- Email: [email protected]
Clinical Significance of Tumor-Infiltrating Lymphocytes in Lung Neoplasms
GENERAL THORACIC SURGERY: The Annals of Thoracic Surgery CME Program is located online at http://cme.ctsnetjournals.org. To take the CME activity related to this article, you must have either an STS member or an individual non-member subscription to the journal.
Clinical Significance of Tumor-Infiltrating Lymphocytes in Lung Neoplasms Enrico Ruffini, MD, Sofia Asioli, MD, Pier Luigi Filosso, MD, Paraskevas Lyberis, MD, Maria Cristina Bruna, MD, Luigia Macrì, MD, Lorenzo Daniele, MD, and Alberto Oliaro, MD Departments of Thoracic Surgery and Pathology, University of Torino, Torino, Italy
Background. Tumor-infiltrating lymphocytes (TIL) are considered important in anticancer immunosurveillance, although their role has not been clearly established yet. We examined prevalence, correlations, and prognostic significance of TIL among our patient population of resected lung neoplasms. Methods. From 1993 to 2006, the presence of TIL was retrospectively evaluated in 1,290 patients operated on for primary lung neoplasms. Tumor-infiltrating lymphocytes were defined as those intraepithelial lymphocytes located within the cancer cell nests. Results. Tumor-infiltrating lymphocytes were detected in 294 patients (23%). A significant difference was found between prevalence in non–small cell lung carcinomas versus neuroendocrine tumors (290 of 1,208, 24% versus 4 of 82, 5%; p ⴝ 0.0001). Prevalence was similar in adenocarcinomas, squamous-cell carcinomas, and large-cell anaplastic carcinomas. Logistic regression analysis indicates that TIL correlate with grading (odds ratio, 1.27; 95% confidence interval, 1.04 to 1.55; p ⴝ 0.02), tumor dimension (odds ratio, 0.86; 95% confidence interval, 0.79 to
0.94; p ⴝ 0.0008), and vascular invasion (odds ratio, 1.62; 95% confidence interval, 1.21 to 2.16; p ⴝ 0.0009). A not significantly better survival in the presence of TIL was observed overall (p ⴝ 0.20), becoming significant in squamous-cell carcinomas (p ⴝ 0.03). In patients with stage I disease, TIL is associated with a significant survival advantage in squamous-cell carcinomas (p ⴝ 0.03). The survival advantage increases with the duration of follow-up and is more evident after 4 to 6 years. Conclusions. Tumor-infiltrating lymphocytes are observed in about one fourth of resected lung neoplasms: they are rare in neuroendocrine tumors. Tumor-infiltrating lymphocytes are more frequent in poorly differentiated tumors and in tumors with microscopic vascular invasion. The presence of TIL correlates with an improved survival in squamous cell carcinomas, particularly at early stage. The survival advantage increases with the duration of follow-up.
L
role in anticancer immunosurveillance. Among the mature human T-cell population, three subtypes seem to play a major role in immunosurveillance against tumor growth: CD8⫹, CD4⫹, and T regulatory (Treg) cells. Cytotoxic (CD8⫹) T cells are the major subset of T cells that constructively mediate an effective antitumor response: they are able to induce tumor killing on direct recognition of peptide antigens presented by the tumor’s MHC class I molecules. CD4⫹ T cells also migrate to the tumor site, but their role in antitumor activity is dualistic: some CD4⫹ T cells seemingly hinder the effector function of CD8⫹ T cells and therefore indirectly promote tumor growth whereas other subtypes of CD4⫹ T cells might aid in the activation of CD8⫹ T cells. Finally, a third subpopulation of T cells, namely the Treg cells, has recently been identified as a subtype of CD4⫹ T cells (also known as naturally occurring CD4⫹ CD25⫹ T regulatory cells); Treg cells may downregulate the immune response to tumors by attenuating the host’s anti-
ung cancer is the leading cause of cancer-related deaths among men and women. Despite recent improvement in early diagnosis and treatment, the overall 5-year survival rate is approximately 15% [1], well below that of other solid-organ tumors. Given the high recurrence rate of the tumor, many authors have looked for clinicopathologic markers that can reliably be used as prognostic indicators, as well as a guide to identify patients likely to benefit from adjunct therapies. In a variety of human solid tumors, tumor-infiltrating lymphocytes (TIL) are considered to play an important
Accepted for publication Oct 16, 2008. Presented at the Forty-fourth Annual Meeting of The Society of Thoracic Surgeons, Fort Lauderdale, FL, Jan 28 –30, 2008. Address correspondence to Dr Ruffini, Thoracic Surgery Service, University of Torino, Ospedale Molinette, Dipartimento di Fisiopatologia Clinica, Sezione di Chirurgia Toracica, 3, Via Genova, Torino, 10126, Italy; e-mail: [email protected].
© 2009 by The Society of Thoracic Surgeons Published by Elsevier Inc
(Ann Thorac Surg 2009;87:365–72) © 2009 by The Society of Thoracic Surgeons
0003-4975/09/$36.00 doi:10.1016/j.athoracsur.2008.10.067
GENERAL THORACIC
ORIGINAL ARTICLES: GENERAL THORACIC
GENERAL THORACIC
366
RUFFINI ET AL TUMOR-INFILTRATING LYMPHOCYTES IN LUNG NEOPLASMS
tumor T cells, thus allowing unrestricted tumor growth and inhibiting systemic immunosurveillance. Definition of TIL varies among different series. In the present study, TIL were considered as those intraepithelial infiltrating lymphocytes located within the cancer cell nests; phenotypically they were almost entirely CD8⫹ T lymphocytes. Although the favorable prognostic significance of TIL has been reported for a variety of human solid cancers, including colorectal cancer [2], esophageal cancer [3], prostatic cancer [4], renal cell carcinoma [5], and pancreatic cancer [6], the role of TIL in non–small cell lung cancer (NSCLC) is still debatable. Some authors have reported that the presence of TIL gives no survival advantage [7], whereas other authors reached the opposite conclusion [8]. The aim of the present study was to investigate the prevalence of TIL in our population of patients with lung neoplasms who underwent surgical resection, to examine possible correlations with other clinicopathologic variables, and finally to assess their prognostic significance, particularly in early stage (stage I) disease.
Patients and Methods Approval for this study was provided by our institutional review board, and patient consent was waived. A retrospective review was undertaken among our population of patients with primary lung neoplasms operated on at our institution between 1993 and 2006. In this period, of a total of 2,200 patients, 1,290 were evaluated for the presence of TIL on the surgical specimen. The remaining 910 were not considered for the analysis either as a result of fulfilling the exclusion criteria (see below) or because of incomplete pathologic information.
Preoperative Evaluation Baseline demographics, histopathologic data, overall survival, and pathologic specimens preserved in paraffin were available for all patients. Patients receiving preoperative induction therapy were excluded. On the basis of above clinicopathologic data and the presence or absence of TIL, defined as CD8⫹ intraepithelial lymphocytes morphologically identified within the cancer cell nests, patients were collected in two different group, TIL(⫹) and TIL(⫺). Patient demographics and tumor features in the two patient groups are detailed in Table 1. All patients were staged postsurgically according to the most recent TNM staging system [9]. For the analysis, the following histopathologic variables were considered: histologic grading of differentiation (well, moderately, and poorly differentiated, respectively G1, G2, and G3), tumor size, microscopic vascular invasion, and microscopic perineural invasion. Follow-up information of all patients was obtained through clinic follow-up notes, direct patient or family contact, or contact with the patient’s primary care physician.
Ann Thorac Surg 2009;87:365–72
Table 1. Prevalence Differences of Clinicopathologic Variables Between Patients With and Without Tumor-Infiltrating Lymphocytes on the Surgical Specimen Variable Patients Sex Male Female Age (y), mean (range) Histology Adenocarcinoma Squamous cell Adenosquamous Large cell anaplastic BAC Neuroendocrine Grading G1 G2 G3 Vascular invasion Yes No Perineural invasion Yes No Tumor dimension Stage I II IIIa N factor N0 N1 N2
TIL(⫹)
TI(⫺)
294
996
253 41 64 (43–80)
831 165 64 (25–82)
p Value
0.28
0.36b
133 (27%) 134 (24%) 2 (9%) 14 (25%) 7 (8%) 4 (5%)
362 (73%) 415 (76%) 20 (91%) 42 (75%) 79 (92%) 78 (95%)
30 (18%) 145 (28%) 145 (28%)
143 (82%) 368 (72%) 362 (72%)
0.002c
164 (30%) 130 (18%)
395 (70%) 601 (82%)
0.0001
30 (22%) 264 (23%) 3.8
107 (78%) 889 (77%) 4.2
0.8
165 (23%) 57 (21%) 49 (23%)
549 (77%) 208 (79%) 165 (77%)
0.9
188 (22%) 63 (28%) 42 (21%)
670 (78%) 161 (72%) 162 (79%)
0.7
Versus adenocarcinoma ⫹ squamous cell. c G1 versus G2 ⫹ G3. cell.
a
b
0.08a 0.92a 0.0003a 0.0001a
0.03
Versus squamous
TIL ⫽ tumor-infiltrating lymphocytes.
