Immune cell infiltration in head and neck squamous cell carcinoma and patient outcome: a retrospective study

Abstract

Background

Human papillomavirus (HPV) prevalence in oropharynx squamous cell carcinoma (OPSCC) is on the rise. HPV-linked OPSCCs represent a distinct clinical entity with a better treatment response and patient survival compared to tumors not linked to HPV. An emerging role in treatment response has been attributed to immune cell infiltration in human tumors. In this study, we investigated immune cell infiltration in human SCC of the head and neck region and its relation to overall survival after treatment with surgery (with or without radiotherapy) or concomitant chemo (or cetuximab)-radiotherapy.

Materials and methods

Paraffin-embedded tumor samples of 136 patients with SCC of the larynx, hypopharynx, oral cavity and oropharynx were processed for immunohistochemical detection of CD3⁺ T-cells, CD8⁺ cytotoxic T-cells, CD20⁺ B-cells and CD163⁺ M2 macrophages within the tumor infiltrated area. Clinico-pathological data were analyzed as a function of tumor location and p16-status. Immune cell infiltration was represented as stained area on the whole tumor infiltrated area, compared for the different tumor locations and correlated to patient survival.

Results

Patients with oropharynx tumors expressing significant p16 levels (p16-sg) had a 5-year overall survival of 85% compared to 43% for patients with no significant p16 (p16-ns) expression (HR: 0.3 – 95% CI: 0.1–0.6). Median immune cell infiltration (T- and B-lymphocytes) was significantly elevated in p16-sg oropharyngeal tumors, compared to p16-ns oropharyngeal tumors and to all other head and neck tumor locations. No difference in CD163⁺ macrophage infiltration was observed across the different patient groups. In the whole population, a high infiltration by CD3⁺ T-lymphocytes was associated to a significantly (p = .03; HR: 0.6, 95% CI: 0.4–0.97) better overall survival.

Conclusion

Oropharynx cancer with significant p16 expression showed an increased overall survival and elevated T- and B-lymphocyte infiltration, which suggests a prognostic relevance of immune cell infiltration.

Introduction

Over the past decades, a significant increase in the incidence of oropharyngeal squamous cell carcinoma (OPSCC) has been observed [1,2]. This observation has been associated with the human papillomavirus (HPV)-induced tumorigenesis [3], which prevalence in OPSCC is on the rise throughout the world, but mainly in developed countries [4]. HPV-linked (referred to as HPV⁺) OPSCC represents a distinct biological and clinical entity compared to tobacco and alcohol-related OPSCC [5]. A better response to treatment has been demonstrated for the former leading to improved prognosis and survival for HPV⁺ patients compared to their not HPV-linked (referred to as HPV⁻) counterparts [6,7]. The reason for the superior response to radiotherapy in HPV⁺ OPSCC is not yet fully understood.

An emerging role has been attributed to the immune infiltration in human tumors that is associated to patient survival [8]. In particular T-lymphocyte infiltration has been extensively studied and established as a prognostic factor for several human cancer types including head and neck (HN) SCC [8]. HNSCC are highly immune-infiltrated, and in particular oropharyngeal tumors have been reported to contain higher levels of T-cell infiltration, compared to tumors at other sites of the head and neck [9,10]. There are conflicting data on the relevance of immune infiltration in HPV⁺ versus HPV^- OPSCC. Studies reported either increased immune infiltration in HPV⁺ tumors [11,12], enhanced overall survival and disease-free survival in highly infiltrated tumors irrespective of their HPV-status [13] or the absence of difference between HPV⁺ and HPV^- tumors but correlation between higher immune infiltration and viral load in the HPV⁺ tumor group [14]. In addition, it has been reported that patients with HPV⁺ OPSCC, which are poorly infiltrated by tumor infiltrating lymphocytes (TILs) respond less to treatment, than HPV⁺ OPSCC with high number of TILs [12].

