TRK fusion-positive cancers and TRK inhibitor therapy

INTRODUCTION — The tropomyosin receptor kinase (TRK) family of transmembrane receptor proteins (TRKA, TRKB, and TRKC) is encoded by the neurotrophic tyrosine receptor kinase (NTRK) genes (NTRK1, NTRK2, and NTRK3, respectively). All three TRK proteins can become targets of structural rearrangement caused by an NTRK gene fusion that results in a chimeric TRK fusion protein that drives uninterrupted downstream signaling, and as a result, these fusions have emerged as important targets for cancer therapy. Selective inhibitors of these fusion proteins have been shown to have robust and durable activity in TRK fusion-positive solid tumors, and two of these agents are now approved in the United States and elsewhere for treatment of TRK fusion-positive refractory solid tumors, regardless of the site of disease origin. As such, these represent "tissue-agnostic" drug approvals.

This topic will review the prevalence, diagnosis, and management of TRK fusion-positive cancers.

MOLECULAR ALTERATIONS OF NTRK GENES IN VARIOUS TUMOR TYPES

Basic biology — The TRK receptor family comprises three transmembrane proteins, TRKA, TRKB, and TRKC, which are encoded, respectively, by the neurotrophic tyrosine receptor kinase (NTRK) 1 (which maps to chromosome 1q21-q22), NTRK2 (which maps to chromosome 9q22.1), and NTRK3 (which maps to chromosome 15q25) genes [1]. These receptors are expressed in neuronal tissue and play a central role in nervous system development during embryogenesis [2]. Following birth, the TRKs are primarily expressed by neuronal tissue and are involved in pain sensing, memory, weight homeostasis, and proprioception [3].

Fusions involving one of the NTRK genes were among the first gene translocations described in cancer [4]. These structural rearrangements result in a TRK fusion oncoprotein that leads to overexpression of the kinase domain or constitutive activity of the kinase function, both of which drive uninterrupted downstream signaling messages and oncogenic transformation and tumor growth [5-7]. Gene fusions represent the main molecular alterations in NTRK genes with known oncogenic and transforming potential [5].

Although dozens of unique NTRK fusions have been identified, they generally have shared features, including the following:

●An upstream fusion partner in the 5' (upstream) position that is typically expressed in the tissue type in which it is found

●Inclusion of the full-length TRK kinase domain in the 3' (downstream) position

Although fusions may occur in any of the three NTRK genes, most of those identified to date involve either NTRK3 or NTRK1 [5,8,9].

This topic will preferentially use the term "TRK fusion-positive cancers" to define those tumors that are identified as having either an NTRK fusion gene or a TRK fusion oncoprotein, depending on the diagnostic assay that is used. (See 'Available diagnostic assays' below.)

Prevalence — TRK fusions are distinct from other oncogenic kinase fusions in their broad distribution across a variety of both adult and pediatric cancers (table 1). This distribution follows two general patterns:

Rare cancers enriched for TRK fusions — TRK fusions are enriched in several rare cancer types (occurring at estimated frequencies as high as >90 percent in select series). These histologies include the following:

●Congenital infantile fibrosarcoma [10-14]

●Congenital mesoblastic nephroma (cellular subtype) [11,15]

●Secretory breast carcinoma [16,17]

●Mammary analogue secretory carcinoma (MASC) of the salivary gland [18-20]

In general, these cancers present in children (including infants) and young adults, although secretory breast carcinoma and MASC of the salivary gland predominately occur in adolescence or later in life, respectively [17,21,22]. (See "Malignant salivary gland tumors: Treatment of recurrent and metastatic disease", section on 'Secretory (NTRK gene fusion positive)' and "Pathology of breast cancer", section on 'Secretory carcinoma'.)

In some cases, the specific TRK fusion can help establish the diagnosis where routine pathology might be ambiguous. As an example, MASC of the salivary gland, which may be initially classified as "acinic cell carcinoma," can be diagnostically reclassified as MASC on the basis of identification of the recurrent translocation t(12;15)(p13;q25), which results in fusion of the ETS variant 6 (ETV6) gene on chromosome 12 and the NTRK3 gene on chromosome 15 to form the fusion gene ETV6-NTRK3. The identification of this fusion is pathognomonic for MASC, as it has not been demonstrated in any other salivary gland tumors [18,23,24].