Immunohistochemistry and Quantification of Tumor-Infiltrating Lymphocytes The main objective of the present study was a descriptive and prognostic evaluation of TIL, and for that purpose the TIL population was considered as a whole for the clinical analysis on 1,290 patients. However, to better understand the relative importance of TIL subpopulations and to see whether our TIL population distribution were in accordance with that reported in the literature, immunohistochemistry and quantification of TIL were performed on a sample of 21 patients (7 adenocarcinomas, 7 squamous cell carcinomas, 3 bronchioloalveolar carcinoma (BAC), 1 typical carcinoid, 1 atypical carcinoid, 1 large cell neuroendocrine carcinoma, 1 small cell lung carcinoma). Immunohistochemical analysis was performed on resected, paraffin-embedded lung cancer tissues. After microtome sectioning (4 m) and slide labeling, immuno-
histochemical analysis was performed using the Ventana automated immunostainer (BenchMark AutoStainer; Ventana Medical Systems, Tucson, AZ) with primary antibodies directed to the T-cell and B-cell surface glycoprotein CD3 (DAKO, Glostrup, Denmark; antigen retrieval with citrate buffer pH 6, dilution 1:150), CD4 (Novocastra, Newcastle, United Kingdom; antigen retrieval with citrate buffer pH 6, dilution 1:20), CD8 (DAKO; antigen retrieval with citrate buffer pH 6, dilution 1:50), and CD20 (DAKO; antigen retrieval with citrate buffer pH 6, dilution 1:200), respectively. Each entire tumor section was evaluated for TIL by using a ⫻40 objective lens and 10 independent fields, with the most abundant TIL selected, digitally photographed at a size of 0.0625 mm2, and counted manually by two different pathologists (A.S., L.D.) blinded to patients’ outcomes. Immunohistochemical staining was define as negative (0 ⫽ no staining or positive staining in ⬍20% of cells) or positive (1 ⫽ positive staining in ⱖ20% of cells) for each patient. This dichotomous classification was adopted because it has customarily been used in our pathology report since 1993. Areas near necrosis were excluded from the analysis. In the present study, TIL were considered as those intraepithelial infiltrating lymphocytes located within the cancer cell nests; phenotypically they were almost entirely CD8⫹ T lymphocytes.
Statistical Analysis Comparison between proportion differences was undertaken using 2 test (Fisher’s exact test when appropriate). Survival was defined as the time from surgery to death or last follow-up. Survival curves were estimated with the Kaplan–Meier method. Significance between survival curves was assessed with the log-rank test. A probability value less than 0.05 was considered to be significant. The Cox proportional hazard model [10] was applied for identifying prognostic factors. Hazard ratios along with 95% confidence intervals (95% CI) were provided. To investigate the relationship between some possible causal (independent) variables and the presence or absence of TIL (considered as a binary output variable), a logistic regression model was used, with the assumption that the events were independent and the relationship plausibly log-linear. The probability value was calculated on the Wald statistic, and 95% CI were provided. All statistical analysis was undertaken using software packages (STATISTICA, release 7.1, 2005; StatSoft, Vigonza, PD, Italy).
Results Overall, TIL were detected on the surgical specimen in 294 patients (23%) operated on for primary lung neoplasms. The typical lymphocytic infiltration pattern around the tumor as emerged from our samples can be described as follows: (1) a peripheral lymphocytic infiltration composed of prevalent B lymphocytes (CD20⫹) organized in lymphoid follicles mixed with T lymphocytes (both CD3⫹ and CD4⫹); (2) a lymphocytic infiltra-
RUFFINI ET AL TUMOR-INFILTRATING LYMPHOCYTES IN LUNG NEOPLASMS
367
tion within cancer stroma (mainly CD3⫹ T lymphocytes); and (3) intraepithelial lymphocytes within the cancer cell nests mostly represented by CD8⫹ T lymphocytes (the true TIL population for the present study). In addition, lymphocytic infiltration along the invasive margin is common, and areas within the cancer are observed usually sparsely in the lymphocytic response as a result of immune evasion mechanisms.
Prevalence of Tumor-Infiltrating Lymphocytes Prevalence of TIL among the different neoplasms was as follows (Table 1): adenocarcinoma, 27% (133 of 495); squamous cell carcinoma, 24% (134 of 549); BAC, 8% (7 of 86); typical carcinoid, 0% (0 of 34); atypical carcinoid, 7% (1 of 14); adenosquamous carcinoma, 9% (2 of 22); large cell anaplastic carcinoma, 25% (14 of 56); large cell neuroendocrine carcinoma, 19% (3 of 16); and small cell carcinoma, 0% (0 of 18). A significant difference was found between prevalence in NSCLC and neuroendocrine carcinomas (290 of 1,208, 24% versus 4 of 82, 5%; p ⫽ 0.0001). Among NSCLC, prevalence was similar in adenocarcinoma, squamous cell carcinoma, and large cell anaplastic carcinoma; conversely, adenosquamous carcinoma and BAC had a lower prevalence of TIL, which was significant in BAC (p ⫽ 0.0003). The prevalence of TIL was also investigated in association with different clinicopathologic variables including grading, microscopic vascular and perineural invasion, tumor dimension, stage, and TNM factor. The presence of TIL was associated with poorly differentiated tumors, with tumors of small size, and with the presence of microscopic vascular invasion (Table 1). We further analyzed the two populations of patients with adenocarcinoma and squamous cell carcinoma with regard to the presence or absence of TIL. In adenocarcinomas, TIL prevalence among patients with microscopic vascular invasion was very high (35%), whereas no difference was found in TIL prevalence among different grades of differentiation (G1, G2, and G3) or tumor dimension. Conversely, in squamous cell carcinomas, TIL prevalence was less associated with microscopic vascular invasion, whereas TIL were found more frequently in poorly differentiated tumors. It therefore seems evident from our data that the presence of TIL in adenocarcinoma and squamous cell carcinoma might have a different presentation and maybe a different significance as resulted from our survival analysis (see later).
Risk Factors for Tumor-Infiltrating Lymphocytes A logistic regression analysis was undertaken using the presence or absence of TIL as a binary dependent variable and the following clinicopathologic independent variables: tumor dimension, grading, vascular invasion, and perineural invasion (Table 2). Of all the independent variables, a significant correlation was found with grading (odds ratio [OR], 1.27; 95% CI, 1.04 to 1.55; p ⫽ 0.02), tumor dimension (OR, 0.86; 95% CI, 0.79 to 0.94; p ⫽ 0.0008), and vascular invasion (OR, 1.62; 95% CI, 1.21 to 2.16; p ⫽ 0.0009).
GENERAL THORACIC
Ann Thorac Surg 2009;87:365–72
GENERAL THORACIC
368
RUFFINI ET AL TUMOR-INFILTRATING LYMPHOCYTES IN LUNG NEOPLASMS
Ann Thorac Surg 2009;87:365–72
Table 2. Logistic Regression Analysis for Tumor-Infiltrating Lymphocytes in Population of Patients With Lung Neoplasms Variable Overall Vascular invasion Grading Tumor dimension Adenocarcinoma Vascular invasion Squamous cell carcinoma Grading Tumor dimension
Odds Ratio
95% Confidence Interval
p Value
1.62 1.27 0.86
1.21–2.16 1.04–1.55 0.79–0.94
0.0009 0.02 0.0008
2.33
1.46–3.71
0.0004
1.53 0.79
1.05–2.23 0.69–0.91
0.03 0.0007
In adenocarcinomas, TIL was associated only with vascular invasion (OR, 2.33; 95% CI, 1.46 to 3.71; p ⫽ 0.0004); in squamous cell carcinomas, TIL was associated with grading (OR, 1.53; 95% CI, 1.05 to 2.23; p ⫽ 0.03) and tumor dimension (OR, 0.79; 95% CI, 0.69 to 0.91; p ⫽ 0.0007).
Prognostic Significance of Tumor-Infiltrating Lymphocytes Prognostic significance was evaluated first by performing a multivariate survival analysis on different independent variables including TIL, vascular invasion, perineural invasion, TNM factors, and stage. Among other covariates, TIL did have an independent prognostic significance overall (hazard ratio, 0.78; 95% CI, 0.64 to 0.98; p ⫽ 0.05) and in squamous cell carcinomas (hazard ratio, 0.68; 95% CI, 0.50 to 0.91; p ⫽ 0.02), but not in adenocarcinomas (hazard ratio, 0.87; 95% CI, 0.64 to 1.18; p ⫽ 0.38). We then compared survivals among patients with different neoplasms according to the presence or absence of TIL. A nonsignificant survival difference between TIL(⫹) and TIL(⫺) groups was observed overall (p ⫽ 0.20; Fig 1) and in adenocarcinomas (p ⫽ 0.6), large cell anaplastic carcinomas (p ⫽ 0.91), and neuroendocrine tumors (p ⫽ 0.39). The survival difference was statistically significant in squamous cell carcinomas (p ⫽ 0.03; Fig 2). Tumor-infiltrating lymphocytes were a negative prognostic index in BAC and adenosquamous carcinomas, although the patient numbers were too small to draw statistically significant conclusions. Finally, survival differences were calculated in patients with stage I and stage II squamous cell carcinoma and adenocarcinoma with and without TIL. In patients with stage I disease, the presence of TIL was associated with a significant survival advantage in squamous cell carcinoma (p ⫽ 0.03; Fig 3) but not in adenocarcinomas (p ⫽ 0.63; Fig 4). In patients with stage II disease, the presence of TIL was not associated with a survival difference in either group (squamous cell carcinomas, p ⫽ 0.86; adenocarcinomas, p ⫽ 0.65). From the analysis of our survival curves, the survival advantage increases with the duration of the follow-up,
Fig 1. Survival according to the presence or absence of tumor-infiltrating lymphocytes (TIL) in total population of patients with lung neoplasms who underwent resection (p ⫽ 0.20).
becoming more pronounced after 4 to 6 years from surgical resection (Figs 1– 4).
