Abstract
Molecular profiling of nonsmall cell lung cancer (NSCLC) contributes to better understanding the different molecular subtypes of this heterogeneous group of diseases. The discovery of oncogenic ALK rearrangements in NSCLC and the subsequent success in their therapeutic targeting with crizotinib reinforces the benefits of a precision approach to systemic anticancer therapy. In addition, the rapid development of crizotinib from first discovery thorough accelerated US Food and Drug Administration approval, and late stage confirmatory clinical trials, exemplifies the success of the drug development strategy of close collaboration between clinicians, industry and regulatory authorities. In this review we describe the identification of ALK rearranged NSCLC, clinical characteristics of such patients, and clinical outcomes when treated with crizotinib.
Introduction
The discovery of oncogenic driver mutations in nonsmall cell lung cancer (NSCLC) and recognition of different molecular subtypes of disease have revolutionized the treatment and prognosis of this disease. The identification of somatic epidermal growth factor receptor (EGFR) mutations sensitive to EGFR tyrosine kinase inhibitors (TKI) and the superiority of EGFR TKIs over chemotherapy in this patient population has changed the management of advanced NSCLC [Mok et al. 2009; Popat et al. 2014]. Identification of a further mutually exclusive molecular subtype of lung adenocarcinoma with ALK rearrangements that is remarkably sensitive to the TKI, crizotinib [Soda et al. 2007], has again changed clinical practice and emphasizes the importance of somatic molecular analysis alongside histopathological diagnosis [Travis et al. 2013].
Discovery of ALK rearrangements in NSCLC
Following the discovery of chromosomal translocations involving the ALK locus in anaplastic large cell lymphoma and inflammatory myofibroblastic tumours, analysis of the transcriptome of a lung adenocarcinoma specimen first described a fusion product of the ALK and EML4 loci [Soda et al. 2007]. The product of the 2p chromosome translocation is transforming secondary to anaplastic lymphoma kinase (ALK) activity, which is redistributed to the cytoplasm from the cell membrane.
Subsequently, several fusion variants have been identified involving the ALK locus and EML4. ALK fusions with the TFG [Rikova et al. 2007] and KIF5B loci [Takeuchi et al. 2009] have also been identified (Figure 1). Functional ALK rearrangements are estimated to occur in approximately 4–8% of NSCLC and are in the vast majority mutually exclusive of other oncogenic driver mutations such as EGFR and KRAS [Gainor et al. 2013; Kris et al. 2014]. Screening of specimens from a cohort of 754 NSCLC patients identified EML4-ALK fusions in 4.24% (32 patients). The E13:A20 (variant 1, 19 cases) and the E6a/b:A20 (variant 3a/b, 10 cases) fusions were the most frequently identified. Other variants, such as E18:A20 and E20:A20, were less commonly identified. The case series did not detect any TGF-ALK or KIF5B-ALK fusions, suggesting the rarity of these gene fusions [Soda et al. 2012]. Patients with ALK-positive tumours tend to be younger than average for lung cancer and have a history of light or no tobacco exposure [Kwak et al. 2010].
Variants of EML4-ALK, TFG-ALK and KIF5B-ALK fusions identified in NSCLC.
Therapeutic targeting of ALK rearrangements
Crizotinib, a competitive inhibitor of both ALK and c-Met kinase, has shown remarkable efficacy in ALK-positive NSCLC since early phase development. A high-throughput kinase inhibitor screen against 602 cancer cell lines identified that crizotinib reduced the viability of the NSCLC cell line NCI-H3122, in which an EML4-ALK variant 1 fusion mRNA is detectable, by suppression of Akt and Erk signalling [McDermott et al. 2008]. In xenograft models of anaplastic large cell lymphoma, crizotinib administration caused dose-dependent tumour responses and tumour cell apoptosis [Christensen et al. 2007].
