Advanced sequencing techniques have revolutionized the understanding of how cancer has developed and enabled treatment selection based on genomic characterization of the tumor.
Over the past few decades, our knowledge of the role of gene mutations in oncogenesis has transformed cancer care.1 Advanced sequencing techniques have revolutionized our understanding of how cancer has developed and enabled treatment selection based on genomic characterization of the tumor.2 An actionable mutation or genomic event refers to a detected DNA change that predicts a patient's treatment response to a particular agent. An actionable mutation may be an oncologic driver, or it may be relevant to a target that can be inhibited pharmacologically.2 Assessing response to a drug among a molecularly defined patient subset across cancer types is increasingly common with the use of umbrella or basket trials.2 In 2017, the FDA approved the first oncologic therapy based on genomic biomarkers independent (or agnostic) of cancer site or histology.3The discovery and targeting of gene fusions was an important oncologic breakthrough in the last century.4 Gene fusions arise from genomic rearrangements, which include chromosomal inversions, duplications, deletions, or translocations.4
Gene fusions were first discovered in the 1970s in patients with chronic myeloid leukemia, from whom circulating tumor cells containing the t(9;22)(q34;q11) translocation were identified.4 Interchromosomal translocations were subsequently recognized in salivary gland adenomas, such as t(3;8)(p21;q12), and sarcomas, such as t(11:22)(p24;q12).4 In the decade after these discoveries, gene fusions were increasingly described in solid tumors.5 As of February 2018, 11,207 gene fusions have been identified in more than 68,000 patient cases of malignancies.6Gene fusions may serve as drivers for both cancer development and progression.4 Chromosomal rearrangements can lead to the fusion of 2 genes, creating chimeric proteins that serve as strong oncogenic drivers.4 The mechanisms by which oncogenic fusions lead to cancer development or progression may include altered transcription and constitutive kinase activation.Altered transcription is one mechanism by which a gene fusion event may drive oncogenesis.4 A fusion event may involve a transcription factor.4 For example, in prostate cancer, the TMPRSS2-ERG fusion protein decreases expression of the androgen receptor, in addition to inhibiting existing androgen receptors, and leads to the disruption of cell differentiation, resulting in the selection for non- androgen-dependent cellular proliferation.4 Altered transcription may also result from the fusion of a promoter to a proto-oncogene, augmenting its expression, as in COL1A1-PDGFB fusion in dermatofibrosarcoma protuberans.4Kinase fusions are commonly targeted oncogenic mechanisms. The fusion of 2 genes can result from chromosomal rearrangements, creating chimeric proteins that are strong oncogenic drivers. One partner in these types of fusions is often a kinase.4 Signal transduction in eukaryotic cells is mostly mediated by protein kinases.7 Through protein phosphorylation, kinases play a critical role in intercellular communication and in mediating physiological responses.7 Protein kinases control many cellular processes by modifying substrate activity. These cellular processes include transcription, cell cycle progression, apoptosis, and differentiation, to name a few.7 During genomic fusion events, the kinase activity is often preserved. Hence, kinase fusions result in constitutive activation and amplified downstream signaling.4 Tyrosine kinase fusions, including ALK, NTRK1/2/3, ROS1, and RET have been identified in a variety of cancer types.4,8
Serine-threonine kinase fusions, including those involving BRAF, CRAF, and MAST1/2, have also been reported.4,8 These kinase fusions lead to amplified signaling in pathways involved in cell growth and survival.4
AKT indicates v-akt murine thymoma viral oncogene homologue (also known as protein kinase B); BDGF, brainderived growth factor; DAG, diacylglycerol; ERK, extracellular signal-regulated kinase; GAB1, GRB2-associated protein 1; GRB2, growth factor receptor-bound protein 2; IP3, inositol trisphosphate; MAPK, mitogen-activated protein kinase; NGF, nerve growth factor; NTF-3, neurotrophin 3; PI3K, phosphatidylinositol-4,5-bisphosphate 3-kinase; PIP2, phosphatidylinositol 4,5-bisphosphate; PKC, protein kinase C; PLC, phospholipase C; RAF, rapidly accelerated fibrosarcoma kinase; RAS, rat sarcoma kinase; SHC, Src homology 2 domain.
Reprinted with permission. Amatu A, Sartore-Bianchi A, Siena S. NTRK gene fusions as novel targets of cancer therapy across multiple tumour types. ESMO Open. 2016;1(2):e000023. doi:10.1136/esmoopen-2015-000023.
A number of tyrosine kinase inhibitors have been developed or are in development. 4 The majority of these compounds are multikinase inhibitors with activity against more than 1 kinase.4 One of these kinase fusions, neurotrophic tropomyosin receptor kinase (NTRK) gene rearrangement, has emerged as a novel target for cancer therapy.The tropomyosin receptor kinase (TRK) receptor family consists of 3 transmembrane proteins: TRKA, TRKB, and TRKC receptors. TRKA is encoded by the NTRK1 gene, whereas TRKB is encoded by the NTRK2 gene, and TRKC by the NTRK3 gene.9
In normal physiology, TRKA, TRKB, and TRKC receptors are abundantly expressed in neuronal tissues and regulate cell differentiation, proliferation, apoptosis, and survival of neurons in the central as well as peripheral nervous systems through intracellular signaling pathways.9
All 3 TRK receptors have an extracellular domain for ligand binding, a transmembrane domain, and an intracellular region containing a kinase domain.10 TRKA, TRKB, and TRKC receptors are activated by neurotrophins. 9 These neurotrophins include nerve growth factor (the primary ligand for TRKA), brain-derived neurotrophic factor (the primary ligand for TRKB), and neurotrophin 3 (the primary ligand for TRKC). Binding of these neurotrophin ligands may be neurotrophin concentration and cell-type-dependent, and the TRK receptors do not always bind exclusively to their primary ligands.9 For example, neurotrophin 3, the primary ligand for TRKC, may also activate TRKA and TRKB.9 Ligand binding leads to oligomerization of the TRK receptors and specific tyrosine residue phosphorylation. 10 Tyrosine phosphorylation leads to signal transduction through downstream pathways, resulting in transcription and other cell programs that regulate synaptic plasticity, neurite repair, repair or prevention of neurodegeneration, cell proliferation, neuron maintenance, or apoptosis.9
MAPK indicates mitogen-activated protein kinase; PI3K, phosphatidylinositol-4,5-bisphosphate 3-kinase; PLCg, phospholipase C-g; TM, transmembrane; TRK, tropomyosin receptor kinase. Reprinted with permission from Farago AF, Le LP, Zheng Z, et al. Durable clinical response to entrectinib in NTRK1-rearranged non-small cell lung cancer. J Thorac Oncol. 2015;10(12):1670-1674.