Comment The aim of the present study was to investigate the prevalence of TIL in our population of patients with lung
Fig 2. Survival according to the presence or absence of tumor-infiltrating lymphocytes (TIL) in population of patients for lung squamous cell carcinoma who underwent resection (p ⫽ 0.03).
Fig 3. Survival according to the presence or absence of tumor-infiltrating lymphocytes (TIL) in population of patients with stage I lung squamous cell carcinoma who underwent resection (p ⫽ 0.03).
neoplasms who underwent surgical resection, to examine possible correlations with other clinicopathologic variables, and to assess their prognostic significance. The retrospective nature of the study design represents a weakness and a limitation of the present study, and it has to be taken into account in interpretation of the results. The results of our study indicate that (1) TIL are not infrequent in lung neoplasms, being observed in approximately 25% of the patients; among different histologic types, they are exceedingly rare in neuroendocrine tumors, BAC, and adenosquamous carcinomas; (2) TIL are more frequent in poorly differentiated tumors and tumors with microscopic vascular invasion; and (3) TIL are an independent positive prognostic factor, and they are associated with a significant survival advantage for long-term survival in squamous cell carcinomas, particularly at an early stage. In these patients the survival advantage increases with the duration of follow-up, being most pronounced after 4 to 6 years of follow-up. Lymphocytes infiltrate many human tumors, and the presence of TIL has been considered in some studies a favorable prognostic variable [11]. The TIL population is composed of several subsets of lymphocytes, of which the main representatives are CD8⫹ T cells, CD4⫹ T cells, CD20⫹ B cells, and CD4⫹ CD25⫹ Treg cells. The pattern of distribution of TIL along and inside the tumor was classified by Naito and associates [2] in colorectal cancer into three groups: (1) those infiltrating within the cancer cell nests (intraepithelial TIL); (2) those distributed in the cancer stroma; (3) those present along the invasive margin (tumor– host interface). Other authors [8], however, were not able to demonstrate a similar distribution in NSCLC: in their series, TIL were predominantly observed
RUFFINI ET AL TUMOR-INFILTRATING LYMPHOCYTES IN LUNG NEOPLASMS
369
within the cancer stroma. In the present study, the most frequently observed lymphocytic infiltration pattern around the tumor included a peripheral lymphocytic infiltration composed of CD20⫹ B cells, CD3⫹ and CD4⫹ T cells organized in lymphoid follicles, a lymphocytic infiltration within cancer stroma (mainly CD3⫹ T cells), and intraepithelial lymphocytes within the cancer cell nests (true TIL in the present study) mostly represented by CD8⫹ T cells. Although in our population immunohistochemistry and TIL quantification were performed only on a sample of 21 patients as the main purpose of the study was the clinical impact of TIL on the population as a whole, our preliminary observations from immunohistochemistry are in accordance with the literature. The first result that comes from our study is that TIL are not infrequent in lung neoplasms, occurring in about one quarter of the population of patients who underwent surgical resection. Other authors found higher prevalences, ranging from 51% to 83% [12, 13]. It is of note that, differently from other series, we have not adopted a scoring system scale for the presence of TIL, but we used a negative–positive system with a cutoff of 20% of cells in an average of 10 fields. Tumor-infiltrating lymphocyte prevalence differs among histologic types: a higher prevalence was found in NSCLC versus neuroendocrine tumors, in which TIL were very infrequently detected (5%). Among NSCLC, adenocarcinomas, squamous cell carcinomas, and large cell anaplastic carcinomas have a similar prevalence, whereas BAC and adenosquamous carcinomas have a very low prevalence of TIL. Two studies in the literature about prevalence of TIL in different NSCLC report a higher prevalence of TIL in squamous cell carcinoma than adenocarcinoma [7, 12].
Fig 4. Survival according to the presence or absence of tumor-infiltrating lymphocytes (TIL) in population of patients with stage I lung adenocarcinoma who underwent resection (p ⫽ 0.63).
GENERAL THORACIC
Ann Thorac Surg 2009;87:365–72
GENERAL THORACIC
370
RUFFINI ET AL TUMOR-INFILTRATING LYMPHOCYTES IN LUNG NEOPLASMS
A second result that emerges from our study is that the presence of TIL correlates with some histopathologic characteristics: TIL were more frequent in poorly differentiated tumors and they were associated with the presence of microscopic vascular invasion. Mori and associates [7] found that in NSCLC the number of CD8⫹ T cells within cancer cell nests was related to the histologic type (large cell or squamous cell carcinoma ⬎ adenocarcinoma), and the degree of differentiation (undifferentiated type ⬎ differentiated type). This suggests that immune cell reactions are more pronounced as the tumor dedifferentiation and biologic malignant behavior progress. Similar results were obtained in NSCLC by other authors [8, 14] and in other human solid-organ tumors including colorectal cancers [15], kidney [5], and esophagus [3]. It seems evident from our and other studies, therefore, that there is a difference in immunogenicity in different histotypes of human cancers: in NSCLC, poorly differentiated cancer or squamous cell carcinomas may express larger amounts of tumorassociated antigens than other histotypes. The third result from our study is the prognostic impact of TIL on survival, which corroborates similar observations from other human solid-organ neoplasms. The favorable prognostic significance of the presence of TIL (mostly CD8⫹) has been reported in a variety of cancers, and according to most authors a larger number of TIL usually signified a strong immune reaction against the tumor, and hence a better prognosis. In colorectal carcinomas, CD8⫹ TIL was a favorable independent prognostic factor [2, 16], and similar results were observed in esophageal carcinomas [3], pancreatic cancers [6], gallbladder cancers [17], bile duct cancers [18], renal cell carcinomas [5], and prostatic adenocarcinomas [4]. By contrast, in NSCLC the prognostic role of TIL is still controversial. Mori and colleagues [7] found that the number of CD8⫹ TIL had no statistically significant impact on survival, and similar observations were made by other authors [14]. Hiraoka and colleagues [8] recently evaluated infiltration of CD8⫹ and CD4⫹ TIL in 109 patients who underwent resection for NSCLC and concluded that only the concurrent infiltration by CD8⫹ and CD4⫹ TIL significantly had an impact on survival. When examining their survival curves, however, a trend for a survival advantage in patients with either CD8⫹ and CD4⫹ TIL was evident. It seems therefore evident in their report that TIL indeed conferred a survival advantage. Petersen and associates [13] recently examined the relationship between Treg cells (which seem to promote tumor growth and metastases) and total TIL to determine a possible correlation with recurrence after resection for stage I NSCLC. They found that patients who have a higher proportion of Treg cells relative to total TIL had a significantly higher risk of recurrence, thus confirming the complexity of the host immune response to NSCLC. The reasons for these conflicting conclusions may partially result from the observation that the antitumor effect of TIL may be circumvented by various mechanisms within the tumor cells, including a lack of adequate T-cell costimulation [19], downregulation of cell-surface MHC class I ex-
Ann Thorac Surg 2009;87:365–72
pression [20], and secretion of immunosuppressive factors including transforming growth factor-beta [21]. The results of our study indicate that the presence of TIL in NSCLC is an independent prognostic factor in a multivariate analysis. The survival advantage associated with the presence of TIL is statistically significant in squamous cell carcinomas both overall and in stage I tumors. In a recent report, Ikeda and associates [12] found similarly in NSCLC that in squamous cell carcinomas the presence of CD8⫹ TIL was associated with a significantly better prognosis. Of particular importance in our survival analysis, we observed that the beneficial effect of TIL increases with the duration of the follow-up, being extremely evident after 4 to 6 years from surgery, particularly in stage I disease. Interestingly, similar findings have been reported in colorectal carcinoma [16]: the authors suggest that one of the effects of TIL, in addition to the antitumor effect against the primary tumor, is the immunosurveillance against the growth of micrometastasis. These considerations may partly explain the different pattern of relapse occurring in patients with earlystage lung carcinoma, in which there are some patients who never recur after resection, others who die shortly after resection of distant metastases, and those who did well after resection, but succumb after a few years of metastatic spread of the disease. It may be speculated that although a host response suppressing metastatic tumor growth is constantly at work, in some a tight surveillance is maintained throughout life, whereas in others tumor progression overwhelms the immune system. Tumor-infiltrating lymphocytes are only one mechanism the host uses to counteract tumor progression, and we are just starting to understand the complex interactions among the different characters playing a role in the human immune system. Strategies to effectively alter the delicate balance between protumoral versus antitumoral microenvironment in favor of promoting the antitumor effector T cells may be an exciting new way to enhance the effectiveness of immune responses against local and distant spread of solid tumors. In this respect the study of TIL infiltrates will surely play a critical role in the future to provide prognostic information and to select the appropriate immunotherapeutic strategy for each patient. In conclusion, our study indicates that TIL may be expected in approximately 25% of all lung neoplasms. Among different histotypes, they are exceedingly rare in neuroendocrine tumors, BAC ,and adenosquamous carcinomas. The presence of TIL correlates with microscopic vascular invasion and tumor dedifferentiation. The presence of TIL is associated with improved survival in squamous cell carcinomas, particularly at early stage, and it seems to reach the maximal survival advantage after 4 to 6 years of follow-up.