B-lymphocytes represent another important fraction of TILs that have been correlated with good prognosis in several human cancer types [15]. In patients with HNSCC, large number of B-lymphocytes in lymph-node metastases was associated with better prognosis [16]. Regarding head and neck primary tumors, Russel et al. [17] reported higher B-lymphocyte infiltration in HPV⁺ compared to HPV^- tumors. In contrast, for oropharyngeal primary tumors, no difference in B-lymphocyte infiltration between HPV⁺ and HPV^- tumors was detected [14].

One of the most frequent innate immune cells in the tumor microenvironment is the tumor associated macrophage (TAM) of the pro-tumoral M2 phenotype that promotes tissue remodeling, tumor progression, angiogenesis and metastasis [review in ref. 18]. High M2-like macrophage infiltration is generally associated with poor prognosis [18,19]. In HNSCC, M2-like macrophages were shown to have mutual interactions with squamous carcinoma cells, triggering both M2 polarization of TAMs and proliferation, invasiveness and angiogenesis [20]. However, differences in infiltration with M2 TAMs between HPV⁺ and HPV^- OPSCC have not yet been examined.

In the present study, we investigated the tumor infiltration by CD3⁺ T-lymphocytes, CD8⁺ cytotoxic T-lymphocytes, CD20⁺ B-lymphocytes and CD163⁺ M2-like macrophages in HPV⁺ and HPV⁻ oropharyngeal SCC, and in SCC of the hypopharynx, larynx, and oral cavity, and the relation to patient survival.

Materials and methods

Patient population

One hundred thirty-six patients with metastasis-free SCC of the larynx, hypopharynx, oral cavity or oropharynx treated between 2007 and 2011 at Saint-Luc University Hospital in Brussels (Belgium) were included into the study. Patients were treated by radiotherapy with or without (w/wo) concomitant chemotherapy or cetuximab, or surgery followed by radiotherapy (w/wo concomitant chemotherapy in case of R1 resection and/or extracapsular extension). Clinical characteristics of the patients included gender, age at diagnosis, smoking status (never smoker, active smoker, former smoker (stopped >1 year before diagnosis)), tumor location, tumor (T) stage, nodal (N) stage, disease stage, p16-status, treatment, and date of death or last contact.

This study was approved on 15th April 2013 by the ‘Commission d’Éthique Biomédicale Hospitalo-Facultaire’ of the medical school of the Université catholique de Louvain (approval # 2013/15AVR/168).

p16 and HPV evaluation

p16 expression in tumor cells was estimated by immunohistochemistry following the XT ultraView procedure with heat-induced epitope retrieval (HIER) in a dedicated EDTA-borate buffer (pH 8.0) on a Ventana Benchmark XT module (Roche, Basel, Switzerland), with diaminobenzidine (DAB) as chromogen, according to manufacturer’s instructions. Primary antibody routinely used in the pathology department was clone 16P04 (Neomarkers’ #MS-887-P, Immucor Belgium) from 2007 to 2011. Several tumors were not analyzed at that time, but only during the course of the study using clone G175-405 (#SS1153, BD Pharmingen) in 2015 and clone E6H4 (CINtec^® p16 Histology #725-4713, Roche Ventana) in 2017. p16 expression was assessed by a head and neck pathologist (EM); intense and diffuse p16 staining in at least 70% of neoplastic cells was considered to be significant (p16-sg) and the other ones nonsignificant (p16-ns). In p16-sg samples, the presence of HPV was investigated from about ten 7-μm-thick dewaxed, rehydrated and proteinase K-treated tumor sections using PCR after DNA extraction with QIAMP DNA mini Kit (Qiagen, Valencia, CA) according to manufacturer’s instructions [see reference 21]. In addition to HPV types 16, 18, 31, and 33 [see reference 21], samples were also screened for HPV-6 and -11 (primers: sens 5′-TTGCTGTGGATGTGACAGCA-3′; anti-sens 5′-GGTGCGCaGATGGGACACAC-3′).

Immunohistochemical staining of tumor biopsies

Five-μm-thick sections from paraffin-embedded tumor biopsies were provided by the tumor library of the King Albert II Cancer Institute (St-Luc University Hospital/Université catholique de Louvain), Brussels.

