Real-World Experience of NTRK Fusion–Positive Thyroid Cancer

Introduction

The tropomyosin receptor kinase (Trk) receptors, TrkA, TrkB, and TrkC, encoded by the NTRK1, NTRK2, and NTRK3 genes, respectively, are transmembrane proteins that play an important role in the normal development and function of the nervous system. Aberrant fusions of NTRK genes lead to the production of chimeric Trk receptors, which are constitutively activated with subsequent activation of downstream signaling pathways including mitogen-activated protein kinase (MAPK) and phosphatidylinositol 3-kinase (PI3K) pathways.¹ Such NTRK fusions have been found to be oncogenic drivers in multiple solid tumors including thyroid cancer.²

Selective Trk inhibitors, larotrectinib and entrectinib, demonstrated excellent efficacies with high and durable responses across the NTRK fusion–positive pediatric and adult solid tumors in several small basket trials.^3,4 Only a few patients with thyroid cancer were included in the published studies because of the rarity of the NTRK fusions in thyroid cancer. Furthermore, the frequencies and the types of NTRK fusions in thyroid cancer are widely variable in different studies.^5-10 Herein, we describe our real-world experience from four patients with NTRK fusion–positive thyroid cancer treated with larotrectinib. We also report the frequencies and the types of NTRK gene alterations in thyroid cancer from available public databases and a real-world data set from Tempus.

Case Presentations

A case series of four patients with NTRK fusion–positive thyroid cancer treated with larotrectinib is summarized in Figure 1. One patient had anaplastic thyroid cancer (ATC), one patient had poorly differentiated thyroid cancer (PDTC), and two patients had papillary thyroid cancer (PTC). The study was approved by the institutional review board of University of California, San Francisco (IRB #20-31865). Patient consent for the study was waived as the study did not involve any identifiable data. Consent to publish images was obtained from patient 2.

FIG 1.

Baseline clinicopathologic characteristics of four patients with NTRK fusion harboring thyroid cancer who were treated with larotrectinib, and waterfall plot for best response. ATC, anaplastic thyroid cancer; CR, complete response; PDTC, poorly differentiated thyroid cancer; PR, partial response; PTC, papillary thyroid cancer; RAI, radioactive iodine; SD, stable disease.

Patient 1 with ATC harboring SQSTM1-NTRK3 presented with a rapidly enlarging neck mass and multiple lung nodules. He underwent total thyroidectomy and central neck dissection; pathology showed small multifocal PTCs in thyroid and 9.5-cm mixed anaplastic and PDTC in left central neck. Because of complicated postoperative course, larotrectinib was initiated instead of intensive chemoradiation. The patient had 16% reduction in tumor burden after 2 months but progressed with enlarging parotid and neck masses after 6 months. Biopsy of the progressing lesion showed no gatekeeper mutations¹¹ or additional alterations.

Patient 2 with PTC harbored novel ETV6-NTRK2 fusion not previously described in other solid tumors. The novel fusion has breakpoints in ETV6 exon 4 and NTRK2 exon 16 with preserved ETV6 PNT domain and NTRK2 kinase domain leading to constitutive activation of TrkB kinase (Fig 2). The patient has a remote history of PTC treated with surgery. She was found to have multiple brain metastases, obstructive hydrocephalus caused by a cerebellar mass, and pleural effusion with pleural masses. Pleural biopsy and cerebellar resection specimens confirmed metastatic PTC with ETV6-NTRK2 fusion and TERT c.-124C>T mutation. Thyrogen-stimulated I-123 scan showed uptake only in the chest. After receiving stereotactic body radiation to brain metastases and cerebellar resection bed, larotrectinib was initiated, resulting in ongoing partial response (PR) in the pleural metastases for more than 18 months (Fig 3) without evidence of recurrence in the brain.

FIG 2.

Novel fusion between 5′ breakpoint in ETV6 exon 4-5 and 3′ breakpoint in NTRK2 exon 14-15. This fusion preserves the ETV6 PNT domain and the NTRK2 kinase domain, leading to constitutive activation of the NTRK2 kinase.

FIG 3.

Patient 2 with metastatic PTC harboring ETV6-NTRK2. Computed tomography chest images demonstrate dramatic response after 1 month treatment with larotrectinib. PTC, papillary thyroid cancer.