References 1. Jemal A, Murray T, Ward E, et al. Cancer statistics, 2005. CA Cancer J Clin 2005;55:10 –30. 2. Naito Y, Saito K, Shiiba K, et al. CD8⫹ T cells infiltrated within cancer cell nests as a prognostic factor in human colorectal cancer. Cancer Res 1998;58:3491– 4.
3. Schumacher K, Haensch W, Roefzaad C, et al. Prognostic significance of activated CD8⫹ T cell infiltrations within esophageal carcinomas. Cancer Res 2001;61:3932– 6. 4. Vesalainen S, Lipponen P, Talja M, et al. Histological grade, perineural infiltration, tumour-infiltrating lymphocytes and apoptosis as determinants of long-term prognosis in prostatic adenocarcinomas. Eur J Cancer 1994;30A:1797– 803. 5. Nakano O, Sato M, Naito Y, et al. Proliferative activity of intratumoral CD8⫹ T-lymphocytes as a prognostic factor in human renal cell carcinoma: clinicopathologic demonstration of antitumor immunity. Cancer Res 2001;61:5132– 6. 6. Fukunaga A, Miyamoto M, Cho Y, et al. CD8⫹ tumourinfiltrating lymphocytes together with CD4⫹ tumourinfiltrating lymphocytes and dendritic cells improve the prognosis of patients with pancreatic adenocarcinomas. Pancreas 2004;28:e26 –31. 7. Mori M, Ohtani H, Naito Y, et al. Infiltration of CD8⫹ T cells in non-small cell lung cancer is associated with dedifferentiation of cancer cells, but not with prognosis. Tohoku J Exp Med 2000;191:113– 8. 8. Hiraoka K, Miyamoto M, Cho Y, et al. Concurrent infiltration by CD8⫹ T cells and CD4⫹ T cells is a favourable prognostic factor in non-small cell carcinoma. Br J Cancer 2006;94:275– 80. 9. Sobin LH, Wittekind C (eds). International Union Against Cancer (UICC) TNM classification of malignant tumors, 6th ed. New York: Wiley-Liss, 2002;191:99 –103. 10. Cox DR. Regression models and life tablets (with discussion). J R Statist Soc Ser 1972;34:187–220. 11. Yu P, Yang-Xin F. Tumor-infiltrating T lymphocytes: friends or foes? Lab Invest 2006;86:231– 45. 12. Ikeda S, Funakoshi N, Inagaki M, et al. Clinicopathologic roles of tumor-infiltrating lymphocytes and CD8-positive lymphocytes in lung cancer imprint smears in squamous cell carcinoma and adenocarcinomas. Acta Cytol 2006;50:423–9.
RUFFINI ET AL TUMOR-INFILTRATING LYMPHOCYTES IN LUNG NEOPLASMS
371
13. Petersen RP, Campa MJ, Sperlazza MJ, et al. Tumor infiltrating FOXP3⫹ regulatory T cells are associated with recurrence in pathologic stage I NSCLC patients. Cancer 2006;107: 2866 –72. 14. Wakabayashi O, Yamazaki K, Oizumi S, et al. CD4⫹ T cells in cancer stroma, not CD8⫹ T cells in cancer cell nests, are associated with favorable prognosis in human non-small cell lung cancers. Cancer Sci 2003;94:1003–9. 15. Ohtani H. Focus on TILs: prognostic significance of tumorinfiltrating lymphocytes in human colorectal cancer. Cancer Immun 2007;7:4. 16. Chiba T, Ohtani H, Mizoi T, et al. Intraepithelial CD8⫹ T-cell count becomes a prognostic factor after a longer follow-up period in human colorectal carcinoma: possible association with suppression of micrometastasis. Br J Cancer 2004;91: 1711–7. 17. Nakakubo Y, Miyamoto M, Cho Y, et al. Clinical significance of immune cell infiltration with gallbladder cancer. Br J Cancer 2003;89:1736 – 42. 18. Oshikiri T, Miyamoto M, Shichinohe T. Prognostic value of intratumoral CD8⫹ T lymphocyte in extrahepatic bile duct carcinoma as essential immune response. J Surg Oncol 2003;84:224 – 8. 19. Melero I, Bach N, Chen L. Costimulation, tolerance and ignorance of cytolytic T lymphocytes in immune responses to tumour antigens. Life Sci 1997;60:2035– 41. 20. Imboden M, Murphy KR, Rakhmilevich AL, et al. The level of MHC class I expression on murine adenocarcinomas can change the antitumor effector mechanism of immunocytokine therapy. Cancer Res 2001;61:1500 –7. 21. Inge TH, Hoover SK, Susskind BM, et al. Inhibition of tumor-specific cytotoxic T-lymphocyte responses by transforming growth factor beta 1. Cancer Res 1992;52:1386 –92.
DISCUSSION DR DAVID H. HARPOLE (Durham, NC): Did you notice any correlation between the presence of necrosis or the presence of a histiocytic infiltrate and TILs?
The second question, yes, the necrosis, as I said, has a strong impulse to reject a lymphoid infiltrate. What is the first question?
DR RUFFINI: Yes. This is probably one of the reasons that in squamous cell carcinoma we found such a difference, because squamous cell carcinoma is very frequently associated with necrosis. It may be that necrosis freed some antigens that probably stimulate the lymphoid infiltrates, but our results on immunohistochemistry will probably shed some light on this.
DR TURNA: The threshold for the TIL positivity.
DR AKIF TURNA (Istanbul, Turkey): I want to congratulate you on this very outstanding study. I would like to ask three short questions, if I may. First, what is the threshold for any patient to be TIL-positive? What is the rate that you choose to be the threshold? Second, can you tell us the heterogeneity of TIL positivity? We found that there is a big heterogeneity in the tumor mass. When you look at the necrotic tumor mass, you can find very high rates of TIL, but the other sites could be TIL-negative. Third, how did you effect the pathologic examination? When you extract the tumor from the specimen, sometimes pathologists can get angry. Can you tell us about this phenomenon? DR RUFFINI: As for the third question, actually this was the standard histologic examination. We send the specimen to the pathologists, and the pathologists, in addition to all other variables, they give us the TIL, the presence of lymphoid infiltrate. So we don’t have to extract any tissue. It’s the pathologists that examine all the tissue, of course.
DR RUFFINI: It’s 20 per high-power field, 20 cells for the high-power field in 10 fields. DR MARK I. BLOCK (Hollywood, FL): I have a couple of questions. First, I think it’s important to be careful about what we mean by TIL. I think a lot of us think of TIL as lymphocytes that are extracted from the tumor and grown up to treat melanoma patients the way Rosenberg first did it. So I think when you just look at lymphocytes in a specimen, it’s important to be able to define exactly what we’re talking about. So you did some special stains. Are these B cells, T cells, CD4 cells, CD8 cells? Do you have a sense about what population of lymphocytes you’re talking about? Second, it seemed a little bit strange that you had a higher correlation of TIL with invasion and degree of differentiation, but that also seemed to correlate with improved survival. At least I think that’s the way that you showed the data. How do you explain that? DR RUFFINI: I think the answer to this question is that we just see the TIL population as a whole by standard histopathology, but, as I said, the TIL population is composed of different subsets and not all these populations act against tumor growth. There are some acting against, some acting pro. And we did immuno-
GENERAL THORACIC
Ann Thorac Surg 2009;87:365–72
GENERAL THORACIC
372
RUFFINI ET AL TUMOR-INFILTRATING LYMPHOCYTES IN LUNG NEOPLASMS
histochemistry only on very few cases. But the message I would like is just to see clinically at conventional histopathology what the significance of TIL is. Of course, this is just the preliminary results. The answer will be immunohistochemistry just to differentiate the different subpopulations. CD8 acts, of course, against tumor growth, but CD4, for example, sometimes acts pro and sometimes acts against tumor growth. So I know that this could be a limitation of the study, but this was not the aim of our study. Our study was just to examine the TIL population as a whole, I mean just in a conventional histopathology. DR PETER GOLDSTRAW (London, UK): We have recently found that a simple peripheral blood neutrophil/lymphocyte
Ann Thorac Surg 2009;87:365–72
ratio is an independent prognostic factor. Did you look at peripheral blood lymphocyte counts in your TILs to see whether there was any correlation? DR RUFFINI: No. I have read about this, but the focus of our study was just to see the lymphocytic infiltrate in the specimen, not in the peripheral blood. DR GOLDSTRAW: You can go back and do that very easily as the data would be in the case record. DR RUFFINI: Yes. That’s a good suggestion. Thank you.
