The past 2 decades have been transformative for the treatment of non–small cell lung cancer; two-thirds of patients have been found to harbor molecular drivers, and a growing proportion of these are now targetable with FDA-approved drugs.
The past 2 decades have been transformative for the treatment of non—small cell lung cancer (NSCLC); two-thirds of patients have been found to harbor molecular drivers,1 and a growing proportion of these are now targetable with FDA-approved drugs.
Among these drivers are chromosomal rearrangements of the ROS1 receptor tyrosine kinase gene, which result in oncogenic fusions of portions of ROS1 and a second gene. Although ROS1-positive tumors (those with ROS1 rearrangements) are uncommon, homology between ROS1 and anaplastic lymphoma kinase (ALK) and parallels between ROS1- and ALK-positive NSCLC have led to ALK inhibitors being repurposed in the setting of ROS1-positive disease.
Crizotinib (Xalkori), already approved for the treatment of patients with ALK fusions, became the first drug specifically indicated for patients with ROS1-positive NSCLC in 2016.2-4 Meanwhile, novel clinical trial designs facilitated the subsequent approval of entrectinib (Rozlytrek) for ROS1-positive disease in 2019.5,6
Table. Select Clinical Trials Examining ROS1-Targeting Agents (Click to Enlarge)
Efforts to tackle these 2 key issues are driving significant forward motion, with the development of next-generation ROS1 inhibitors designed to have improved CNS penetration and activity against a broad spectrum of ROS1 resistance mutations.3,9,11
Lorlatinib (Lorbrena) was developed to address these challenges in ALK-positive NSCLC12 and is already approved in that setting.13 Recently published results from a phase 1/2 trial suggest that it might soon be added to the treatment armamentarium for ROS1-positive disease as well.
The newer drugs repotrectinib (TPX-0005) and taletrectinib (formerly DS-6051b) better address ROS1 mechanisms of resistance and have shown preliminary activity against the most common ROS1 resistance mutation,14 G2032R.15,16 In the ongoing TRIDENT-1 study (NCT03093116), repotrectinib demonstrated a promising intracranial response rate (IC-ORR) of 100% in 3 tyrosine kinase inhibitor (TKI)-naïve patients with CNS disease.17 Results of the expanded phase 2 cohort may help determine whether repotrectinib is the highly sought-after answer to both CNS progression and resistance to crizotinib.
Ongoing clinical development of TKIs targeting ROS1-positive cancers, primarily in NSCLC, comprises FDA-approved drugs and novel agents, according to a search of ClinicalTrials.gov (Table).
A Rare Driver With Parallels to ALK
Located on chromosome 6q22,9 the ROS1 gene encodes an orphan receptor tyrosine kinase for which an activating ligand has yet to be described. Little is known about the physiological functions of the ROS1 protein; it is found in a variety of embryonic tissues and organs but displays limited expression in adults. ROS1 is the sole member of its own subfamily within the broader insulin receptor family and is closely related to ALK, another kinase in this family.11
The capacity of ROS1 activity to drive cancerous transformation in vitro was uncovered in the 1980s, when it was found to be the oncogene product of a chicken sarcoma virus.18-20 However, it was not until 2003 that chromosomal rearrangements involving the ROS1 gene were linked to the development of human cancer.21
That initial report involved a fusion between FIG and ROS1 genes in cell lines derived from a patient with glioblastoma.21 Since then, a range of ROS1 gene fusions have been described across numerous cancer types.3,4,9 Notably, many have been found in patients with NSCLC,4,11 most commonly the CD74-ROS1 fusion.22 Overall, however, ROS1 fusions are rare, with a frequency of up to approximately 3% in patients with adenocarcinoma histology and less than 1% in those with non-adenocarcinomas.23,24
The precise mechanisms underlying the activation of ROS1 in these gene fusions are not fully understood.4 The resulting fusion proteins have constitutive ROS1 kinase activity, which upregulates downstream signaling pathways, including the MAPK, PI3K/AKT, and JAK pathways, which play key roles in numerous hallmarks of malignant transformation (Figure 1).3,9