Common cancers with low to intermediate frequency of TRK fusions — TRK fusions are also found in many common cancers (table 1), albeit at a much lower frequency (often <1 percent [25-28]):

●Non-small cell lung cancer

●Breast cancer other than secretory cancer

●Colorectal and appendiceal cancer

●Pancreatic cancer

●Renal cell cancer

●Head and neck cancer other than MASC

●Sarcoma, not otherwise specified

●Melanoma [29-31]

●Glioblastoma multiforme

●Cholangiocarcinoma

●Gastrointestinal stromal tumors

Intermediate frequencies — Some cancers may harbor TRK fusions at intermediate frequencies (1 percent or greater but less than 25 to 35 percent of cases) [32]:

●Thyroid cancer, particularly in children (frequency 2 to 28 percent, depending on the series [25,33-36])

●Glioma (particularly select pediatric high-grade gliomas [37])

●Specific sarcomas, such as inflammatory myofibroblastic tumor [38]

●Spitzoid neoplasms

In addition, these fusions can very rarely be found in hematologic malignancies, such as acute lymphoblastic/myeloid leukemia, multiple myeloma, histiocytosis, and dendritic cell neoplasms [39].

Diagnosis — The US Food and Drug Administration has approved a specific test that uses tissue next-generation sequencing for detection of NTRK fusion genes [40]. There are a variety of other diagnostic modalities available for the identification of TRK fusions, each with different operating characteristics and tradeoffs. Here, we will review both the available diagnostic assays and the recommended testing algorithms that incorporate these methodologies.

Available diagnostic assays

Tissue next-generation sequencing — Given the rarity of NTRK gene fusions in most cancers, assays that evaluate exclusively for the presence of NTRK gene fusions are impractical in most cases. Instead, next-generation sequencing (NGS, also called massively parallel sequencing) tests capable of evaluating for all classes of potentially actionable genomic alterations (single-nucleotide variants, indels, copy number alterations, structural rearrangements) across dozens or even hundreds of genes simultaneously are favored. This allows the assessment of NTRK gene fusions in conjunction with other tumor-agnostic biomarkers (ie, high levels of microsatellite instability/deficient mismatch repair) as well as any relevant tumor-specific biomarkers. (See "Tissue-agnostic cancer therapy: DNA mismatch repair deficiency, tumor mutational burden, and response to immune checkpoint blockade in solid tumors".)

NGS tests can be performed using DNA, RNA, or both. Although sequencing of either DNA or RNA can detect NTRK gene fusions, multiple technical considerations make RNA-based methods superior for this purpose. Specifically:

●The sensitivity of DNA-based NGS for NTRK gene fusion detection is lower, particularly for fusions involving NTRK2 or NTRK3 other than those involving ETV6. (See 'Basic biology' above.)

●These limitations of DNA-based NGS are the result of both large and repetitive introns of NTRK2 and NTRK3, as well as the diversity of potential fusion partners. By comparison, some RNA-based sequencing methodologies are capable of detecting any fusion involving NTRK1, NTRK2, or NTRK3.

●DNA-based sequencing may detect complex or ambiguous structural rearrangements of uncertain significance involving the NTRK genes. These typically manifest as structural rearrangements that do not appear to include the TRK kinase domain or in which the upstream partner cannot be resolved. By comparison, RNA-based sequencing preferentially detects rearrangements that result in expressed transcripts, thus resulting in improved specificity.

Plasma next-generation sequencing — Currently available, plasma-based, cell-free DNA assays (sometimes referred to as "liquid tumor biopsies") are not optimal for the detection of NTRK gene fusions. Although the identification of an NTRK gene fusion using plasma NGS can help select patients for TRK inhibitor therapy, the absence of an NTRK gene fusion in plasma should not be interpreted as definitive evidence that an NTRK gene fusion is not present in a patient's cancer. In this situation, tumor-based testing is recommended. (See 'Testing algorithms' below.)

Immunohistochemistry — Similar to its use in other cancers (eg, anaplastic lymphoma kinase [ALK] immunohistochemistry [IHC] in ALK fusion-positive lung cancers), IHC can be performed as a potential screening tool for TRK fusion oncogenes based on the fact that NTRK gene fusions can result in moderate to high levels of TRK oncoprotein expression, while many non-central nervous system (CNS) normal tissues do not express TRK receptors to a meaningful degree.

A positive test in the right context thus serves as a surrogate for the presence of an NTRK gene fusion. A pan-TRK antibody (EPR17341, Ventana, Roche) is available as an in vitro diagnostic device for the detection of NTRK gene fusions [41]. The epitope for this antibody is located in a conserved region of the C-terminus of TRKA, TRKB, and TRKC. Preliminary studies have demonstrated excellent sensitivity and specificity for this antibody in the detection of TRK fusion oncogenes [42]. It is important, however, to consider that TRK expression is normal and expected in certain organs (specifically in the CNS). As such, IHC positivity can occasionally be seen in cancers without TRK fusions [43] including, in our experience, tumors with neuroendocrine differentiation. For this reason, we recommend that, in tumors in which TRK fusions are rare, NGS be used in IHC-positive cases, when available, to confirm the positive result [43]. (See 'Testing algorithms' below.)

