WO2016050623A1 - Prognosis markers in lung cancer - Google Patents

Abstract

The present invention provides two gene expression signatures indicative of poor prognosis for lung cancer subjects, preferably in NSCLC, more preferably stage I NSCLC, and even more preferably stage I lung adenocarcinoma. These signatures were found to be very efficient to predict patient's outcome. Thus, they allow the practitioner to accurately distinguish lung tumors associated with poor prognosis so as to target the patients who would really benefit from adjuvant therapy and spare unnecessary treatment to patients with a good prognosis, preferably in NSCLC, more preferably stage I NSCLC, and even more preferably stage I lung adenocarcinoma.

Description

Prognosis markers in lung cancer.

Field of the Invention

The present invention relates to the field of medicine, in particular of oncology. It provides new prognostic markers in lung cancer.

Background of the Invention

Lung cancer is the leading cause of cancer-related deaths worldwide, accounting for 1.3 million deaths annually. It is defined as cancer that forms in the tissues of the lung, usually in the cells lining air passages, and is divided into two main subtypes: small cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC). NSCLC is the predominant subtype form and accounts for about 85% of all lung cancers. Among NSCLC, two major subtypes can be distinguished: squamous cell carcinoma (SCC) and adenocarcinoma (ADC). The 5- year survival for patients with regional disease is approximately 26%, which decreases to 3.9% for patients with metastatic disease.

Although the management of metastatic NSCLC has been profoundly modified by the identification of actionable molecular traits (e.g. EGFR mutation, ALK translocation, etc.), decision making for early-stage NSCLC still relies on the tumor stage only. For patients diagnosed with early stage disease (i.e., stages I and II) or even stage III, surgery offers the best option for survival and cure. Unfortunately, most of these patients recur within the 5 years after the tumor resection (Stage I: 30%, Stage II: 50%>, Stage III: 70%>), leading to their death by cancer in most cases.

Currently, the methods to determine prognosis and select patients for adjuvant therapy rely mainly on pathological and clinical staging (i.e. the TNM AJCC classification). Most of stage I patients do not receive adjuvant therapy after surgery, while stage II and stage III patients typically do, with the objective of preventing tumor relapse. However, it is very difficult to predict which resected localized tumor will eventuate in recurrence, which invariably leads to cancer death. Improving the prediction of probability of overall survival in patients with early stage - as especially stage I lung cancer is thus critical, in order to better identify the stage I patients at high risk of recurrence or death that could benefit of an adjuvant therapy and the stage II-III patients at low risk of recurrence or death that could be spared unnecessary treatment. Therefore, there is a great need for the identification of prognostic markers that can accurately distinguish tumors associated with poor prognosis including high probability of metastasis, early disease progression, increased disease recurrence or decreased patient survival, from the others. Using such markers, the practitioner would be able to predict the patient's prognosis and effectively target the individuals who would most likely benefit from adjuvant therapy.

Summary of the Invention

The understanding of the molecular basis of cancer has advanced tremendously with the identification of mutations in the genome of tumor cells. Yet, while numerous studies support a major role for genetic events in cancer susceptibility, in particular for lung cancers, this genetic contribution alone does not explain the clinical complexity and heterogeneity of cancers, therefore suggesting that other mechanisms may contribute to the process of tumorigenesis and aggressiveness. Abnormal gene expression in cancer cells and how they could promote tumorigenesis is of major interest in that respect.

The inventors discovered two new gene expression signatures, one comprising POU2F1, HSD3B1, ING3, PDE6H, RARRES3, and TIMP2, the other further comprising FOSL2. These signatures were found to be very efficient to predict patient's outcome. Thus, they allow the practitioner to accurately distinguish stage I NSCLC associated with poor prognosis so as to target the patients who would really benefit from adjuvant therapy and spare unnecessary treatment to patients with a good prognosis.

Accordingly, in a first aspect, the present invention concerns an in vitro method for predicting clinical outcome of a subject affected with a lung cancer, preferably NSCLC, wherein the method comprises the step of determining the expression levels of POU2F1 (POU class 2 homeobox 1), HSD3B1 (Hydroxy-delta-S-steroid dehydrogenase, 3 beta- and steroid delta-isomerase 1), ING3 (Inhibitor of Growth family 3), PDE6H (Phosphodiesterase 6 H cGMP-specific, cone, gamma), RARRES3 (Retinoic Acid Receptor Responder 3), andTIMP2 (TIMP Mettalopeptidase inhibitor 2) in a lung cancer sample from said subject, a dysregulation of the gene expression signature being indicative of a poor prognosis.

Preferably, the method further comprises the determination of the expression level of FOSL2 (FOS-Like Antigen 2), a dysregulation of the gene expression signature being indicative of a poor prognosis. More preferably, a high expression level of POU2F1, ING3 and RARRES3 and a low expression level of HSD3B1, PDE6H, TIMP2 and optionally FOSL2 are indicative of a poor prognosis.

In a second aspect, the present invention also concerns an in vitro method for selecting a subject affected with a lung cancer, preferably NSCLC, for a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy, or determining whether a subject affected with a lung cancer, preferably NSCLC, is susceptible to benefit from a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy, wherein the method comprises the step of determining the expression levels of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, and TIMP2 in a lung cancer sample from said subject, , a dysregulation of the gene expression signature indicating that a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy, is recommended.

Preferably, the method further comprises the determination of the expression level of FOSL2, a dysregulation of the gene expression signature indicating that a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy, is recommended.

More preferably, a high expression level of POU2F1, ING3 and RARRES3 and a low expression level of HSD3B1, PDE6H, TIMP2 and optionally FOSL2 indicating that a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy, is recommended.

Preferably, the subject of above mentioned methods has previously been subjected to a tumor resection.

Preferably, the subject of above mentioned methods has not been previously treated by an adjuvant therapy and/or a neoadjuvant therapy.

Preferably, the expression levels of genes in above mentioned methods are determined by measuring the quantity of corresponding proteins or mRNA.

