Tumor Immune-Infiltrate Landscape After Chemo-Radiotherapy in a Case Series of Patients with Non-small Cell Lung Cancer: Pretreatment Predictors and Correlation With Outcome.

INTRODUCTION: Data on tumor immune-milieu after chemo-radiation (CT-RT) are scarce. Noninvasive tools are needed to improve the treatment of non-small cell lung cancer (NSCLC), especially in the locally advanced (LA) setting.
METHODS: We collected a series of superior-sulcus (SS)- patients with NSCLC referred to our Institute (2015-2019), eligible for a preoperative CT-RT. We characterized tumor-infiltrating immune cells (TIICs), determined PD-L1-TPS and the residual viable tumor cells (RVTC). Radiological and metabolic responses were reviewed. We calculated pre-surgery neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR).
RESULTS: Eight patients were included. Radiological responses were 6 disease stabilities (SD) and 2 partial responses (PR). Metabolic responses were 4 SD and 4 PR. CD68+-TIICs were correlated with metabolic response and lower RVTC. CD68+-TIICs were associated with higher PLR. Higher PLR values seemed linked with lower RVTC.
CONCLUSIONS: These preliminary results could be useful for consolidation treatment selection for patients with LA-NSCLC without evaluable baseline PD-L1 and higher PLR values.

Introduction

Tumor immune-infiltrate characterization of non–small cell lung cancer (NSCLC) has risen the interest in the latest years, for its prognostic role in patients stratification and as a predictor of treatment outcome with immune-checkpoint inhibitors (ICIs).

ICIs have changed lung cancer treatment, first in the advanced/metastatic setting, more recently for the locally advanced (LA) unresectable stage, as described in the PACIFIC trial.[1]

This trial brought to the fore the unresolved issue of predictive biomarkers in the decision-making of lung cancer care. In fact, in the PACIFIC trial, patients were enrolled regardless of tumor PD-L1 expression. European Medicines Agency (EMA) requested post hoc exploratory analysis of survival rates, using PD-L1 expression-level cutoff of 1%. Given the stronger benefit experienced by patients with PD-L1 positive, EMA approved durvalumab after chemo-radiotherapy (CT-RT) only for this subset of patients,[2] firing several controversies. The analysis was not preplanned, therefore not powered enough to produce statistically strong results. PD-L1 expression was not available in over 37% of cases and patients with an unknown PD-L1 status also seem to benefit from durvalumab. Finally, CT-RT could enhance immunogenicity even in “immune-cold” tumors,[3,4] leaving no biological basis to recommend durvalumab consolidation to selected cases. Theoretically, tumor re-biopsy during or after CT-RT might help redefine the sensitivity to ICI. However, a new biopsy is not always practically feasible, nor tolerable for patients.

Few data on immune-infiltrate and checkpoint expression after CT-RT are available. The primary aim of our exploratory analysis was to describe histological features of tumor specimens after CT-RT, to correlate them with the response, and to find new tools for the prediction of tumor immune-milieu, which may overlook tissue re-biopsy in LA-NSCLC.

We decided to focus on Superior-Sulcus(SS)-NSCLCs: this way we had the chance to analyze samples of patients treated with radical-intent CT-RT,[5] without the risk of tissue inadequacy, as it was obtained from surgical specimens.

Materials and Methods

We retrospectively collected a series of patients with SS-NSCLC referred to our Center from 2015 to 2019 and deemed eligible for a preoperative CT-RT. The backbone of chemotherapy was a carboplatin-based regimen plus a third-generation agent. Concurrent radiotherapy was administered 5 fractions per week, 200 cGy per fraction. Surgery was planned after 3 weeks since the conclusion of radiotherapy.

Pre- and post-treatment whole-body CT scans and FDG-PET-CT were performed. Radiologic and metabolic responses were evaluated according to RECIST 1.1 and EORTC criteria, respectively.

Tissue serial sections were stained using monoclonal antibodies anti-CD8, anti-CD3, anti-CD4, anti-CD68, and anti-PD-L1. Immunoreactivity for CD3-, CD4-, and CD8-lymphocytes, and for CD68-macrophages was expressed as a score 0-3.[6] The rate of residual viable tumor cells (RVTC), fibrosis, and necrosis were also determined. The pathological response was assessed according to the IASLC recommendations.[7]

Finally, we calculated the pre-surgery neutrophil-to-lymphocyte ratio (NLR, absolute neutrophil count/absolute lymphocyte count) and platelet-to-lymphocyte ratio (PLR, platelet count/lymphocyte count), using pre-surgery blood counts data.

Statistical correlations and distributions among clinical and pathological variables were evaluated using Pearson’s correlation coefficient (PCC), chi-square test, and Mann-Whitney test as appropriate. The statistical significance level was set at P < .05 for all tests.

Results

Eleven patients diagnosed with SS-NSCLC were treated with preoperative CT-RT followed by surgery. Three patients were not included, due to incomplete radiological data needed to perform the analyses, so a total of 8 cases were tested. Data about gender, age, smoking status, tumor biopsy, histology, pre-surgery median NLR and PLR, radiologic and metabolic responses are depicted in Table 1.

Table 1.

Patients’ characteristics.