Tissue sections were deparaffinized and rehydrated in serial baths of methanol and toluol. Endogenous peroxidases were inhibited 20 min with 3% hydrogen peroxide in methanol. Sections were then subjected to antigen retrieval by heating to 95 °C for 60 min in 10 mM citrate buffer, pH 5.7. Subsequent steps were performed at room temperature in an automatic slide stainer (Dako Autostainer Plus, Glostrup, Denmark). Sections were blocked in 1% human immunoglobulins (from 3% Sandoglobulins reconstituted fluid, CSL Behring), 2% dry milk (Nonfat Dry Milk, #9999, Cell Signaling Technology), 5% BSA (Bovine Serum Albumin, #A2153, Sigma), 0.1% Tween 20 (#8.22184.0500, Merck Schuchardt OHG, Hohenbrunn, Germany) in TBS for 45 min. Primary antibodies used to detect epithelial cells (including squamous carcinoma cells) and the immune cells are listed in Supplementary Table 1. Antibodies were incubated for 30–60 min and revealed with Dako EnVision detection system (appropriate secondary antibodies are listed in Supplementary Table 1) followed by chromogen detection with DAB (#K3468, Dako). After counterstaining with hematoxylin (#S3301, Dako), slides were dehydrated and mounted with mounting medium (#CS703, Dako). No staining was observed when primary antibody was omitted. Lymph node sections were included in each batch as a positive control.

Image analysis

Stained slides were digitalized using a Leica slide scanner (SCN400, Leica Microsystems GmbH, Wetzlar, Germany) at 20x magnification. Whole scanned slides were analyzed using the image analysis tool Author version 2017.1 6.9.2 (Visiopharm, Hørsholm, Denmark). Immune cell infiltration was determined as percentage of stained area on the whole tumor infiltrated area, as previously described [22]. This protocol has been adapted to the Visiopharm Author software by CB.

The tumor infiltrated biopsy samples were first identified on each slide by a pathologist (EM). Limits of each biopsy sample infiltrated by the tumor were then accurately delineated using the annotation tool of the software and tissue was automatically detected from background at an intermediate digital magnification (10x) using a threshold classification method based on the ‘RGB-R’ image feature. In the same algorithm, empty spaces (vascular structures, rips), bubbles and folds were excluded based on the ‘HDAB-H’ image feature of the software (Supplementary Figure 1A). When required, manual corrections were applied. CD3 (T-lymphocytes), CD8 (cytotoxic T-lymphocytes), CD20 (B-lymphocytes) or CD163 (M2 macrophages) immunostained pixels were subsequently detected within the whole previously delineated tumor infiltrated area at high resolution (20x) using a threshold classification method based on pre-processing steps highlighting the DAB staining (‘HDAB-DAB’ matrix of the software). Threshold was adjusted on representative stained versus unstained regions. Detected areas smaller than 5 µm² and/or presenting a circularity smaller than 0.95 (1 = circle, 0 = line) were considered as artifacts and discarded. The parameters were kept constant for all slides. Representative images for immune cell detection are presented in Supplementary Figure 1B. The results were expressed as percentages, that is, stained area divided by whole tumor infiltrated area ×100 per biopsy sample. One biopsy sample of the p16-sg OPSCC group was not quantifiable.

Statistical analysis

Overall survival (OS) was determined from the date of biopsy to the date of death (from any cause) or the date of last contact. Immune staining on tumors was classified into high (above median percentage of staining for the total patient population) and low (below median percentage of staining for the total patient population).

Statistical analyses were conducted using Prism 7 (GraphPad Software Inc., La Jolla, USA) by KS and SAS software (SAS Institute Inc., 100 SAS Campus Drive, Cary, NC 27513-2414, v. 9.4). To compare two independent groups of non-normally distributed markers, non-parametric Wilcoxon–Mann–Whitney test was used. For comparison of multiple groups, Kruskal–Wallis nonparametric test was applied. Fisher’s exact test and chi-square test were used to compare proportions of clinical parameters in the patient population. Analyses of overall survival (OS) were performed using the Kaplan–Meier method including log-rank test. Hazard ratios (HR) for the whole survival curves were mentioned including 95% Confidential Interval (CI). For all analyses a p value < .05 was considered to be statistically significant and was adjusted for multiple comparisons according to the applied test.

