581
PD-1 immune cells >1%	.840	0.930	0.456–1.893
Supratentorial localization	.440	1.267	0.695–2.313
Whole cohort, survival from brain surgery
Later stages	.161	1.595	0.831–3.062
Lack of chemotherapy	.224	1.430	0.804–2.543
Solitary metastasis	.010	0.656	0.476–0.905
Lack of MR	.050	1.731	0.578–2.994
Medium- and high-expressor (POPLAR)	.102	1.604	0.910–2.825
PD-1 immune cells >5%	.840	0.941	0.524–1.691
PD-L1 tumor cells >50%	.090	0.562	0.288–1.095
Supratentorial localization	.357	0.739	0.388–1.407
Subgroup 1, OS
Later stages	.375	1.634	0.552–4.837
Lack of chemotherapy	.999	2.717	0.829–8.901
Lack of MR	.027	3.054	1.137–8.202
No expressor (POPLAR)	.731	0.829	0.285–2.414
PD-1 immune cells >1%	.526	0.722	0.264–1.974

Early clinical stages (I–II) at the diagnosis of LUAD, surgical resection of the primary tumor, and chemotherapy showed a significant association with better OS with the univariate log-rank test (P < .001, P < .001, P = .027, respectively). Multivariate Cox regression analysis identified both surgery and the lack of chemotherapy as independent prognostic factors (P = .004, hazard ratio [HR] = 0.318, CI = 0.145–0.701; P < .001, HR = 3.139, CI = 1.715–5.745, respectively). These parameters did not significantly correlate with survival after brain surgery by log-rank test (P = .064, P = .063, P = .151, respectively), and Cox regression analysis only revealed a tendency toward worse survival from brain surgery for patients in advanced clinical stages (Table 4).

Patients with solitary brain metastasis showed better survival after brain metastasis surgery compared with patients with multiple brain metastases (P = .012; log-rank), which remained significant with multivariate Cox regression analysis (P = .010, HR = 0.656, CI = 0.476–0.905). Regarding OS, however, no significant difference was observed (P = .095; log-rank; and P = .174, HR = 0.796, CI = 0.573–1.106; Cox).

Patients also showed better OS and survival after brain metastasis surgery in cases of supratentorial solitary metastasis compared with infratentorial solitary metastasis (P = .012 and .001, respectively; log-rank). With multivariate Cox regression analysis, however, the location of the brain metastasis did not appear to be an independent prognostic factor.

Younger patients (≤60 y) showed better OS and survival after brain metastasis surgery (P = .029 and P = .019, respectively; log-rank). In the case of OS, age remained significant with Cox analysis (P = .032, HR = 0.554, CI = 0.323–0.949). Gender had no influence on OS or survival after brain metastasis surgery (data not shown).

Correlation of the Amount of Mononuclear Cells with Survival

The lack of peritumoral mononuclear ring was associated with worse survival after brain metastasis surgery (P = .01; log-rank) (Fig. 1), and a borderline significant tendency toward worse survival remained with multivariate Cox regression analysis (P = .050, HR = 1.731, CI = 0.578–2.994). Cases with ≥20% of stromal ICs only showed a tendency for better OS (P = .066; log-rank).

Fig. 1

Kaplan–Meier analysis. The presence of mononuclear ring (MR) around the brain metastasis of lung adenocarcinoma is associated with better survival after brain metastasis surgery. The presence of MR is also associated with a tendency for better OS in patients with solitary brain metastasis who underwent primary tumor resection (subgroup 1).

Due to the strong influence of surgical resection of the primary tumor, chemotherapy, and multiplicity of brain metastasis on survival, we also performed multivariate Cox regression analysis on the clinically more homogeneous 4 subgroups described in the Materials and Methods section. In subgroup 1 variables POPLAR “no versus any expressor” category, PD-1 IC “<1% versus ≥1%” category, presence of MR, stage category, and chemotherapy were included in the Cox analysis for OS (Table 4). The lack of mononuclear rings correlated with worse OS with a multivariate model (P = .027, HR = 3.054, CI = 1.137–8.202), while Kaplan–Meier analysis only showed a tendency for better OS (P = .060) (Fig. 1). No significant association was found in the other subgroups.