Alternatively, the quantities of corresponding proteins in above mentioned methods are measured by immunohistochemistry, semi-quantitative Western-Blot, or protein or antibody arrays.

Alternatively, the quantities of corresponding mRNA in above mentioned methods are measured by quantitative or semi-quantitative RT-PCR, or by real time quantitative or semiquantitative RT-PCR or by transcriptome approaches.

Preferably, the above-mentioned methods further comprise the step of comparing expression levels of genes to reference expression levels and optionally the step of determining whether expression levels of these genes are high compared to said reference expression levels. Preferably, a poor prognosis in above mentioned methods is a decreased patient survival and/or an early disease progression and/or an increased disease recurrence and/or an increased metastasis formation, preferably an increased disease recurrence.

In addition, the present invention concerns the use of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, and TIMP2 as prognosis markers in lung cancer, preferably NSCLC, as markers for selecting a subject affected with a lung cancer for a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy, or for determining whether a subject affected with a lung cancer is susceptible to benefit from a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy.

Preferably, FOSL2 is also used in combination with POU2F1, HSD3B1, ING3, PDE6H, RARRES3, and TIMP2 as prognosis markers in lung cancer, as markers for selecting a subject affected with a lung cancer for a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy, or for determining whether a subject affected with a lung cancer is susceptible to benefit from a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy.

Preferably, the lung cancer of above mentioned methods and above mentioned uses is a NSCLC (Non Small Cell Lung Cancer), more preferably an early stage NSCLC, still more preferably a stage I NSCLC, and even more preferably a stage I adenocarcinoma (ADC). Optionally, the lung cancer of above mentioned methods and above mentioned uses is a stage I squamous cell cancer (SCC). More particularly, POU2F1, HSD3B1, ING3, PDE6H, RARRES3, TIMP2 and optionally FOSL2 are associated with prognosis of stage I NSCLC and more preferably with stage I lung adenocarcinoma.

The present invention also concerns a kit (a) for predicting clinical outcome of a subject affected with a lung cancer; and/or (b) for selecting a subject affected with a lung cancer for a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy, or determining whether a subject affected with a lung cancer is susceptible to benefit from a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy, wherein the kit comprises (i) at least one antibody specific to each of the following: POU2F1; HSD3B1, ING3, PDE6H, RARRES3, and TIMP2 and/or (ii) at least one probe specific to the mRNA or cDNA of each of the following: POU2F1, HSD3B1, ING3, PDE6H, RARRES3, and TIMP2 and/or (iii) at least one nucleic acid primer pair specific to each of the following: POU2F1, HSD3B1, ING3, PDE6H, RARRES3, and TIMP2 mRNA or cDNA and optionally, a leaflet providing guidelines to use such a kit. Preferably, the kit further comprises (i) at least one antibody specific to FOSL2 and/or (ii) at least one probe specific to FOSL2 mRNA or cDNA and/or (iii) at least one nucleic acid primer pair specific to FOSL2 mRNA or cDNA. Finally, the present invention concerns the use of a kit as disclosed above, (a) for predicting clinical outcome of a subject affected with a lung cancer; and/or (b) for selecting a subject affected with a lung cancer for a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy, or determining whether a subject affected with a lung cancer is susceptible to benefit from a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy.

Preferably, the lung cancer of above mentioned kits and above mentioned use is a NSCLC, more preferably an early stage NSCLC, still more preferably a stage I NSCLC and even more preferably a stage I lung ADC.

Brief Description of the Drawings

Figure 1 : Model performance measured by ROC AUC (Area Under the Curve for the Receiver Operating Characteristics curves, left panel), and the probability of overall survival (right panel) estimated by Kaplan Meier of high- and low-risk groups of patients determined by the six genes expression signature in the merged validation datasets. To note, there was 307 patients with known 3 -year survival included in the ROC- AUC, and 496 patients included in the overall survival estimation by Kaplan Meier.

Figure 2: Model performance measured by ROC AUC in the five independent validation datasets (Zhu et al, Rousseaux et al, Raponi et al, TCGA LUSC and TCGA LUAD) for the six genes expression signature.

Figure 3: Probability of overall survival estimated by Kaplan Meier for high- vs. low-risk predicted determined by the six genes expression signature in the five independent validation datasets (Zhu et al, Rousseaux et al, Raponi et al, TCGA LUSC and TCGA LUAD).

Figure 4: Model performance measured by ROC AUC (Area Under the Curve for the Receiver Operating Characteristics curves, left panel), and the probability of overall survival (right panel) estimated by Kaplan Meier of high- and low-risk groups of patients determined by the seven genes expression signature in the merged validation datasets. To note, there was 312 patients with known 3 -year survival included in the ROC-AUC, and 495 patients included in the overall survival estimation by Kaplan Meier.

Figure 5: Model performance measured by ROC AUC in the five independent validation datasets (Zhu et al, Rousseaux et al, Raponi et al, TCGA LUSC and TCGA LUAD) for the seven genes expression signatures.

Figure 6: Probability of overall survival estimated by Kaplan Meier for high- vs. low-risk predicted determined by the seven genes expression signature in the five independent validation datasets (Zhu et al, Rousseaux et al, Raponi et al, TCGA LUSC and TCGA LUAD).

Figure 7: Model performance measured by ROC AUC (Area Under the Curve for the Receiver Operating Characteristics curves, left panel), and the probability of overall survival (right panel) estimated by Kaplan Meier of high- and low-risk groups of patients determined by the seven genes expression signature in the merged validation dataset restricted to stage I lung adenocarcinoma patients. To note, there was 156 patients with known 3 -year survival included in the ROC-AUC, and 297 patients included in the overall survival estimation by Kaplan Meier.

Figure 8: Model performance measured by ROC AUC in three independent validation datasets restricted to stage I lung adenocarcinoma patients (Zhu et al, Rousseaux et al. and TCGA LUAD) for the seven genes expression signature.

Figure 9: Probability of overall survival estimated by Kaplan Meier for high- vs. low-risk patients determined by the seven genes expression signature in three independent validation datasets restricted to stage I lung adenocarcinoma patients (Zhu et al, Rousseaux et al. and TCGA LUAD).