Clinical Significance of Tumor-Infiltrating Lymphocytes in Lung Neoplasms Enrico Ruffini, MD, Sofia Asioli, MD, Pier Luigi Filosso, MD, Paraskevas Lyberis, MD, Maria Cristina Bruna, MD, Luigia Macrì, MD, Lorenzo Daniele, MD, and Alberto Oliaro, MD Departments of Thoracic Surgery and Pathology, University of Torino, Torino, Italy
Background. Tumor-infiltrating lymphocytes (TIL) are considered important in anticancer immunosurveillance, although their role has not been clearly established yet. We examined prevalence, correlations, and prognostic significance of TIL among our patient population of resected lung neoplasms. Methods. From 1993 to 2006, the presence of TIL was retrospectively evaluated in 1,290 patients operated on for primary lung neoplasms. Tumor-infiltrating lymphocytes were defined as those intraepithelial lymphocytes located within the cancer cell nests. Results. Tumor-infiltrating lymphocytes were detected in 294 patients (23%). A significant difference was found between prevalence in non–small cell lung carcinomas versus neuroendocrine tumors (290 of 1,208, 24% versus 4 of 82, 5%; p ⴝ 0.0001). Prevalence was similar in adenocarcinomas, squamous-cell carcinomas, and large-cell anaplastic carcinomas. Logistic regression analysis indicates that TIL correlate with grading (odds ratio, 1.27; 95% confidence interval, 1.04 to 1.55; p ⴝ 0.02), tumor dimension (odds ratio, 0.86; 95% confidence interval, 0.79 to
0.94; p ⴝ 0.0008), and vascular invasion (odds ratio, 1.62; 95% confidence interval, 1.21 to 2.16; p ⴝ 0.0009). A not significantly better survival in the presence of TIL was observed overall (p ⴝ 0.20), becoming significant in squamous-cell carcinomas (p ⴝ 0.03). In patients with stage I disease, TIL is associated with a significant survival advantage in squamous-cell carcinomas (p ⴝ 0.03). The survival advantage increases with the duration of follow-up and is more evident after 4 to 6 years. Conclusions. Tumor-infiltrating lymphocytes are observed in about one fourth of resected lung neoplasms: they are rare in neuroendocrine tumors. Tumor-infiltrating lymphocytes are more frequent in poorly differentiated tumors and in tumors with microscopic vascular invasion. The presence of TIL correlates with an improved survival in squamous cell carcinomas, particularly at early stage. The survival advantage increases with the duration of follow-up.
L
role in anticancer immunosurveillance. Among the mature human T-cell population, three subtypes seem to play a major role in immunosurveillance against tumor growth: CD8⫹, CD4⫹, and T regulatory (Treg) cells. Cytotoxic (CD8⫹) T cells are the major subset of T cells that constructively mediate an effective antitumor response: they are able to induce tumor killing on direct recognition of peptide antigens presented by the tumor’s MHC class I molecules. CD4⫹ T cells also migrate to the tumor site, but their role in antitumor activity is dualistic: some CD4⫹ T cells seemingly hinder the effector function of CD8⫹ T cells and therefore indirectly promote tumor growth whereas other subtypes of CD4⫹ T cells might aid in the activation of CD8⫹ T cells. Finally, a third subpopulation of T cells, namely the Treg cells, has recently been identified as a subtype of CD4⫹ T cells (also known as naturally occurring CD4⫹ CD25⫹ T regulatory cells); Treg cells may downregulate the immune response to tumors by attenuating the host’s anti-
ung cancer is the leading cause of cancer-related deaths among men and women. Despite recent improvement in early diagnosis and treatment, the overall 5-year survival rate is approximately 15% [1], well below that of other solid-organ tumors. Given the high recurrence rate of the tumor, many authors have looked for clinicopathologic markers that can reliably be used as prognostic indicators, as well as a guide to identify patients likely to benefit from adjunct therapies. In a variety of human solid tumors, tumor-infiltrating lymphocytes (TIL) are considered to play an important
Accepted for publication Oct 16, 2008. Presented at the Forty-fourth Annual Meeting of The Society of Thoracic Surgeons, Fort Lauderdale, FL, Jan 28 –30, 2008. Address correspondence to Dr Ruffini, Thoracic Surgery Service, University of Torino, Ospedale Molinette, Dipartimento di Fisiopatologia Clinica, Sezione di Chirurgia Toracica, 3, Via Genova, Torino, 10126, Italy; e-mail: [email protected].
© 2009 by The Society of Thoracic Surgeons Published by Elsevier Inc
(Ann Thorac Surg 2009;87:365–72) © 2009 by The Society of Thoracic Surgeons
0003-4975/09/$36.00 doi:10.1016/j.athoracsur.2008.10.067
GENERAL THORACIC
ORIGINAL ARTICLES: GENERAL THORACIC
GENERAL THORACIC
366
RUFFINI ET AL TUMOR-INFILTRATING LYMPHOCYTES IN LUNG NEOPLASMS
tumor T cells, thus allowing unrestricted tumor growth and inhibiting systemic immunosurveillance. Definition of TIL varies among different series. In the present study, TIL were considered as those intraepithelial infiltrating lymphocytes located within the cancer cell nests; phenotypically they were almost entirely CD8⫹ T lymphocytes. Although the favorable prognostic significance of TIL has been reported for a variety of human solid cancers, including colorectal cancer [2], esophageal cancer [3], prostatic cancer [4], renal cell carcinoma [5], and pancreatic cancer [6], the role of TIL in non–small cell lung cancer (NSCLC) is still debatable. Some authors have reported that the presence of TIL gives no survival advantage [7], whereas other authors reached the opposite conclusion [8]. The aim of the present study was to investigate the prevalence of TIL in our population of patients with lung neoplasms who underwent surgical resection, to examine possible correlations with other clinicopathologic variables, and finally to assess their prognostic significance, particularly in early stage (stage I) disease.
Patients and Methods Approval for this study was provided by our institutional review board, and patient consent was waived. A retrospective review was undertaken among our population of patients with primary lung neoplasms operated on at our institution between 1993 and 2006. In this period, of a total of 2,200 patients, 1,290 were evaluated for the presence of TIL on the surgical specimen. The remaining 910 were not considered for the analysis either as a result of fulfilling the exclusion criteria (see below) or because of incomplete pathologic information.
Preoperative Evaluation Baseline demographics, histopathologic data, overall survival, and pathologic specimens preserved in paraffin were available for all patients. Patients receiving preoperative induction therapy were excluded. On the basis of above clinicopathologic data and the presence or absence of TIL, defined as CD8⫹ intraepithelial lymphocytes morphologically identified within the cancer cell nests, patients were collected in two different group, TIL(⫹) and TIL(⫺). Patient demographics and tumor features in the two patient groups are detailed in Table 1. All patients were staged postsurgically according to the most recent TNM staging system [9]. For the analysis, the following histopathologic variables were considered: histologic grading of differentiation (well, moderately, and poorly differentiated, respectively G1, G2, and G3), tumor size, microscopic vascular invasion, and microscopic perineural invasion. Follow-up information of all patients was obtained through clinic follow-up notes, direct patient or family contact, or contact with the patient’s primary care physician.
Ann Thorac Surg 2009;87:365–72
Table 1. Prevalence Differences of Clinicopathologic Variables Between Patients With and Without Tumor-Infiltrating Lymphocytes on the Surgical Specimen Variable Patients Sex Male Female Age (y), mean (range) Histology Adenocarcinoma Squamous cell Adenosquamous Large cell anaplastic BAC Neuroendocrine Grading G1 G2 G3 Vascular invasion Yes No Perineural invasion Yes No Tumor dimension Stage I II IIIa N factor N0 N1 N2
TIL(⫹)
TI(⫺)
294
996
253 41 64 (43–80)
831 165 64 (25–82)
p Value
0.28
0.36b
133 (27%) 134 (24%) 2 (9%) 14 (25%) 7 (8%) 4 (5%)
362 (73%) 415 (76%) 20 (91%) 42 (75%) 79 (92%) 78 (95%)
30 (18%) 145 (28%) 145 (28%)
143 (82%) 368 (72%) 362 (72%)
0.002c
164 (30%) 130 (18%)
395 (70%) 601 (82%)
0.0001
30 (22%) 264 (23%) 3.8
107 (78%) 889 (77%) 4.2
0.8
165 (23%) 57 (21%) 49 (23%)
549 (77%) 208 (79%) 165 (77%)
0.9
188 (22%) 63 (28%) 42 (21%)
670 (78%) 161 (72%) 162 (79%)
0.7
Versus adenocarcinoma ⫹ squamous cell. c G1 versus G2 ⫹ G3. cell.
a
b
0.08a 0.92a 0.0003a 0.0001a
0.03
Versus squamous
TIL ⫽ tumor-infiltrating lymphocytes.