It is also important to recognize that overall expression intensity, as well as the specific staining localization, can be dictated by the upstream fusion partner. In our experience, ETV6-NTRK3 fusions in particular appear to be often associated with relatively faint staining, and a false negative result on immunohistochemistry [44]. Similarly, pan-TRK IHC is challenging to interpret in gastrointestinal stromal tumors, and we recommend against its use in this setting. Thus, both experienced pathologists and proper controls are required for proper interpretation of this assay.

Fluorescence in situ hybridization — Fluorescence in situ hybridization (FISH) break-apart assays can be used to detect the presence of an NTRK gene fusion. Three separate FISH tests need to be performed to detect fusions involving NTRK1, NTRK2, and NTRK3. This increases the amount of tissue required to successfully perform all three assays. For this reason, the primary indication for FISH testing has been the evaluation of NTRK gene fusions in tumor types that harbor a specific pathognomonic fusion (for example, infantile fibrosarcoma, which nearly always harbors ETV6-NTRK3 fusions). The interpretation of a positive FISH result is also subject to interobserver variability. Another limitation is that FISH break-apart tests generally cannot identify the upstream fusion partner. Finally, FISH break-apart assays may be falsely negative for fusions resulting from small intrachromosomal deletions/rearrangements, as is often the case with NTRK1 fusions.

Reverse transcriptase polymerase chain reaction — Reverse transcriptase polymerase chain reaction (RT-PCR) probes are designed to detect specific fusion events. For example, probes can be designed to detect ETV6-NTRK3, but these probes do not allow the detection of other fusions. This makes the use of this test relatively limited.

Testing algorithms — The complexity of diagnostic assays and the variable frequency of these events across cancer types have led to the development of various testing algorithms to detect TRK fusions [22,25,45,46]. We recommend the following approach, which is outlined in the algorithm (algorithm 1):

Tumors with a high frequency of TRK fusions (>20 percent) — For tumor types with a high frequency of TRK fusions, we recommend dedicated screening for TRK fusions, typically using pan-TRK IHC. FISH is a reasonable alternative to pan-TRK IHC.

●Positive result – In these high-TRK-prevalence tumor types, pan-TRK IHC alone may be considered sufficient evidence of a TRK fusion oncoprotein, and confirmatory orthogonal testing could be deferred.

●Negative result – Given the high prevalence of TRK fusions in these tumor types and the possibility of false negatives with pan-TRK IHC or FISH, we recommend proceeding to secondary screening, ideally using RNA-based NGS.

Tumors with a low or intermediate frequency of TRK fusions (<20 percent) — In this setting, we recommend broad screening with NGS, preferably using a methodology that incorporates both DNA and RNA sequencing (if available):

●Positive result – A positive result is considered to be any positive test using RNA-based methods, or a DNA-based test that resolves an upstream partner and is predicted to be in frame and to contain the full-length TRK kinase domain. These results are sufficient, and no additional testing is necessary.

●Ambiguous result – Structural variants detected using DNA-based testing that do not meet the criteria specified above should be reflected to either pan-TRK IHC or RNA-based sequencing.

●Negative result – If the assay included RNA sequencing, we recommend no additional testing. If the assay included only DNA sequencing, we recommend the following:

•For tumors in which DNA sequencing identified a tissue-relevant driver alteration (eg, epidermal growth factor receptor [EGFR], ALK, c-ROS oncogene 1 [ROS1], KRAS, NRAS, or BRAF), we do not recommend additional testing.

•For tumor types known to harbor TRK fusions (eg, non-small cell lung cancer, thyroid cancer, pancreatic cancer, cholangiocarcinoma, sarcoma) with no concurrent driver alteration, we recommend reflexing to RNA-based screening (when available). In one proof-of-concept study, RNA sequencing in "driver-negative" non-small cell lung cancers identified occult kinase fusions in 11 percent of cases, including several TRKB and TRKC fusions [47].

In summary, there is no "one size fits all" approach to the detection of NTRK gene fusions and TRK fusion oncogenes. Testing algorithms are based on the pretest probability of TRK fusion positivity and the assays available to the pathologist or ordering clinician. With the approval of TRK inhibitors, we expect that testing providers will continue to iterate their assays to improve TRK fusion detection. Optimal detection using NGS-based assays requires inclusion of RNA sequencing. Current blood-based testing should be considered reliable when positive but generally insufficient for screening purposes, even for tests that explicitly include TRK detection in their assay design.