Specifically, binding of TRKA leads to activation of the Ras/mitogen-activated protein kinase pathway, resulting in extracellular signal-regulated kinase (ERK) signaling and increased cellular growth and proliferation. Phospholipase C-gamma (PLC-gamma) and phosphatidylinositol 3-kinase (PI3K) pathways are also activated. TRKB ligand binding leads to activation of the Ras-ERK, PI3K, and PLC-gamma pathways, culminating in neuronal differentiation and survival. TRKC receptor activation leads to activation of the PI3K/protein kinase B (AKT) pathway, which prevents apoptosis and potentiates cell survival. Figure 1 shows these 3 major pathways involved in TRK receptor-mediated cell differentiation and survival.10
In summary, ligand binding and activation of TRK receptors lead to phosphorylation and activation of signal transduction pathways, bringing about proliferation, differentiation, and survival in neuronal cells.9 TRK oncogenic fusions may use many of the same downstream signal transduction cascades to effect cancer driver function, because the kinase domain and tyrosine binding sites are often preserved in the fusion event.9In NTRK oncogenic fusions, the 5' region of a fusion partner gene is joined with the 3' region of the NTRK gene via interchromosomal or intrachromosomal rearrangement, resulting in an oncogenic chimera.10 The oncogenic chimera typically contains a constitutively overexpressed or activated kinase.10 The oncogenic fusion partner gene contains a dimerization domain.11 In the absence of a ligand (Figure 2), normal TRK proteins are unable to dimerize and therefore cannot activate downstream signaling pathways.11 In a partner-TRK fusion protein, however, dimerization occurs in a ligand-independent manner, as the fusion partner contains the dimerization domain.11 Dimerization leads to constitutive kinase activation and downstream signaling, resulting in cell proliferation and survival.11 Therefore, the chimeric partner-TRK fusion protein, or the fusion oncoprotein, can be aberrantly expressed and constitutively active, resulting in initiation of downstream pro-oncogenic pathways.
NTRK fusions are typically found independent of other oncogenic events and are considered oncogenic drivers.8The first NTRK gene fusion was identified in colon cancer and contained sequences from nonmuscle tropomyosin (TPM3).9 Increasing amounts of tumorsequencing studies bring attention to the wide variety of cancer types associated with NTRK gene fusions, such as lung adenocarcinoma, glioblastoma, and papillary thyroid carcinoma, among others (Table 1).9,10
Although NTRK fusions occur infrequently in many common adult and pediatric cancers, they are found at high frequencies in certain rare pediatric tumors, such as infantile fibrosarcoma and papillary thyroid cancer.12In specific patient populations, routine testing for oncogenic fusions is a part of their clinical care, because the presence of an actionable genetic alternation has inherent therapeutic implications.4 For example, the FDA has approved companion diagnostic tests for the detection of ALK and ROS1 fusions in patients with non-small cell lung cancer (NSCLC).13-15 FDA-approved companion diagnostics are not yet available, however, to address other malignancies.13
The availability of improved technology has enabled genomic profiling to detect actionable oncogenic alterations, so that patients may be able to be matched to genotype-driven targeted therapy.16 Data supporting genomic profiling to inform targeted therapy are emerging. Among patients assessed with tumor genomic profiling, 30% to 49% have an actionable genetic alteration.16,17 In 500 consecutive patients with advanced cancer from various disease sites, analyses of archived specimens showed that 30% of patients had actionable mutations.16 In a prospective study of patients with locally advanced or metastatic cancer, 49% had actionable gene targets.17
The frequent presence of actionable mutations has also been reported in pediatric patients. Among pediatric patients with a variety of central nervous system (CNS) and non-CNS solid tumors, 39% had a clinically relevant alteration of some kind, such as a medically actionable mutation.18 In yet another multicenter study, among pediatric patients with refractory, recurrent, high-risk extracranial cancer, 31% had an actionable mutation indicating potential utility of a matched targeted therapy.19 Tumor profiling led to individualized cancer therapy recommendations in these patients.19
These data suggest that when genomic profiling technologies are available, actionable gene alterations are commonly found and are of clinical significance. The encouraging results of genomic profiling are shifting the paradigm of cancer treatment toward precision oncology. In this new model, rather than using empiric chemotherapy, therapy is selected using oncogenic mutations as predictive biomarkers.9 The presence of a predictive genomic biomarker may help select patients with a high likelihood of benefiting from an oncogene-targeted therapy.9 For example, targeted therapy with oral crizotinib in patients with ALK-positive NSCLC is superior to chemotherapy for progression-free survival.20 Improved outcomes with oncogene-targeted therapies suggest that even low-frequency oncogenes should be investigated as therapeutic targets.9Currently, NTRK gene fusions can be detected by several testing platforms.
