Predictors of radioiodine (RAI)-avidity restoration for NTRK fusion-positive RAI-resistant metastatic thyroid cancers

Context Two-thirds of metastatic differentiated thyroid cancer (DTC) patients have radioiodine (RAI)-resistant disease, resulting in poor prognosis and high mortality. For rare NTRK and RET fusion-positive metastatic, RAI-resistant thyroid cancers, variable success of re-induction of RAI avidity during treatment with NTRK or RET inhibitors has been reported. Case presentation and results We report two cases with RAI-resistant lung metastases treated with larotrectinib: an 83-year-old male presenting with an ETV6::NTRK3 fusion-positive tumor with the TERT promoter mutation c.-124C>T, and a 31-year-old female presenting with a TPR::NTRK1 fusion-positive tumor (and negative for TERT promoter mutation). Post larotrectinib treatment, diagnostic I-123 whole body scan revealed unsuccessful RAI-uptake re-induction in the TERT-positive tumor, with a thyroid differentiation score (TDS) of −0.287. In contrast, the TERT-negative tumor exhibited successful I-131 reuptake with a TDS of −0.060. Conclusion As observed for RAI-resistance associated with concurrent TERT and BRAF mutations, the co-occurrence of TERT mutations and NTRK fusions may also contribute to re-sensitization failure.


Introduction
Papillary thyroid carcinoma (PTC) typically has an excellent clinical prognosis.However, 10% of cases of differentiated thyroid cancers (DTC) will progress to a metastatic stage after receiving initial treatment (1).Two-thirds of metastatic DTCs lose their ability to uptake radioiodine (RAI) due to oncogene driver mutations silencing thyroid iodide-metabolizing genes, causing them to be RAI-resistant or refractory (1,2,3).Patients with RAI-refractory metastatic DTC have a poor prognosis, with a mortality of approximately 90% within 10 years of diagnosis (4).
Data from The Cancer Genome Atlas (TCGA) investigating nearly 500 PTCs confirmed activation of the mitogenactivated protein kinase (MAPK) pathway for the large majority of PTCs (5).Among this cohort, BRAF V600E mutation was present in nearly 60% of PTCs, followed by HRAS and NRAS mutations in 10% of cases and RET fusions in ~5% of cases (5).Therapeutic approaches that inhibit specific mediators, first demonstrated with selumetinib, a MAPK pathway inhibitor, and subsequently with BRAF mutation inhibitors, have demonstrated the ability to restore RAI uptake for patients with metastatic RAI-refractory thyroid cancer (6,7).Recently, this strategy was extended to rarer oncogenic fusion genes like RET and NTRK1/2/3 rearrangements using the RET inhibitor selpercatinib and the NTRK inhibitor larotrectinib (6,8,9).The first instance of RAI uptake restoration leading to a structural response to RAI treatment was reported in a 64-year-old female adult with an EML4:NTRK3 rearrangement-positive PTC with synchronous lymph and lung metastases (8).Restoration of RAI uptake was also observed in a metastatic TPR::NTRK1 fusion-positive pediatric PTC; however, the patient did not undergo RAI therapy due to the protocol followed in the context of participation in a clinical trial (10).Recently, larotrectinib was reported to re-induce RAI uptake and subsequent RAI treatment led to a partial response according to RECIST at 3 months post RAI treatment in two of three female adults with metastatic, acquired and primary RAI-resistant NTRKrearranged PTC (11).The reasons for the unsuccessful RAI re-sensitization in the third patient are unknown.In BRAF or RAS mutated or wildtype metastatic, RAIresistant thyroid cancers treated with BRAF inhibitors and MEK inhibitors, re-induction of RAI uptake can only be achieved in 50-71% according to a recent systematic review of seven studies (12) and 60-95% in another recent study (13,14).In a few patients, mutations in SWI/SNF genes were identified as the reason for the failure of RAI re-induction upon MAPK inhibition (15).
There is a compelling need to identify patients most likely to fail RAI re-sensitization treatments, as opposed to the current practice of empirically treating BRAF, RAS mutated, or RET or NTRK-rearranged metastatic RAI-resistant patients with RAI after MAPK inhibition irrespective of the tumor genotype.TERT promoter mutations were shown to be strongly associated with loss of RAI avidity and impairment of the iodide-metabolizing machinery in recurrent PTC (16).Therefore, we have analyzed the implication of TERT promoter mutations and the expression of iodide metabolizing genes in two patients with successful and unsuccessful RAI re-sensitization after treatment with the NTRK inhibitor larotrectinib.