Our approach is consistent with an ASCO provisional clinical opinion on somatic genomic testing in patients with advanced cancer, which stated that NTRK fusion testing should be performed in patients with metastatic or advanced solid tumors who may be candidates for TRK-inhibitor therapy, considering the prevalence of NTRK fusions in individual tumor types [48].

TREATMENT WITH TRK INHIBITORS — All patients with TRK fusion-positive cancers should be offered treatment with a TRK inhibitor. For patients of all ages, we favor the use of larotrectinib over entrectinib, where available. Although comparative studies are not available, entrectinib carries a number of unique safety warnings and precautions not found with larotrectinib, including heart failure, skeletal fracture, hyperuricemia, and QT interval prolongation.

The accurate and timely identification of TRK fusions is of immediate therapeutic importance given that TRK inhibitors have been shown to have robust and durable activity in TRK fusion-positive solid tumors. Specifically, two agents have been approved for the treatment of TRK fusion-positive cancers in the United States and elsewhere:

●Larotrectinib, a selective inhibitor of TRKA/B/C

●Entrectinib, a multikinase inhibitor of TRKA/B/C, c-ROS oncogene 1 (ROS1), and anaplastic lymphoma kinase (ALK)

In the United States, larotrectinib and entrectinib are both approved for the treatment of adult and pediatric patients with solid tumors that have a neurotrophic tyrosine receptor kinase (NTRK) gene fusion without a known acquired resistance mutation, are metastatic or where surgical resection is likely to result in severe morbidity, and have no satisfactory alternative treatments or have progressed following treatment. An oral solution formulation of larotrectinib is available to support administration for young children as well as adults unable to swallow capsules. Information on the US Food and Drug Administration-approved tests to identify appropriate candidates is available.

The European Medicines Agency has approved larotrectinib for treatment of advanced TRK fusion-positive solid tumors. In Japan, entrectinib has been approved for treatment of 10 tumor types with an NTRK gene fusion (cholangiocarcinoma, neuroendocrine tumors, gynecologic malignancies, non-small cell lung, breast, thyroid, colorectal, and pancreatic cancer, salivary gland tumors, and soft tissue sarcoma). Neither drug is yet approved in the United Kingdom [49].

Efficacy of first-generation TRK inhibitors — The approvals of both larotrectinib and entrectinib in the United States, Europe, and elsewhere were based on integrated analyses of consecutively enrolled patients with TRK fusion-positive cancers across multiple clinical trials. The patient populations and efficacy data from two pooled analyses of patients treated with larotrectinib or entrectinib are shown in the table (table 2). These data can be summarized as follows:

●Demographics – Both programs enrolled approximately 50 patients. All patients enrolled in the entrectinib program were adults, whereas 20 percent of those enrolled in the larotrectinib program were children. A higher proportion of patients enrolled in the entrectinib program had known brain metastases at entry compared with the larotrectinib program (22 versus 2 percent, respectively).

●Treatment outcomes – Key treatment outcomes for each drug are as follows:

•Larotrectinib – In an early combined analysis of three phase I to II studies, 55 patients with various malignancies treated previously with chemotherapy received larotrectinib [8]. The overall objective response rate across all tumor types was 76 percent (95% CI 61-85), including a 22 percent complete response rate. Responses were observed regardless of age, tumor type, or fusion type. Treatment was well tolerated, with no discontinuation of therapy due to adverse events. The most common grade ≥3 adverse events included increased alanine aminotransferase (ALT) or aspartate aminotransferase (AST), fatigue, vomiting, dizziness, nausea, diarrhea, constipation, and cough. (See 'Side effects' below.)

An updated report of antitumor efficacy in an expanded cohort of 159 adults and children with TRK fusion-positive cancers (including the original 55 patients described above) treated with larotrectinib confirmed these results and more fully characterized the durability of disease control [50]. In this analysis with nearly three times the original number of patients, a similar objective response rate of 79 percent (95% CI 72-85) and a 16 percent complete response rate were achieved across tumor types. The median duration of response was 35.2 months, and the median progression-free survival was 28.3 months. No new safety signals were identified with longer treatment durations; the most common grade 3 or 4 toxicities were increased alanine aminotransferase (n = 8, 4 percent), and neutropenia (n = 5, 3 percent).