DNA Fluorescence in Situ Hybridization. Fluorescence in situ hybridization (FISH) is among the most commonly used laboratory methods for detecting chromosomal rearrangements.21 In this method, fluorescent-labeled probes bind to a specific complementary sequence in a biological sample.4
For solid tumors, FISH has been the method of choice for detecting recurrent mutations.22 Several NTRK fusions have been discovered using FISH.23,24 The advantages of FISH include high sensitivity and the ability to test formalin-fixed paraffin-embedded tissue sections in interphase, and good sensitivity of the assay.21,22
FISH, however, requires highly trained personnel to score rearrangements using fluorescent microscopy. 21 The presence of complicated rearrangements in different loci often requires multiple probes to completely identify the translocations fully.22 Although FISH may also represent a good method of recognizing chromosomal translocations, it is less than ideal for detecting intrachromosomal events.4 Furthermore, FISH cannot distinguish whether or not DNA rearrangements have been transcribed (oncogenic) or translated, nor can it identify the partner gene involved.25
Immunohistochemistry. Immunohistochemistry (IHC) detects specific proteins (antigens) in biological tissues by imaging the antibodies bound to these antigens.26 IHC is a well-established method for screening for NTRK fusions. It has high sensitivity and specificity and is more efficient and less expensive than other molecular tests.25 In addition, only transcribed and translated fusions are identified with IHC, rather than all DNA rearrangements including those that may not result in a transcribed (oncogenic) fusion.25 In a recent study, pan-TRK (TRK A, B, C) IHC appears to be both tissuesparing and tissue-efficient as well as time efficient for screening NTRK fusions.25 Although a pan-TRK monoclonal antibody is currently commercially available, it has not yet been approved for clinical use by the FDA, and IHC assay is not yet available for TRK proteins.25Next-generation sequencing (NGS) is a group of methods with high sequencing throughput.22 NGS can be used to generate data on the whole genome, data on an exome (genome's coding sequences), or specific genes of interest. Whole genome coverage can be combined with targeted sequencing of specific regions of interest to dramatically increase sensitivity of NGS in detecting genome alterations.22 The ability to examine multiple loci with minimal cost confers an advantage in malignancies where mutations and rearrangements are likely found in multiple genes.21 In a recent study, NGS showed sensitivity and specificity equal to that of FISH.21 One disadvantage of NGS is a slower turnaround time than that of IHC.
AFAP1 indicates actin filament-associated protein 1; AGBL2, ATP/GTP-binding protein-like 2; BCAN, brevican; BTBD1, BTB (POZ) domain containing 1; CD74, CD74 molecule; ETV6, ets variant 6; LMNA, lamin A/C; MPRIP, myosin phosphatase Rho interacting protein; NACC2, NACC family member 2; NFASC, neurofascin; PAN3, PAN3 poly(A) specific ribonuclease subunit; QKI, KH domain containing RNA binding; RABGAP1L, RAB GTPase activating protein 1-like; RBPMS, RNA-binding protein with multiple splicing; SQSTM1, sequestosome 1; TFG, TRK-fused gene; TP53, tumor protein p53; TPM3, tropomyosin 3; TPR, translocated promoter region, nuclear basket protein; TRIM24, tripartite motif containing 24; VCL, vinculin. Reprinted with permission. Vaishnavi A, Le AT, Doebele RC. TRKing down an old oncogene in a new era of targeted therapy. Cancer Discov. 2015;5(1):25-34; and Amatu A, Sartore-Bianchi A, Siena S. NTRK gene fusions as novel targets of cancer therapy across multiple tumour types. ESMO Open. 2016;1(2):e000023.
Reprinted with permission. Amatu A, Sartore-Bianchi A, Siena S. NTRK gene fusions as novel targets of cancer therapy across multiple tumour types. ESMO Open. 2016;1(2):e000023. doi:10.1136/esmoopen-2015-000023.
In addition, missed fusions may occur as a result of large introns or the presence of repetitive regions.25 NGS is the most comprehensive test of the molecular assays available and can identify NTRK gene fusions and other actionable alterations simultaneously.27 Not all NGS assays can detect NTRK gene fusions, however. Examples of NGS test platforms that detect all 3 NTRK gene fusions (NTRK1, NTRK2, and NTRK3) include15,28-37:
In addition to commercial approaches, major medical centers may also have in-house NGS assays capable of detecting a wide variety of actionable genetic alterations, including NTRK gene fusions. For example, the Stanford Anatomic Pathology and Clinical Laboratories offer a Stanford Solid Tumor Actionable Mutation Panel (STAMP).38 STAMP uses NGS to target genomic regions of interest based on their potential impact as medically actionable targets, mutation recurrence frequencies, and prognostic value.38 STAMP can detect all 3 NTRK gene fusions.38
Plasma-Based Molecular Profiling. A recently developed approach for detecting somatic alterations involves molecular profiling of circulating tumor cells and cellfree DNA.4 These "liquid biopsies" offer a noninvasive detection approach for patients who may be unable to tolerate invasive examinations or who may not be able to provide an adequate tissue sample for other molecular tests.39 In this approach, circulating cell-free tumor DNA is isolated from the patient's blood and then sequenced using quantitative polymerase chain reaction or NCS assays. FISH can be used to analyze DNA in circulating tumor cells but not circulating cell-free tumor DNA.4
The disadvantage of plasma-based molecular profiling is that quantitative polymerase chain reaction requires insights regarding the specific breakpoint prior to sequencing. NGS can detect specific fusions only if the breakpoint is within the preselected introns and exons for amplification.4 With NGS, long introns may be excluded due to their high cost, and therefore some fusions may not be detected with this approach.4
Analyses of circulating tumor DNA for various fusion oncogenes, including ALK and ROS1, have been reported.40 In addition, plasma-based molecular profiling has also been used to monitor drug resistance due to novel NTRK1 genetic alterations in a patient receiving a pan-TRK kinase inhibitor.41 Commercial platforms using plasma-based molecular profiling approaches to detect NTRK fusions have not been developed, however.Companion diagnostics for NTRK fusions are in development. A pan-TRK IHC test is being developed jointly by Bayer and LOXO Oncology and Ventana Medical Systems as a companion diagnostic to identify patients who are candidates for the NTRK inhibitor, larotrectinib, across multiple tumor types. The development goal is to validate this assay for FDA premarket approval as a companion diagnostic. An NGS-based multigene panel companion diagnostic for broad tumor profiling is also undergoing validation to become an FDA-approved diagnostic in conjunction with specific genetic-based treatment. This NGS-based pancancer companion diagnostic is jointly developed by Bayer and LOXO Oncology and Illumina and is expected to detect genetic alterations in 170 genes commonly associated with solid tumors, including NTRK gene fusions and RET gene alterations.