Immunohistochemistry and Quantification of Tumor-Infiltrating Lymphocytes The main objective of the present study was a descriptive and prognostic evaluation of TIL, and for that purpose the TIL population was considered as a whole for the clinical analysis on 1,290 patients. However, to better understand the relative importance of TIL subpopulations and to see whether our TIL population distribution were in accordance with that reported in the literature, immunohistochemistry and quantification of TIL were performed on a sample of 21 patients (7 adenocarcinomas, 7 squamous cell carcinomas, 3 bronchioloalveolar carcinoma (BAC), 1 typical carcinoid, 1 atypical carcinoid, 1 large cell neuroendocrine carcinoma, 1 small cell lung carcinoma). Immunohistochemical analysis was performed on resected, paraffin-embedded lung cancer tissues. After microtome sectioning (4 m) and slide labeling, immuno-
histochemical analysis was performed using the Ventana automated immunostainer (BenchMark AutoStainer; Ventana Medical Systems, Tucson, AZ) with primary antibodies directed to the T-cell and B-cell surface glycoprotein CD3 (DAKO, Glostrup, Denmark; antigen retrieval with citrate buffer pH 6, dilution 1:150), CD4 (Novocastra, Newcastle, United Kingdom; antigen retrieval with citrate buffer pH 6, dilution 1:20), CD8 (DAKO; antigen retrieval with citrate buffer pH 6, dilution 1:50), and CD20 (DAKO; antigen retrieval with citrate buffer pH 6, dilution 1:200), respectively. Each entire tumor section was evaluated for TIL by using a ⫻40 objective lens and 10 independent fields, with the most abundant TIL selected, digitally photographed at a size of 0.0625 mm2, and counted manually by two different pathologists (A.S., L.D.) blinded to patients’ outcomes. Immunohistochemical staining was define as negative (0 ⫽ no staining or positive staining in ⬍20% of cells) or positive (1 ⫽ positive staining in ⱖ20% of cells) for each patient. This dichotomous classification was adopted because it has customarily been used in our pathology report since 1993. Areas near necrosis were excluded from the analysis. In the present study, TIL were considered as those intraepithelial infiltrating lymphocytes located within the cancer cell nests; phenotypically they were almost entirely CD8⫹ T lymphocytes.
Statistical Analysis Comparison between proportion differences was undertaken using 2 test (Fisher’s exact test when appropriate). Survival was defined as the time from surgery to death or last follow-up. Survival curves were estimated with the Kaplan–Meier method. Significance between survival curves was assessed with the log-rank test. A probability value less than 0.05 was considered to be significant. The Cox proportional hazard model [10] was applied for identifying prognostic factors. Hazard ratios along with 95% confidence intervals (95% CI) were provided. To investigate the relationship between some possible causal (independent) variables and the presence or absence of TIL (considered as a binary output variable), a logistic regression model was used, with the assumption that the events were independent and the relationship plausibly log-linear. The probability value was calculated on the Wald statistic, and 95% CI were provided. All statistical analysis was undertaken using software packages (STATISTICA, release 7.1, 2005; StatSoft, Vigonza, PD, Italy).
Results Overall, TIL were detected on the surgical specimen in 294 patients (23%) operated on for primary lung neoplasms. The typical lymphocytic infiltration pattern around the tumor as emerged from our samples can be described as follows: (1) a peripheral lymphocytic infiltration composed of prevalent B lymphocytes (CD20⫹) organized in lymphoid follicles mixed with T lymphocytes (both CD3⫹ and CD4⫹); (2) a lymphocytic infiltra-
RUFFINI ET AL TUMOR-INFILTRATING LYMPHOCYTES IN LUNG NEOPLASMS
367
tion within cancer stroma (mainly CD3⫹ T lymphocytes); and (3) intraepithelial lymphocytes within the cancer cell nests mostly represented by CD8⫹ T lymphocytes (the true TIL population for the present study). In addition, lymphocytic infiltration along the invasive margin is common, and areas within the cancer are observed usually sparsely in the lymphocytic response as a result of immune evasion mechanisms.
Prevalence of Tumor-Infiltrating Lymphocytes Prevalence of TIL among the different neoplasms was as follows (Table 1): adenocarcinoma, 27% (133 of 495); squamous cell carcinoma, 24% (134 of 549); BAC, 8% (7 of 86); typical carcinoid, 0% (0 of 34); atypical carcinoid, 7% (1 of 14); adenosquamous carcinoma, 9% (2 of 22); large cell anaplastic carcinoma, 25% (14 of 56); large cell neuroendocrine carcinoma, 19% (3 of 16); and small cell carcinoma, 0% (0 of 18). A significant difference was found between prevalence in NSCLC and neuroendocrine carcinomas (290 of 1,208, 24% versus 4 of 82, 5%; p ⫽ 0.0001). Among NSCLC, prevalence was similar in adenocarcinoma, squamous cell carcinoma, and large cell anaplastic carcinoma; conversely, adenosquamous carcinoma and BAC had a lower prevalence of TIL, which was significant in BAC (p ⫽ 0.0003). The prevalence of TIL was also investigated in association with different clinicopathologic variables including grading, microscopic vascular and perineural invasion, tumor dimension, stage, and TNM factor. The presence of TIL was associated with poorly differentiated tumors, with tumors of small size, and with the presence of microscopic vascular invasion (Table 1). We further analyzed the two populations of patients with adenocarcinoma and squamous cell carcinoma with regard to the presence or absence of TIL. In adenocarcinomas, TIL prevalence among patients with microscopic vascular invasion was very high (35%), whereas no difference was found in TIL prevalence among different grades of differentiation (G1, G2, and G3) or tumor dimension. Conversely, in squamous cell carcinomas, TIL prevalence was less associated with microscopic vascular invasion, whereas TIL were found more frequently in poorly differentiated tumors. It therefore seems evident from our data that the presence of TIL in adenocarcinoma and squamous cell carcinoma might have a different presentation and maybe a different significance as resulted from our survival analysis (see later).
Risk Factors for Tumor-Infiltrating Lymphocytes A logistic regression analysis was undertaken using the presence or absence of TIL as a binary dependent variable and the following clinicopathologic independent variables: tumor dimension, grading, vascular invasion, and perineural invasion (Table 2). Of all the independent variables, a significant correlation was found with grading (odds ratio [OR], 1.27; 95% CI, 1.04 to 1.55; p ⫽ 0.02), tumor dimension (OR, 0.86; 95% CI, 0.79 to 0.94; p ⫽ 0.0008), and vascular invasion (OR, 1.62; 95% CI, 1.21 to 2.16; p ⫽ 0.0009).
GENERAL THORACIC
Ann Thorac Surg 2009;87:365–72
GENERAL THORACIC
368
RUFFINI ET AL TUMOR-INFILTRATING LYMPHOCYTES IN LUNG NEOPLASMS
Ann Thorac Surg 2009;87:365–72
Table 2. Logistic Regression Analysis for Tumor-Infiltrating Lymphocytes in Population of Patients With Lung Neoplasms Variable Overall Vascular invasion Grading Tumor dimension Adenocarcinoma Vascular invasion Squamous cell carcinoma Grading Tumor dimension
Odds Ratio
95% Confidence Interval
p Value
1.62 1.27 0.86
1.21–2.16 1.04–1.55 0.79–0.94
0.0009 0.02 0.0008
2.33
1.46–3.71
0.0004
1.53 0.79
1.05–2.23 0.69–0.91
0.03 0.0007
In adenocarcinomas, TIL was associated only with vascular invasion (OR, 2.33; 95% CI, 1.46 to 3.71; p ⫽ 0.0004); in squamous cell carcinomas, TIL was associated with grading (OR, 1.53; 95% CI, 1.05 to 2.23; p ⫽ 0.03) and tumor dimension (OR, 0.79; 95% CI, 0.69 to 0.91; p ⫽ 0.0007).
Prognostic Significance of Tumor-Infiltrating Lymphocytes Prognostic significance was evaluated first by performing a multivariate survival analysis on different independent variables including TIL, vascular invasion, perineural invasion, TNM factors, and stage. Among other covariates, TIL did have an independent prognostic significance overall (hazard ratio, 0.78; 95% CI, 0.64 to 0.98; p ⫽ 0.05) and in squamous cell carcinomas (hazard ratio, 0.68; 95% CI, 0.50 to 0.91; p ⫽ 0.02), but not in adenocarcinomas (hazard ratio, 0.87; 95% CI, 0.64 to 1.18; p ⫽ 0.38). We then compared survivals among patients with different neoplasms according to the presence or absence of TIL. A nonsignificant survival difference between TIL(⫹) and TIL(⫺) groups was observed overall (p ⫽ 0.20; Fig 1) and in adenocarcinomas (p ⫽ 0.6), large cell anaplastic carcinomas (p ⫽ 0.91), and neuroendocrine tumors (p ⫽ 0.39). The survival difference was statistically significant in squamous cell carcinomas (p ⫽ 0.03; Fig 2). Tumor-infiltrating lymphocytes were a negative prognostic index in BAC and adenosquamous carcinomas, although the patient numbers were too small to draw statistically significant conclusions. Finally, survival differences were calculated in patients with stage I and stage II squamous cell carcinoma and adenocarcinoma with and without TIL. In patients with stage I disease, the presence of TIL was associated with a significant survival advantage in squamous cell carcinoma (p ⫽ 0.03; Fig 3) but not in adenocarcinomas (p ⫽ 0.63; Fig 4). In patients with stage II disease, the presence of TIL was not associated with a survival difference in either group (squamous cell carcinomas, p ⫽ 0.86; adenocarcinomas, p ⫽ 0.65). From the analysis of our survival curves, the survival advantage increases with the duration of the follow-up,
Fig 1. Survival according to the presence or absence of tumor-infiltrating lymphocytes (TIL) in total population of patients with lung neoplasms who underwent resection (p ⫽ 0.20).
becoming more pronounced after 4 to 6 years from surgical resection (Figs 1– 4).