Correlation of PD-L1 and PD-1 Expression with Survival

No correlation was observed by log-rank test between PD-L1 expression of TCs or ICs, and PD-1 expression of ICs at any cutoff level and OS or survival after brain metastasis surgery (data not shown). Similarly, no correlation was detected with OS or survival after brain metastasis surgery in the groups used in the POPLAR study, and the results remained nonsignificant with multivariate Cox regression analysis (data not shown).

Performing our analysis separately on the previously described subgroups, we found the following correlations for patients who underwent removal of the primary tumor and had multiple brain metastases (subgroup 2). Less than 10% PD-1 expression of ICs was associated with a tendency toward better OS with Kaplan–Meier analysis (P = .055) (Supplementary Figure S2); however, the variable did not meet the proportional hazards assumption criteria. In addition, lower PD-L1 expression of TCs at any cutoff level in subgroup 2 was associated with better survival after brain surgery, but Kaplan–Meier analyses revealed significantly better post brain metastasis surgery survival only by using a 1% cutoff level (P = .023) (Supplementary Figure S2). The analysis of different PD-L1 expressor groups used in the POPLAR study showed better survival after brain metastasis surgery in the “non-expressors” compared with the “any-expressors” (P = .008; log-rank) (Fig. 2). Kaplan–Meier analyses also showed better OS in the “lower-expressors” compared with the “high-expressors” (P = .046) (Fig. 2).

Fig. 2

Kaplan–Meier analysis of the different PD-L1 expressor groups used in the POPLAR study. Among patients who underwent primary tumor resection and had multiple brain metastases (subgroup 2), “non-expressors” showed better survival after brain metastasis surgery compared with the “any-expressors”; and “lower-expressors” showed better OS compared with the “high-expressors.”

Open in new tab Download slide

Discussion

In this study we presented a comprehensive analysis of immune cell infiltration, PD-1 expression of immune cells, and PD-L1 expression of tumor cells and immune cells and their clinical relevance on a large cohort of lung adenocarcinoma patients with brain metastases. The importance of analyzing LUAD and squamous cell carcinoma separately in this context is emphasized by results that showed that PD-L1 expression has a significant effect on survival limited to the adenocarcinoma only, which was not present in squamous cell carcinoma.⁴³

This is the first study demonstrating that mononuclear cell infiltration, especially in peritumoral location, is associated with better survival in LUAD brain metastasis, therefore reactivation of the immune response by immune checkpoint inhibitors will likely add further clinical benefit. We also showed, however, that a significant portion of LUAD brain metastasis is without IC infiltrates, suggesting that a significant number of patients lack any significant local immune response in association with their brain metastasis. While several prior publications on brain metastasis of mixed primary origin reported a higher proportion of brain metastasis with ICs present, our results suggest that this question should be reevaluated on more homogeneous cohorts. Agents acting on the PD-L1/PD-1 axis will be unlikely to recruit immune cells into the metastases⁷ and thus will probably not provide clinical benefit, while other agents, such as cytotoxic T lymphocyte antigen (CTLA-4) inhibitors, increase T-cell recruitment into tumors. Therefore, the relatively high number of cases without IC infiltrates into brain metastasis of LUAD may suggest that combination of PD-L1 and CTLA-4 inhibitors may be a better choice of treatment.

Even if TILs are present in primary tumors or brain metastases, immune checkpoint pathways are regularly targeted by tumors to escape immune-mediated destruction, therefore the mere presence of TILs may not always be sufficient to ensure better survival.²³ Since tumor cells often inactivate immune cells by activation of the PD-1 pathway, we decided to evaluate the prognostic value of PD-1 and PD-L1 expression in brain metastasis of LUAD patients. It is generally accepted that in tumors where the PD-1 activation dampens the otherwise potentially active antitumor immune response, and therefore portends worse prognosis, anti–PD-1/PD-L1 therapy might be rather effective by reactivating the already present therapeutic response. In light of this, it is important to point out that PD-1 or PD-L1 expression had no prognostic power in the entire cohort, therefore anti–PD-1/PD-L1 therapy alone may have limited clinical value in brain metastasis LUAD patients in general. Considering the significant impact of removal of the primary tumor and solitary versus multiple brain metastasis on survival, we subdivided the patients into 4 different groups based on these criteria and reanalyzed the prognostic value of PD-1/PD-L1 expression in those subcohorts separately as well. In subgroup 2, where the primary tumor was removed and the patients had multiple brain metastases, low expression of PD-1 and PD-L1 was associated with significantly better survival. This suggests that in these cases, activating the PD-1/PD-L1 axis may actively suppress antitumor immune suppression, therefore these cases may benefit the most from treatments containing checkpoint inhibitors targeting this regulatory pathway.