Patients Case 1
Case 1 is an 83-year-old male.In June 2017, the patient had a chest X-ray following an upper respiratory tract infection, which incidentally revealed multiple pulmonary nodules.Subsequent [18F] Fluorodeoxyglucose ( 18 F-FDG) PET-CT revealed ~40 bilateral lung metastases, ranging from 5 mm to 14 mm.Synchronous thyroid nodules were discovered, and a 4.1 × 2.8 cm thyroid nodule biopsy was positive for follicular variant PTC.He is in otherwise good health.He has not been exposed to radiation or chemotherapy, nor does his family have any documented cases of hereditary cancer syndrome.In November 2017, a total thyroidectomy was performed with minimal extrathyroidal extension toward the trachea and positive for vascular invasion but no lymphatic or perineural invasion.Cancer staging of the mixed follicular, classical variant PTC was determined to be T4bNxM1 in accordance with the American Joint Committee on Cancer (AJCC), 8th edition (17), reflecting lung metastases and categorizing the patient as having a high risk of recurrence.
In January 2018, he received 150 mCi (5550 MBq) of 131 I for thyroid ablation with subsequent post-131 I therapy whole-body scan(WBS) indicating RAI accumulation in the thyroid bed and a low-grade RAI accumulation within the lungs.In August 2018, a chest CT scan exhibited a general decrease in the size of the lung metastases, concomitant with a decrease in Tg levels from 3.3 µg/L in February 2018 to 1.4 µg/L in August 2018.However, between August 2018 and September 2020, Tg levels began to uptrend from a nadir of 1.4 µg/L to 5.0 µg/L (Fig. 1).Additionally, there was ~0.25 cm growth within the indexed pulmonary metastases detected on chest CT, while a neck ultrasound was negative for recurrence.

Case 2
A previously healthy 31-year-old female was diagnosed with PTC detected at 20 weeks' gestation.She observed a neck lump 3 years prior, with no size alterations since its discovery.During her pregnancy, the patient had bilateral FNA biopsies of the thyroid performed due to a right lower lobe nodule measuring 3.7 × 2.3 cm, which showed microcalcifications and hypervascularity, as well as a 1.3 × 0.9 cm lymph node on the left side of her neck confirming classical PTC.
In March 2016, 1 month post partum, she underwent a total thyroidectomy, central and bilateral cervical lymph node dissection, and partial tracheal resection.Pathology evaluation confirmed classical variant PTC, with the largest tumor on the right lobe being 3.3 cm.Postsurgery pathology evaluation revealed 9 of 18 resected lymph nodes examined were positive for metastatic PTC, the largest metastasis being 2.0 cm.Moreover, pathology also confirmed microscopic extrathyroidal extension into skeletal muscles and vascular invasion.She received 100 mCi (3700 MBq) of 131 I in June 2016.
Initially, in 2016, case 2 demonstrated a negative 131 I WBS post RAI therapy as the scan displayed focal RAI uptake in the thyroid bed but no RAI-avid distant metastases.However, an increase in Tg prompted a comprehensive CT scan of the whole body, including a targeted scan of the head and neck with 18 F-FDGPET in March 2018 to investigate further disease progression.The scans did not reveal any significant 18 F-FDG uptake indicative of active disease, but they did identify numerous bilateral pulmonary nodules suspicious of metastatic spread, characterized by only low-grade 18 F-FDG uptake.These nodules were scattered, primarily located in the lower lobes, among which two target lesions were identified.The first target lesion (target lesion 1) was situated at the medial basal region of the right lower lobe, measuring 1.2 × 1.1 cm, while the second lesion (target lesion 2) was the largest in the left lower lobe located at the anteromedial basal region measuring 0.8 × 0.8 cm.A core biopsy of the right lung nodules in June 2018 confirmed metastatic PTC.This led to a re-staging of the PTC to T4aN1bM1, in accordance with the AJCC, 8th edition (17), reflecting lung metastases.
Further, ThyroSPEC study of the primary tumor demonstrated a TPR::NTRK1 fusion-positive tumor (without TERT promoter mutations).With evidence of a TPR::NTRK1 fusion-positive tumor, the patient began treatment with larotrectinib at a dose of 100 mg twice daily in June 2022.

