Brain Metastases

Patients with metastases to the brain were allowed in all ALK inhibitor trials. Patients were eligible to participate in the study if the metastases were neurologically stable for at least 2 weeks before enrollment and patients had no ongoing requirement for glucocorticoids. The proportion of patients with brain metastases enrolled in the randomized trials of ALK-TKIs varies from 26% for

592 EML4-ALK Fusion Gene and Therapy with ALK-Targeted Agents in NSCLC

the initial crizotinib trial [25] to 59% in the cetirinib trial. As patients with brain metastases are very likely to respond to ALK-TKI, local therapy with either surgery or radiation can be deferred, if clinically stable and unlikely to deteriorate during initial ALK-TKI therapy.

The frequency of intracranial disease progression seems to be similar with crizotinib and chemotherapy [25]. However, in the alectinib studies, there was a significant improvement seen in PFS for patients with brain metastases compared to crizotinib for an impressive HR of 0.08 in the Japanese study and 0.40 in the international study (95% CI 0.01-0.61 and 0.25-0.64; respectively); the duration of response was also much better for patients receiving alectinib (17.3 vs. 5.5 months) [61, 62]. Brigatinib also demonstrated superior efficacy versus crizotinib, demonstrating a lower intracranial disease progression of 9% versus 19% [56]. Cetirinib was also superior to chemotherapy with an intracranial response of 73% versus 27%. Lorlatinib, as previously discussed, was developed to overcome ALK-resistance and improved CNS penetration resulting in it being efficacious for patients with progressive disease and CNS metastases. Finally, in the phase 2 trial, alectinib demonstrated an intracranial response rate of 63% and a median duration of intracranial response of 14.5 months [64]. For patients progressing to alectinib in the CNS, a case report suggested that an increase in the dose to 900 mg twice daily can be an effective therapeutic approach [52].

ROS1-Positive Non-Small Cell Lung Cancer

The c-ROS oncogene 1 (ROS1) gene belongs to the subfamily of tyrosine kinase insulin receptor genes. It was first discovered in 1986 as the homolog of the chicken c-ros, which is the proto-oncogene. The ROS1 encodes an orphan receptor tyrosine kinase similar to ALK, along with members of the insulin-receptor family [74]. ROS1 and ALK are phylogenetically related [75, 76]. Both proteins display an extracellular ligand-binding domain, a transmembrane-spanning region, and an intracellular tyrosine kinase domain. They mediate signaling via the JAK/STAT/RAS/MAPK and PI3K pathways.

ROS1 was first discovered as the oncogene product of an avian sarcoma RNA tumor virus [77]. ROS1 is activated by chromosomal

ROSl-Positive Non-Small Cell Lung Cancer 593

rearrangement in a variety of human cancers, including NSCLC, cholangiocarcinoina, gastric cancer, ovarian cancer, and glioblastoma [78-82]. Rearrangements lead to the fusion of a portion of ROS1 that includes the entire tyrosine kinase domain with 1 of 12 different partner proteins [83]. The resulting ROS1 fusion kinases are constitutively activated and drive cellular transformation. ROS1 rearrangements occur in 1% to 2% of NSCLC via a genetic translocation between ROS1 and other genes, the most common of which is CD74 [84-86].

Mechanism of ROS1 Activation

ROS1 activates several downstream signaling pathways related to cell differentiation, proliferation, growth, and survival. It phosphorylates and activates predominantly STAT3, which is required for ROS1 anchorage-independent growth [87]. It also phosphorylates VAV2, which is a guanine-nucleotide exchange factor for Rho GTPases that regulates actin dynamics and gene expression [88]. It also activates the PI3K/AKT/mT0R signaling pathway [89, 90].

Clinical and Pathologic Features of ROS1 NSCLC

The typical demographic features of ROSl-mutated NSCLC include younger patients, and never smokers and adenocarcinoma histology. In a pivotal single-arm phase 2 trial that enrolled 50 patients, the median age was 53 years (range 25-77). The males included in the study represented 44% of the population while females were 56%. No race differences were observed. Never smokers represented 78% of the cohort and former smokers 22%. The predominant histology was adenocarcinoma (98% vs. 2%, squamous cell carcinomas) [85].

Searching for ROS1 Fusions

Several methodological approaches for ROS1 fusion searching have been reported. FISH is the gold standard in the clinical setting. FISH is the technique that can detect almost all known and unknown ROS1 rearrangements [91]. Immunohistochemistry (IHC), using the commercially available anti-ROSl antibody D4D6 clone is also

594 EML4-ALK Fusion Gene and Therapy with ALK-Targeted Agents in NSCLC

an efficient method of evaluation [92-94]. IHC is user friendly and less costly. It appears to have a sensitivity approaching 100% but variable specificity (from 73-100%) depending on the cutoff value adopted and the complementary technique used to confirm a possible rearrangement. Reverse transcriptase PCR (RT-PCR) can be used to identify ROS1 rearrangements [95].