•Entrectinib – In a pooled analysis of 54 adult patients with advanced/metastatic TRK fusion-positive solid tumors (10 tumor types, >19 histopathologies) who were treated with entrectinib in the STARTRK-2, STARTRK-1, and ALKA-372-001 trials, the overall objective response rate was 57 percent across all tumor types (95% CI 43-71), including a 7 percent complete response rate [51]. Responses were observed across tumor types and fusion types. The median duration of response was 10.4 months and the median progression-free survival was 11 months. In a later analysis of this combined experience, with an expanded cohort of 121 adults with 14 tumor types and over 30 different histologies, the overall objective response rate was 61 percent (16 percent complete) with a median duration of response of 20 months (95% CI 13.0-38.2) [52].

Timing of therapy for patients with advanced disease — Both larotrectinib and entrectinib are approved for the treatment of advanced TRK fusion-positive cancers that "have no satisfactory alternative treatments or that have progressed following treatment." However, given the very high response rates and durable disease control achieved with TRK inhibitors, we recommend that TRK inhibitors be considered for first-line therapy in patients with advanced, fusion-positive advanced cancers. This follows paradigms established in other driver-positive cancers (eg, epidermal growth factor receptor [EGFR]-mutant, ALK/ROS1 fusion-positive non-small cell lung cancers) in which first-line targeted therapy has proven to be more efficacious than existing alternatives [53,54]. (See "Overview of the initial treatment of advanced non-small cell lung cancer", section on 'Driver mutation present'.)

Benefit of upfront or neoadjuvant therapy — In addition to advanced/metastatic disease, both larotrectinib and entrectinib have also been approved for the treatment of locally advanced disease "where surgical resection is likely to result in severe morbidity." The most significant experience with neoadjuvant use of a TRK inhibitor has been in the preoperative use of larotrectinib for pediatric sarcomas [55,56]:

●Among five pediatric patients with locally advanced sarcomas, all underwent successful curative-intent surgical resection, with three achieving complete or near-complete (>98 percent treatment effect) pathologic responses with upfront larotrectinib. These three patients remained off larotrectinib between 7 and 15 months postoperatively.

●All four patients with initially inoperable limb sarcomas were able to undergo limb-sparing surgery.

While small case series such as this demonstrate the potential promise of neoadjuvant therapy with a TRK inhibitor, the main obstacle to broader use of this approach has been the timely preoperative identification of the presence of an NTRK gene or a TRK oncogene fusion. This has been easier to achieve in tumor types with high rates of TRK fusions, such as infantile fibrosarcoma. Anecdotally, we have administered larotrectinib preoperatively to patients with locally advanced pancreatic cancer and esophageal cancer (one each) and achieved marked downstaging in both cases, permitting subsequent definitive therapy for both. Others have reported the use of upfront larotrectinib in a patient with stage II TRK fusion-positive pediatric thyroid cancer to allow more effective delivery of radioactive iodine [57].

Use in patients with central nervous system tumors — TRK inhibitors have antitumor activity in patients with TRK fusion-positive central nervous system (CNS) metastases as well as primary CNS tumors.

A subset of TRK fusion-positive cancers are histologic types with significant rates of brain metastasis, most commonly non-small cell lung cancer and melanoma. Thus, the activity of TRK inhibitors in the CNS is relevant for a subset of patients with TRK fusion-positive cancers. In addition, a subset of pediatric and adult primary brain tumors may also harbor TRK fusions and may benefit from a TRK inhibitor.

The following summarizes what is known about the activity of the available TRK inhibitors in both primary and metastatic brain tumors:

Larotrectinib

●Brain metastasis

•Various cancers – In the expanded 159-patient dataset that led to the approval of larotrectinib, 13 patients (8 percent) were known to harbor CNS metastasis at baseline [50]. In this dataset [58,59], nine of 12 (75 percent) evaluable patients had overall responses (extracranial and intracranial) with larotrectinib. Of these patients, three had RECIST version 1.1 measurable intracranial disease: one had a complete intracranial response, one had a partial intracranial response, and one had stable intracranial disease, with the remaining patients achieving stable disease according to investigator assessment.

•Lung cancers – Additional information is available from a report of 20 patients with TRK fusion-positive lung cancers; of the eight with evaluable CNS metastases, five (63 percent) had a partial response, with response duration ranging from 0.03 to 20.2 months [60].

●Primary CNS tumors

•The expanded 159-patient dataset that led to the approval of larotrectinib included 18 patients with primary brain tumors; 14 had RECIST version 1.1 evaluable disease, and five of these (36 percent) had objective responses, including two complete responses [58,59]. No patient had progressive disease as their best response. Median progression-free survival was 11 months. An exploratory subset analysis suggested that outcomes may be more favorable in pediatric brain tumor patients compared with their adult counterparts.