Detailed description of the Invention

Recent data have uncovered a number of alterations in gene expression in the context of cancer. How they occur and connect with other genetic alterations during the development of this disease is currently a major issue. High-dimensional molecular data (gene expression data) for lung adenocarcinoma patients are now available. In this context, the inventors have analyzed publically available datasets (eight of them) of NSCLC patients to search for gene expression patterns associated to a high frequency of recurrence or death.

Identification of prognostic markers that can accurately distinguish lung tumors associated with poor prognosis represents a significant step forward which can help to classify lung cancer types, particularly NSCLC types, and help to effectively target the individuals who would most likely benefit from adjuvant therapy. This is particularly important when considering the complexity and heterogeneity encountered in lung cancers. A poor prognosis includes a high probability of metastasis, early disease progression, increased disease recurrence or decreased patient survival.

In the present study, the inventors unraveled new gene expression signatures that robustly predicts the outcome of patients diagnose with lung cancer. They discovered a first signature and from this preliminary signature, they selected the six most influent markers: POU2F1, HSD3B1, ING3, PDE6H, RARRES3 and TIMP2. This signature was found very efficient to predict patient's outcome. They also discovered a second signature and from this preliminary signature, they selected the seven most influent markers: POU2F1, HSD3B1, ING3, PDE6H, RARRES3, TIMP2 and FOSL2. This signature was found very efficient to predict patient's outcome.

Definitions

As used herein, the term "POU2F1" refers to the POU (referring to the first three transcription factors described in this family: Pit- 1 , Oct-1 and Unc-86) class 2 homeobox 1 gene. Unigene Cluster for POU2F1 is Hs.283402. Representative mRNA and protein sequences are AY113189.1 and NP_002688.3, respectively.

As used herein, the term "HSD3B1" refers to the Hydroxy-delta-S-steroid dehydrogenase, 3 beta- and steroid delta-isomerase 1 gene. Unigene Cluster for HSD3B1 is Hs.364941. Representative mRNA and protein sequences are NM 000862.2 and NP_000853.1, respectively.

As used herein, the term "ING3" refers to the Inhibitor of Growth family 3 gene. Unigene Cluster for ING3 is Hs.489811. Representative mRNA and protein sequences are NM_019071.2 or NM_198267.1 and NP_061944.2, respectively.

As used herein, the term "PDE6H" refers to the Phosphodiesterase 6 H cGMP- specific, cone, gamma gene. Unigene Cluster for PDE6H is Hs.54471. Representative mRNA and protein sequences are NM 006205.2 and NP 006196.1, respectively. As used herein, the term "RARRES3" refers to the Retinoic Acid Receptor Responder 3 gene. Unigene Cluster for RARRES3 is Hs.17466. Representative mRNA and protein sequences are NM_004585.3 and NP_004576.2 1, respectively.

As used herein, the term "TIMP2" refers to the Tissue Inhibitors of Metalloproteinases 2 (or TIMP Mettalopeptidase inhibitor 2) gene. Unigene Cluster for TIMP2 is Hs.633514. Representative mRNA and protein sequences are NM_003255.4 and NP_003246.1, respectively.

As used herein, the term "FOSL2" refers to the FOS-like Antigen 2 gene. Unigene Cluster for FOSL2 is Hs.220971. Representative mRNA and protein sequences are NM_005253.3 and NP_005244.1, respectively.

The term "cancer" or "tumor", as used herein, refers to the presence of cells possessing characteristics typical of cancer-causing cells, such as uncontrolled proliferation, and/or immortality, and/or metastatic potential, and/or rapid growth and/or proliferation rate, and/or certain characteristic morphological features. This term refers to any type of malignancy (primary or metastases) in any type of subject, male or female. In particular, the term encompasses lung cancer and more particularly NSCLC at any stage of progression. Cancer staging is defined as follows (TNM 6th edition):

Stage I: the tumor is smaller than or equal to 5 cm in maximum diameter and has not spread to any other tissues or lymph nodes.

Stage II: the tumor is either between 3 and 7 cm or it has spread to the lymph nodes, or it has invaded the tissues surrounding the lung, or it has started to invade the large bronchial tubes.

Stage IIIA: the tumor can be bigger than 7 cm or has spread to the lymph nodes in the mediastinum or has spread to the rib cage, heart, esophagus or to the trachea.

Stage IIIB: the tumor has spread to lymph nodes on the other side of the mediastinum or to the lymph nodes above or behind the clavicle. Stage IIIB also includes large tumors that have spread to the rib cage, heart, esophagus or to the trachea when there is involvement of the mediastinal lymph nodes.

Stage IV: the cancer has spread outside of the chest or has spread/ metastasized to a critical location or has caused some complication. Possible complications are that the cancer has caused fluid to collect around the lung or heart (called a malignant effusion), or it has spread to the opposite side of the chest, has spread to outside the chest such as "early stage" (stage I and II). As used herein, the term "treatment", "treat" or "treating" refers to any act intended to ameliorate the health status of patients such as therapy, prevention, prophylaxis and retardation of the disease.

The term "therapy", as used herein, refers to any type of treatment of cancer (i.e., antitumoral therapy), including an adjuvant therapy. Therapy comprises radiotherapy and therapies, preferably systemic therapies such as hormone therapy, chemotherapy, immunotherapy and monoclonal antibody therapy.

The term "adjuvant therapy", as used herein, refers to any type of treatment of cancer given as additional treatment, usually after surgical resection of the primary tumor, in a patient affected with a cancer that is at risk of metastasizing and/or likely to recur. The aim of such an adjuvant treatment is to improve the prognosis. Adjuvant therapies comprise radiotherapy and therapy, preferably systemic therapy, such as hormone therapy, chemotherapy, immunotherapy and monoclonal antibody therapy.

