Nonsmall cell lung cancer (NSCLC) is a heterogeneous disease, of which our knowledge regarding specific molecular subtypes is rapidly increasing. Oncogenic mutations in the epidermal growth factor receptor (EGFR) and KRAS genes account for nearly 15% and 22%, respectively, of all cases of adenocarcinoma, the most common histological subtype.1 A recent addition to this list is rearrangements in the anaplastic lymphoma kinase (ALK) gene, which are noted in approximately 7% of patients with adenocarcinoma of the lung.1, 2 The discovery of ALK gene rearrangement has been rapidly followed by development of ALK inhibitors and has recently culminated in the approval of crizotinib by the US Food and Drug Administration (FDA) for the treatment of patients with advanced stage, ALK-positive NSCLC. ALK gene rearrangements are associated with certain clinical phenotypes such as never-smoker, adenocarcinoma histology, and younger age.3 In particular, the presence of signet-ring features on light microscopy is more common with this subtype of NSCLC.4 Beyond this, our knowledge of the clinical characteristics of ALK-positive patients is limited and is only beginning to emerge.
In this issue of Cancer, Doebele and colleagues report on an important study that compared the clinical characteristics of patients with ALK-positive NSCLC to those of patients with EGFR or KRAS mutation or none of the 3 events (triple negative).5 Patients with ALK-positive tumors had nearly 5-fold higher prevalence of pleural and/or pericardial involvement compared with patients in the other groups. In addition, the total number of metastatic lesions was higher for ALK-positive patients. The data were collected retrospectively from a limited sample size of approximately 200 patients in total, of whom 41 had ALK-positive disease. In addition, the types of imaging studies obtained in these patients were variable and reflected the real world practice outside of a clinical trial protocol. It is unlikely that the detection of pleural or pericardial metastases would be much different between a conventional computed tomography (CT) scan and a positron emission tomography/CT scan. However, the observed prevalence of brain metastasis could have been influenced by the type of imaging study used. In addition, a higher proportion of patients with triple-negative NSCLC had not undergone an imaging study of the brain. Therefore, it is not possible to make conclusions about differences in brain metastasis based on the genotype of lung cancer. Despite these limitations, the results are important for several reasons. Pleural and pericardial involvement is often associated with disabling symptoms in patients with lung cancer. Malignant involvement of the pleura is often difficult to measure by Response Evaluation Criteria in Solid Tumors and could be a confounding factor for determination of objective response in patients treated with ALK inhibitors. Appropriate management of pleural or pericardial disease by either interventional approaches and/or initiation of systemic therapy with crizotinib would be helpful for effective palliation. These clinical features could also form a part of a clinical enrichment strategy for screening for ALK status in NSCLC patients. Although we believe that routine testing of all lung cancer specimens for common treatable oncogenic events will be done in the not too distant future, a clinical enrichment strategy will continue to play an important role in the daily practice in many parts of the world. The presence of retinal metastasis in 2 patients with ALK-positive disease is another finding that needs further evaluation in larger cohorts of patients. Given that visual disturbances are known to occur with crizotinib treatment, it is important for practicing physicians to be aware of the possibility of disease involvement of the retina as a potential etiological factor. Undoubtedly, ongoing large randomized studies in ALK-positive patients will provide a better estimate of the actual burden of these clinical features.
The development of appropriate diagnostic tools to identify ALK positivity has become a major area of investigation since the demonstration of clinical benefit in patients with ALK-positive NSCLC. The initial clinical trials used a fluorescence in situ hybridization (FISH) assay to detect ALK-positive tumors.6 This assay uses probes to hybridize to each end of the ALK gene breakpoint. In patients with ALK-positive disease, the signals from the 2 probes are noted as being further apart or have a strong red signal indicating the site of inversion. The FISH assay has been approved as a companion diagnostic by the FDA for treatment with crizotinib. Using the FISH assay for detection of ALK positivity has its own challenges, because the positive signal could be subtle or missed altogether. A positive result is indicated by the presence of a split signal or dominant red fluorescent signal in at least 15% of the nuclei. In another paper published in this issue of Cancer, Camidge and colleagues report on the correlations between the percentage of tumor cells that harbor the ALK rearrangement, ALK signal copy number, and the efficacy of crizotinib therapy.7 Ninety specimens with an ALK-positive signal by the FISH assay were included in the analysis, of which 30 were from patients treated with crizotinib. An increase in the red signal copy number was correlated with a higher percentage of ALK-positive cells. Only about 55% of the cells had ALK positivity, and only 54% had the split signal. A third of the patient samples had a single red pattern, and a small proportion had both. In the subset of patients who received crizotinib therapy, there was no correlation between the number of cells with positive signals or the ALK gene copy number and the degree of clinical response. The authors appropriately conclude that the lack of detection of an ALK-positive signal in 100% of the cells is likely related to technical factors rather than a true biological effect. It is needless to say that the observations are derived from a relatively small cohort of patient samples, and further work in a larger patient population is necessary. Regardless, this important work highlights the limitations of the FISH assay and also calls for the need to identify more sensitive tools to select patients for therapy with crizotinib. Early reports using immunohistochemistry (IHC) and reverse transcriptase polymerase chain reaction (RT-PCR) have demonstrated promising results.8 The IHC method is easy to perform and correlates with FISH positivity in patients with a strong protein expression. The RT-PCR method has the added advantage of detecting ALK translocation with other known fusion partners, but requires high-quality tissue specimens.
The paper by Camidge and colleagues is timely. It illustrates the need to rapidly validate the testing methods and ensure that every patient with ALK-positive NSCLC has an opportunity to receive appropriate therapy. In the case of EGFR inhibitors, therapeutic agents were developed before the appropriate target population was known. This provided us additional opportunities to test the utility of EGFR inhibitors in a broader patient population, which led to knowledge regarding the prognostic and predictive value of EGFR mutations. By contrast, the ALK example illustrates a challenge at the other end of the spectrum. Identification of the target and an effective therapeutic strategy in short order limits the ability to perform large studies to understand the natural history of the disease, prognostic value of the biomarker, and therapeutic potential of non–ALK-targeted agents. For these reasons, the work done by Doebele et al and Camidge et al provide valuable information regarding ALK-driven adenocarcinoma of the lung.