In summary, various detection methods for NTRK fusions are available, each with their advantages and drawbacks. NGS is increasingly being utilized to simultaneously detect a wide range of gene fusions recognized as oncogenic drivers.4 Plasma-based methods may serve as a complementary approach for select patients with insufficient tissue samples or those who are unable to undergo biopsy procedures. Companion diagnostics for genetic-based therapies are increasingly being developed, and some are undergoing validation for FDA premarket approval.Growing evidence of the role of NTRK gene fusions across multiple cancer types led to the effort of targeting NTRK gene fusions for cancer therapy. Residues in the kinase domain are highly conserved among TRKA, TRKB, and TRKC proteins.10 As a result, it may be beneficial to produce broader antitumor activity by inhibiting all 3 TRK proteins' gene fusions. Therefore, many TRK inhibitors in development are designed to be pan-inhibitors of all 3 TRK receptor isoforms.10
Several TRK inhibitors are undergoing evaluation in clinical trials (Table 2).10,12,46,49-52,62-81 Published results are available for entrectinib (RXDX-101) and larotrectinib (LOXO-101).Entrectinib (formerly RXDX-101) is an oral inhibitor of ALK, TRKA, TRKB, TRKC, and ROS1.10 Entrectinib crosses the blood-brain barrier, making it potentially useful as a treatment for glioblastoma and brain metastases containing gene fusions of ALK, NTRK, or ROS1.10 Initial clinical experience with entrectinib as well as case reports of patients treated in single-patient protocols have been published.11,42,43
ALKA and STARTRK-1. Interim results of 2 phase I trials of entrectinib in patients with locally advanced or metastatic solid tumors, including patients with stable, asymptomatic CNS involvement, have been reported (ALKA and STARTRK-1).44 In these 2 trials, patients did not need to have ALK, NTRK1/2/3, or ROS1 molecular alterations to enter the study, but the presence of these alterations was evaluated.44 Of 119 total patients in both protocols, most patients (n = 114, or 98%) had an Eastern Cooperative Oncology Group (ECOG) performance status of 0 or 1, and the majority (83%) had been previously treated with 3 or more anticancer therapies, including 27% who had previously received ROS1 or ALK inhibitors.42 Included patients had a wide range of solid tumor types, including head and neck, brain, breast, sarcoma, melanoma, neuroendocrine, ovarian, and renal cell carcinomas.42 The most prevalent tumor was NSCLC (60%), followed by gastrointestinal tract tumors (15%).44
Patients were assigned sequentially to receive escalating doses of entrectinib according to a standard 3 + 3 study design. Entrectinib was administered orally and treatment continued until disease progressed without clinical benefit became evident, unacceptable toxicity developed, or consent was withdrawn.42 A variety of dosing strategies were employed, including fasted, fed, flat, body surface area (BSA)-based, continuous, and intermittent. Initial dosing of entrectinib was based on BSA, with doses ranging from 100 mg/m2 to 1600 mg/ m2; it eventually transitioned to flat dosing in the range of 600 mg to 800 mg daily.42
Tumor assessments were performed by CT or by MRI after the first or second cycle of treatment and every 8 weeks thereafter. Tumor response was evaluated according to RECIST version 1.1.42 Patients were included in safety and efficacy analyses if they had received at least 1 dose of entrectinib.42
The most commonly encountered treatment-related adverse events (AEs) of any grade were fatigue/asthenia (46%), dysgeusia (42%), paresthesia (29%), nausea (28%), or myalgia (23%).42 Most AEs were grade 1 or 2 in severity.42 Dose reduction occurred in 18 patients (15%).42 Doselimiting toxicities in STARTRK-1 included 1 case of grade 3 cognitive disturbance and one case of grade 3 fatigue at a daily entrectinib dose of 800 mg, both of which resolved with dose interruption.42 An additional patient experienced grade 4 eosinophilic myocarditis following 2 doses of entrectinib, which fully resolved after the patient discontinued the study drug and withdrew from the trial. No grade 5 adverse events were observed.42
Of 119 patients enrolled, 60 had gene fusions involving ALK, NTRK1/2/3, or ROS1. Of the remaining patients, 53 had molecular alterations other than gene fusions, and 6 had no known alteration of ALK, NTRK1/2/3, or ROS1.42 All but one patient with tumors negative for ALK, NTRK1/2/3, or ROS1 gene fusions had no objective responses to treatment with entrectinib. In 25 patients with ALK and ROS1 gene fusions who had previously received an ALK inhibitor (alectinib, ceritinib or crizotinib) or a ROS1 inhibitor (crizotinib), no objective responses to entrectinib were observed. Objective responses (ORs) were only observed in patients with tumors harboring a gene fusion involving ALK, NTRK1/2/3, or ROS1 and who were naive to tyrosine kinase inhibitors. Patients meeting this definition and who had received at least 600 mg of entrectinib daily were considered "phase II eligible" and underwent additional analyses.42
Of the 30 patients considered "phase II eligible," 25 could be evaluated.44 The objective response rate (ORR) was 100% (95% CI, 44%-100%) in 3 patients with NTRK1/2/3 solid tumors. The ORR was 86% (95% CI, 60%-96%) in 14 patients with ROS1 gene fusions. Thirteen of the 14 patients with ROS1 gene fusions had NSCLC. The ORR was 58% (95% CI, 25% -84%) in 7 patients with ALK-rearranged solid tumors.
Initial responses were detected as early as 4 to 8 weeks after initiation of treatment (first or second cycle scans).42 The median duration of response (DOR) was 17.4 months (95% CI, 12.7-not reached) for ROS1-rearranged cancers and 7.4 months (95% CI, 3.7-not reached) for ALK-rearranged cancers. In the 3 patients with NTRK-rearranged cancers, the duration of response ranged from 2.6 months to 15.1 months, with the final patient ongoing as of data cutoff. The median duration of follow-up was 15 months.42 In patients with ROS1 gene fusions (n = 14), the median progression-free survival (PFS) was 19.0 months (95% CI, 6.5-not reached). The median PFS was 8.3 months (95% CI, 4.6-12) in patients with ALK-rearranged cancers (n = 7). The median PFS has not been reached in patients with NTRK1/2/3 fusions (n = 4).