Gene fusions in ROS1 and ALK in NSCLC were first described in the same landmark study in 2007.25 ROS1 shares significant sequence homology with ALK; 64% of the amino acids within their kinase domains are the same, as are 84% of those within the adenosine triphosphate (ATP)-binding domain. ROS1- and ALK-positive NSCLC also have similar clinical characteristics: patients are usually younger, light or never smokers, and have adenocarcinoma histology.3,4,9,11
Crizotinib Breaks New Ground
The successful development of ALK inhibitors has positioned gene fusions as an important class of actionable driver alterations in NSCLC. Because of the homology between the 2 proteins, most of the currently available ALK inhibitors also target ROS1.11 Crizotinib was originally developed as an inhibitor of the MET kinase but was subsequently found to be a potent inhibitor of both ALK and ROS1.26
Click to Enlarge
In the PROFILE 1001 expansion cohort, crizotinib demonstrated an objective response rate (ORR) of 72% and median progression-free survival (PFS) of 19.2 months.2 Based on these results, crizotinib received FDA approval in 2016 for the treatment of ROS1-positive NSCLC.30 After a median follow-up of 62.6 months, findings from an updated analysis of PROFILE 1001 demonstrated a median overall survival (OS) of 51.4 months.31
Despite initial responses, most patients treated with crizotinib eventually develop progressive disease,3,9 and retrospective studies have demonstrated that many also develop brain metastases or experience progression of preexisting intracranial disease,14,32 although prospective data are limited. Crizotinib has been shown to poorly penetrate the protective BBB,8 and this has been proposed to be a limiting factor to its efficacy in patients with NSCLC, in whom CNS progression is common.7
In ALK-positive NSCLC, following up crizotinib with second-generation ALK inhibitors is standard, and the same strategy has been pursued in patients with ROS1 fusions. Like crizotinib, entrectinib is an ATP-competitive inhibitor of ROS1. It also targets ALK and NTRK1/2/3 activity.3,4,9 In preclinical trials, entrectinib demonstrated penetration of the BBB,10 suggesting it could provide better intracranial disease control than crizotinib.
Expanding Frontline Standard of Care
The development of entrectinib exemplifies innovative clinical trial strategies aimed at testing novel drugs more effectively in patients with rare driver alterations, such as in ROS1. The drug was evaluated in a phase 2 basket trial that enrolled patients on the basis of specific molecular alterations rather than tumor histology.
To date, entrectinib has not shown activity in patients previously treated with ALK or ROS1 inhibitors.33 In contrast, an integrated efficacy analysis of 53 patients with TKI-naïve ROS1-positive NSCLC treated with entrectinib across 3 clinical trials (2 phase 1 trials and a phase 2 basket trial) showed an ORR of 77% and median PFS of 19 months.5
In a subset of 23 patients with CNS metastasis at baseline, the ORR was 74% and median PFS was 13.6 months, compared with 80% and 26.3 months in the 20 patients without CNS disease. The IC-ORR was 55% in the CNS disease subset, and the median intracranial duration of response was 12.9 months. Among 134 patients evaluated for safety (a larger group that also included non—TKI-naïve patients), the most common treatment-related adverse events (TRAEs) included dysgeusia, dizziness, constipation, diarrhea, weight gain, fatigue, paresthesia, and nausea. Most were grade 1 or 2; less common serious AEs were managed with dose alteration.5
On the basis of these data, entrectinib received regulatory approval in August 2019 for the treatment of ROS1-positive NSCLC.6 Comparison of crizotinib or entrectinib with chemotherapy in randomized phase 3 trials is challenging because of the rarity of ROS1 fusions, but prospective clinical trials and retrospective analyses strongly suggest the superiority of ROS1 TKIs in the frontline setting, and they have become the standard of care.34
The choice of optimal TKI is currently unclear, but the improved CNS penetration of entrectinib would argue for its superiority.34 Phase 3 randomized trials comparing crizotinib and entrectinib are unlikely due to the small numbers of patients with ROS1-rearranged NSCLC, and direct comparison from clinical trial data is difficult because of differences in trial design. However, the results of a recent innovative “virtual clinical trial” offer some insight.