Sulcus superior patients with NSCLC			N° of patients = 8
Gender	Female		1 (12.5%)
	Male		7 (87.5%)
Age	Median (years, range)		60.5 (37.2-77.1)
Smoking status	Former/current		8 (100.0%)
Tumor biopsy route	CT-guided fine needle biopsy		6 (75.0%)
	Bronchoscopy		2 (25.0%)
Histology	Adenocarcinoma		6 (75.0%)
	Squamous cell ­carcinoma		1 (12.5%)
	NOS		1 (12.5%)
Chemotherapy (carboplatin based)	8 cycles qw		6 (75.0%)
	3 cycles q3w		2 (25.0%)
Radiotherapy	Median (range; gray)		50 (44-64)
Radiological response	Stable disease		6 (75.0%)
	Partial response		2 (25.0%)
Metabolic response	Stable disease		4 (50.0%)
	Partial response		4 (50.0%)
Pre-surgery NLR	Median (range)		3.5 (1.7-4.6)
Pre-surgery PLR	Median (range)		199 (132-264)
RVTC	<10%		4 (50.0%)
	10%-20%		2 (25.0%)
	20%-50%		0 (0.0%)
	>50%		2 (25.0%)
Tumor immune infiltrate	CD3+	1+	2 (25.0%)
		2+	4 (50.0%)
		3+	2 (25.0%)
	CD4+	1+	7 (87.5%)
		2+	1 (12.5%)
	CD8+	1+	1 (12.5%)
		2+	4 (50.0%)
		3+	3 (37.5%)
	CD68+	1+	1 (12.5%)
		2+	3 (37.5%)
		3+	4 (50.0%)
Fibrosis	Median (%, range)		58 (20-95)
Necrosis	Median (%, range)		20 (5-50)

NLR, neutrophil-to-lymphocyte ratio; NSCLC, non–small cell lung cancer; PLR, platelet-to-lymphocyte ratio; RVTC, residual viable tumor cells.

The inflammatory cell characterization was performed on surgical samples, as it was not representative of the overall microenvironment in the diagnostic biopsy samples. PD-L1 status could be assessed in 5 tumor biopsies: 2 were PD-L1 negative and 3 were PD-L1 positive. Three biopsy samples were insufficient for molecular testing. PD-L1 determination was possible only in 6 surgical samples and resulted negative in all evaluable cases.

Median RVTC value was 10% (0%-70%), with 4 patients achieving a major pathological response.[7] Median fibrosis and necrosis rates reached 58% (20%-95%) and 20% (5%-50%), respectively.

The presence of CD3-lymphocytes was associated with stronger CD8-infiltrate (PCC of 0.8, P = .017). No correlations were found among other subsets of immune infiltrate. A stronger presence of CD68-macrophages was linked to lower RVTC (PCC = −0.708, P = .049; Fig. 1A).

Figure 1.

Correlation between stronger presence of CD68+ macrophages and lower RVTC values (A) and between CD68+ infiltrate and a superior metabolic response (B). Emblematic case of a patient with a superior sulcus adenocarcinoma showing strong CD68+ infiltrate, who achieved major pathological response and a partial metabolic response (C). A nodular necrotic area is evident (yellow arrow) (C.A). The tumor bed shows fibrosis and dense lymphoplasmacytic inflammatory infiltrate with cholesterol clefts (C.B; hematoxylin and eosin stain, original magnification ×50). Fibrotic area with neovascularization in which some residual neoplastic glands (black arrows) are present (C.C; hematoxylin and eosin stain, original magnification ×50). The CD68 immunoreaction highlights numerous macrophages infiltrating the tumor bed (C.D.; CD68 immunoperoxidase staining, original magnification ×100). On the right, 18-FDG PET-CT scan images show the metabolic response after treatment.

Immune infiltrate did not correlate with radiologic response to treatment. A correlation between CD68-infiltrate and better metabolic response was observed (P = .025, chi-square test) (Fig. 1B).

Pre-surgery NLR and PLR values were not linked to radiologic or metabolic responses. PLR values were associated both to stronger CD68-infiltrate (PCC = 0.8, P = .016; Supplementary Fig. 1A) and to lower RVTC values (PCC = −0.7, P = .047; Supplementary Fig. 1B).

Discussion

The diagnostic-therapeutic pathway of LA-NSCLC is complex and heterogeneous because of the different staging subgroups included and the need for a multidisciplinary treatment approach. Recently, survival of not resectable cases was improved by the advent of 12 months consolidation with durvalumab in responder patients after CT-RT. In this setting, durvalumab improved progression-free and overall survival (OS) rates, with an astonishing 60-month OS rate of 42.9% over a 33.4% for the placebo arm.[1]

As previously mentioned, patients’ selection for durvalumab treatment is currently based on pre-treatment evaluation of PD-L1 expression, which does not reflect the tumor-milieu after CT-RT.

Most patients with NSCLC are diagnosed using small biopsy specimens, often inadequate for a complete characterization.[8] Moreover, concordance of PD-L1 positivity between biopsy and resected tumor specimens varied according to reports, due to spatial heterogeneity and differences in biopsy methods.[9]

Our case series confirmed these data, showing a considerable percentage of cases with NOS histology (12.5%) and without evaluable PD-L1 expression (>30%). Considering the inadequacy of most diagnostic samples for PD-L1 evaluation, it is tempting to speculate that a rebiopsy could be helpful. However, our analysis of surgical specimens showed low RVTC with about half of patients achieving a major pathological response, underlining the difficult feasibility and clinical application of a rebiopsy after CT-RT.

In our exploratory analysis, pre-surgery higher PLR levels seemed to predict lower RVTC and higher CD68-infiltrate after CT-RT.

Additionally, CD68-infiltrate, as described for other malignancies,[10] was associated with a stronger pathological response and with a better metabolic response.

We acknowledge the limitations of our study. Pretreatment samples were not sufficient for a complete immune-infiltrate evaluation and a comparison with surgical specimens. Furthermore, it is a single-institutional, retrospective, and small-sized study; therefore correlations between histological and clinical parameters should be interpreted with caution. However, we think that these data, suggesting PLR as a potential circulating biomarker identifying best responder patients to CT-RT, deserve further prospective investigation in a wider patient population, to explore durvalumab consolidation benefit for those without evaluable baseline PD-L1 and higher PLR values.

Click here for additional data file.