Because entrectinib crosses the blood-brain barrier, responses in patients with malignancies involving the brain were also analyzed. Among 25 evaluable "phase II eligible" patients, 8 (32%) had a primary tumor or metastatic disease involving the brain. Responses were observed in 5 of these 8 patients (63%), including 4 patients with NSCLC (1 with NTRK1-, 2 with ROS1-, and 1 with ALK-1 gene fusions) and 1 patient with ALK-rearranged colorectal cancer.42
In summary, although the number of treated patients with analyzable data was small, ALKA and STARTRK-1 provided proof-of-concept data supporting further investigation of tyrosine kinase inhibitors in patients with tumors harboring gene fusions involving ALK, ROS1, and NTRK1/2/3. Treatment responses with entrectinib were observed as early as 4 weeks, and the longest DOR approached 2.5 years as of data cutoff.42 As patients who have been previously treated with other tyrosine kinase inhibitors did not respond to entrectinib, potential resistance mutations may be at play. These studies emphasized a new paradigm of using molecular biomarkers independent of tumor site or histology for targeted therapies in patients with genomic alterations.42
Case Reports. Case reports of individual patients demonstrating response to entrectinib are also available. In a study determining the frequency of NTRKI gene fusions in NSCLC, 1 patient with NSCLC found to have an NTRK1 gene fusion was enrolled into the ongoing STARTRK-1 trial (NCT 02097810).11 After 26 days of treatment, a partial response of -47% (per RECIST criteria) was observed. At day 155, further tumor reduction of -77% was observed.11 Complete response (CR) of all brain metastases was also observed.11
*Approved for the treatment of progressive, metastatic medullary thyroid cancer ALK indicates anaplastic lymphoma receptor tyrosine kinase; AXL, AXL receptor tyrosine kinase; c-kit, mast/stem cell growth factor receptor; CNS, central nervous system; DDR2, discoidin domain receptor 2; Eph, ephrin receptor tyrosine kinase; MER, MER receptor tyrosine kinase; MERTK, tyrosine-protein kinase MER precursor; MET, hepatocyte growth factor receptor; MKNK1/2, MAP kinase interacting serine/threonine kinase 1/2; MST1R, macrophage-stimulating 1 receptor; NCT, National Clinical Trial identifier number; PDGFR, platelet-derived growth factor receptor; RET, rearranged during transfection; ROS1, ROS proto-oncogene 1; TRK, tropomyosin-related kinases (also known as TRKA, TRKB, and TRKC for kinase A, B and C); VEGFR, vascular endothelial growth factor receptor.
Reprinted with permission. Amatu A, Sartore-Bianchi A, Siena S. NTRK gene fusions as novel targets of cancer therapy across multiple tumour types. ESMO Open. 2016;1(2):e000023. doi:10.1136/esmoopen-2015-000023.
During molecular screening for the European phase I study of entrectinib (EudraCT number: 2012-000148-88), the investigators identified a novel gene fusion involving LMNA-NTRK1 in a patient with metastatic colorectal cancer.43 After 4 weeks of entrectinib treatment, a repeat CT scan showed a partial response (PR) of -30%.4 STARTRK-2. A phase II basket study of entrectinib is underway for patients with locally advanced or metastatic solid tumors harboring ALK, NTRK1/2/3, or ROS1 gene rearrangements.44 Study arms include patients with NSCLC, metastatic colorectal cancer, or other solid tumors additionally divided according to each of the 3 gene fusions.44 The primary outcome will be ORR at 24 months according to RECIST version 1.1 criteria. Secondary outcome measures include DOR, time to response, intracranial tumor response, CNS and overall PFS, and overall survival.44
In 2015, entrectinib was granted FDA Orphan Drug designation for the treatment of NSCLC, colorectal cancer, and neuroblastoma harboring NTRK1/2/3, ALK, or ROS1 gene fusions.45Larotrectinib is an oral pan-TRK inhibitor. Larotrectinib blocks the ATP-binding site of the TRK family of receptors, with activity against all 3 TRK proteins (TRKA, TRKB, and TRKC).46,47 It is highly selective for TRK kinases, with virtually no inhibitory activity against non-TRK kinases.46 Larotrectinib's development program includes patients with any tumor type and of any age (a "tumor- and age-agnostic" therapy). The program includes 3 clinical studies47:
Patients in these 3 multicenter open-label studies had locally advanced or metastatic solid tumors with an ECOG score of 0 to 3 and had previously received standard therapy if available. Adequate organ functions were also part of the entrance criteria.47 Only 1 patient had been previously treated with a kinase inhibitor with anti-TRK activity; [he or she, as appropriate] was enrolled prior to an early protocol amendment, which would have excluded enrollment.47
Oral larotrectinib was dosed at 100 mg twice daily for adults, adolescents, and children with a body surface area of 1 m2 or more. In children with a body surface area of under 1 m2, larotrectinib was dosed at 100 mg /m2 twice daily.47 Treatment continued until disease progression, unacceptable adverse events, or patient withdrawal. Tumor assessments by imaging were performed at baseline, every 8 weeks for the first year, and every 12 weeks thereafter.47
The primary endpoint for the integrated analysis was the ORR, assessed using RECIST version 1.1 criteria. Secondary endpoints included ORR assessed by the investigator, DOR, PFS, and safety. Not all patients from the 3 studies had an NTRK gene fusion. The presence of an NTRK fusion was only required in the phase II basket study (NAVIGATE).47
NTRK gene fusions were assessed by NGS or FISH according to the analytic procedures at local laboratories. 47 Only the first 55 patients (adults and children) with identified NTRK fusions from these 3 studies who had a non-CNS primary tumor assessable by RECIST version 1.1, and who had received at least 1 dose of larotrectinib, were included in the integrated analysis.47 Among 55 included patients ranging from 4 months to 76 years old, 43 were 15 years or older.47 Most patients (n = 51) had an ECOG performance status score of 0 or 1. About half (n = 28) underwent 2 or more previous systemic chemotherapies. The population included 17 unique cancer diagnoses, with mammary analogue secretory carcinoma of the salivary gland being the most common (n = 12), followed by soft tissue sarcoma (n = 11), infantile fibrosarcoma (n = 7), thyroid tumor (n = 5), colon (n = 4), lung tumor (n = 4), melanoma (n = 4), and other tumors (n = 8). Most patients did not have CNS metastases (n = 54). NTRK1 fusions were found in 25 patients (45%), NTRK2 in 1 patient (2%), and NTRK3 in 29 patients (53%).47
The ORR was 75% (95% CI, 61%-85%), including 7 patients (13%) with a complete response and 34 (62%) with partial response.47 The median time to response (TTR) was 1.8 months (range, 0.7-6.4 months).47 In 2 pediatric patients with locally advanced infantile fibrosarcoma, tumor shrinkage was sufficient during treatment to allow for limb-sparing surgery with curative intention.47 In addition, responses were independent of which fusion partner was involved or whether fusion involved NTRK1, 2, or 3.47