According to findings presented at the 2019 American Society of Clinical Oncology Annual Meeting, investigators searched a large database and found deidentified electronic patient records for patients with ROS1-positive NSCLC treated with crizotinib (n = 69) who matched the enrollment requirements for 3 published entrectinib trials (n = 53).5 Although there are challenges to this approach, the comparison demonstrated a longer time to treatment discontinuation and improved PFS for patients treated with entrectinib.35
Guidelines from numerous professional societies recommend testing for ROS1 fusions in patients with advanced lung adenocarcinoma.36-38 The Oncomine Dx Target Test, a multibiomarker analysis of alterations in 23 NSCLC-associated genes including ROS1, is approved as a companion diagnostic for crizotinib treatment.39 There is no FDA-approved companion diagnostic test available for entrectinib, although Foundation Medicine is expected to seek an entrectinib indication for its FoundationOneCDx assay, which includes ROS1 aberrations and rearrangements in its 324-gene sequencing panel.40 Immunohistochemistry, fluorescence in situ hybridization, and next-generation sequencing can all be used to identify ROS1 rearrangements, each with its own benefits and limitations (Figure 238).3,9,38
Another oral ATP-competitive ROS1 inhibitor with improved BBB penetration is ceritinib (Zykadia).3,4,9 This agent is also an inhibitor of ALK and is approved for the treatment of both TKI-naïve and crizotinib-resistant ALK-positive NSCLC.41 In a clinical trial (NCT01283516), ceritinib demonstrated activity in both crizotinib-sensitive and crizotinib-resistant patients with ALK-positive NSCLC.42
In a phase 2 clinical trial of ceritinib (NCT01964157) among 30 crizotinib-naïve patients with ROS1-positive NSCLC, ORR was 67% and median PFS was 19.3 months, similar to the results of phase 2 trials of crizotinib and entrectinib. Among 8 patients with CNS disease, the IC-ORR was 25% and the intracranial disease control rate was 63%. Unfortunately, there were no responses in the 2 patients previously treated with crizotinib.43 Ceritinib is not currently approved by the FDA for patients with ROS1-positive NSCLC, but it could offer another frontline treatment option in the future.
The Challenge of Resistance
Although clinical trials of entrectinib and ceritinib have included limited patient numbers, these studies’ results indicate that neither agent has activity in crizotinibresistant patients,33,42 suggesting a clear need for drugs with even greater CNS penetration. In addition, the lack of response highlights another significant challenge to the clinical efficacy of ROS1 inhibitors: the development of resistance.
Analogous to findings in ALK-positive NSCLC, point mutations in ROS1 have been identified that drive resistance to crizotinib in patients with ROS1-positive disease; a retrospective study found ROS1 mutations in approximately 50% of crizotinib-resistant tumors.14 Those identified to date include a gatekeeper mutation (L2026M), which affects the accessibility of the ATP-binding pocket, and several solvent front mutations.3,9
Solvent front mutations are particularly recalcitrant and include the most common ROS1 resistance mutation, G2032R.14 Most of the currently available ROS1 inhibitors have an extra chemical group that extends into the solvent front, an area just outside the kinase domain, and mutations that result in the substitution of large, bulky amino acids in this area can impede the binding of these ROS1 inhibitors.15
Although it has not been tested directly, entrectinib is not expected to be effective against the G2032R mutation based on its poor activity against an analogous ALK mutation.10,44
Lorlatinib is a next-generation inhibitor of ROS1 and ALK approved for second- or thirdline treatment of ALK-positive NSCLC.13 It has demonstrated some ability to cross the BBB,12,45 and in preclinical studies, it inhibited cells with the L2026M and G2032R ROS1 mutations.46
In a recent phase 1/2 clinical trial (NCT01970865), 69 patients with ROS1-positive NSCLC were treated with lorlatinib; 21 were TKI naïve, 40 were previously treated with crizotinib, and 8 were previously treated with another ROS1 TKI or 2 or more ROS1 TKIs. Objective responses were observed in 62%, 35%, and 13% of patients, respectively, and median duration of response was 25.3 months, 13.8 months, and 5.6 months.47
One patient with the K1991E mutation and 1 with the S1986F mutation both experienced a partial response; however, the best response in the 6 patients with the ROS1 G2032R mutation was stable disease, although 4 experienced some tumor shrinkage. Intracranial responses occurred in 64% of TKI-naïve patients and 50% of crizotinib-resistant patients with CNS metastases.