Comment The aim of the present study was to investigate the prevalence of TIL in our population of patients with lung
Fig 2. Survival according to the presence or absence of tumor-infiltrating lymphocytes (TIL) in population of patients for lung squamous cell carcinoma who underwent resection (p ⫽ 0.03).
Fig 3. Survival according to the presence or absence of tumor-infiltrating lymphocytes (TIL) in population of patients with stage I lung squamous cell carcinoma who underwent resection (p ⫽ 0.03).
neoplasms who underwent surgical resection, to examine possible correlations with other clinicopathologic variables, and to assess their prognostic significance. The retrospective nature of the study design represents a weakness and a limitation of the present study, and it has to be taken into account in interpretation of the results. The results of our study indicate that (1) TIL are not infrequent in lung neoplasms, being observed in approximately 25% of the patients; among different histologic types, they are exceedingly rare in neuroendocrine tumors, BAC, and adenosquamous carcinomas; (2) TIL are more frequent in poorly differentiated tumors and tumors with microscopic vascular invasion; and (3) TIL are an independent positive prognostic factor, and they are associated with a significant survival advantage for long-term survival in squamous cell carcinomas, particularly at an early stage. In these patients the survival advantage increases with the duration of follow-up, being most pronounced after 4 to 6 years of follow-up. Lymphocytes infiltrate many human tumors, and the presence of TIL has been considered in some studies a favorable prognostic variable [11]. The TIL population is composed of several subsets of lymphocytes, of which the main representatives are CD8⫹ T cells, CD4⫹ T cells, CD20⫹ B cells, and CD4⫹ CD25⫹ Treg cells. The pattern of distribution of TIL along and inside the tumor was classified by Naito and associates [2] in colorectal cancer into three groups: (1) those infiltrating within the cancer cell nests (intraepithelial TIL); (2) those distributed in the cancer stroma; (3) those present along the invasive margin (tumor– host interface). Other authors [8], however, were not able to demonstrate a similar distribution in NSCLC: in their series, TIL were predominantly observed
RUFFINI ET AL TUMOR-INFILTRATING LYMPHOCYTES IN LUNG NEOPLASMS
369
within the cancer stroma. In the present study, the most frequently observed lymphocytic infiltration pattern around the tumor included a peripheral lymphocytic infiltration composed of CD20⫹ B cells, CD3⫹ and CD4⫹ T cells organized in lymphoid follicles, a lymphocytic infiltration within cancer stroma (mainly CD3⫹ T cells), and intraepithelial lymphocytes within the cancer cell nests (true TIL in the present study) mostly represented by CD8⫹ T cells. Although in our population immunohistochemistry and TIL quantification were performed only on a sample of 21 patients as the main purpose of the study was the clinical impact of TIL on the population as a whole, our preliminary observations from immunohistochemistry are in accordance with the literature. The first result that comes from our study is that TIL are not infrequent in lung neoplasms, occurring in about one quarter of the population of patients who underwent surgical resection. Other authors found higher prevalences, ranging from 51% to 83% [12, 13]. It is of note that, differently from other series, we have not adopted a scoring system scale for the presence of TIL, but we used a negative–positive system with a cutoff of 20% of cells in an average of 10 fields. Tumor-infiltrating lymphocyte prevalence differs among histologic types: a higher prevalence was found in NSCLC versus neuroendocrine tumors, in which TIL were very infrequently detected (5%). Among NSCLC, adenocarcinomas, squamous cell carcinomas, and large cell anaplastic carcinomas have a similar prevalence, whereas BAC and adenosquamous carcinomas have a very low prevalence of TIL. Two studies in the literature about prevalence of TIL in different NSCLC report a higher prevalence of TIL in squamous cell carcinoma than adenocarcinoma [7, 12].
Fig 4. Survival according to the presence or absence of tumor-infiltrating lymphocytes (TIL) in population of patients with stage I lung adenocarcinoma who underwent resection (p ⫽ 0.63).
GENERAL THORACIC
Ann Thorac Surg 2009;87:365–72
GENERAL THORACIC
370
RUFFINI ET AL TUMOR-INFILTRATING LYMPHOCYTES IN LUNG NEOPLASMS
A second result that emerges from our study is that the presence of TIL correlates with some histopathologic characteristics: TIL were more frequent in poorly differentiated tumors and they were associated with the presence of microscopic vascular invasion. Mori and associates [7] found that in NSCLC the number of CD8⫹ T cells within cancer cell nests was related to the histologic type (large cell or squamous cell carcinoma ⬎ adenocarcinoma), and the degree of differentiation (undifferentiated type ⬎ differentiated type). This suggests that immune cell reactions are more pronounced as the tumor dedifferentiation and biologic malignant behavior progress. Similar results were obtained in NSCLC by other authors [8, 14] and in other human solid-organ tumors including colorectal cancers [15], kidney [5], and esophagus [3]. It seems evident from our and other studies, therefore, that there is a difference in immunogenicity in different histotypes of human cancers: in NSCLC, poorly differentiated cancer or squamous cell carcinomas may express larger amounts of tumorassociated antigens than other histotypes. The third result from our study is the prognostic impact of TIL on survival, which corroborates similar observations from other human solid-organ neoplasms. The favorable prognostic significance of the presence of TIL (mostly CD8⫹) has been reported in a variety of cancers, and according to most authors a larger number of TIL usually signified a strong immune reaction against the tumor, and hence a better prognosis. In colorectal carcinomas, CD8⫹ TIL was a favorable independent prognostic factor [2, 16], and similar results were observed in esophageal carcinomas [3], pancreatic cancers [6], gallbladder cancers [17], bile duct cancers [18], renal cell carcinomas [5], and prostatic adenocarcinomas [4]. By contrast, in NSCLC the prognostic role of TIL is still controversial. Mori and colleagues [7] found that the number of CD8⫹ TIL had no statistically significant impact on survival, and similar observations were made by other authors [14]. Hiraoka and colleagues [8] recently evaluated infiltration of CD8⫹ and CD4⫹ TIL in 109 patients who underwent resection for NSCLC and concluded that only the concurrent infiltration by CD8⫹ and CD4⫹ TIL significantly had an impact on survival. When examining their survival curves, however, a trend for a survival advantage in patients with either CD8⫹ and CD4⫹ TIL was evident. It seems therefore evident in their report that TIL indeed conferred a survival advantage. Petersen and associates [13] recently examined the relationship between Treg cells (which seem to promote tumor growth and metastases) and total TIL to determine a possible correlation with recurrence after resection for stage I NSCLC. They found that patients who have a higher proportion of Treg cells relative to total TIL had a significantly higher risk of recurrence, thus confirming the complexity of the host immune response to NSCLC. The reasons for these conflicting conclusions may partially result from the observation that the antitumor effect of TIL may be circumvented by various mechanisms within the tumor cells, including a lack of adequate T-cell costimulation [19], downregulation of cell-surface MHC class I ex-
Ann Thorac Surg 2009;87:365–72
pression [20], and secretion of immunosuppressive factors including transforming growth factor-beta [21]. The results of our study indicate that the presence of TIL in NSCLC is an independent prognostic factor in a multivariate analysis. The survival advantage associated with the presence of TIL is statistically significant in squamous cell carcinomas both overall and in stage I tumors. In a recent report, Ikeda and associates [12] found similarly in NSCLC that in squamous cell carcinomas the presence of CD8⫹ TIL was associated with a significantly better prognosis. Of particular importance in our survival analysis, we observed that the beneficial effect of TIL increases with the duration of the follow-up, being extremely evident after 4 to 6 years from surgery, particularly in stage I disease. Interestingly, similar findings have been reported in colorectal carcinoma [16]: the authors suggest that one of the effects of TIL, in addition to the antitumor effect against the primary tumor, is the immunosurveillance against the growth of micrometastasis. These considerations may partly explain the different pattern of relapse occurring in patients with earlystage lung carcinoma, in which there are some patients who never recur after resection, others who die shortly after resection of distant metastases, and those who did well after resection, but succumb after a few years of metastatic spread of the disease. It may be speculated that although a host response suppressing metastatic tumor growth is constantly at work, in some a tight surveillance is maintained throughout life, whereas in others tumor progression overwhelms the immune system. Tumor-infiltrating lymphocytes are only one mechanism the host uses to counteract tumor progression, and we are just starting to understand the complex interactions among the different characters playing a role in the human immune system. Strategies to effectively alter the delicate balance between protumoral versus antitumoral microenvironment in favor of promoting the antitumor effector T cells may be an exciting new way to enhance the effectiveness of immune responses against local and distant spread of solid tumors. In this respect the study of TIL infiltrates will surely play a critical role in the future to provide prognostic information and to select the appropriate immunotherapeutic strategy for each patient. In conclusion, our study indicates that TIL may be expected in approximately 25% of all lung neoplasms. Among different histotypes, they are exceedingly rare in neuroendocrine tumors, BAC ,and adenosquamous carcinomas. The presence of TIL correlates with microscopic vascular invasion and tumor dedifferentiation. The presence of TIL is associated with improved survival in squamous cell carcinomas, particularly at early stage, and it seems to reach the maximal survival advantage after 4 to 6 years of follow-up.