The median follow-up was 8.3 months for response assessment and 9.9 months for PFS. At the time of the study, the median DOR and the median PFS had not been reached. As of the date of data cutoff, 86% of patients with a response were continuing treatment with larotrectinib or had undergone surgery with curative intention.47
The majority (93%) of AEs were of grade 1 or 2. Among grade 3 adverse events, regardless of attribution, the most common were anemia (11%), an increase in aspartate aminotransferase (AST) or alanine aminotransferase (ALT) (7%), weight gain (7%), and a decrease in neutrophil count (7%). There were no grade 4 or 5 treatment-related adverse events. Among grade 3 treatment-related adverse events, the most common were an increase in AST/ALT (5%), dizziness (2%), nausea (2%), anemia (2%), and a decrease in neutrophil count (2%). Larotrectinib dose reduction was necessary in 8 patients (15%) due to increased AST/ALT levels, dizziness, or a decrease in absolute neutrophil count. No patients required larotrectinib discontinuation due to an adverse event.47
Six patients (11%) had primary resistance to larotrectinib, defined as progressive disease as a best response.47 One patient, previously treated with another TRK inhibitor, displayed an NTRK3 G623R mutation in the kinase domain's ATP-binding site prior to larotrectinib administration. The NTRK3 G623R mutation and its paralogue, NTRK1 G595R, are called solvent-front mutations because they lead to sterical interference with larotrectinib binding by altering a hydrophilic solvent-exposed portion of the kinase domain's nucleotide- binding loop. Thus, the presence of NTRK3, G623R, or NTRK1 G595R mutations reduces larotrectinib's potency as a TRK inhibitor.47
For the remaining 5 patients with primary resistance to larotrectinib, tumor samples were available from 3 patients for re-analyses. In these 3 patients, IHC analysis did not confirm the presence of an expressed TRK fusion. The investigators suggested the possibility that the TRK fusion test performed at the local laboratory could have had a false-positive result. It is also possible that the fusion identified at the molecular level was not expressed at the protein level, which would explain the lack of response in these patients.47
Ten patients had acquired resistance to larotrectinib, defined as disease progression after objective response or at least 6 months of stable disease with treatment. In 9 of these patients, kinase domain mutations were identified from samples obtained after progression. These kinase domain mutations included solvent-front mutations (NTRK1 G595R or NTRK3 G623R) mutations at the gatekeeper position (NTRK1 F589L) and the xDFG position (NTRK1 G667S or NTRK3 G696A). Both the solvent-front and xDFG mutations interfere with the binding of larotrectinib. All 3 mutations are paralogues of acquired-resistance mutations that have been reported for other classes of kinase inhibitors.47
In summary, a rapid and durable antitumor response was observed with larotrectinib in a diverse pediatric and adult patient population with a variety of tumor types harboring NTRK fusions. Response to larotrectinib was "tumor agnostic" (not dependent on tumor type), confirming that NTRK fusions were important therapeutic targets in cancer. Additional data from ongoing studies will provide further information on the durability of response with larotrectinib and its safety profile.
Phase I/II Study in Pediatric Solid Tumors. Safety data from the phase I dose-escalation component of the pediatric phase I/II trial, Oral TRK Inhibitor LOXO-101 (larotrectinib) for Treatment of Advanced Pediatric Solid or Primary Central Nervous System Tumors (SCOUT) have also been separately reported.12
In SCOUT, infants, children, and adolescents aged 1 month to 21 years with locally advanced or metastatic solid tumors or primary CNS tumors were included.
Participants were nonresponsive to available therapies and were enrolled regardless of the presence of molecular TRK fusion.12 After efficacy was observed in patients with infantile fibrosarcoma, the FDA requested a protocol amendment to include patients with locally advanced infantile fibrosarcoma who would have otherwise required disfiguring surgery to attain a complete surgical resection.12 Participants aged >16 years had Karnofsky performance status scores of 50 or more, whereas those <16 years had Lansky performance status scores of 50 or more. Participants were determined to have adequate organ functions and had fully recovered from previous cancer therapies.12
Prior to enrollment, TRK fusion status was assessed locally.12 Patients were enrolled to receive 1 of 3 dose regimens, according to a modified rolling six design.12 Oral larotrectinib (in capsule or liquid formulation) was administered twice daily, in increasing doses adjusted for age and/or body weight, on a continuous 28-day schedule. The 3 dosing regimens were:
Dose interruptions of up to 21 days were allowed for grade 3 or 4 adverse events. Upon recovery, treatment was restarted at a lower protocol-defined dose. The primary endpoints of the phase I dose escalation were safety and dose-limiting toxicity of larotrectinib. All patients who had received at least 1 dose of larotrectinib were included in the safety analyses. Secondary endpoints included pharmacokinetic profile of larotrectinib, appropriate dose for phase II investigations, and OR as assessed using RECIST version 1.1.12
Among 24 enrolled patients with a median age of 4.5 years (range, 1 month to 18 years), 17 had tumors with TRK fusions.12 Among the 17 patients with TRK fusions (NTRK1, n = 9; NTRK2, n = 1; NTRK3, n = 7), 8 had infantile fibrosarcoma (NTRK1, n = 2; NTRK3, n = 6), 7 had other soft tissue sarcomas (NTRK1, n = 6; NTRK2, n = 1), and 2 had papillary thyroid cancer (NTRK1, n = 1; NTRK3, n = 1).12
At a median follow-up of 5.6 months, most AEs were of grade 1 or 2 severity, occurring in 21 of 24 patients (88%).12 The most common treatment-related AEs of all grades were increased ALT (42%), increased AST (42%), decreased leucocyte count (21%), decreased neutrophil count (21%), and vomiting (21%). No patients had a grade 4 or 5 TRAE. Two grade 3 serious adverse events included a case of grade 3 nausea and a case of grade 3 reduced ejection fraction while on doxorubicin at the 28-day follow-up after discontinuing larotrectinib. One patient in cohort 3 had a grade 3 increase in ALT elevation and discontinued therapy. Another patient in cohort 3 had a dose reduction due to neutropenia. The maximum tolerated dose was not reached. After safety analysis, pharmacokinetics, and objective responses, the recommended dose for phase II studies for pediatric patients was 100 mg/m2 twice daily (maximum of 100 mg per dose).12
Among 24 enrolled patients, 22 had measurable disease according to RECIST version 1.1 at enrollment and were evaluable for objective response.12 After a median of 8.2 months, 14 (93%) of the 15 patients with TRK fusions and measurable disease achieved an objective response (2 with complete response and 12 with a partial response per independent radiology review).