The dose escalation phase of an open-label phase I study of crizotinib in molecularly unselected advanced NSCLC (PROFILE 1001) initially demonstrated impressive radiological responses in two patients who were sub-sequently shown to harbour ALK rearrangements [Kwak et al. 2010]. Subsequently, approximately 1500 samples were prospectively screened using fluorescence in situ hybridization (FISH) for ALK rearrangements to recruit an expanded cohort of 82 ALK-positive patients. Although no subtypes of NSCLC were excluded from molecular screening, the vast majority (96%) of ALK rearrangements were identified in adenocarcinomas; 48% of ALK-positive adenocarcinoma cases had solid histological pattern with mucin and 38% were mixed-pattern type adenocarcinoma. All cases within this cohort were negative for EGFR mutation or MET amplification. These patients received a dose of 250 mg crizotinib twice daily (BID). The overall radiological response rate to crizotinib was 57% [95% confidence interval (CI) 46–68] and 87% of patients had controlled disease (stable or responding to therapy) at 8 weeks, notable in a patient cohort in which 94% had previously treated NSCLC. Later updated data from this trial of 143 patients with ALK-positive advanced NSCLC concluded a 60.8% (95% CI 52.3–68.9) objective response rate, with a median duration of response of 49.1 weeks (95% CI 37.3–75.4); 84% of subjects had received at least one previous systemic treatment prior to crizotinib [Camidge et al. 2012].
Initial findings from the PROFILE 1005 phase II single arm study of crizotinib in pretreated ALK-positive NSCLC were presented at the American Society of Clinical Oncology meeting in June 2011 [Crino et al. 2011] and were updated at the World Lung Conference in July 2011 [Kim et al. 2011]. In this study, 93% of patients treated with crizotinib had received at least two prior lines of treatment. Of the 76 patients with an evaluable response, 83% had target lesion shrinkage on crizotinib and 41 of 76 patients had a response according to Response Evaluation Criteria In Solid Tumors (RECIST) criteria.
Following the interim analysis of this phase II study and the PROFILE 1001 results, the US Food and Drug Administration (FDA) granted accelerated approval for crizotinib in August 2011 [Jefferson, 2011]. Evaluation in January 2012 of the first 261 patients enrolled in PROFILE 1005 showed a progression-free survival (PFS) of 8.5 months (95% CI 6.2–9.9) [Kim et al. 2012], way in excess of the 2.7 months previously observed for docetaxel monotherapy [Shepherd et al. 2000]; 77% of evaluable patients remained on treatment with a median duration of treatment of 25 weeks and median duration of response of 43 weeks (95% CI 36–50). This study also reported a clinically relevant improvement in quality of life on crizotinib (p < 0.05) [Blackhall et al. 2012].
Crizotinib phase III study
The PROFILE 1007 randomized phase III trial was opened to objectively demonstrate crizotinib efficacy versus standard-of-care chemotherapy in the second-line advanced NSCLC setting. All participants had ALK-positive advanced NSCLC centrally verified by FISH and had received standard-of-care platinum doublet chemotherapy. Participants were randomized 1:1 to crizotinib or either single agent docetaxel or pemetrexed chemotherapy, receiving docetaxel if tumour had squamous histology or in case of prior pemetrexed therapy [Shaw et al. 2013] and, on progression, patients randomized to chemotherapy crossed over and became eligible for PROFILE 1005.
At the time of data analysis, both patient cohorts had been on study for just over 12 months. The primary endpoint of PROFILE 1007 was of PFS, and this was 7.7 months for crizotinib and 3.3 months for chemotherapy [hazard ratio (HR) =0.49, 95% CI 0.37–0.64, p < 0.0001]. The high response rate seen in earlier trials was reproducible with 65% response rate (95% CI 58–72) on crizotinib compared with 9% (95% CI 2–16) with docetaxel and 29% (95% CI 21–39) with pemetrexed, demonstrating a differential improved efficacy with pemetrexed over docetaxel for ALK-positive NSCLC.
At time of primary analysis, 49 deaths (28%) had occurred on crizotinib and 47 (27%) on chemotherapy. Only 40% of the number of deaths required for analysis of overall survival (OS) had occurred and there was no significant difference between the two arms (HR =1.02, 95% CI 0.58–1.54, p = 0.54) with median survival of 20.3 months (for crizotinib) versus 22.8 months for chemotherapy. Of note, 64% of chemo-therapy patients went on to receive crizotinib on progression.