Lorlatinib was generally well tolerated. Forty-three percent of patients experienced grade 3 or 4 TRAEs, the most common of which were hypertriglyceridemia (19%) and hypercholesterolemia (14%). Serious TRAEs were observed in 7% of patients.47
Brigatinib (Alunbrig), another FDA-approved ALK inhibitor for second-line treatment of ALK-positive NSCLC,48 is active against many ALK mutations associated with crizotinib resistance,49 several of which are analogous to ROS1 resistance mutations. In a phase 1/2 study in patients with advanced malignancies, there were 3 patients with ROS1-positive NSCLC. Two of these patients, 1 of whom had been previously treated with crizotinib, experienced an objective response.50
FIGURE 2. Testing for ROS1 Fusions38 (Click to Enlarge)
Preclinical studies have suggested that cabozantinib potently inhibits the growth of cell lines with several ROS1 solvent front mutations, including G2032R, and recent case reports have indicated activity in patients with ROS1-positive NSCLC who have resistance to crizotinib and ceritinib.4,51-53
The Next Generation
Two newer ROS1 inhibitors have recently shown promise in phase 1 trials in patients with crizotinib-resistant disease. Taletrectinib was formerly being developed by Daiichi Sankyo as DS-6051b; in late 2018, the drug was licensed to AnHeart Therapeutics.54 It was recently reported that the company had received clearance from China’s Center for Drug Evaluation to begin 2 phase 2 trials and that it also plans to initiate phase 2 trials in Japan and the rest of the world, including the United States.55
Taletrectinib is a dual inhibitor of ROS1 and NTRK1/2/3. It dramatically inhibited the growth of G2032R-mutant cancers in vitro and in vivo and showed efficacy in crizotinibresistant ROS1-positive cancers with the G2032R mutation in preclinical models.16 A firstin- human study in patients with advanced solid tumors included 6 patients with ROS1-positive NSCLC who had previously been treated with crizotinib and were evaluable for response. Two of these patients experienced a partial response and 2 achieved stable disease.56
Turning Point Therapeutics developed repotrectinib, an inhibitor of ROS1, ALK, and NTRK1/2/3 that was specifically designed to overcome solvent front mutations; since it is more compact, it fits more tightly in the ATP-binding pocket. It is also designed to have better CNS drug—like properties and improved BBB penetration.57 In preclinical models, repotrectinib had potent activity against the G2032R and D2033N solvent front mutations in ROS1.15
Preliminary phase 1 results from the ongoing TRIDENT-1 trial (NCT03093116) were recently reported. In this trial, 33 patients with ROS1-positive NSCLC, including 18 patients with CNS metastases, received escalating doses of repotrectinib (40 mg once daily to 200 mg twice daily). The ORR was 82% in the 11 TKI-naïve patients, and the IC-ORR was 100% in the subset of 3 with measurable CNS disease at baseline; these values were 39% (7/18) and 75% (3/4), respectively, in patients who had been treated with 1 prior TKI. Among 5 patients with the G2032R mutation, the ORR was 40%.
Safety was evaluated in the larger cohort of 83 patients, which included non-NSCLC solid tumors and rearrangements of ALK or NTRK, in addition to ROS1. The most common TRAEs were dizziness, dysgeusia, dyspnea, fatigue, constipation, paresthesia, anemia, and nausea; most were grade 1 or 2 and manageable. There were 4 dose-limiting toxicities at the 160-mg twice daily and 240-mg once daily doses and 4 grade 5 treatment-emergent AEs, including 1 sudden death that was potentially related to treatment.17
Jane de Lartigue, PhD, is a freelance medical writer and editor based in Gainesville, Florida.
Cabozantinib is unique in that it was not originally designed as an inhibitor of ALK or ROS1, although it includes the latter among its many targets. This multitargeted kinase inhibitor has antiangiogenic properties due to its inhibition of the VEGF and MET pathways. It is approved for use in patients with medullary thyroid cancer (marketed as Cometriq) and in patients with advanced renal and hepatocellular carcinomas (marketed as Cabometyx).
Clinical evaluation in patients with ALK-positive NSCLC culminated in the approval of crizotinib for this indication by the FDA in 2011.27 As case reports of crizotinib efficacy in ROS1-positive NSCLC began to emerge the subsequent year, the ongoing PROFILE 1001 phase 1 clinical trial of crizotinib in advanced ALK-rearranged NSCLC was amended to include an expansion cohort of patients with ROS1 fusions (NCT00585195).3,28,29
Disease progression to the central nervous system (CNS) presents a major challenge for patients with NSCLC.7 Crizotinib efficacy is limited by its inability to effectively cross the protective blood-brain barrier (BBB).8 Although entrectinib has better CNS activity, it faces a second major hurdle: the development of ROS1 resistance mutations.9,10