•In a later combined analysis of a larger group of 33 patients with TRK fusion-positive primary CNS tumors previously treated with standard chemotherapy and enrolled in the two clinical trials (including 19 high-grade gliomas, 8 low-grade gliomas), larotrectinib was associated with a 30 percent objective response rate, the median time to response was 1.9 months, and the duration of treatment ranged from 1.2 to 31.3+ months [61]. Although treatment-related adverse events developed in 24 percent of treated patients overall, they were severe (grade ≥3) in only 3 percent.

Entrectinib — Entrectinib is also active in patients with primary CNS tumors and CNS metastases [51,62].

●In the 54-patient dataset that led to the approval of entrectinib, 11 patients (22 percent) had CNS metastasis at baseline [51]. Of these, intracranial responses were observed in six (55 percent) by blinded central review. In a later expanded cohort of this experience that included 26 patients with CNS metastases at baseline, there were 21 objective responses by blinded central review (58 percent), which included two complete responses [52]. Median duration of response was 17.2 months (95% CI 6-29.4).

●The phase II portion of the STARTRK-NG trial of entrectinib in children and young adults with solid extracranial or primary CNS tumors included 26 individuals with a target fusion in NTRK, ROS1, or ALK; 16 had primary CNS tumors [62]. There were eight objective responses (50 percent), four complete.

Dosing

Larotrectinib — For adults and children with a body surface area (BSA) of at least 1.0 m², the recommended dose of larotrectinib is 100 mg orally twice daily, with or without food, until disease progression or unacceptable toxicity. For smaller children, the recommended dose is 100 mg/m² orally twice daily, and a liquid formulation is available.

Entrectinib — The recommended dose of entrectinib in adults is 600 mg orally once daily, with or without food, until disease progression or unacceptable toxicity. For children aged ≥12 years, dosing is based on BSA:

●BSA >1.5 m²: 600 mg once daily

●BSA 1.11 to 1.5 m²: 500 mg once daily

●BSA 0.91 to 1.10 m²: 400 mg once daily

The drug is not approved for children under the age of 12 years, and a liquid formulation is not available.

Side effects — A summary of the overall safety profile in expanded cohorts of patients, as reported in the United States prescribing information for larotrectinib and entrectinib, is shown in the table (table 3).

On-target adverse effects — The TRK pathway controls appetite, balance, and pain sensitivity and TRK inhibitors can cause adverse effects related to these on-target effects [63]. As a group, these adverse effects have been underrecognized and underreported. Data on the clinical presentation, frequency, and natural history of weight gain, dizziness, and withdrawal pain are available from a single-institution series of 96 patients who received at least one dose of a tyrosine kinase inhibitor (TKI) with potent anti-TRK activity [63].

●Weight gain – In this same series, weight gain was reported in 53 percent (51 of 96 patients); median time to onset was one month, and the frequency and severity of weight gain increased with longer time on therapy [63]. Weight gain was not due to worsening fluid retention, but to increased adipose tissue, as assessed radiographically. Eight of the ten patients who received pharmacologic management (glucagon-like peptide 1 analogs, metformin, bupropion, topiramate, sibutramine, or phentermine, either alone or in combination) lost weight or stopped gaining weight. The contribution of diet and exercise to weight control could not be assessed, although these were recommended by the investigators.

Recommendations include serial monitoring of weight on TRK inhibitor therapy, paying careful attention to diet and increased exercise, pharmacologic management (potentially weigh the help of an endocrinologist or weight loss specialist), and ultimately TRK inhibitor dose reduction for moderate to severe and refractory weight gain.

●Dizziness – Dizziness was observed in 41 percent (39 of 96) of the patients in this series; six had concurrent ataxia [63]. The median time to onset was two weeks, (range 3 days to 16 months). Symptoms were described as positional light-headedness most often (33 percent), but also as imbalance (18 percent), vertigo (8 percent), or mixed (5 percent). Dizziness was accompanied by orthostatic hypotension in 21 percent (8 of 39 patients), although signs of volume depletion were absent.

Recommendations include carefully characterizing the type of dizziness that patients experience as supportive care measures can vary depending on the presentation. For example, meclizine seemed beneficial for individuals with vertigo, and midodrine was considered for patients with orthostasis. Ultimately, the most effective intervention was dose reduction in refractory cases, which led to symptom resolution in 7 of 9 patients.

●Withdrawal pain – Given that TRK inhibitors can modulate pain receptors, chronic inhibition of TRK signaling may be associated with an increased setpoint for pain neurotransmitter signaling [64]. Drug discontinuation can thus result in withdrawal pain.