A positive correlation could be demonstrated between the amount of stromal ICs and the presence of mononuclear ring, which suggests a potentially similar mechanism for attracting ICs to different areas of brain metastasis.

We also found positive correlations between the amount of PD-1+ ICs and PD-L1+ TCs and ICs, arguing for a dynamic communication among tumor cells and immune cells and supporting the concept that exhausted PD-1 expressing T cells would induce PD-L1 expression in the microenvironment.⁴⁴

Considering the increasing importance of immune response in cancer therapy, it is important to clarify the impact of steroid and chemotherapy treatment on the status of the ICs in brain metastasis. Steroid administration before brain metastasis surgery had no significant influence either on stromal immune cell infiltration or on peritumoral mononuclear rings in our LUAD cases, which is concordant with earlier findings on cohorts with mixed histology,^15–17^,³² indicating that steroids have no major effect on IC recruitment. Likewise, chemotherapy prior to brain metastasis surgery appears not to affect IC recruitment. In recent studies steroid was found to have no influence on PD-L1 expression in brain metastasis associated with various types of primary tumors.¹⁷^,³² In our cohort, steroid administration before brain metastasis surgery significantly correlated with less PD-L1 expression of ICs at 10% cutoff, suggesting that steroid may have some impact on PD-L1 expression of ICs in brain metastasis LUAD; however, it did not affect PD-L1 expression of TCs and PD-1 expression of ICs. The effect of chemotherapy on PD-1/PD-L1 has been investigated in vitro with contrary results depending on the agent and cell type used.^45–47 Interestingly, our brain metastasis cases with or without prior chemotherapy showed similar PD-L1 and PD-1 expressions, suggesting that systemic chemotherapy may not lead to PD-L1/PD-1 expression change in LUAD brain metastasis.

We have analyzed our data with the scoring system suggested by the POPLAR study,⁴² because their PD-L1 biomarker selection measuring the protein on both TC and IC could be biologically more relevant. In our LUAD brain metastasis cohort, 26.4% of the cases were “non-expressors” (see Supplementary File 2), in which neither TCs nor ICs expressed PD-L1. According to the results of the POPLAR study, these patients are unlikely to benefit from anti–PD-L1 therapy. In contrast, 41.7% of our cases were “high-expressors,” who probably could benefit from anti–PD-L1 therapy. According to our findings, “higher PD-L1 expressors” showed significant association with the presence of mononuclear ring and ≥20% stromal ICs compared with the remaining “lower-” or “non-expressors.” In addition, “higher PD-L1 expressors” were characterized by more PD-1+ ICs. No association was found with OS or survival after brain metastasis surgery in the different PD-L1 expressor groups in the whole cohort; however, better survival after brain metastasis surgery was seen in the “non-expressors” compared with the “any-expressors” in subgroup 2. Moreover, Kaplan–Meier analyses also showed better OS in the “lower-expressors” compared with the “high-expressors” in subgroup 2.

In summary, this is the first study demonstrating the independent prognostic value of peritumoral mononuclear cells in LUAD brain metastasis cases. Besides, massive stromal ICs in brain metastasis also showed a strong tendency for favorable prognosis. Steroid may have some impact on reducing PD-L1 expression of ICs without affecting their recruitment. Chemotherapy appears not to disturb either the density of ICs or PD-1/PD-L1 expression in LUAD brain metastasis.

Our findings in the largest cohort of LUAD brain metastasis cases indicate a significant impact of the host immune system on controlling brain metastasis, thus modern immunotherapies could support the ICs in tumor destruction in a subset of patients. However, appropriate patient selection, especially in LUAD cases, for a potentially successful immunotherapy is still debated; moreover, consensus criteria for testing PD-1 and/or PD-L1 as predictive biomarkers have yet to be settled.

Supplementary Material

Supplementary data are available at Neuro-Oncology online.