The term "neoadjuvant therapy", as used herein, refers to any type of treatment of cancer given prior to surgical resection of the primary tumor, in a patient affected with a cancer. The most common reason for neoadjuvant therapy is to reduce the size of the tumor so as to facilitate a more effective surgery. Neoadjuvant therapies comprise radiotherapy and therapy, preferably systemic therapy, such as hormone therapy, chemotherapy, immunotherapy and monoclonal antibody therapy.

As used herein, the term "chemotherapeutic treatment" or "chemotherapy" refers to a cancer therapeutic treatment using chemical or biochemical substances, in particular using one or several antineoplastic agents.

The term "radiotherapeutic treatment" or "radiotherapy" is a term commonly used in the art to refer to multiple types of radiation therapy including internal and external radiation therapies or radioimmunotherapy, and the use of various types of radiations including X-rays, gamma rays, alpha particles, beta particles, photons, electrons, neutrons, radioisotopes, and other forms of ionizing radiations.

The term "immunotherapy" refers to a cancer therapeutic treatment using the immune system to reject cancer. The therapeutic treatment stimulates the patient's immune system to attack the malignant tumor cells. It includes immunization of the patient with tumoral antigens (eg. by administering a cancer vaccine), in which case the patient's own immune system is trained to recognize tumor cells as targets to be destroyed, or administration of molecules stimulating the immune system such as cytokines, or administration of therapeutic antibodies as drugs, in which case the patient's immune system is recruited to destroy tumor cells by the therapeutic antibodies. In particular, antibodies are directed against specific antigens such as the unusual antigens that are presented on the surfaces of tumors. As illustrating example, one can cite Trastuzumab or Herceptin antibody which is directed against HER2 and approved by FDA for treating breast cancer.

The term "monoclonal antibody therapy" refers to any antibody that functions to deplete tumor cells in a patient. In particular, therapeutic antibodies specifically bind to antigens present on the surface of the tumor cells, e.g. tumor specific antigens present predominantly or exclusively on tumor cells. Alternatively, therapeutic antibodies may also prevent tumor growth by blocking specific cell receptors.

The term "hormone therapy" or "hormonal therapy" refers to a cancer treatment having for purpose to block, add or remove hormones. For instance, in breast cancer, the female hormones estrogen and progesterone can promote the growth of some breast cancer cells. So in these patients, hormone therapy is given to block estrogen and a non-exhaustive list commonly used drugs includes: Tamoxifen, Fareston, Arimidex, Aromasin, Femara, Zoladex/Lupron, Megace, and Halotestin.

As used herein, the term "poor prognosis" refers to a decreased patient survival and/or an early disease progression and/or an increased disease recurrence and/or an increased metastasis formation.

As used herein, the term "subject" or "patient" refers to an animal, preferably to a mammal, even more preferably to a human, including adult, child and human at the prenatal stage. However, the term "subject" can also refer to non-human animals, in particular mammals such as dogs, cats, horses, cows, pigs, sheep and non-human primates, among others, that are in need of treatment.

The term "sample", as used herein, means any sample containing cells derived from a subject, preferably a sample which contains nucleic acids. Examples of such samples include fluids such as blood, plasma, saliva, urine and seminal fluid samples as well as biopsies, organs, tissues or cell samples. Preferably, the sample is a lung biopsy from the subject or is part of the lung tumor resection from the subject. The sample may be treated prior to its use.

The term "cancer sample" refers to any sample containing tumoral cells derived from a patient, preferably a sample which contains nucleic acids. Preferably, the sample contains only tumoral cells. The term "normal sample" refers to any sample which does not contain any tumoral cells.

The methods of the invention as disclosed below, may be in vivo, ex vivo or in vitro methods, preferably in vitro methods. ^ ^

The present invention relates to a method for predicting or monitoring clinical outcome of a subject affected with a lung cancer, wherein the method comprises the step of determining the expression levels of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, and TIMP2, in a lung cancer sample from said subject, a dysregulation of the gene expression signature being indicative of a poor prognosis.

In a preferred embodiment, the method further comprises determining the expression level of FOSL2.

It is important to note that the expression levels of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, TIMP2 and FOSL2, are significant prognosis markers for clinical outcome taken alone.

In an embodiment of the above mentioned method, a high expression level of POU2F1, ING3 and RARRES3 and a low expression level of HSD3B1, PDE6H, TIMP2 and optionally FOSL2 are indicative of a poor prognosis.

In a further aspect, the present invention concerns a method for selecting a subject affected with a lung cancer for a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy, or determining whether a subject affected with a lung cancer is susceptible to benefit from a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy, wherein the method comprises the step of determining the expression level of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, and TIMP2 in a cancer sample from said subject, a dysregulation of the gene expression signature indicating that a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy, is required.

In a preferred embodiment, the method further comprises determining the expression level of FOSL2.

A dysregulation of the gene expression signature of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, TIMP2 and optionally FOSL2, indicate a decreased patient survival and/or an early disease progression and/or an increased disease recurrence and/or an increased metastasis formation. Accordingly, this type of lung cancer associated with poor prognosis has to be treated with a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy, in order to improve the patient's chance for survival. The type of therapy is chosen by the practitioner. It includes radiotherapy, chemotherapy, hormonal therapy, immunotherapy and monoclonal antibody therapy. However, these therapies are usually aggressive and cause several side effects. By using the method according to the invention it is therefore possible to limit adjuvant therapy to subjects who really need them and spare a large subgroup of subjects (those identified as having a good prognosis) of a harmful, useless and expensive treatment. More particularly, improving the prediction of probability of overall survival in patients with lung cancer allows to better identify the stage I-II patients at high risk of recurrence or death that could benefit of an adjuvant therapy and the stage II-III patients at low risk of recurrence or death that could be spared unnecessary treatment.

In an embodiment of the above mentioned method, a high expression level of POU2F1, ING3 and RARRES3 and a low expression level of HSD3B1, PDE6H, TIMP2 and optionally FOSL2 are indicative that a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy, is recommended. Preferred adjuvant therapies may be selected among the following combinations: cisplatin-gemcitabine, cisplatin-vinorelbine and carboplatin- paclitaxel.

In an embodiment of these above mentioned methods, the method further comprises the step of providing a lung cancer sample from the subject.