Patient 3 with PTC harboring ETV6-NTRK3 fusion and TERT c.-146C>T mutation presented with a spine metastasis. He underwent total thyroidectomy, neck dissection, and metastasectomy of the spine lesion, followed by radioactive iodine treatment (RAI-T; 100 mCi) and radiation to the spine and neck lymph nodes. After 2 years, he developed multiple new bone and pulmonary metastases with a recurrence in the ipsilateral neck. He started larotrectinib and achieved PR ongoing for 7 months.

Patient 4 with PDTC harboring TPM3-NTRK1 fusion developed mediastinal nodal metastases after initial thyroidectomy. She received RAI-T (155 mCi) after Thyrogen stimulation following surgery, and the post-treatment scan did not show any iodine uptake. After another year, she developed multiple hilar, mediastinal, and pulmonary metastases and started larotrectinib. She achieved complete resolution of enlarged lymph nodes and pulmonary nodules consistent with complete response (CR) in 2 months. Thyroglobulin (TG) rose from 329 to 1,588 ng/mL within 1 month of larotrectinib associated with a radiographic response. TG gradually decreased over the next 8 months but remained higher than the baseline before larotrectinib.

Types and frequencies of NTRK gene alterations.

Of 2,362 thyroid cancer specimens identified in the American Association for Cancer Research (AACR) Genie, The Cancer Genome Atlas (TCGA), and Tempus databases, NTRK1 or NTRK3 gene fusions were found in 51 patients (2.2%): 28 of 1,133 in the AACR Genie data set (2.4%), 12 of 482 in the TCGA data set (2.5%), and 11 of 747 (1.5%) in the Tempus data set. No NTRK2 gene fusions were identified in any of the databases (Table1).

TABLE 1.

Identified NTRK Gene Fusion Alterations in Thyroid Cancers From GENIE, TCGA, and Tempus Databases

We identified 10 different 5′ fusion partner genes; ETV6-NTRK3 fusion was the most common, accounting for 43% of all NTRK fusions identified in thyroid cancer, followed by TPM3-NTRK1 fusion (18%) and TPR-NTRK1 fusion (14%). TERT promoter mutations were the most frequent coalteration, found in 15 cases (29%), followed by TP53 (8%). Among cases from the Tempus cohort whose RNA expression data are available, overexpression of genes related to MAPK/ERK signaling pathway and cell-cycle regulation, and receptor tyrosine kinase genes were observed. We explored other relevant genomic alterations of NTRK genes and identified 24 cases of NTRK1/2/3 single-nucleotide alterations, two cases of NTRK1 amplification, and a splice variant of NTRK1 in both differentiated and medullary thyroid cancers (Table2). More than half (58%) of the point mutations were predicted to be pathogenic,¹² but the majority of non–fusion-altered NTRK cases also harbored well-established driver mutations such as BRAF/KRAS/HRAS mutations or RET/ALK gene fusions.

TABLE 2.

Identified NTRK Gene Nonfusion Alterations in Thyroid Cancers From GENIE and TCGA Databases

Discussion

We report a single-institution experience of four consecutive patients with advanced thyroid cancer harboring NTRK gene fusions, treated with larotrectinib, a selective Trk inhibitor. Three patients with PTC or PDTC achieved durable radiographic responses, and all of them have remained on larotrectinib. This is consistent with the data from prior phase I and II Trk inhibitor studies in solid tumors, demonstrating lower overall response rate (ORR) in patients with ATC compared to patients with DTC. In the combined analysis of phase I/II basket trials of larotrectinib including 28 patients with NTRK fusion–positive advanced thyroid cancer (22 DTCs and six ATCs), the ORR was 75% with two CRs and 19 PRs, 90% in DTC and 29% in ATC.¹³ Entrectinib was designed to cross the blood-brain barrier¹⁴ and demonstrated an ORR of 55% among patients with known brain metastases.³ Patient 2 with brain metastases started larotrectinib before approval of entrectinib. In the pooled analysis, two in four larotrectinib-treated thyroid cancer patients with CNS metastases had decreases in measurable brain lesions.¹³