Funding

This work was supported by the Research and Technology Innovation Fund (KTIA_NAP_13-2014-0021 to L.R., L.V., Z.S., SNN114490 to B.D.); Hungarian Science Foundation (OTKA-PD115792 to L.R., OTKA-K116151 to L.R., J.T., L.V., B.D., OTKA-K112371 to J.T., OTKA-K109626 to B.D., OTKA-K108465 to B.D.); Semmelweis University Start-Up grant (40148-11658 to B.D.), and the Vienna Fund for Innovative Interdisciplinary Cancer Research (to B.D.); Breast Cancer Research Foundation and the Novo Nordisk Foundation Interdisciplinary Synergy Programme Grant (NNF15OC0016584 to Z.S.). C.S. is a senior Cancer Research UK clinical research fellow. Support also came from the Francis Crick Institute, which receives its core funding from Cancer Research UK (FC001169), the UK Medical Research Council (FC001169), and the Wellcome Trust (FC001169); by the UK Medical Research Council (grant reference MR/FC001169/1); C.S. is funded by Cancer Research UK (TRACERx), the CRUK Lung Cancer Centre of Excellence, Stand Up 2 Cancer (SU2C), the Rosetrees Trust, Novo Nordisk Foundation (ID16584), EU-FP7 (projects PREDICT and RESPONSIFY, ID:259303), the Prostate Cancer Foundation, the Breast Cancer Research Foundation, the European Research Council (THESEUS). Support was provided to C.S. by the National Institute for Health Research, the University College London Hospitals Biomedical Research Centre, and the Cancer Research UK University College London Experimental Cancer Medicine Centre.

Conflict of interest statement. None declared.

Acknowledgment

We thank Anna Tamási, Mónika Szilágyiné Paulusz, and Zsuzsanna Kaminszky for excellent technical assistance.

References

1

Dela Cruz

CS

Tanoue

LT

Matthay

RA

.

Lung cancer: epidemiology, etiology, and prevention

.

Clin Chest Med

.

2011

;

32

(

4

):

605

–

644

.

2

Barnholtz-Sloan

JS

Sloan

AE

Davis

FG

Vigneau

FD

Lai

P

Sawaya

RE

.

Incidence proportions of brain metastases in patients diagnosed (1973 to 2001) in the Metropolitan Detroit Cancer Surveillance System

.

J Clin Oncol

.

2004

;

22

(

14

):

2865

–

2872

.

3

Peters

S

Bexelius

C

Munk

V

Leighl

N

.

The impact of brain metastasis on quality of life, resource utilization and survival in patients with non-small-cell lung cancer

.

Cancer Treat Rev

.

2016

;

45

:

139

–

162

.

4

Brahmer

JR

Tykodi

SS

Chow

LQ

et al.

Safety and activity of anti-PD-L1 antibody in patients with advanced cancer

.

N Engl J Med

.

2012

;

366

(

26

):

2455

–

2465

.

5

Topalian

SL

Hodi

FS

Brahmer

JR

et al.

Safety, activity, and immune correlates of anti-PD-1 antibody in cancer

.

N Engl J Med

.

2012

;

366

(

26

):

2443

–

2454

.

6

Herbst

RS

Soria

JC

Kowanetz

M

et al.

Predictive correlates of response to the anti-PD-L1 antibody MPDL3280A in cancer patients

.

Nature

.

2014

;

515

(

7528

):

563

–

567

.

7

Antonia

SJ

Vansteenkiste

JF

Moon

E

.

Immunotherapy: beyond anti-PD-1 and anti-PD-L1 therapies

.

Am Soc Clin Oncol Educ Book

.

2016

;

35

:

e450

–

e458

.

8

Rozek

LS

Schmit

SL

Greenson

JK

et al.

Tumor-infiltrating lymphocytes, Crohn’s-like lymphoid reaction, and survival from colorectal cancer

.

J Natl Cancer Inst

.

2016

;

108

(

8

).

9

Lee

N

Zakka

LR

Mihm

MC

Jr

Schatton

T

.

Tumour-infiltrating lymphocytes in melanoma prognosis and cancer immunotherapy

.

Pathology

.

2016

;

48

(

2

):

177

–

187

.

10

Nguyen

N

Bellile

E

Thomas

D

et al.

Tumor infiltrating lymphocytes and survival in patients with head and neck squamous cell carcinoma

.

Head Neck

.

2016

;

38

(

7

):

1074

–

1084

.

11

Zeng

DQ

Yu

YF

Ou

QY

et al.