In a preferred embodiment of these above mentioned methods, the lung cancer is a NSCLC, preferably an early stage NSCLC (i.e. stage I or II) and more preferably a stage I NSCLC. Even more preferably, the lung cancer can be adenocarcinoma (ADC), preferably a stage I ADC.

In a particularly preferred embodiment, the expression levels of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, TIMP2 and optionally FOSL2 are associated with prognosis of stage I NSCLC, preferably stage I ADC.

The expression levels of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, TIMP2 and FOSL2 can be determined from a lung cancer sample by a variety of techniques. In an embodiment, the expression levels of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, TIMP2 and FOSL2 are determined by measuring the quantity of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, TIMP2 and FOSL2 proteins or POU2F1, HSD3B1, ING3, PDE6H, RARRES3, TIMP2 and FOSL2 mRNA.

In a particular embodiment of these above mentioned methods, the expression levels of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, TIMP2 and FOSL2 are determined by measuring the quantity of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, TIMP2 and FOSL2 proteins. The quantity of these proteins may be measured by any methods known by the skilled person. Usually, these methods comprise contacting the sample with binding partners capable of selectively interacting with the POU2F1, HSD3B1, ING3, PDE6H, RARRES3, TIMP2 and FOSL2 proteins present in the sample. The binding partners are generally polyclonal or monoclonal antibodies, preferably monoclonal. Antibodies are commercially available. In addition, the methods for producing antibodies are well-known in the art. The quantity of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, TIMP2 and FOSL2 proteins may be measured by semi-quantitative Western blots, enzyme-labeled and mediated immunoassays, such as ELISAs, biotin/avidin type assays, radioimmunoassay, Immunoelectrophoresis or immunoprecipitation or by protein or antibody arrays. The proteins expression level may be assessed by immunohistochemistry on a tissue section of the lung cancer sample (e.g. frozen or formalin-fixed paraffin embedded material). The reactions generally include revealing labels such as fluorescent, chemiluminescent, radioactive, enzymatic labels or dye molecules, or other methods for detecting the formation of a complex between the antigen and the antibody or antibodies reacted therewith. Preferably, the quantity of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, TIMP2 and FOSL2 proteins are measured by immunohistochemistry or semi-quantitative western-blot.

In another embodiment of these above mentioned methods, the expression levels of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, TIMP2 and FOSL2 are determined by measuring the quantity of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, TIMP2 and FOSL2 mRNA. Methods for determining the quantity of mRNA are well known in the art. For example, the nucleic acid contained in the lung sample (e.g., cell or tissue prepared from the patient) is first extracted according to standard methods, for example using lytic enzymes or chemical solutions or extracted by nucleic-acid-binding resins following the manufacturer's instructions. The extracted mRNA is then detected by hybridization (e. g., Northern blot analysis) and/or amplification (e.g., RT-PCR). Preferably quantitative or semi-quantitative RT-PCR is preferred. Real-time quantitative or semi-quantitative RT-PCR is particularly advantageous. Preferably, primer pairs were designed in order to overlap an intron, so as to distinguish cDNA amplification from putative genomic contamination. Primers may be easily designed by the skilled person. Other methods of Amplification include ligase chain reaction (LCR), transcription-mediated amplification (TMA), strand displacement amplification (SDA) and nucleic acid sequence based amplification (NASBA). Preferably, the quantity of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, TIMP2 and FOSL2 mRNA are measured by quantitative or semi-quantitative RT-PCR or by real-time quantitative or semi-quantitative RT-PCR or by transcriptome approaches.

In an embodiment of these above mentioned methods, the method further comprises the step of comparing the expression levels of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, TIMP2 and FOSL2 to a reference expression level.

In particular, the reference expression level can be the expression levels of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, TIMP2 and FOSL2 in a normal sample. The normal sample is a non-tumoral sample, preferably a normal lung sample. The normal sample may be obtained from the subject affected with the lung cancer or from another subject, preferably a normal or healthy subject, i.e. a subject who does not suffer from a cancer. Preferably, the normal sample is obtained from the same subject than the lung cancer sample. Expression levels obtained from lung cancer and normal samples may be normalized by using expression levels of proteins which are known to have stable expression such as RPLPO (acidic ribosomal phosphoprotein PO), TBP (TATA box binding protein), GAPDH (glyceraldehyde 3 -phosphate dehydrogenase) or β-actin.

Alternatively, the reference expression level may be the expression level of a gene having a stable expression in lung cancer samples. Such genes include for example, RPLPO, TBP, GAPDH or β-actin. Preferably, the reference expression level is the expression level of the RPLPO gene. The use of the human acidic ribosomal phosphoprotein PO (RPLPO) gene as reference was described in the article of de Cremoux et al. (de Cremoux et al., 2004). In a preferred embodiment, the quantities of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, TIMP2 and FOSL2 mRNA are normalized according to the quantity of RPLPO mRNA. The quantity of RPLPO mRNA is used as reference quantity (i.e. 100%). The quantities of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, TIMP2 and FOSL2 mRNA are expressed as relative quantities with respect to the quantity of RPLPO mRNA.

In a further embodiment of these above mentioned methods, the method further comprises the step of determining whether the expression levels of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, TIMP2 and FOSL2 are dysregulated compared to the reference expression level.

The expression levels of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, TIMP2 and FOSL2 in the cancer sample are considered as significantly different (i.e. dysregulated) compared to the reference expression levels in a normal sample, if, after normalization, differences are in the order of 2.5-fold higher for POU2F1, ING3 and RARRES3 than the expression levels in the normal sample or more, and in the order of 2.5 -fold lower for HSD3B1, PDE6H, TIMP2 and FOSL2 than the expression levels in the normal sample or less. Preferably, the expression levels of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, TIMP2 and FOSL2 in the cancer sample are considered as dysregulated if the levels are at least 3-fold higher, or 4, 5 or 6-fold higher for POU2F1, ING3 and RARRES3 than the expression levels in the normal sample, and if the levels are at least 3-fold lower, or 4, 5 or 6- fold lower for HSD3B1, PDE6H, TIMP2 and FOSL2 than the expression levels in the normal sample. In an embodiment of these above mentioned methods, the subject affected with a lung cancer has previously been subjected to a tumor resection.