Withdrawal pain was observed in 28 of 81 patients (34 percent) treated with at least one dose of a TKI with potent anti-TRK activity who discontinued treatment at some point [63]. Symptoms were described as diffuse achiness, muscle pain, and/or allodynia often accompanied by headache. The median time to onset was two days, and severity was grade 1, 2, or 3 in 12, 11, and 11 percent, respectively. For patients who permanently discontinued therapy, the median duration of flare was 14 days (range 10 to 26 days). For those who temporarily discontinued therapy, the median duration was three days (range one to seven days), which was consistent with the median duration of drug hold (three days). Patients who had been taking the TRK inhibitor for longer than six months were more likely to experience withdrawal pain (63 versus 13 percent), and opioid administration prior to drug withdrawal was not beneficial.

Pharmacologic interventions including opioids and gabapentin are of modest benefit in many patients. By contrast, for patients who temporarily discontinue therapy, restarting the TRK inhibitor results in rapid and often complete relief. For patients who have to discontinue therapy permanently, down-titration of TRK inhibitor therapy (25 percent decrease in dose every seven days) was beneficial in at least one patient.

Hepatotoxicity — Both agents can cause elevation in liver enzymes, which may be severe. The United States prescribing information for both drugs recommends that liver tests, including AST and ALT, be monitored every two weeks during the first month of treatment, then monthly thereafter, as clinically indicated. Recommendations are provided for dose modifications for severe hepatotoxicity during therapy.

Neurotoxicity — Neurologic adverse events are common with both drugs (>50 percent) but are severe (grade 3 or 4) in fewer than 10 percent. The majority develop within the first three months of treatment.

●Larotrectinib – Among the grade 3 or worse neurologic toxicities occurring with larotrectinib are delirium, dysarthria, dizziness, gait disturbance, and paresthesia.

●Entrectinib – A broader spectrum of CNS effects has been seen with entrectinib, which is not surprising since the drug was designed to cross the blood-brain barrier [65]. These include cognitive impairment (27 percent all grade, including confusional state, disturbance in attention, memory impairment, aphasia, mental status changes, hallucinations, and delirium), mood disorders (10 percent all grade, including anxiety, depression, and agitation), dizziness (38 percent all grade), sleep disturbance (14 percent all grade), and vision disorders (21 percent all grade, including blurred vision, photophobia, diplopia, visual impairment, and vitreous floaters).

Dose modification guidelines are available in the United States prescribing information for both drugs for patients who develop severe neurotoxicity during therapy.

Other toxicities specific to entrectinib — In general, the rate, number, and severity of adverse events all appear to be generally more favorable with larotrectinib. Entrectinib carries a number of unique safety warnings and precautions not found with larotrectinib, including congestive heart failure, skeletal fracture, hyperuricemia, and QT interval prolongation.

●Heart failure – Across clinical trials, congestive heart failure has occurred in 3.4 percent of patients treated with entrectinib, including grade 3 heart failure (2.4 percent). The median time to onset was two months. The United States prescribing information for entrectinib recommends assessment of left ventricular ejection fraction prior to initiation of entrectinib in patients with heart failure or known risk factors for congestive heart failure, and that patients be closely monitored for clinical signs and symptoms of congestive heart failure, including shortness of breath and edema.

●QT interval prolongation – Across clinical trials, 3 percent of patients treated with entrectinib who had at least one electrocardiogram (ECG) during treatment had prolongation of the corrected QT (QTc) interval, including some >500 milliseconds. Monitor patients who have or who are at significant risk for QTc interval prolongation, including those with known long QT syndrome, clinically significant bradyarrhythmia, and severe or uncontrolled heart failure and those taking other medicinal products associated with QT prolongation. The United States prescribing information for entrectinib recommends assessment of the QT interval and electrolytes at baseline and periodically during treatment, particularly for those with known long QT syndrome, clinically significant bradyarrhythmia, and severe or uncontrolled heart failure or who are taking other medicinal products associated with QT prolongation.

●Skeletal fractures – Entrectinib increases the risk of fractures (5 percent of adult patients and 25 percent of pediatric patients in combined clinical trial data reported in the United States prescribing information for entrectinib). In children, these have occurred mainly in the setting of minimal or no trauma, although some patients have had radiologic abnormalities suggesting tumor involvement.

●Hyperuricemia – Nearly 1 in 10 patients treated with entrectinib across clinical trials has experienced hyperuricemia, which was grade 4 in 1.7 percent and included one fatality due to tumor lysis syndrome. The United States prescribing information for entrectinib recommends assessment of serum uric acid levels prior to initiating therapy and periodically during treatment, and that patients be monitored for signs and symptoms of hyperuricemia.