Quality of life was a secondary endpoint and analysis confirmed improvements in quality of life compared with chemotherapy. The time to deterioration of cough, chest pain or dyspnea was significantly longer on crizotinib than chemotherapy (5.6 months versus 1.4 months; HR = 0.54, p < 0.0001). A significant improvement in global quality of life (p < 0.0001) was also observed on crizotinib, in contrast to no improvement on chemotherapy.
This study validated the previous accelerated FDA approval of crizotinib based on PROFILE 1001 and 1005 in ALK-positive advanced NSCLC [Kazandjian et al. 2014] and led to European Medicines Agency (EMA) approval in ALK-positive advanced NSCLC previously treated with one line of platinum-based therapy [Committee for Medicinal Products for Human Use].
Crizotinib in treatment-naïve advanced ALK-positive nonsquamous NSCLC
PROFILE 1005 allowed recruitment of patients without previous chemotherapy and the activity of crizotinib reported led to accelerated FDA approval for crizotinib to include patients that had not failed platinum-doublet chemotherapy. However, the randomized evidence that crizotinib is superior to platinum doublet chemotherapy, came from the PROFILE 1014 trial [Mok et al. 2014]. Here, patients with locally advanced, recurrent or metastatic ALK-positive nonsquamous NSCLC were randomized 1:1 to crizotinib 250 mg BID or up to six cycles of platinum-pemetrexed chemotherapy. Similar proportions of patients had stable treated brain metastases in both study arms (26% versus 28%). The median time from diagnosis in both arms was 1.2 months. Never smokers made up 62% of the crizotinib cohort and 65% of the chemotherapy cohort. The vast majority of patients enrolled had adenocarcinoma (92% crizotinib arm, 93% chemotherapy arm). The 343 patient study showed superior median PFS of 10.9 months on crizotinib versus 7.0 months on chemotherapy (HR 0.454, 95% CI 0.35–0.61, p< 0.0001) and a higher overall response rate of 74% (95% CI 67–81) versus 45% (95% CI 37–53, p< 0.0001). Median time to response was quicker with crizotinib (6.1 weeks, range 2.7-41.4) than with chemotherapy (12.1 weeks, range 5.1–36.7) and median duration of response was prolonged with crizotinib [49 weeks, 95% CI 35.1–60 versus 22.9 weeks (95% CI 18.0-25.1)] with chemotherapy).
At the point of data analysis, 120 (71%) chemotherapy patients had crossed over to crizotinib following progressive disease on chemotherapy and 68% of all participants were still in follow up. No significant improvement in median OS survival with crizotinib has been shown as yet (HR 0.821, 95% CI 0.54–1.26, p= 0.1804), likely due to crossover. Toxicities with crizotinib were as in previous trials with this compound.
Even though survival data from this study are yet to mature and will be biased by crizotinib crossover, given the large PFS and response rate advantage coupled with a favourable safety profile, crizotinib is likely to become a standard first choice therapy for patients with ALK-positive advanced NSCLC.
Summary of key clinical trials with crizotinib in NSCLC.
CI, confidence interval; HR, hazard ratio; NE, not evaluated; NSCLC, nonsmall cell lung cancer.
Survival benefit with crizotinib
Formal evidence for an OS advantage for crizotinib has been difficult to prove given the accelerated nature of the drug development programme, rendering it unethical to withhold drug from ALK-positive patients. Therefore, OS results of PROFILE 1007 have been biased by marked crossover. In order to try to quantify the OS benefit of crizotinib, retrospective analysis of patients screened for the (1001) phase I study was performed comparing survival outcomes for ALK-positive patients that received crizotinib with those that did not [Shaw et al. 2011]. Here, crizotinib-treated ALK-positive patients had not yet reached median OS at the time of analysis, with a 1-year survival of 71% (95% CI 58–91) and a 2-year survival of 57% (95% CI 40–71), and was comparable with OS of a separate control cohort of ALK-negative EGFR mutant NSCLC treated with EGFR-TKI [1 year OS 74% (95% CI 61–83), 2 year OS 52% (95% CI 38–65)], compared with crizotinib-naïve ALK-positive untreated patients, with a median survival of 20 months (95% CI 13–16).