In another embodiment of these above mentioned methods, the subject has not been previously treated by an adjuvant therapy and/or a neoadjuvant therapy.

In yet another embodiment of these above mentioned methods, prognosis analysis can be completed by a tumor staging study.

The present invention further concerns the use of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, and TIMP2, as prognosis markers in lung cancer. Preferably, FOSL2 is also used in combination as a prognosis marker in lung cancer. In a preferred embodiment, these markers, are used as prognosis markers in an early stage lung cancer, i.e. stages I or II, or stage III lung cancer, preferably stage Ilia, without local or systemic invasion. As used herein, the term "prognosis markers" refers to combination of compounds, i.e. combination of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, and TIMP2, used to predict clinical outcome of a subject affected with a lung cancer. Preferably, the term "prognosis markers" also refers to FOSL2.

The present invention also concerns the use of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, and TIMP2, as markers for selecting a subject affected with a lung cancer for a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy, or determining whether a subject affected with a lung cancer is susceptible to benefit from a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy. Preferably, FOSL2 is also used in combination with these markers as a marker in lung cancer for selecting a subject affected with a lung cancer for a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy, or determining whether a subject affected with a lung cancer is susceptible to benefit from a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy.

In a preferred embodiment of the above mentioned uses, the lung cancer is a NSCLC, preferably an early stage NSCLC, more preferably a stage I NSCLC, and even more preferably a stage I lung ADC.

In another aspect, the present invention further concerns a kit, and its use,

(a) for predicting clinical outcome of a subject affected with a lung cancer; and/or

(b) for selecting a subject affected with a lung cancer for a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy, or determining whether a subject affected with a lung cancer is susceptible to benefit from a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy;

wherein the kit comprises

(i) at least one antibody specific to each of the followings : POU2F1, HSD3B1, ING3, PDE6H, RAR ES3, and TIMP2 and, optionally, means for detecting the formation of the complexes between POU2F1, HSD3B1, ING3, PDE6H, RARRES3, and TIMP2 ; and/or

(ii) at least one probe specific to mRNA or cDNA of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, and TIMP2, and, optionally, means for detecting the hybridization of said at least one probe on POU2F1, HSD3B1, ING3, PDE6H, RARRES3, and TIMP2 mRNA or cDNA; and/or

(iii) at least one nucleic acid primer pair specific to POU2F1, HSD3B1, ING3, PDE6H, RARRES3, and TIMP2 mRNA or cDNA and, optionally, means for amplifying and/or detecting said mRNA or cDNA ; and,

(iv) optionally, a leaflet providing guidelines to use such a kit.

In a preferred embodiment, the kit comprises

(i) at least one antibody specific to each of the followings : POU2F1, HSD3B1, ING3, PDE6H, RARRES3, TIMP2 and FOSL2 and, optionally, means for detecting the formation of the complexes between POU2F1, HSD3B1 , ING3, PDE6H, RARRES3, TIMP2 and FOSL2; and/or

(ii) at least one antibody specific to each of the followings : POU2F1, HSD3B1, ING3, PDE6H, RARRES3, TIMP2 and FOSL2 and, optionally, means for detecting the formation of the complexes between POU2F1, HSD3B1 , ING3, PDE6H, RARRES3, TIMP2 and FOSL2; and/or

(iii) at least one probe specific to mRNA or cDNA of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, TIMP2 and FOSL2, and, optionally, means for detecting the hybridization of said at least one probe on POU2F1, HSD3B1, ING3, PDE6H, RARRES3, and TIMP2 and FOSL2 mRNA or cDNA; and/or

(iv) at least one nucleic acid primer pair specific to POU2F1, HSD3B1, ING3, PDE6H, RARRES3, TIMP2 and FOSL2 mRNA or cDNA and, optionally, means for amplifying and/or detecting said mRNA or cDNA ; and,

(v) optionally, a leaflet providing guidelines to use such a kit. In an embodiment of the above mentioned kit, the lung cancer is a NSCLC, preferably an early stage NSCLC, more preferably a stage I NSCLC and even more preferably a stage I lung ADC.

Further aspects and advantages of the present invention will be described in the following examples, which should be regarded as illustrative and not limiting.

Examples

Patients and methods:

Patients

Gene expression datasets of patients diagnosed with early stage NSCLC were selected using a stringent criteria: each dataset had to be publicly available either as raw data (CEL files and clinical summary) or normalized data (MAS5, RMA, etc.); clinical covariates had to include pathological stage, adjuvant or neoadjuvant antineoplastic treatment status and overall survival information.

A literature search for studies of NSCLC patients who underwent molecular assessment of tumor tissue revealed 26 studies. Of these studies, the inventors selected and downloaded the data of eight gene expression datasets that were publicly available, had adequate clinical information including stage and sufficient numbers of patients that were not treated with chemotherapy or radiation:

• Directors challenge data (available on caArray at https://arrav.nci.nih.gov/caarray/proiect/iacob-00182)

• Zhu et al (GEO: GSE14814)

• Hou et al (GEO: GSE19188)

• TCGA LUSC Lung Squamous carcinoma (TCGA at tcgadata.nci.nih.gov/tcga)

• TCGA LUAD Lung Adenocarcinoma (TCGA at https://tcgadata.nci.nih.gov/tcga),

• Bhattacharjee et al. (available at http://www.broadinstitute.Org/m.pr/l.ung/),

• Raponi et al. (GEO: GSE4573)

• Rousseaux et al. (GEO : GSE30219)

To ensure uniformity across the different datasets, patients that received any adjuvant or neoadjuvant treatment, patients without any overall survival information, as well as patients with stage II, III or IV were excluded from this analysis. In addition, only patients with complete surgical resection and negative surgical margins have been retained for analyses. Patients who died within the 3 months after chemotherapy were also excluded, assuming this was due to perioperative complications. A minimum follow up of 36 months was required.