Specific dose reduction guidelines are available for patients who develop these toxicities.

Resistance and options for second-line therapy — As with other targeted therapies, the onset of acquired resistance may limit the efficacy of TRK inhibitors. Resistance to these agents appears to be mediated either by mutation of the TRK kinase domain (called "on-target" resistance) or, alternatively, by activation of bypass or downstream pathways (called "off-target" resistance) [66-68].

Currently, there are no approved drugs for the management of acquired resistance that develops on larotrectinib or entrectinib. Two next-generation TRK inhibitors are currently in development to address patients with on-target resistance: selitrectinib (BAY 2731954, LOXO-195) and repotrectinib (TPX-0005). Both agents have shown preliminary activity in this setting [69-71], and clinical trials (including some with expanded access) are ongoing (NCT03215511, NCT03206931, NCT03093116, NCT04094610).

We recommend that any patient with a TRK fusion-positive tumor progressing after an initial response to larotrectinib or entrectinib have a repeat molecular profile by tumor biopsy and/or cell-free DNA testing and be considered for referral to a participating clinical trial site. When a clinical trial option is not available, standard-of-care therapies that are specific to the patient's cancer type should be considered.

SUMMARY AND RECOMMENDATIONS

●Molecular biology and prevalence of TRK fusions

•Structural rearrangements involving one of the neurotrophic tyrosine receptor kinase (NTRK) genes result in a tropomyosin receptor kinase (TRK) fusion oncoprotein that leads to overexpression of the kinase domain or constitutive activity of the kinase function, both of which drive uninterrupted downstream signaling messages and oncogenic transformation and tumor growth. (See 'Basic biology' above.)

•TRK fusion oncogenes are distinct from other oncogenic kinase fusions in their broad distribution across a variety of both adult and pediatric cancers (table 1). This distribution follows two general patterns: rare cancers that are enriched for TRK fusions (eg, congenital infantile fibrosarcoma, secretory breast cancer, mammary analogue secretory carcinoma [MASC] of the salivary gland) and common cancers that have low to intermediate frequencies of TRK fusions. (See 'Prevalence' above.)

●Diagnostic evaluation

•The diagnostic evaluation for TRK fusions is challenging and depends on the frequency with which TRK fusions are found in a given tumor type (table 1).

•Our suggested approach to diagnostic testing is outlined in the algorithm (algorithm 1). (See 'Diagnosis' above.)

●Treatment with TRK inhibitors

•The presence of a TRK fusion defines a new diagnostic category for solid tumors, independent of site of origin, that share sensitivity to TRK inhibitors. Two such agents (larotrectinib, entrectinib) are now approved in the United States and elsewhere for treatment of TRK fusion-positive refractory solid tumors, regardless of the site of disease origin. (See 'Treatment with TRK inhibitors' above.)

•All patients with TRK fusion-positive cancers should be offered treatment with a TRK inhibitor:

-For patients of all ages, we suggest larotrectinib over entrectinib, where available (Grade 2C). Although comparative studies are not available, entrectinib carries a number of unique safety warnings and precautions not found with larotrectinib, including heart failure, skeletal fracture, hyperuricemia, and QT interval prolongation. (See 'Other toxicities specific to entrectinib' above and 'Side effects' above.)

-Among patients with metastatic disease, TRK inhibitor therapy should be considered as soon as a TRK fusion is identified. (See 'Timing of therapy for patients with advanced disease' above.)

-For patients with locally advanced disease in whom immediate surgical resection would result in significant morbidity, preoperative use of TRK inhibition has been associated with downstaging and favorable outcomes in limited series. (See 'Benefit of upfront or neoadjuvant therapy' above.)

-TRK inhibitors have antitumor activity in patients with central nervous system (CNS) metastases as well as primary CNS tumors. (See 'Use in patients with central nervous system tumors' above.)

•On-target adverse events such as weight gain and dizziness during therapy, and withdrawal pain with therapy discontinuation should be monitored carefully. (See 'On-target adverse effects' above.)

•The onset of acquired resistance may limit the efficacy of TRK inhibitors. Any patient with a TRK fusion-positive tumor progressing after an initial response to a TRK inhibitor should have a repeat molecular profile by tumor biopsy and/or cell-free DNA testing and, if possible, should be referred to a participating clinical trial site for second-line therapy options. (See 'Resistance and options for second-line therapy' above.)

ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges David Hyman, MD, who contributed to an earlier version of this topic review.