Prognostic and predictive value of tumor-infiltrating lymphocytes for clinical therapeutic research in patients with non-small cell lung cancer

.

Oncotarget

.

2016

;

7

(

12

):

13765

–

13781

.

12

Gooden

MJ

de Bock

GH

Leffers

N

Daemen

T

Nijman

HW

.

The prognostic influence of tumour-infiltrating lymphocytes in cancer: a systematic review with meta-analysis

.

Br J Cancer

.

2011

;

105

(

1

):

93

–

103

.

13

Ingold Heppner

B

Loibl

S

Denkert

C

.

Tumor-infiltrating lymphocytes: a promising biomarker in breast cancer

.

Breast Care (Basel)

.

2016

;

11

(

2

):

96

–

100

.

14

Ameratunga

M

Asadi

K

Lin

X

et al.

PD-L1 and tumor infiltrating lymphocytes as prognostic markers in resected NSCLC

.

PLoS One

.

2016

;

11

(

4

):

e0153954

.

15

Berghoff

AS

Lassmann

H

Preusser

M

Höftberger

R

.

Characterization of the inflammatory response to solid cancer metastases in the human brain

.

Clin Exp Metastasis

.

2013

;

30

(

1

):

69

–

81

.

16

Berghoff

AS

Fuchs

E

Ricken

G

et al.

Density of tumor-infiltrating lymphocytes correlates with extent of brain edema and overall survival time in patients with brain metastases

.

Oncoimmunology

.

2016

;

5

(

1

):

e1057388

.

17

Harter

PN

Bernatz

S

Scholz

A

et al.

Distribution and prognostic relevance of tumor-infiltrating lymphocytes (TILs) and PD-1/PD-L1 immune checkpoints in human brain metastases

.

Oncotarget

.

2015

;

6

(

38

):

40836

–

40849

.

18

Berghoff

AS

Inan

C

Ricken

G

et al.

Tumor-infiltrating lymphocytes (TILs) and PD-L1 expression in non-small cell lung cancer brain metastases (BM) and matched primary tumors (PT)

.

Annals of Oncology

.

2014

;

25

(

suppl 4

):

iv465

–

iv466

.

Crossref

19

Dong

H

Chen

L

.

B7-H1 pathway and its role in the evasion of tumor immunity

.

J Mol Med (Berl)

.

2003

;

81

(

5

):

281

–

287

.

20

Blank

C

Gajewski

TF

Mackensen

A

.

Interaction of PD-L1 on tumor cells with PD-1 on tumor-specific T cells as a mechanism of immune evasion: implications for tumor immunotherapy

.

Cancer Immunol Immunother

.

2005

;

54

(

4

):

307

–

314

.

21

Socinski

MA

.

Incorporating immunotherapy into the treatment of non-small cell lung cancer: practical guidance for the clinic

.

Semin Oncol

.

2015

;

42

(

Suppl 2

):

S19

–

S28

.

22

Melosky

B

Chu

Q

Juergens

R

Leighl

N

McLeod

D

Hirsh

V

.

Pointed progress in second-line advanced non-small-cell lung cancer: the rapidly evolving field of checkpoint inhibition

.

J Clin Oncol

.

2016

;

34

(

14

):

1676

–

1688

.

23

Velcheti

V

Schalper

K

.

Basic overview of current immunotherapy approaches in cancer

.

Am Soc Clin Oncol Educ Book

.

2016

;

35

:

298

–

308

.

24

Kluger

HM

Goldberg

SB

Sznol

M

et al.

Safety and activity of pembrolizumab in melanoma patients with untreated brain metastases

. [abstr 9009]

J Clin Oncol

.

2015

;

33

(

suppl

).

25

Goldberg

SB

Gettinger

SN

Mahajan

A

et al.

Activity and safety of pembrolizumab in patients with metastatic non-small cell lung cancer with untreated brain metastases

. [abstr 8035]

J Clin Oncol

.

2015

;

33

(

suppl

).

26

Shukuya

T

Yamanaka

T

Seto

T

et al.

Nedaplatin plus docetaxel versus cisplatin plus docetaxel for advanced or relapsed squamous cell carcinoma of the lung (WJOG5208L): a randomised, open-label, phase 3 trial

.

Lancet Oncol

.

2015

;

16

(

16

):

1630

–

1638

.

27

Kerr

KM

Nicolson

MC

.