Results

Patients’ characteristics

The study included 136 cases of primary HNSCC that were treated either primarily by surgery (followed by radiotherapy or concomitant chemo-radiotherapy) or primarily by radiotherapy (w/wo concomitant chemotherapy or cetuximab). Patients’ characteristics are presented in Table 1. Patients with oropharyngeal SCC were divided into patients with a significant staining for p16 (named ‘p16-sg’) and patients without significant staining for p16 (named ‘p16-ns’).

Table 1. Patients’ characteristics.

	Larynx (n = 22) Number (%)	Hypopharynx (n = 16) Number (%)	Oral cavity (n = 36) Number (%)	p16-sg Oropharynx (n = 20) Number (%)	p16-ns Oropharynx (n = 42) Number (%)
Gender
Female	3 (13.6)	4 (25)	9 (25)	7 (35)	9 (21.4)
Male	19 (86.4)	12 (75)	27 (75)	13 (65)	33 (78.6)
Age at diagnosis
Mean (range)	65 (47–84)	62 (51–81)	60 (31–83)	59 (41–77)	60 (37–83)
Smoking status
Active	18 (81.8)	11 (68.8)	21 (58.3)	9 (45)	30 (71.4)
Former	4 (18.2)	2 (12.5)	6 (16.7)	5 (25)	7 (16.7)
Never	0 (0)	3 (18.8)	9 (25)	6 (30)	5 (11.9)
p16 status
p16-sg	1 (4.5)	2 (12.5)	1 (2.8)	20 (100)	0 (0)
p16-ns	21 (95.5)	14 (87.5)	35 (97.2)	0 (0)	42 (100)
T stage
1	11 (50)	1 (6.3)	4 (11.1)	6 (30)	8 (19)
2	4 (18.2)	5 (31.3)	15 (41.7)	10 (50)	15 (35.7)
3	2 (9.1)	1 (6.3)	3 (8.3)	2 (10)	3 (7.1)
4	5 (22.7)	9 (56.3)	14 (38.9)	2 (10)	16 (38.1)
N stage
0	20 (90.9)	7 (43.8)	18 (50)	3 (15)	16 (38.1)
1	0 (0)	4 (25)	11 (30.6)	1 (5)	8 (19)
2	2 (9.1)	5 (31.3)	7 (19.4)	16 (80)*	17 (40.5)
3	0 (0)	0 (0)	0 (0)	0 (0)	1 (2.4)
Stage
I	11 (50)	0 (0)	3 (8.3)	0 (0)	2 (4.8)
II	4 (18.2)	3 (18.8)	7 (19.4)	2 (10)	11 (26.2)
III	2 (9.1)	1 (6.3)	9 (25)	1 (5)	8 (19)
IV	5 (22.7)	12 (75)	17 (47.2)	17 (85)#	21 (50)
Treatment
RT	22 (100)	15 (93.8)	10 (27.8)	20 (100)	42 (100)
RT alone	15 (68.2)	3 (18.8)	4 (11.1)	4 (20)	19 (45.2)
RT-CH	7 (31.8)	10 (62.5)	6 (16.7)	12 (60)	21 (50)
RT-CTX	0 (0)	2 (12.5)	0 (0)	4 (20)	2 (4.8)
SRT	0 (0)	1 (6.3)	26 (72.2)§	0 (0)	0 (0)

Former: stopped smoking at least 1 year before diagnosis. T stage, N stage, and Stage: according to UICC-AJCC 7th edition. RT: primary radiotherapy, without (RT alone) or with concomitant chemotherapy (RT-CH) or cetuximab (RT-CTX). SRT: primary surgery (followed by radiotherapy w/wo chemotherapy). *p ≤ .01 (N2 in p16-sg oropharyngeal tumors compared to all other tumor locations); #p ≤ .01 (stage IV in p16-sg oropharyngeal tumors compared to all other tumor locations); §p ≤ .0001 (SRT in oral cavity tumors compared to all other tumor locations).