References 1. Jemal A, Murray T, Ward E, et al. Cancer statistics, 2005. CA Cancer J Clin 2005;55:10 –30. 2. Naito Y, Saito K, Shiiba K, et al. CD8⫹ T cells infiltrated within cancer cell nests as a prognostic factor in human colorectal cancer. Cancer Res 1998;58:3491– 4.
3. Schumacher K, Haensch W, Roefzaad C, et al. Prognostic significance of activated CD8⫹ T cell infiltrations within esophageal carcinomas. Cancer Res 2001;61:3932– 6. 4. Vesalainen S, Lipponen P, Talja M, et al. Histological grade, perineural infiltration, tumour-infiltrating lymphocytes and apoptosis as determinants of long-term prognosis in prostatic adenocarcinomas. Eur J Cancer 1994;30A:1797– 803. 5. Nakano O, Sato M, Naito Y, et al. Proliferative activity of intratumoral CD8⫹ T-lymphocytes as a prognostic factor in human renal cell carcinoma: clinicopathologic demonstration of antitumor immunity. Cancer Res 2001;61:5132– 6. 6. Fukunaga A, Miyamoto M, Cho Y, et al. CD8⫹ tumourinfiltrating lymphocytes together with CD4⫹ tumourinfiltrating lymphocytes and dendritic cells improve the prognosis of patients with pancreatic adenocarcinomas. Pancreas 2004;28:e26 –31. 7. Mori M, Ohtani H, Naito Y, et al. Infiltration of CD8⫹ T cells in non-small cell lung cancer is associated with dedifferentiation of cancer cells, but not with prognosis. Tohoku J Exp Med 2000;191:113– 8. 8. Hiraoka K, Miyamoto M, Cho Y, et al. Concurrent infiltration by CD8⫹ T cells and CD4⫹ T cells is a favourable prognostic factor in non-small cell carcinoma. Br J Cancer 2006;94:275– 80. 9. Sobin LH, Wittekind C (eds). International Union Against Cancer (UICC) TNM classification of malignant tumors, 6th ed. New York: Wiley-Liss, 2002;191:99 –103. 10. Cox DR. Regression models and life tablets (with discussion). J R Statist Soc Ser 1972;34:187–220. 11. Yu P, Yang-Xin F. Tumor-infiltrating T lymphocytes: friends or foes? Lab Invest 2006;86:231– 45. 12. Ikeda S, Funakoshi N, Inagaki M, et al. Clinicopathologic roles of tumor-infiltrating lymphocytes and CD8-positive lymphocytes in lung cancer imprint smears in squamous cell carcinoma and adenocarcinomas. Acta Cytol 2006;50:423–9.
RUFFINI ET AL TUMOR-INFILTRATING LYMPHOCYTES IN LUNG NEOPLASMS
371
13. Petersen RP, Campa MJ, Sperlazza MJ, et al. Tumor infiltrating FOXP3⫹ regulatory T cells are associated with recurrence in pathologic stage I NSCLC patients. Cancer 2006;107: 2866 –72. 14. Wakabayashi O, Yamazaki K, Oizumi S, et al. CD4⫹ T cells in cancer stroma, not CD8⫹ T cells in cancer cell nests, are associated with favorable prognosis in human non-small cell lung cancers. Cancer Sci 2003;94:1003–9. 15. Ohtani H. Focus on TILs: prognostic significance of tumorinfiltrating lymphocytes in human colorectal cancer. Cancer Immun 2007;7:4. 16. Chiba T, Ohtani H, Mizoi T, et al. Intraepithelial CD8⫹ T-cell count becomes a prognostic factor after a longer follow-up period in human colorectal carcinoma: possible association with suppression of micrometastasis. Br J Cancer 2004;91: 1711–7. 17. Nakakubo Y, Miyamoto M, Cho Y, et al. Clinical significance of immune cell infiltration with gallbladder cancer. Br J Cancer 2003;89:1736 – 42. 18. Oshikiri T, Miyamoto M, Shichinohe T. Prognostic value of intratumoral CD8⫹ T lymphocyte in extrahepatic bile duct carcinoma as essential immune response. J Surg Oncol 2003;84:224 – 8. 19. Melero I, Bach N, Chen L. Costimulation, tolerance and ignorance of cytolytic T lymphocytes in immune responses to tumour antigens. Life Sci 1997;60:2035– 41. 20. Imboden M, Murphy KR, Rakhmilevich AL, et al. The level of MHC class I expression on murine adenocarcinomas can change the antitumor effector mechanism of immunocytokine therapy. Cancer Res 2001;61:1500 –7. 21. Inge TH, Hoover SK, Susskind BM, et al. Inhibition of tumor-specific cytotoxic T-lymphocyte responses by transforming growth factor beta 1. Cancer Res 1992;52:1386 –92.
DISCUSSION DR DAVID H. HARPOLE (Durham, NC): Did you notice any correlation between the presence of necrosis or the presence of a histiocytic infiltrate and TILs?
The second question, yes, the necrosis, as I said, has a strong impulse to reject a lymphoid infiltrate. What is the first question?
DR RUFFINI: Yes. This is probably one of the reasons that in squamous cell carcinoma we found such a difference, because squamous cell carcinoma is very frequently associated with necrosis. It may be that necrosis freed some antigens that probably stimulate the lymphoid infiltrates, but our results on immunohistochemistry will probably shed some light on this.
DR TURNA: The threshold for the TIL positivity.
DR AKIF TURNA (Istanbul, Turkey): I want to congratulate you on this very outstanding study. I would like to ask three short questions, if I may. First, what is the threshold for any patient to be TIL-positive? What is the rate that you choose to be the threshold? Second, can you tell us the heterogeneity of TIL positivity? We found that there is a big heterogeneity in the tumor mass. When you look at the necrotic tumor mass, you can find very high rates of TIL, but the other sites could be TIL-negative. Third, how did you effect the pathologic examination? When you extract the tumor from the specimen, sometimes pathologists can get angry. Can you tell us about this phenomenon? DR RUFFINI: As for the third question, actually this was the standard histologic examination. We send the specimen to the pathologists, and the pathologists, in addition to all other variables, they give us the TIL, the presence of lymphoid infiltrate. So we don’t have to extract any tissue. It’s the pathologists that examine all the tissue, of course.
DR RUFFINI: It’s 20 per high-power field, 20 cells for the high-power field in 10 fields. DR MARK I. BLOCK (Hollywood, FL): I have a couple of questions. First, I think it’s important to be careful about what we mean by TIL. I think a lot of us think of TIL as lymphocytes that are extracted from the tumor and grown up to treat melanoma patients the way Rosenberg first did it. So I think when you just look at lymphocytes in a specimen, it’s important to be able to define exactly what we’re talking about. So you did some special stains. Are these B cells, T cells, CD4 cells, CD8 cells? Do you have a sense about what population of lymphocytes you’re talking about? Second, it seemed a little bit strange that you had a higher correlation of TIL with invasion and degree of differentiation, but that also seemed to correlate with improved survival. At least I think that’s the way that you showed the data. How do you explain that? DR RUFFINI: I think the answer to this question is that we just see the TIL population as a whole by standard histopathology, but, as I said, the TIL population is composed of different subsets and not all these populations act against tumor growth. There are some acting against, some acting pro. And we did immuno-
GENERAL THORACIC
Ann Thorac Surg 2009;87:365–72
GENERAL THORACIC
372
RUFFINI ET AL TUMOR-INFILTRATING LYMPHOCYTES IN LUNG NEOPLASMS
histochemistry only on very few cases. But the message I would like is just to see clinically at conventional histopathology what the significance of TIL is. Of course, this is just the preliminary results. The answer will be immunohistochemistry just to differentiate the different subpopulations. CD8 acts, of course, against tumor growth, but CD4, for example, sometimes acts pro and sometimes acts against tumor growth. So I know that this could be a limitation of the study, but this was not the aim of our study. Our study was just to examine the TIL population as a whole, I mean just in a conventional histopathology. DR PETER GOLDSTRAW (London, UK): We have recently found that a simple peripheral blood neutrophil/lymphocyte
Ann Thorac Surg 2009;87:365–72
ratio is an independent prognostic factor. Did you look at peripheral blood lymphocyte counts in your TILs to see whether there was any correlation? DR RUFFINI: No. I have read about this, but the focus of our study was just to see the lymphocytic infiltrate in the specimen, not in the peripheral blood. DR GOLDSTRAW: You can go back and do that very easily as the data would be in the case record. DR RUFFINI: Yes. That’s a good suggestion. Thank you.