None of the patients with TRK fusion—negative tumors had an objective response. Except for 1 patient, all 17 patients with TRK fusions continued treatment or had undergone surgery intended to be curative. The median DOR had not been reached as of data publication.
Reductions in tumor burden were seen in all 15 patients. Responses were observed in patients with each of the NTRK fusions, NTRK1 (n = 7), NTKR2 (n = 1), and NTRK3 (n = 6) and in those with both infantile fibrosarcoma and other soft tissue sarcomas.12
The only patient with NTRK who experienced tumor progression while on larotrectinib was a 23-month-old girl with infantile fibrosarcoma, after an objective response to treatment. A solvent-front resistance mutation, TRKC G623R, was identified in the recurrent tumor.12,51
In summary, larotrectinib was well tolerated in pediatric patients and demonstrated antitumor activity in patients with solid tumors harboring TRK fusions. This study highlights the need to identify TRK fusions in infants, children, and adolescents with advanced cancers. Tumor genetic profiling earlier in the disease course may also be considered.12
Case Reports. Response to larotrectinib has been reported in 2 cases of breast cancer. In a patient with advanced triple-negative secretory breast cancer harboring an ETV6-NTRK3 fusion, larotrectinib led to a rapid clinical and radiographic response. A 37-year-old patient presented with bilateral lung metastases as well as vertebral lesions of breast origin after declining adjuvant chemotherapy or radiotherapy following a mastectomy.52 She underwent palliative radiation therapy to the spine but declined chemotherapy. The patient became pregnant and required several thoracentesis procedures for her pleural effusions. Because the patient continued to refuse chemotherapy, an archived biopsy sample was sent for tumor profiling, which revealed an ETV6- NTRK3 fusion. After delivery of her infant, the patient presented with increasing dyspnea and required pain management, with a performance status of 3. A CT scan revealed bilateral pleural effusions, peritoneal infiltration with ascites, scattered bone lesions, and multifocal lymphadenopathy. The CA-125 tumor marker was 2521 U/mL. Because the patient had not undergone any standard systemic chemotherapy, she did not qualify for participation in clinical trials involving larotrectinib. The drug was provided under compassionate access, however, and larotrectinib 100 mg twice daily was initiated.52
The patient’s dyspnea improved within days of initiating larotrectinib. There was also a rapid reduction in the size of her cervical lymph nodes. Her performance status score decreased to 1 following 2 weeks of treatment with larotrectinib.52 After 6 weeks of treatment, the patient reported only minimal exertional dyspnea and needing only few analgesics. Her axillary and cervical lymph nodes were impalpable at physical examination. Furthermore, the ascites in the abdominal cavity had disappeared. Re-imaging revealed a more than 80% reduction by RECIST version 1.1. The patient continued to show clinical response at 8 weeks of treatment with the CA-125 tumor marker normalized to 19 U/mL. The dose of larotrectinib was reduced to 75 mg twice daily due to grade 1 dizziness. The patient continued to receive larotrectinib at the time of report at 6 months.52
A second case involved a 14-year-old girl who had secretory breast carcinoma treated with lumpectomy.53 She developed lung metastases not responsive to multiple chemotherapy agents.53 Because ETV6-NTRK3 fusion was considered the pathognomonic genomic alteration in secretory breast carcinoma, molecular testing for ETV6-NTRK3 fusion was performed.53
Presence of ETV6-NTRK3 fusion was confirmed, and imaging revealed chest masses as well as pulmonary and bone metastases. A single-patient use protocol for larotrectinib was approved, and larotrectinib 100 mg twice daily was administered. Rapid reduction in the left chest mass was achieved within 1 week of therapy, and near-complete resolution was achieved at 2 months. A near-complete resolution of pulmonary metastases was also noted. The patient’s response was ongoing at 4 months. AEs included 2 episodes of grade 1 to 2 dizziness.53
Larotrectinib was granted FDA orphan drug designation for the treatment of NTRK fusion—positive solid tumors. In addition, the FDA granted breakthrough therapy designation to larotrectinib “for the treatment of unresectable or metastatic solid tumors with NTRK-fusion proteins in adult and pediatric patients who require systemic therapy and who have either progressed following prior treatment or who have no acceptable alternative treatments.” A rolling new drug application was initiated in December 2017 and completed in March 2018.54There are other multikinase inhibitors with in vitro activity against TRK. These include altiratinib (DCC- 2701), cabozantinib, F17752, DS-6051b, entrectinib, PLX7486, sitravatinib (MGCD516), and TSR-011.10 Clinical trials of altiratinib and PLX7486 have been terminated.55-57Resistance to TRK inhibitors has been reported and may limit the efficacy of these agents. Primary resistance leads to lack of response to therapy, whereas acquired (secondary) resistance may be the culprit in a patient who has initial response to therapy but later experiences tumor progression.
Resistance to Entrectinib. Two case reports of acquired resistance and acquisition of mutations in the TRK kinase domain after treatment with entrectinib have been published.