Among these eight datasets the inventors defined:

- Three training sets: Directors Challenge (n=198); Hou et al (n=37); Bhattacharjee et al (n=70)

- Five independent validation sets: Zhu et al (JBR.10 trial) (n=31), Rousseaux et al (n=122), Raponi et al (n=59), TCGA LUSC (n=35) and Luad et al (n=60). Importantly, these validation sets were hold out during the statistical learning process to avoid any resubstitution bias and ensure the true reproducibility assessment of the results.

Table 1 : distribution of important clinical variables across studies for the six genes expression signature

Dir Hou Bhat All Zhu Rous Raponi LCGA TCGA (n=198) (n=37) (n=70) Training (n=31 ) (n=122) (n=59) Lsuc LUAD

(n=305) (n=35) (n=60)

Histology ADC 198 22 70 290 (95%) 19 73 0 0 60

(100%) (59%) (100%) (61 %) (60%) (0%) (0%) (100%) sec 0 15 0 15 12 49 59 35 0

(0%) (41 %) (0%) (5%) (39%) (40%) (100%) (100%) (0%)

Stage 1 A 96 9 33 138 (45%) 0 1 12 21 6 22

(48%) (24%) (47%) (0%) (92%) (36%) (17%) (37%) 1 B 102 28 37 167 (55%) 31 10 38 29 38 (52%) (76%) (53%) (100%) (8%) (64%) (83%) (63%)

Gender Female 99 10 40 149 (49%) 1 1 16 22 9 32

(50%) (27%) (57%) (35%) (13%) (37%) (26%) (53%)

Probability (%) 19% 40% 29% 24% 23% 29% 39% 57% 30% of 3yr

survival Table 2: distribution of important clinical variables across studies for the seven genes expression signature

survival

Methods

Pre processing

Gene expression data were normalized using RMA (Robust Multi-array Average) method. The set of genes expression variables was restricted to the 8492 gene expression variables present in all datasets. Across datasets, the genes were rescaled to have the same means and standard deviations in order to make datasets generated by different platforms and technologies comparable.

Feature selection process

A weighted logistic regression model was applied to each learning set (Dir, Hou, Bhat), for each of the 8492 features (i.e. genes). The SO variables were then selected when they satisfied the two following conditions: (i) Wald's p value < 0.05 for each training dataset, (ii) the sign of the regression coefficient is identical across the training datasets. The underlying assumption is that if a variable has a true impact on the 3 -year survival (os3yr), it should be visible and coherent across the three training datasets.

Additionally a weighted Cox regression model was applied to each learning set (Dir, Hou, Bhat), for each of the 8492 features (i.e. genes). The SO' variables were then selected when they satisfied the two following conditions: (i) Wald's p value < 0.1 for each training dataset, (ii) the sign of the regression coefficient is identical across the training datasets. The underlying assumption is that if a variable has a true impact median survival time, it should be seen consistently across the three training datasets. To retain the most influential variables, a GLM (Generalized Linear Model) using only the SO or the S0+S0' variables was run on the merged dataset including the three training sets (DirHouBhat) and only the variables verifying p value <0.05 in the multivariate model were selected in the SI (6 genes from the SO variables) or SI ' (six genes from the SO variables and 1 gene from the SO' variables) lists.

Statistical learning

Using the SI or SI ' variables, a regression model (Logistic Model) was trained in DirHouBhat restricted to ADC patients only. The fitted model was then applied to the validation datasets separately and to a validation set resulting from the merge of all validation sets. A separate analysis was also performed on ADC patients only.

To assess the performance of the models, the AUC (Area Under the Curve) of the receiver operating characteristics curves (ROC) were computed. The probability of overall survival was further estimated using the Kaplan Meier method. A log rank test was performed to test the difference between high- and low-risk groups. High and low-risk prognostic groups were defined according to the cut-off of returned probabilities of 0.5 : high risk having a returned probability > 0.5 of death within the first three years and low risk having a returned probability < 0.5 of death within the first three years.

Results:

Regarding the feature selection process, a total of nineteen SO variables was found to satisfy the conditions (p < 0.05 and identical regression sign, wald test). The Cox regression enabled to identify one additional variable FOSL2 which was not already identified in the SO list. Six SI variables, the most influential among the SO variables (p < 0.05, wald test), were identified secondarily: HSD3B1, ING3, PDE6H, POU2F1, RARRES3, TIMP2. Seven SI ' variables were identified, the six SI variables and the most influential among the SO' variables (p < 0.1, wald test): FOSL2. A Logistic Model was trained in the merged training datasets (DirHouBhat, n = 305) using only the six SI variables (first set of experiments, figures 1-3) or the seven SI ' variables (second set of experiments, figures 4-9). The found models were then applied to the validation sets.

When applying the model obtained with the 6 genes expression signature to the merged validation datasets (Figure 1), a ROC-AUC of 0.64 was observed (with a Confidence Interval (CI) of 95 %, 0.57-0.70).

When applying the model obtained with the 7 genes expression signature to the merged validation datasets (Figure 4), a ROC-AUC of 0.65 was observed (with a Confidence Interval (CI) of 95 %, 0.59-0.72).

Second, when applying both models in each dataset separately the same trends were observed in all of them (Figure 2 and 3 for the 6 genes expression signature and figures 5 and 6 for the seven genes expression signatures). Model performance in TCGA LUSC and Raponi et al. is lower, possibly due to different composition of population (100 % of SCC (Squamous Cell Carcinoma), whereas only 5 % of SCC in total learning data).

When applying the model obtained with the 7 genes expression signature to the subset of patients with stage I lung adenocarcinoma patients (Figure 7), a ROC AUC of 0.73 was observed (with a Confidence Interval (CI) of 95% 0.64-0.82). The median survival difference between high- and low- risk groups identified by the model was >20 months, p<10-3 (Figure 5). Model performance was consistently good in three independent datasets with ADC patients only (Rousseaux, Zhu and TCGA LUAD) (Figure 8 and Figure 9).

