Future directions of lung cancer: Histology-based treatment, molecular markers

HemOnc Today recently convened its well-respected Section Editors to present on ‘hot’ or controversial topics they deemed important to their subspecialty area. Presented here is Joseph Aisner, MD’s, discussion of his hit list of topics that he believed fulfilled this qualification. HemOnc Today will continue to follow and expand coverage of these topics in the ensuing months. We hope you find these executive overviews informative.

The field of lung cancer, as in most cancers, is moving in the direction of individualized medicine. In the last several decades, we have seen some important accomplishments.

First of all, we have seen the five-year overall lung cancer survival increase from about 5% to 15%. For stages IIa–IIIa disease, postoperative chemotherapy has decreased the risk for recurrence and improved survival by about one-third. For regionally advanced disease (stages IIIa-b), we nearly doubled the survival due to combined concurrent modality therapy.

In advanced disease, we also learned that the traditional chemotherapy agents are usually given as doublets, and among these doublets, there has been no major distinctions as far as their effect on disease progression and survival — although there are differences in toxicity. More recently, initial antiangiogenic therapy, as well as epidermal growth factor receptor inhibiting for salvage therapy, have been found helpful in stage IV disease, and both are being tested in earlier stage disease.

In a large trial by Sandler and colleagues, data indicated that the addition of bevacizumab (Avastin, Genentech) increased the response and the median duration of survival in the bevacizumab-treated population compared with a bevacizumab-nontreated population. However, squamous tumors and patients with metastasis to the brain were excluded. A regimen including bevacizumab has now become a standard of care for bevacizumab-eligible patients in the United States.

All of these discoveries came from a unitary approach to all lung cancer, and we have, up to this point, treated all lung cancer the same. But is lung cancer really just one disease?

Until recently, the approach for treating lung cancer has been empiric, and there has been no rational guidance for the use of many of the drugs we have available.

However, there are some data that indicate this situation is likely to change.

Histology

The first data set comes from a large Japanese autopsy series. The series included more than 5,000 autopsies in patients with lung cancer.

This autopsy series showed that patients with well differentiated squamous carcinoma died predominantly (>90%) of locoregional disease as compared with the other subtypes who died predominantly (>90%) of disseminated disease. Although these data have been available for more than 50 years, we have not paid a lot of attention. Yet, this early data suggests that the therapy approach might vary by histology subtype.

Based on this autopsy series, we might wish to focus therapy for squamous histology on improving local control. This concept is well supported by past EORTC and RTOG trials.

Joseph Aisner

In these trials, enhanced local control methods (eg, hyperfraction) appeared better for the squamous subtypes. Recently, we have heard suggestions that pemetrexed (Alimta, Eli Lilly) may be more useful for the subtype for adenocarcinoma. Based on fairly scanty data, pemetrexed has become widely used in practice as a first approach to adenocarcinoma. Most of the data come from subanalyses of trials, and, at best, none of these trials were powered to answer the question.

A recent publication by Scagliotti and colleagues, which was better powered to answer this question, found a small advantage favoring pemetrexed when compared with gemcitabine in patients with adenocarcinoma. Although justifying a pemetrexed-based combination as first-line therapy for adenocarcinoma, the advantage does not mandate this approach, and it is not known if a pemetrexed-based combination is comparatively better than other regimens. Histology has also become the basis for adding bevacizumab as squamous tumors are currently excluded.

New molecular markers

Another developing area of considerable interest in lung cancer is the use of molecular markers to help guide therapeutics. Most prominent, in terms of implications for treatment approaches, are EGFR mutation status or gene copy number. Currently, EGFR targeting agents are used in the United States for salvage therapy and some practitioners have started using them as first-line therapy.

The mounting question is, should EGFR targeted therapy be used as first-line treatment, and in whom, and how will assessing EGFR mutation status impact upon this. In the IPASS study, researchers addressed this question in a meaningful way.

This study randomly assigned patients to therapy with either chemotherapy or gefitinib (Irissa, AstraZeneca). It showed that the group whose tumors did not express EGFR mutation did far better with chemotherapy than the tyrosine kinase inhibitor, whereas the group whose tumors carried an EGFR mutation did better with the EGFR TKI than chemotherapy. Although this needs to be validated in a Western population, it is already having an effect in the clinic in terms of first-line use of the TKIs and is probably to great advantage in the appropriate patients. However, this does not justify the use of this agent as first-line therapy in the elderly or poor performance patients in the absence of appropriate data.

There are additional molecular markers that have not yet spread to wide-scale use but have the potential for enormous impact, and even more markers are being studied. ERCC-1 and RRM1 are two such markers for which preliminary data suggest a possible role for selecting appropriate therapy.

ERCC-1, or Excision Repair Cross Complementation group 1, plays an important role in the repair of interstrand cross-links in DNA and, as such, recognizes and removes cisplatin-induced DNA adducts. This enzyme complex thereby produces resistance to cisplatin. Thus, its expression is predictive of cisplatin failure. In addition, ERCC-1 expression is also a prognostic factor, and its presence is favorable for survival.

The IALT trial showed the predictive value of ERCC-1 dramatically. The group that had ERCC-1 low-expressing tumors actually had the benefit of the adjuvant chemotherapy, whereas the group that had high expression did not benefit from the platinum-based chemotherapy.

The confounding issue in evaluating the impact of ERCC-1 as a predictive factor is that ERCC is a strong prognostic factor for survival. The population whose tumors express ERCC-1 have better survival than the patients who have ERCC-1 low-expression tumors. This is one of the reasons why the selection trials, while showing a response advantage, have not been able to show a survival advantage for therapy selected based on the ERCC-1 expression. This confusion resulted from comparing those treated with high vs. low ERCC-1 levels.

However, if we consider that patients who would, by prognostic factors, have fared worse than a comparative group but by rationally assigning therapy bring them back to a survival comparable with the better prognostic group, we would need to consider that something beneficial occurred. Thus, while we have not been able to show comparative survival benefit by selection, we must wonder if the right comparator was used.

Another marker is RRM1, which exerts its impact on gemcitabine. RRM1 catalyzes the reduction of ribonucleoside diphosphatase to corresponding deoxyrubonucleotides. Gemcitabine targets exactly this process, so low levels of RRM1 are associated with an increased response to gemcitabine, whereas high levels are associated with resistance. This marker shares the same issues as ERCC-1.

RRM1 is also a powerful predictive and prognostic factor. If we are going to compare a good prognostic group to a bad prognostic group, the survival differences by treatment assignment may again be strongly influenced by comparison groups.

In addition, the RRM1 gene is situated close to the ERCC-1 gene. They are often co-expressed. Although prospective selection trials of RRM1 are not as mature as the ERCC-1 trials, one would suspect that the same kind of confusion will prevail with RRM1 as do with ERCC-1.

There are also several other predictive biomarkers that may impact therapy selection moving forward, including VEGF polymorphisms, KRAS mutations status, p53 mutation status, BRAF mutation states and EML4-ALK translocation.

Cardinal to the adequate investigation and implementation of these and other markers in assisting and guiding therapeutic decisions is the need to obtain adequate tissue and/or developing technology to identify these markers in the small specimens obtained by the less invasive procedures.

The data, already developed, show that we are on the verge of a much more rational approach for selecting appropriate therapy and excitingly offer the possibility that we will identify different cohorts and build enormous gains in this deadly disease.

Joseph Aisner, MD, is Chief Medical Officer at the Cancer Institute of New Jersey, and is Section Editor of the HemOnc Today Lung Cancer section.