For the whole cohort, male patients were over-represented (77%; no significant difference between groups), median age at diagnosis was 60 years (range: 31–84 years; no significant difference between groups), 89% of patients were active smokers (no significant difference between groups), and 46% (62 out of 136) had an oropharyngeal primary tumor.

Patients with oropharyngeal SCC p16-sg had significantly more N2 disease (80%; UICC-AJCC 7th ed.) compared to the other groups (p < .01, Fisher’s exact test) and significantly higher stage IV disease (85%; UICC-AJCC 7th ed.) compared to the other groups (p ≤ .01, Fisher’s exact test) except the hypopharyngeal SCC.

Patients with oral cavity tumors were preferentially treated by surgery and post-operative radiotherapy w/wo concomitant chemotherapy) (72%, p < .0001, Fisher’s exact test) while patients with laryngeal, hypopharyngeal and oropharyngeal SCC were preferentially treated by radiotherapy w/wo chemotherapy or cetuximab. For laryngeal SCC, primary radiotherapy alone was the main treatment modality (68%), reflecting the lower disease stage.

Regarding the p16 staining, 24 patients were classified as p16-sg, 1 with a laryngeal SCC, 2 with a hypopharyngeal SCC, 1 with an oral cavity SCC and 20 with an oropharyngeal SCC. No HPV was detected in the 3 p16-sg laryngeal or hypopharyngeal tumors, whereas HPV was detected in the p16-sg oral cavity tumor. HPV was not found in 3 out of the 20 p16-sg oropharyngeal cases. Consequently, the positive predictive value (PPV) of p16 staining for HPV-association in oropharynx SCC was 85%.

Overall survival

Kaplan–Meier curves for overall survival (OS) are presented in Figure 1 as a function of tumor location. For oropharyngeal tumor, patients with p16-sg and p16-ns were separated. Five-year OS was much better for patients with oropharyngeal p16-sg tumors (85%) compared to patients with laryngeal (68%; HR ns), oral cavity (50%; HR: 0.3–95% CI: 0.2–0.7), oropharyngeal p16-ns (43%; HR: 0.3–95% CI: 0.1–0.6), and hypopharyngeal tumors (38%; HR: 0.3–95% CI: 0.1–0.8). These results did not change when the survival curves were calculated segregating the oropharyngeal patients according to the HPV status, instead of the p16 status. Five-year OS reached 82% for p16-sg/HPV-linked compared to 47% for the other oropharyngeal SCC patients. For the whole population, the analysis of patient survival according to the smoking status did not reveal any difference between never, former and active smokers. Similar results were observed for the patients with p16-sg oropharyngeal SCC.

Figure 1. Kaplan–Meier curves for overall survival as a function of the tumor location; patients with oropharyngeal SCC are separated according to the p16-status.

Immune cell infiltration

In p16-sg oropharyngeal tumors, a higher infiltration by CD3⁺ T-lymphocytes, CD8⁺ cytotoxic T-lymphocytes and CD20⁺ B-lymphocytes was observed in comparison with p16-ns oropharyngeal, laryngeal, hypopharyngeal and oral cavity tumors (Figure 2). The differences were all significant except for CD20⁺ B-lymphocytes infiltration in hypopharyngeal tumors. There was no difference in CD163⁺ macrophage infiltration between the several regions of the head and neck population and the p16 status.

Figure 2. Immune infiltration in laryngeal, hypopharyngeal, oral cavity, p16-significant (p16-sg) and p16-not significant (p16-ns) oropharyngeal SCC. p values: *≤.05; **≤.01; ***≤.001; ****≤.0001. For p16-sg OPSCC, only 19 out of 20 cases were quantifiable.