Non-small cell lung cancer, PD-L1, and the pathologist

.

Arch Pathol Lab Med

.

2016

;

140

(

3

):

249

–

254

.

28

Yang

CY

Lin

MW

Chang

YL

Wu

CT

Yang

PC

.

Programmed cell death-ligand 1 expression is associated with a favourable immune microenvironment and better overall survival in stage I pulmonary squamous cell carcinoma

.

Eur J Cancer

.

2016

;

57

:

91

–

103

.

29

Inman

BA

Sebo

TJ

Frigola

X

et al.

PD-L1 (B7-H1) expression by urothelial carcinoma of the bladder and BCG-induced granulomata: associations with localized stage progression

.

Cancer

.

2007

;

109

(

8

):

1499

–

1505

.

30

Konishi

J

Yamazaki

K

Azuma

M

Kinoshita

I

Dosaka-Akita

H

Nishimura

M

.

B7-H1 expression on non-small cell lung cancer cells and its relationship with tumor-infiltrating lymphocytes and their PD-1 expression

.

Clin Cancer Res

.

2004

;

10

(

15

):

5094

–

5100

.

31

Nomi

T

Sho

M

Akahori

T

et al.

Clinical significance and therapeutic potential of the programmed death-1 ligand/programmed death-1 pathway in human pancreatic cancer

.

Clin Cancer Res

.

2007

;

13

(

7

):

2151

–

2157

.

32

Berghoff

AS

Ricken

G

Widhalm

G

et al.

Tumour-infiltrating lymphocytes and expression of programmed death ligand 1 (PD-L1) in melanoma brain metastases

.

Histopathology

.

2015

;

66

(

2

):

289

–

299

.

33

Duchnowska

R

Pęksa

R

Radecka

B

et al.

Immune response in breast cancer brain metastases and their microenvironment: the role of the PD-1/PD-L axis

.

Breast Cancer Res

.

2016

;

18

(

1

):

43

.

34

Luo

M

Fu

L

.

The effect of chemotherapy on programmed cell death 1/programmed cell death 1 ligand axis: some chemotherapeutical drugs may finally work through immune response

.

Oncotarget

.

2016

;

7

(

20

):

29794

–

29803

.

35

Al-Shibli

KI

Donnem

T

Al-Saad

S

Persson

M

Bremnes

RM

Busund

LT

.

Prognostic effect of epithelial and stromal lymphocyte infiltration in non-small cell lung cancer

.

Clin Cancer Res

.

2008

;

14

(

16

):

5220

–

5227

.

36

Bremnes

RM

Donnem

T

Busund

LT

.

Importance of tumor infiltrating lymphocytes in non-small cell lung cancer?

Ann Transl Med

.

2016

;

4

(

7

):

142

.

37

McLaughlin

J

Han

G

Schalper

KA

et al.

Quantitative assessment of the heterogeneity of PD-L1 expression in non-small-cell lung cancer

.

JAMA Oncol

.

2016

;

2

(

1

):

46

–

54

.

38

Passiglia

F

Bronte

G

Bazan

V

et al.

PD-L1 expression as predictive biomarker in patients with NSCLC: a pooled analysis

.

Oncotarget

.

2016

;

7

(

15

):

19738

–

19747

.

39

Travis

WD

Brambilla

E

Noguchi

M

et al.

International Association for the Study of Lung Cancer/American Thoracic Society/European Respiratory Society international multidisciplinary classification of lung adenocarcinoma

.

J Thorac Oncol

.

2011

;

6

(

2

):

244

–

285

.

40

Tian

T

Ruan

M

Yang

W

Shui

R

.

Evaluation of the prognostic value of tumor-infiltrating lymphocytes in triple-negative breast cancers

.

Oncotarget

.

2016

;

7

(

28

):

44395

–

44405

.

41

Festino

L

Botti

G

Lorigan

P

et al.

Cancer treatment with anti-PD-1/PD-L1 agents: is PD-L1 expression a biomarker for patient selection?

Drugs

.

2016

;

76

(

9

):

925

–

945

.

42

Spira

AI

Park

K

Maziéres

J

et al.

Efficacy, safety and predictive biomarker results from a randomized phase II study comparing MPDL3280A vs docetaxel in 2L/3L NSCLC (POPLAR)

. [abstr 8010]

J Clin Oncol

.

2015

;

33

(

suppl

).