As screening tool, RT-PCR allows identification of specific fusion partners and includes rapid turnaround time and limited tissue requirements. It has high sensitivity and specificity. However, results depend on the RNA quality, which can be degraded in formalin-fixed paraffin-embedded tissues. Next-generation sequencing has emerged as a powerful approach to look for genetic abnormalities. Whole-genome and whole-transcriptome sequencing has been used to identify ROS1 rearrangements [24]. Although NGS is a very attractive option, it is expensive and requires additional infrastructure and time for analysis [96]. Although they are rare mutations, current guidelines recommend to analyze for the presence of ALK and ROS1 in parallel [97, 98].

Management of ROS1 NSCLC

The ROS1 tyrosine kinase is highly sensitive to the tropomyosin receptor kinase (TRK) ROS1 inhibitor entrectinib as well as crizotinib, due to a high degree of homology between the ALK and ROS tyrosine kinase domains [85]. Both drugs entrectinib and crizotinib are approved for patients with the ROS1 translocation, including those who have received chemotherapy and those who are treatment naive.

Entrectinib potently inhibits kinases encoded by the NTRK and ROS1 genes. It achieves therapeutic levels in the CNS with antitumor activity in intra cranial tumor mo dels [99]. The approval of entrectinib is based on the results of three phase 1/2 entrectinib trials with data integrated in a large cohort of adults with ROS1 fusion-positive NSCLC ROSl-positive or NTRKL fusion-positive solid tumors [100]. Across studies, 53 patients with ROS1 mutation positive NSCLC, the ORR was 55% (31.5-76.9%), median duration of response was 12.9 months (5.6 months versus not evaluable), and median PFS was 7.7 months (3.8-19.3 months). Entrectinib induced clinically meaning durable responses, including responses in CNS metastases.

ROSl-Positive Non-Small Cell Lung Cancer 595

Evidence suggests that R0S1 is another therapeutic target of the ALK inhibitor crizotinib. First, the kinase domains of ALK and ROS1 share 77% amino acid homology within the ATP-binding sites. Consistent with this homology, crizotinib binds with high affinity to both ALK and ROS1 [101]. Second, the studies of the inhibition of kinase autophosphorylation in cell-based assays have found that ALK and ROS1 are sensitive to crizotinib with a half-maximal inhibitory concentration of 40 to 60 nM. Third, in cell lines expressing ROS1 fusions, crizotinib potently inhibits ROS1 signaling and cell viability [85,102,103]. Finally, case reports have described marked responses to crizotinib in patients with ROS-1 rearranged NSCLC [104].

In a multicenter international expansion cohort of the phase 1 study of crizotinib, 50 patients with advanced NSCLC and ROS1 rearrangements were enrolled. The patients were treated with crizotinib at the standard oral dose of 250 mg twice daily and assessed for safety, pharmacokinetics, and response to therapy. ROS1 fusion partners were identified with the use of next-generation sequencing or reverse-transcriptase-polymerase-chain-reaction assays. An objective response rate of 72% (95% CI 58% to 84%) was demonstrated; 3 patients achieved complete response and 33 partial responses. The median duration of response was 17.6 months (95% CI 14.5 months to not reached). The median PFS was 19.2 months (95% CI 14.4 months to not reached). No correlation was observed between the type of ROS1 rearrangement and the clinical response to crizotinib. The safety profile of crizotinib was similar to that for the patients enrolled with EML4-ALK mutations [105].

Mechanism of Resistance

Two mechanisms of resistance to crizotinib have been described: a secondary mutation that hinders drug binding [106] and activation of EGFR, which enables cancer cells to bypass crizotinib-mediated inhibition of ROS1 signaling [107].

For those patients with disease progressing on crizotinib, treatment options include lorlatinib, ceritinib, brigatinib, or alectinib. In a phase 2 trial that included 28 evaluable patients with ROSl-rearranged NSCLC, cediranib demonstrated an ORR of 62%, duration of response 21 months, mPFS of 9.3 among all patients and

• 596 EML4-ALK Fusion Gene and Therapy with ALK-Targeted Agents in NSCLC
• 19.3 months for crizotinib, naive patients. The mOS was 24 months [108]. A preliminary report of 47 patients with ROSl-positive NSCLC, of which 34 had received prior crizotinib, treated with lorlatinib, the response rate in the crizotinib-nai've was 62%, while response in those previous exposed to crizotinib was 27% [109].

For the patients who progressed on prior ROSl-targeted therapies, nontargeted approaches can be utilized, including chemotherapy and immunotherapy. Investigational therapies are being tested in early-phase clinical trials in ROS1 NSCLC and include repotrectinib and DS-6051b [110, 111].