Brain as a site of progressive disease on crizotinib
Crizotinib penetrates the blood–brain barrier poorly [Costa et al. 2011] and the brain is the commonest (45%) reported site of new disease in ALK-positive patients treated with crizotinib in the PROFILE 1001 and 1005 clinical trials [Otterson et al. 2012]. A retrospective exploratory analysis of patients recruited to PROFILE 1005 and 1007 has shown that 20% of patients with progressive disease developed new brain metastases [Costa et al. 2015]. Brain as the site of progression in PROFILE 1014 was approximately 15% and crizotinib was numerically superior to platinum-doublet chemotherapy for intracranial time to progression (HR 0.6, 95% CI 0.34–1.05 in all patients; HR 0.45, 95% CI 0.19–1.07 in patients with baseline brain metastases p = not significant) [Solomon et al. 2014]. Intracranial oligoprogressive disease can be treated with palliative radiation or ablative therapy followed by continuation of crizotinib [Takeda et al. 2013], systemic chemotherapy or newer small molecule ALK inhibitors [Shaw et al. 2014]. In an intriguing single case-report, escalation of crizotinib dose to 1000 mg daily has been reported to control oligoprogressive cranial disease in a single patient [Kim et al. 2013].
Isolated central nervous system (CNS) failure can be successfully controlled with local therapy and crizotinib continued beyond progression with a further period of disease stability of at least 4 months [Weickhardt et al. 2012b; Takeda et al. 2013]. Retrospective analysis suggests that continuation of crizotinib beyond progression confers a median survival benefit of 11 months (HR 0.38, 95% CI 0.22–0.66; p = 0.0005) compared with other subsequent systemic therapy. Furthermore, limited early phase trials data suggest that new more potent ALK inhibitors seem to have CNS activity in crizotinib refractory patients [Gadgeel et al. 2014]. The optimum treatment of oligoprogressive cranial disease may therefore evolve over the coming years, as these agents become a more routine part of clinical practice.
Toxicity profile of crizotinib
In the PROFILE 1007 trial, whilst more adverse events were most commonly reported in the crizotinib-treated group after controlling for disease progression (1768 versus 1190 events) [Shaw et al. 2013], the incidence of grade 3–4 adverse events and serious adverse events was similar in both groups, and many crizotinib-associated adverse events were grade 1–2. Adverse events that occurred with over 5% greater incidence in the crizotinib-treated group were elevated liver aminotransferases, oedema, vision disorders, constipation, nausea, vomiting, dizziness, upper respiratory tract infections and taste changes.
Liver dysfunction and hepatitis
Grade 3–4 elevated aminotransferases were observed in 16% of patients on crizotinib in PROFILE 1007, with one grade 5 hepatic toxicity, and in general is easily managed by drug hold and dose reduction. However, only 1% of patients required crizotinib discontinuation due to hepatotoxicity and the EMA gives clear guidelines on the management of this toxicity [Committee for Medicinal Products for Human Use, 2014b].
Hypogonadism secondary to crizotinib
Low serum testosterone levels have been observed in male NSCLC patients receiving crizotinib that were not identified from PROFILE 1001 or 1005 [Weickhardt et al. 2012a]. In a small case series, this has been shown to be independent of age, prior cytotoxic therapy, presence of brain metastases, prior whole brain irradiation, concurrent medications and nutritional status. A marked fall in testosterone is generally noted within days of starting crizotinib and serum testosterone levels change according to dose interruptions and recommencement of therapy in some patients. Falling luteinizing hormone (LH) and follicle-stimulating hormone (FSH) accompany the fall in serum testosterone in some patients, suggesting crizotinib has a central affect, but in others LH and FSH remained elevated or in the normal range, suggesting crizotinib also has an additional direct gonadal effect. Low testosterone levels can contribute to reduced quality of life [Fleishman et al. 2010] and testosterone replacement therapy should be considered in patients with symptomatic, proven hypogonadism on crizotinib therapy.
