The wind of change in the therapy of lung cancer

All authors

Hans-Stefan Hofmann

https://doi.org/10.1586/14737140.6.4.469

Published online:

10 January 2014

Display full size

Lung cancer can still be considered the leading cause of cancer-related deaths in industrialized countries and is closely correlated with cigarette consumption. It has been estimated that smoking kills more than 650,000 people in Europe and costs EU states approximately €100 billion (US$126 billion) per year. The indirect cost of smoking is extremely high and therefore, the most important public health intervention is to reduce lung cancer incidence and deaths by changing smoking habits. It would appear that restrictive measures are the only way of controlling nicotine abuse, thus reducing the incidence of bronchial carcinoma and, therefore, laws forbidding smoking are becoming ever more strict. The most recent example is the general ban of smoking in all pubs and private members’ clubs that has been adopted by the British Parliament.

Within the first 5 years of diagnosis, 86% of lung cancer patients die. Thus, the 169,000 newly diagnosed cases of lung cancer per year leads to more deaths than the combined cumulative death rate associated with breast, colon, prostate and cervical cancer in the USA. Prediction of overall survival is provided by the Tumor Nodes Metastases (TNM) staging system of the American Joint Committee on Cancer and is based almost exclusively on the anatomical extent of the disease. The TNM staging system also determines the choice of initial treatment as well as the stratification of patients in clinical trials.

Surgical resection is the treatment of choice for patients with localized non-small cell lung cancer (NSCLC). Unfortunately, at the time of diagnosis, fewer than 25% of all NSCLC patients are considered candidates for surgical therapy. Despite improvements in surgical techniques, current 5-year survival rates are only slightly better than they were 20–30 years ago, since two-thirds of all recurrences following a complete resection occurr distantly. Therefore, like breast cancer, bronchial carcinoma is a systemic disease.

For almost 15 years, radical lymphadenectomy and/or systematic nodal dissection has been performed in order to improve survival rates after surgery. In the field of thoracic surgical oncology, the usefulness of systematic nodal dissection in the treatment of NSCLC is still a matter of controversy. It is my opinion that the advantage of systematic nodal dissection is its ability to improve staging of the tumor rather than survival benefit, since radical systematic mediastinal lymphadenectomy does not influence disease-free or overall survival in patients with NSCLC and without overt lymph node involvement. However, a small subgroup of patients with limited mediastinal lymph node metastases might benefit from a systematic lymphadenectomy [1]. Given the poor prognosis that NSCLC acquires when it spreads to the mediastinum, the use of systemic therapy before local therapy, generally referred to as induction or neoadjuvant treatment, aims at preventing the growth of systemic disease and, at the same time, shrinking locoregional macroscopic disease, possibly allowing it to be adequately treated by surgery. It also allows early eradication of micrometastasis and prolongs overall survival. Various Phase II studies with new third-generation, platinum-based combinations have confirmed the high activity and very good tolerability of induction therapy that can currently be considered a standard treatment with Stage IIIA–N2 NSCLC [2].

In addition, almost 40% of patients with postsurgical non-small cell Stage I bronchogenic carcinoma developed recurrent cancer, with most (55.6%) being nonregional metastases [3]. Efforts to improve the outcome of these patients with early-stage NSCLC have focused on adjuvant or neoadjuvant chemotherapy. Last year, Winton and colleagues showed that adjuvant treatment with vinorelbine plus cisplatin prolonged disease-free and overall survival among patients with completely resected early-stage NSCLC [4]. While there is growing evidence to support adjuvant chemotherapy following radical surgery, compared with surgery alone, it is still unclear whether neoadjuvant chemotherapy in early-stage NSCLC is equally effective. Thus, the tumor stage does not justify the singular resection of the tumor as an adequate oncological treatment, except for Stage I NSCLC, whereas multimodal therapy concepts lead to better results.

The anatomically based TNM system remains useful for predicting the survival of a patient group, although an individual survival calculation is not possible. Biomarkers can perhaps solve these problems by subdividing the classical tumor stages into subsets, thus allowing a better individual prediction of survival and/or selection of therapy. However, a wide range of potential prognostic markers has been assessed in lung cancer, and the various studies have yielded conflicting results. Significant reasons for this are not only inadequate validations by independent investigators in a larger series of cases, but also the lack of robust and reproducible systems and techniques to accurately quantify the changes in biomarkers [5]. Moreover, there will be no success in reflecting the complex process of tumorigenesis and metastases based on one factor. Several factors will be required for accurate molecular staging in the context of patients’ risk profiling. Currently, serial analysis of gene expression (SAGE) and the high-throughput array method are established techniques that allow expression analyses of many genes at a time. In addition to multiple gene expression analysis by DNA arrays, the simultaneous detection of differentially expressed proteins has recently been established by proteome techniques, which have rapidly become an important tool for research.

Biomarkers could play a significant role prior to cancer diagnosis (screening), at the time of diagnosis and after diagnosis (in monitoring therapy, selecting additional therapy and detecting recurrence). I will now focus on applications of various biomarkers at the time of diagnosis, and especially on targeted therapy.