Notably, the ETV6-NTRK2 fusion found in patient 2 is a novel gene fusion not previously reported for a solid tumor. The fusion was described in a patient with acute myeloid leukemia and was found to have transforming potential in a murine hematopoietic cell line.¹⁵ The patient did not have any abnormal blood counts, and germline sequencing performed on peripheral blood cells did not demonstrate abnormal findings. A good response to larotrectinib in patient 2 adds to the evidence that a selective Trk inhibitor has an efficacy in a tissue-agnostic manner, across the spectrum of NTRK fusion types. Another interesting observation was a rise in serum TG in patient 4 with PDTC harboring TPM3-NTRK1 fusion and durable CR. This suggests a potential role of larotrectinib in redifferentiation, similar to other tyrosine kinase inhibitors that have been used to restore iodine avidity.¹⁶ A recent case report demonstrated enhanced radioactive iodine uptake in a patient with PTC with EML4-NTRK3 fusion after larotrectinib.¹⁷ Among seven patients with thyroid cancer treated with larotrectinib in clinical trials, one patient with PPL-NTRK1 fusion achieved CR.¹⁸ TrkA encoded by NTRK1 is not expressed in normal thyroid tissue, but overexpression was observed in thyroid cancer, with activated Rous sarcoma oncogene and extracellular signal-regulated kinase pathways.¹⁹ Exceptional responses may be related to TrkA's oncogenic role in thyroid cancer.

In search for NTRK alterations in thyroid cancer using AACR Genie, TCGA, and Tempus databases, we identified various alterations in NTRK1 and NTRK3, but none in NTRK2. These fusions were found mostly in PTC, but also in PDTC, MTC, and ATC. ETV6-NTRK3 was the most common fusion found in 22 of 55 cases (40%). The actual frequency of NTRK fusions in thyroid cancer is not known, as some targeted exome sequencing can easily miss fusion event involving introns of certain genes. Studies on frequency of NTRK fusions from a single institution and from the TCGA found NTRK fusion in 10 of 451 (2.2%; four NTRK1 and six NTRK3 fusions) and 12 of 498 (2.4%; five NTRK1 and seven NTRK3 fusions) patient with thyroid cancer, respectively.^20,21 In our study cohort, TERT promoter mutations were found in 29% of the cases: 10 in 42 (23.8%) PTCs and four in five (80%) ATCs. It is not known whether TERT promoter coalteration has any impact on prognosis or response to Trk inhibitor in NTRK-altered thyroid cancers. TERT promoter mutation has been reported in various frequencies in different histologies ranging from 10% in PTD up to 50% in ATC.²² It is associated with more advanced stage and poor prognosis.^22-24

We also explored other genetic alterations of NTRK genes including nonrecurring missense single-nucleotide variations in NTRK1/2/3 and NTRK1 gene amplification. Interestingly, most cases with a missense mutation of NTRK1/2/3 also harbored well-described oncogenic alterations in genes encoding for RAS/RAF pathways, suggesting that these mutations are not likely the main driver for these tumors.

APPENDIX 1. Supplementary Text

Materials and Methods

Patients with advanced thyroid cancer harboring NTRK1/2/3 gene fusions were identified through retrospective review of clinical records at the University of California, San Francisco (UCSF). Presence of NTRK fusions was confirmed with commercially available oncology genomic profiling assays, including the UCSF500 DNA-based next-generation sequencing (NGS) test, which uses capture-based NGS and analyzes the exons of 529 cancer-related genes, as well as select intron of 47 genes, and the Tempus xT DNA, which is a targeted NGS test that detects single-nucleotide variants, indels, and copy-number variants of 648 genes and chromosomal rearrangements in 22 genes, supplemented by whole-transcriptome RNA sequencing for enhanced fusion detection.²⁵ Demographic data, molecular analysis data, treatment history, and treatment responses were obtained from the patient records. The radiographic responses to the treatment were collected from each patient. Patient consent for the study was waived as the study did not involve any identifiable data.

To describe the landscape of NTRK gene alterations in thyroid cancer, the public data generated from American Association for Cancer Research (AACR) Project Genie cohort version 9.0²⁶ and The Cancer Genome Atlas (TCGA) research network²⁷ were reviewed. Among 40 patients identified in AACR Genie and TCGA, median age was 39 years, and 53% of the patients were women. Additionally, a retrospective analysis on deidentified data from the Tempus real-world database was conducted to identify patients with thyroid cancer with NTRK fusions and discern the prevalence of these fusions. For Tempus specimens, gene expression was generated through RNA-seq of formalin-fixed paraffin-embedded tumor samples using an exome capture–based protocol as previously described.²⁸ Demographic information was not available for patients in the Tempus database.

AUTHOR CONTRIBUTIONS

Conception and design: Hyunseok Kang

Administrative support: Hyunseok Kang

Provision of study materials or patients: Chienying Liu, Hyunseok Kang

Collection and assembly of data: All authors

Data analysis and interpretation: Jong Chul Park, Arya Ashok, Hyunseok Kang

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/po/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).