Bradycardia
Another effect that has been described in clinical practice since the early development of crizotinib is a decrease in heart rate. A retrospective analysis of all patients enrolled in PROFILE 1005 and the crizotinib arm of PROFILE 1007 study (1053 participants) concluded that the mean decrease in heart rate was 25 beats/minute [standard deviation (SD) 15.8] and that 41.9% of patients had a least one reading of sinus bradycardia (<60 beats/minute) [Ou et al. 2014]. Of the patients with sinus bradycardia, only 5.9% had a lowest reading below 45 beats/minute. Systolic blood pressure drops of >40 mmHg did not accompany sinus bradycardia in the vast majority of cases (99.1%), although this was slightly more common when there was concomitant use of beta-blockers (1.7% had a >40 mmHg decrease in systolic blood pressure, 2% had a >30 mmHg decrease in diastolic blood pressure). Non-Asian patients and older patients (⩾65 years-old) were more likely to develop sinus bradycardia (p = 0.003 and p = 0.0039, respectively). Analysis of a small patient cohort suggests that sinus bradycardia may correlate with maximum tumour shrinkage (p = 0.0205) and overall response rate (p = 0.0195) [Ou et al. 2013]. Whilst this may reflect drug exposure, this study also found that bradycardic patients remained on crizotinib significantly longer than those with normocardia (52.9 weeks versus 24.6 weeks, p = 0.005), as a potential covariate.
Molecular mechanisms of crizotinib resistance
Several different molecular mechanisms have been observed thus far in crizotinib-resistant NSCLC. As with EGFR TKIs in advanced NSCLC [Kobayashi et al. 2005], secondary mutations in the ALK kinase domain have been discovered in patients who become resistant to crizotinib treatment. The ALK L1196M mutation corresponds to the position of the EGFR T790M mutation and the resulting amino acid change may similarly sterically inhibit inhibitor binding. Other ALK kinase mutations detected in crizotinib-resistant tumours include L1152R, C1156Y and G1269A – another mutation that may interfere with crizotinib binding [Doebele et al. 2012; Choi et al. 2010; Sasaki et al. 2011].
Other molecular resistance mechanisms of crizotinib resistance, which are again analogous to those seen with EGFR TKIs [Sequist et al. 2011], include ALK amplification and addition of other oncogenic mutations such as in EGFR and KRAS. Loss of the ALK fusion gene, in the absence of any additional mutations, has also been reported in crizotinib-resistant disease [Doebele et al. 2012]. In addition, in vitro data suggest that EGFR pathway activation, in the absence of EGFR amplification or mutation, can contribute to crizotinib resistance and that this effect can be overcome with dual EGFR and ALK inhibition [Yamaguchi et al. 2014; Sasaki et al. 2011].
Conclusion
Crizotinib is an effective treatment in ALK-positive advanced nonsquamous NSCLC with a tolerable side effect profile and a remarkable response rate. In addition, the median duration of response exceeds that of platinum-pemetrexed chemotherapy. The superiority of crizotinib over standard-of-care chemotherapy in untreated patients has been demonstrated. Intracranial oligoprogressive disease on crizotinib can be treated using various strategies and does not necessarily mandate a change from crizotinib. Particular side effects of crizotinib such as bradycardia and low testosterone may require monitoring but are well tolerated. The success of targeting ALK rearrangements with crizotinib emphasizes the importance of molecular profiling of advanced NSCLC at diagnosis to allow selection of the most appropriate therapeutic agent for disease palliation.
Footnotes
Conflict of interest statement
The authors declare no conflicts of interest in preparing this article.
Funding
The authors acknowledge NHS funding to the Royal Marsden NHS Foundation Trust/The Institute of Cancer Research NIHR Biomedical Research Centre.