Microarray examinations allow for relatively fast biomarker evaluation, which enables us to differentiate between normal and tumor tissues and/or structures within the individual tumor entities, since the ability of malignant cells to proliferate and metastasize can be ascribed to cellular alterations, including the genetic profile. Our data has demonstrated that a quotient of only two genes, RAGE and Cyclin-B2, is a reliable marker to differentiate normal from malignant lung or lung metastases with a different histological origin [6]. Moreover, a considerable differentiation of histological NSCLC subpopulations (e.g., squamous cell vs adenocarcinoma) is possible on the basis of differential gene expression, as the development of these neoplasias relates to different cells in the lung.

Subdividing traditional tumor classes into subsets behaving differently to each other, using biomarkers, can be performed by unsupervised classification (clusters) of gene expression. Gene expression profiling projects reported that hierarchical clustering analysis of the gene expression patterns could distinguish the major morphological classes of lung cancer (i.e., small-cell carcinoma, squamous cell carcinoma and adenocarcinoma) and also define subgroups of adenocarcinoma tumors [7,8]. However, subclassification of adenocarcinoma appeared to be different within projects, which also has an effect on survival differences. This variability arises as a result of biological differences among the samples and populations considered in the studies, technological differences between the platforms and random technical variation. Quantifying these three sources of variation separately would be helpful, however, this would require further studies.

The initiation of therapies based on mechanisms that target critical molecular pathways of tumors has evoked considerable interest. There are a number potential targets in NSCLC but, curently, there is no obvious critical molecular key, making the development of targeted therapy challenging. Presently, one of the most discussed molecular pathways that drives tumor growth is triggered by the epidermal growth factor receptor (EGFR). Two major classes of EGFR inhibitors are currently under development: the small molecule tyrosine kinase inhibitors (e.g., gefitinib and erlotinib) and monoclonal antibodies (e.g., cetuximab) directed against the EGFR. These EGFR-targeted antibodies are administered in nonsurgical patients with advanced or metastatic NSCLC and often in combination with classical chemotherapeutics. The gefitinib studies, Iressa^® Dose Evaluation in Advanced Lung Cancer (IDEAL)-1 and -2, demonstrated objective tumor response and symptoms improvement [9,10]. However, the Iressa NSCLC Trials Assesing Combination Treatment (INTACT)-1 and -2 failed to demonstrate an improved patient survival after first-line therapy of gefitinib in combination with two different chemotherapy regimens [11,12]. Interestingly, a different gene- xpression profile of lung cancer patients was identified by cDNA microarray analysis, which allowed a separation of responders and nonresponders to gefitinib therapy [13]. Furthermore, Lynch and colleagues found that EGFR mutations correlated with gefitinib response and survival [14]. The tyrosine kinase inhibitor, erlotinib, significantly improved overall and progression-free survival [15]. Although the presence of an EGFR mutation could also increase responsiveness to erlotinib, this was not indicative of a survival benefit [13].

The vascular endothelial growth factor (VEGF) is one of the most efficient growth factors of endothelial cells involved in the process of neoangiogenesis. Data from current trials suggest that anti-VEGF monotherapy was less efficient than standard chemotherapy. However, direct blocking of serum VEGF by means of the monoclonal anti-VEGF antibody, bevacizumab, significantly supported the antitumor effect of cytostatics, including carboplatin and paclitaxel [16]. In this context, recent data from the E4599 randomized Phase II/III trial of the Eastern Cooperative Oncology Group (ECOG) confirmed, at the American Society of Clinical Oncology (ASCO) annual meeting 2005, that the addition of bevacizumab to paclitaxel plus carboplatin in patients with advanced NSCLC provides a statistically and clinically significant survival advantage with tolerable toxicity [17]. This combination therapy of bevacizumab, paclitaxel and carboplatin is the ECOG’s new treatment standard in this patient population.

The list of further potentially targeted therapies affecting a variety of pathways involved in processes of tumorigenesis and metastasizing is long and is beyond the scope of this editorial [18]. Expression profiling analysis by microarrays and proteomic analysis could be crucial in the discovery of new drug targets as well as a source of potential pharmacodynamic end points. However, of potential interest is the possibility that there may be a subset of NSCLC patients who have molecular characteristics that predict them to be responsive. This may not relate simply to the level of expression of the target gene (e.g., EGFR), but could feasibly correlate with the flux through the signal transduction pathway or with the expression of any number of genes that could be detected by microarray profiling [19].

Currently, targeted therapy does not play a role in multimodal concepts with lung cancer surgery. A range of drugs that target the molecular pathology of advanced lung cancer are now undergoing clinical trials; but in the foreseeable future, the success of this initial wave of molecular therapeutics will arrive in neoadjuvant, as well as adjuvant, therapies for our surgically treated patients.

Until that time, our task will be to evaluate patients at a high risk of relapse within a tumor stage. Thus far, molecular staging has been attempted using only one prognosis factor, failing clinically with regard to its reproducibility. I am firmly convinced that complex biological processes in such a heterogeneous tumor entity may only be assessed on the basis of several prognosis factors. To date, this has been achieved in two studies [20] [Hofmann and colleagues, unplublished observations], which identified at risk patients using the expression of at least two genes. Despite these works being the first step towards molecular staging, two genes will certainly not be sufficient. In the near future, tumor-specific chips will hopefully be established, allowing for a risk stratification of patients via several molecular prognosis factors up to pattern recognition (clusters). The capability for making customized gene chips/microarrays is currently widely available. This might prove very helpful for the choice of an individual therapy for each patient, as well as for the development of novel drugs to overcome acquired therapy resistance.