We further analyzed the immune cell infiltration by splitting all cases into high infiltration (above the median value of the whole patient population) and low infiltration (below the median value of the whole patient population). In p16-sg oropharyngeal tumors, the percentage of cases with high infiltration of CD3⁺, CD8⁺ and CD20⁺ cells reached 89% (17 of 19 cases), 89% (17 of 19 cases) and 79% (15 of 19 cases), respectively, whereas in the other tumors corresponding values reached at the maximum 56%, 56% and 64%, respectively (data not shown). All these differences were significant (p ≤ .05, Wilcoxon–Mann–Whitney) except for low versus high CD20⁺ infiltration between oropharyngeal p16-sg SCC and oral cavity or hypopharyngeal SCC.

When the p16-sg/HPV-linked oropharyngeal SCC patient population was compared to the other patients, all the differences in immune cell infiltration and intensity of infiltration remained (data not shown).

Last, no difference was observed in immune cell infiltration of oropharynx tumors between active smokers, former smokers, and never smokers.

Immune cell infiltration and overall survival

In the whole population, a high infiltration by CD3⁺ T-lymphocytes was associated to a significantly (p = .03; HR: 0.6, 95% CI: 0.4–0.97) better overall survival (data not shown). Such distinction was not observed for CD8⁺, CD20⁺ or CD163⁺. In the oropharyngeal p16-sg and p16-ns groups, overall survival was not better in patients expressing high versus low immune cell infiltration neither for CD3⁺, CD8⁺, CD20⁺ or CD163⁺ (data not shown, p > .1). In the p16-sg oropharynx subgroup, possibly the low number of tumors with low infiltration prevented to reach any significance.

Discussion

In the present retrospective study, we investigated the immune cell infiltration by T-lymphocytes, B-lymphocytes and M2 macrophages in oropharyngeal SCC with significant or no significant p16 expression and in laryngeal, hypopharyngeal and oral cavity SCC and explored the correlation to patient survival. Patients with oropharyngeal SCC and significant p16 expression had a better overall survival and expressed more immune cell infiltration than the others.

In our series, 32% of oropharyngeal SCC treated between 2007 and 2011 had a significant p16 expression, indicative of HPV association. Such figure is similar to what has been reported in neighboring countries during similar time periods, for example, 34% p16⁺/HPV⁺ oropharynx tumors in Germany until 2011 [23], 38% HPV⁺ oropharynx tumors in the Netherlands from 2000 to 2009 [24], but it is slightly lower than the mean figure of 49.5% reported for the entire European countries between 2005 and 2014 [4].

As already described, patients with oropharyngeal SCC and a significant p16 expression had a much better overall survival despite the advanced nodal disease and overall stage [6]. In our series, the rather good outcome of patients with laryngeal SCC is likely explained by the early stage of the disease compared to the other tumor locations.

We analyzed infiltration by CD3⁺ cells to get an overview of the global T-lymphocyte infiltration. We then investigated the CD8⁺ T-cell subset as it represents cytotoxic T-lymphocytes that may contribute to better tumor cell clearance after radiotherapy of HPV-linked OPSCC. Further, we explored CD20⁺ immune cells that are B-lymphocytes and produce specific antibodies that mark tumor cells for clearance by T-lymphocytes. Finally, we studied infiltration by macrophages of M2 type (CD163⁺) that have pro-tumor properties and could thus contribute to worse treatment response in OPSCC that are not associated to HPV. Although reported in previous studies, we did not look at regulatory T-lymphocytes (Treg) by FoxP3, as it is not a specific marker for this cell population and it should only be used in combination with a panel of markers to specifically identify Treg [25].