A patient with LMNA-NTRK1—positive metastatic colorectal cancer treated with entrectinib developed acquired resistance.41 Circulating tumor DNA was profiled and revealed 2 new NTRK1 genetic alterations in the kinase domain, G595R and G667C.41 Longitudinal analysis of circulating tumor DNA revealed that these mutated alleles were not present at baseline before treatment but appeared in circulation as early as 4 weeks after treatment initiation with entrectinib. Mutation frequencies in circulating tumor DNA continued to increase and peaked at 16 weeks after treatment initiation, at the same time clinical progression was confirmed radiologically.41
A dose-dependent effect leading to the emergence of the mutations was also explored. In an in vitro model, NTRK1 G677C emerged during exposure to low concentrations of entrectinib and was absent with high doses. At high concentrations of entrectinib, only the NTRK1 G595R mutation, which was absent at low doses, was detected.41 Whether continuous or intermittent dosing affects the emergence or type of acquired mutation requires more studies.10
Further in vitro studies also suggest that in addition to entrectinib, NTRK1 G595R and G677C mutations were also resistant to larotrectinib and TSR-011 in vitro. Of the 2 mutations, G595R appeared to be more potent as compared to G677C in conferring resistance.41
The second reported case of acquired resistance with entrectinib involved a patient with ETV6-NTRK3 fusion—positive metastatic mammary analogue secretory carcinoma of the salivary gland.58 Prior to entrectinib, she was treated with crizotinib, which had modest inhibitory activity against TRKC. She had a rapid partial response to entrectinib within 4 weeks of treatment. Acquired resistance and disease progression was observed after 7 months of entrectinib therapy. A novel NTRK3 G623R mutation was identified and corresponded to the development resistance to entrectinib.58 Structural analysis revealed that the NTRK3 G623R mutation created steric hindrance and reduced the binding of entrectinib with the ATP binding.58 The NTRK3 G623R mutation is homologous to a TRK1 G595R mutation.51,58
Resistance to Larotrectinib. Primary and acquired resistance have been reported in larotrectinib clinical trials and correlated with lack of response and disease progression. These resistance cases have been described in the previous sections discussing clinical trials involving larotrectinib. Briefly, primary resistance to larotrectinib was observed in 1 patient who had previously been treated with another TRK inhibitor prior to larotrectinib. In this patient, an NTRK3 G623R mutation in the kinase domain’s ATP-binding site was identified.47 Other patients with acquired resistance to larotrectinib displayed kinase domain mutations, including solvent front mutations and mutations at the gatekeeper and the xDFG position, after disease progression.47 Both the solvent—front and xDFG mutations interfere with the binding of larotrectinib. All 3 mutations are paralogues of acquired resistance mutations that have been reported for other classes of kinase inhibitors.47
Results from direct mutagenesis experiments also revealed the same mutations observed in patients, in addition to other potential resistance mutations. Structural investigations using x-ray crystallography identified the following mutations at the ATP-binding site, and these mutations may impair larotrectinib binding potency51:
Based on insights on how mutations may structurally affect binding of larotrectinib, LOXO-195 was developed to accommodate these acquired resistance mutations. LOXO-195 is a low-molecular-weight macrocycle (a molecule composed of at least 1 large ring with a minimum of 12 atoms).51,59 LOXO-195 is structurally designed to inhibit TRK resistance mutations.51 In preclinical studies, LOXO-195 shows high oral exposure and a pharmacokinetic profile favorable for human dosing.51 It exhibits potent in vitro highly selective inhibitory activity against wild-type TRKA, TRKB, and TRKC, as well as low activity against resistance mutations TRKA G595R, TRKC G623R, and TRKA G667C.51
Clinical experience with LOXO-195 has been reported for 2 patients who had developed acquired resistance to larotrectinib.51 The first patient was a 55-year-old woman with advanced LMNA-NTRK1 fusion-positive colorectal cancer, which had previously been heavily treated.53 Although the patient achieved a rapid partial response to larotrectinib, disease progression occurred after 6 months of treatment. Post-larotrectinib treatment samples demonstrated TRKA G595R mutations. She received LOXO-195 under a single-patient protocol, with dose escalation determined by real-time intrapatient pharmacokinetics information. LOXO-195 was initiated at 50 mg twice daily and titrated to 100 mg twice daily.
Therapy was well tolerated. Grade 2 dizziness and grade 1 diarrhea were observed initially but resolved despite continued administration of LOXO-195. The patient had a rapid clinical response to LOXO-195, and repeat imaging showed a 38% decrease in tumor burden at 4 weeks. NTRK1 G595R allele fraction decreased to below detectable levels after 2 weeks of treatment. The patient remained on LOXO-195 treatment for more than 6 months.51
The second patient was a 2-year-old girl with ETV6- NTRK3 fusion-positive infantile fibrosarcoma, which recurred despite numerous surgical resections and multiple combination chemotherapies. She achieved 90% tumor regression with larotrectinib, but a repeat biopsy at 8 months identified an acquired TRKC G623R mutation. Similar to the previous patient, she received LOXO-195 under a single-patient protocol, with dose escalation determined by real-time intrapatient pharmacokinetics information. LOXO-195 was initiated at a dose of 20 mg twice daily and titrated to a dose of 100 mg twice daily. LOXO-195 was well tolerated. Grade 2 dizziness was the only adverse event and did not interfere with dosing. The patient had a rapid clinical response to LOXO-195, and repeat imaging demonstrated a 30% reduction in tumor burden at 4 weeks. Repeat imaging at 66 days showed continued response to treatment. One month later, however, she developed a new mediastinal mass and pleural effusion and eventually succumbed to her cancer.51
In summary, LOXO-195 possesses selective activity against all 3 TRK kinases and acquired resistance mutations identified in both preclinical demonstrations and patients. It offers the opportunity to extend the period of disease control in some patients with resistance mutations to larotrectinib. A phase I/II study of LOXO-195 in patients with NTRK fusion-positive cancers previously treated with a TRK inhibitor is currently ongoing (NCT03215511).60New technologies in gene sequencing are bringing to light a landscape of previously unknown gene fusions in a broad multitude of cancers. NTRK fusions are oncologic drivers and are emerging as important targets across multiple tumor types. Clinical experience with larotrectinib, a pan-TRK inhibitor, reveals a rapid and durable antitumor response in a diverse pediatric and adult patient population with a variety of tumor types harboring NTRK fusions. In addition, larotrectinib appears to be well tolerated. Companion NTRK and other gene fusion diagnostics are in development.
Resistance mutations may develop in some patients taking TRK inhibitors; however, agents that target these resistance mutations, such as LOXO-195, are in development. Additional data from the ongoing studies will provide further information on the durability of response and safety profile with larotrectinib and LOXO-195, as well as other kinase inhibitors with activity against TRK.