Discussion:

The management of Stage I NSCLC remains challenging with 30% of patients having recurrence within the 5 years and ultimately dying from cancer. The two powerful gene expression models provide here can robustly predict the prognostic across five different datasets of resected NSCLC, not previously treated by chemotherapy. These two gene expression signature have the potential to change the decision making at bedside: patients with high risk profile could be treated by chemotherapy, whereas patients with low risk profile would be assigned to surveillance only.

Importantly, a number of facts strongly strengthen the prognosis efficiency of these two signatures: (i) the stringent criteria used to select the training and validation datasets (known pathological stage, no adjuvant chemotherapy or radiotherapy, R0 resection, follow- up > 36 months), (ii) the number (n=5) and the different nature of the validation datasets (Agilent CGH arrays, Affymetrix CGH arrays, RNA-Seq data), as well as the fact that (iii) some of the genes composing these signatures were previously associated with cancer outcome in other tumor types (e.g. POU2F1) or to the sensitivity to anticancer agents (e.g. TIMP2).

Claims

Claims

1. An in vitro method for predicting clinical outcome of a subject affected with a lung cancer, wherein the method comprises the step of determining the expression levels of POU2F1 (POU class 2 homeobox 1), HSD3B1 (Hydroxy-delta-S-steroid dehydrogenase, 3 beta- and steroid delta-isomerase 1), ING3 (Inhibitor of Growth family 3), PDE6H (Phosphodiesterase 6 H cGMP-specific, cone, gamma), RARRES3 (Retinoic Acid Receptor Responder 3), and TIMP2 (TIMP Metallopeptidase inhibitor 2) in a lung cancer sample from said subject, a dysregulation of the gene expression signature being indicative of a poor prognosis.

2. The method according to claim 1, wherein the method further comprises the determination of expression level of FOSL2 (FOS-Like Antigen 2).

3. The method according to claim 1 or 2, wherein a high expression level of POU2F1, ING3 and RARRES3 and a low expression level of HSD3B1, PDE6H, TIMP2 and FOSL2 are indicative of a poor prognosis.

4. An in vitro method for selecting a subject affected with a lung cancer for a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy, or determining whether a subject affected with a lung cancer is susceptible to benefit from a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy, wherein the method comprises the step of determining the expression levels of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, and TIMP2 in a lung cancer sample from said subject, a dysregulation of the gene expression signature indicating that a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy, is recommended.

5. The method according to claim 4, wherein the method further comprises the determination of expression level of FOSL2.

6. The method according to claim 4 or 5, wherein a high expression level of POU2F1, ING3 and RARRES3 and a low expression level of HSD3B1, PDE6H, TIMP2 and FOSL2 indicating that a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy, is recommended.

7. The method according to claim 4 to 6, wherein the subject has previously been subjected to a tumor resection.

8. The method according to anyone of claims 4 to 7, wherein the subject has not been previously treated by an adjuvant therapy and/or a neoadjuvant therapy.

9. The method according to anyone of claims 1 to 8, wherein a poor prognosis is a decreased patient survival and/or an early disease progression and/or an increased disease recurrence and/or an increased metastasis formation, preferably an increased disease recurrence.

10. Use of POU2F1, HSD3B1, ING3, PDE6H, RARRES3, and TIMP2 as prognosis markers in lung cancer, as markers for selecting a subject affected with a lung cancer for a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy, or for determining whether a subject affected with a lung cancer is susceptible to benefit from a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy.

11. The use according to claim 10, wherein FOSL2 is also used as a prognosis marker in lung cancer, as a marker for selecting a subject affected with a lung cancer for a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy, or for determining whether a subject affected with a lung cancer is susceptible to benefit from a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy.

12. The method according to anyone of claims 1 to 9 or the use according to claim 10 or 11, wherein the lung cancer is a lung cancer, preferably a NSCLC (Non Small Cell Lung Cancer), more preferably an early stage NSCLC, still more preferably a stage I NSCLC, and even more preferably a stage I lung adenocarcinoma.

13. A kit (a) for predicting clinical outcome of a subject affected with a lung cancer, preferably a NSCLC, more preferably an early stage NSCLC, even more preferably a stage I NSCLC and in the most preferred embodiment a stage I lung adenocarcinoma; and/or (b) for selecting a subject affected with a lung cancer, preferably a NSCLC, more preferably an early stage NSCLC, still more preferably a stage I NSCLC and even more preferably a stage I lung adenocarcinoma, for a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy, or determining whether a subject affected with a lung cancer is susceptible to benefit from a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy ,wherein the kit comprises (i) at least one antibody specific to each of the following: POU2F1, HSD3B1, ING3, PDE6H, RARRES3, and TIMP2; and/or (ii) at least one probe specific to the mRNA or cDNA of each of the following: POU2F1, HSD3B1, ING3, PDE6H, RARRES3, and TIMP2 and/or (iii) at least one nucleic acid primer pair specific to each of the following: POU2F1, HSD3B1, ING3, PDE6H, RARRES3, and TIMP2 mRNA or cDNA and optionally, a leafiet providing guidelines to use such a kit.

14. The kit according to claim 13, wherein the kit further comprises (i) at least one antibody specific to FOSL2; and/or (ii) at least one probe specific to FOSL2 mRNA or cDNA; and/or (iii) at least one nucleic acid primer pair specific to FOSL2 mRNA or cDNA.

15. Use of a kit according to claim 13 or 14, (a) for predicting clinical outcome of a subject affected with a lung cancer, preferably a NSCLC, more preferably an early stage NSCLC, even more preferably a stage I NSCLC and in the most preferred embodiment a stage I lung adenocarcinoma; and/or (b) for selecting a subject affected with a lung cancer, preferably a NSCLC, , more preferably an early stage NSCLC, still more preferably a stage I NSCLC and even more preferably a stage I lung adenocarcinoma for a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy, or determining whether a subject affected with a lung cancer is susceptible to benefit from a therapy, preferably an adjuvant therapy and/or a neoadjuvant therapy.