Previous investigations have demonstrated an increased infiltration with different T-lymphocyte subsets (CD4⁺, CD8⁺, FoxP3⁺, Th1, Th2, Th17) in oropharyngeal SCC compared to laryngeal, hypopharyngeal and oral cavity SCC [9,10,26], as well as an elevated infiltration with T-lymphocytes in HPV-linked compared to not HPV-linked oropharyngeal SCC [12,26–28]. We confirmed the latter finding by detection of increased CD3⁺ and CD8⁺ staining, but using p16 immunohistochemical staining, further validating this marker of HPV association for physiopathological studies. We also demonstrated that p16-sg oropharyngeal SCC contains not only more infiltrating T-lymphocytes (CD3⁺) and more cytotoxic T-lymphocytes (CD8⁺) than the p16-ns oropharyngeal SCC, but also more than the other head and neck sites as already reported by Nguyen [26]. Furthermore, CD20⁺ B-lymphocyte infiltration was increased in the oropharyngeal p16-sg SCC population. An elevated number of tumors with a high B-lymphocyte infiltration has been reported for the HPV⁺ head and neck cancer population compared to their HPV^- counterparts [17] and is in accordance with our results on p16-sg compared to p16-ns oropharynx cancers. In contrast, the group of Rittà [14] did not find such increased immune cell infiltration in HPV⁺ OPSCC. However, it has to be noticed that this patient cohort was quite small (22 OPSCC: 9 HPV⁺, 13 HPV^-). Last, we are the first to report CD163⁺ macrophage infiltration in HNSCC but no difference was observed between p16-sg and p16-ns OPSCC or the whole HNSCC population and no correlation to survival was found. This finding suggests that the pro-tumor activity of type M2 macrophages plays no role in the worse response to radiation treatment in not HPV-related OPSCC.

Although patients with p16-sg OPSCC showed an elevated survival and an increased tumor T- and B-lymphocyte infiltration, we did not find a correlation between the amount of lymphocyte infiltration (high versus low) and the survival in p16-sg oropharyngeal SCC. Unfortunately, the number of patients was likely too small to draw any meaningful conclusion. However, for the total HNSCC population, we could show a significantly increased overall survival for patients bearing tumors with high total T-lymphocyte (CD3⁺) infiltration compared to those with low CD3⁺ infiltration. Former studies confirmed that patients with high T-lymphocyte infiltrated tumors of the head and neck region had an increased survival [reviewed by 29]. Ward et al. [12] demonstrated an increased survival for high T-cell infiltrated HPV⁺/p16⁺ oropharyngeal SCC compared to their low infiltrated counterparts, which suggests a correlation between T-lymphocyte infiltration and survival.

Our patient cohort contained a high number of active and former smokers, including patients with p16-sg OPSCC, in accordance with other study cohorts [30,31]. In contrast to former studies that indicated a strong negative impact of smoking on patient outcome [6,9,31], we did not find significant differences between the groups of active, former and never smokers in overall survival, what might be due to the small number of cases per group. Smoking habit has further been associated with lower T-lymphocyte (including CD8⁺) infiltration in SCC of different head and neck sites [9,32]. However, we did not detect differences in lymphocyte infiltration between active, former and never smokers within our oropharyngeal cancer cohort.

The present study is a single institution retrospective investigation with some limitations. Only patients treated between 2007 and 2011 were included, and in Belgium during that period, the number of HPV⁺ SCC was quite low. Further, we only investigated the presence of tumor infiltrating immune cells in head and neck SCC, in order to recognize a basic distinction in immune association between oropharynx cancer associated with HPV (p16-sg) or not (p16-ns). However, several studies [17,33–36] showed that it might be of importance to further explore the organization, activation status and gene signature of the infiltrating immune cells, their functionality and target specificity, as well as the expression of immune-activating and immune-suppressing genes in the tumor microenvironment in order to fully understand the role of the immune system in better treatment response for patients with HPV⁺ HNSCC.

Acknowledgments

We thank the Laboratoire Luc Olivier (Villers-le-Bouillet) that kindly provided one paraffin-embedded tumor. We also thank the Biobank of the King Albert II Cancer Institute (St-Luc University Hospital/Université catholique de Louvain, Brussels) for providing the sections from all paraffin-embedded tumors. The authors wish to thank Mrs. A. Collard from the statistical unit of the King Albert II Cancer Institute (St-Luc University Hospital/Université catholique de Louvain, Brussels) for her assistance during the analysis of the data.

Disclosure statement

The authors report no conflicts of interest.

Additional information

Funding

This work was supported by the Fonds National de la Recherche Scientifique (F.R.S.-FNRS) under Grant 7.4520.15 (KS, Télévie doctoral fellow) and Grant 7.4648.12F (VB, Télévie postdoctoral fellow); and by the Patrimoine of the Université catholique de Louvain under a doctoral thesis grant (KS).