Larotrectinib treatment
Health Canada has authorized larotrectinib for the treatment of adults and children with solid tumors that exhibit NTRK gene fusion, provided that these tumors do not have a known acquired resistance mutation.It is applicable for patients whose tumors are either metastatic or in cases where surgical removal would likely cause significant harm, and who lack other effective treatment options.
In case 1, the patient began larotrectinib 100 mg twice daily starting in June 2021.With Synthroid 175 mg, TSH suppression was achieved.
In case 2, the patient began larotrectinib 100 mg twice daily starting in June 2022.With Synthroid 125 mg, TSH suppression was achieved.
For both cases, at no point was larotrectinib paused during RAI treatment.Tg was monitored with serial follow-up appointments (Figs. 1 and 2).

ThyroSPEC
ThyroSPEC formalin-fixed, paraffin embedded (FFPE) analysis was performed as previously described by Eszlinger et al. (18) As part of the analysis, DNA and RNA are isolated from FFPE tissue of the primary tumor resected during the total thyroidectomy, pre-larotrectinib treatment.

RNAseq and determination of thyroid differentiation score
For RNAseq, 100 ng total RNA extracted from macrodissected FFPE material of the primary tumors were used to construct RNASeq libraries following the Illumina RNA Prep with Enrichment Tagmentation protocol.RNA was denatured, and first-strand cDNA was synthesized, followed by second-strand synthesis.cDNA was then fragmented using bead-linked transposomes, and adapter sequences were added.Libraries were then cleaned and normalized before hybridizing to exome probes.Hybridized probe libraries were captured, washed, and amplified to enrich the target library.The enriched library was cleaned, and quality metrics were assessed before sequencing on a NextSeq 500 (Illumina).Thyroid differentiation score (TDS) was derived from primary tumor RNAseq.The mean log2 fold changes for mRNASeq read counts for 16 thyroid differentiation genes were determined as described in the TCGA study (5).

Larotrectinib treatment
Case 1: One-year post larotrectinib start, Tg levels decreased from a pre-treatment value of 5.6 µg/L recorded in April 2021 to 1.7 µg/L obtained in October 2023 (Fig. 1), while a 18 F-FDG PET-CT scan demonstrated reduced metabolic activity and size reduction of pulmonary metastases.However, diagnostic 123 I WBS at 3 and 12 months post larotrectinib treatment were negative for re-induction of RAI uptake; thus, the patient did not receive further RAI treatment.
Case 2: Larotrectinib at a dose of 100 mg twice daily started in June 2022 yielded favorable radiological and biochemical responses over the next year.The patient's Tg levels began to steadily decline from a pre-treatment value of 6.0 µg/L in May 2022 to 1.9 µg/L in January 2023 (Fig. 2).
In December 2022, a diagnostic 123 I WBS revealed RAI uptake in the lung metastases ( 123 I WBS on larotrectinib, Fig. 3); this was a significant development compared to the 123 I WBS conducted pre-larotrectinib treatment in December 2021, which showed no RAI uptake in the thyroid bed or in the pulmonary metastases ( 123 I WBS pre-larotrectinib, Fig. 3).This finding supported the decision to administer a further round of RAI therapy, leading to the patient receiving 150 mCi (5550 MBq) 131 I therapy in January 2023 (post-131 I treatment WBS, Fig. 3).
Diagnostic imaging with 123 I SPECT/CT scans was conducted both before and during larotrectinib treatment.The initial 123 I SPECT/CT scan in December 2021 (Figs.4A and 5A) showed no RAI uptake before treatment.However, a follow-up 123 I SPECT/CT B (Figs. 4B and 5B) conducted 6 months into larotrectinib therapy, in December 2022, revealed a significant reduction in the size of target lesions along with Tg values obtained with suppressed TSH recorded over time for case 2. 123 I RAI treatments, 123 I WBS and SPECT/CTs, and FDG PET-CTs are indicated.Case 2 received her first 100 mCi (3700 MBq) of 131 I in June 2016, followed by a diagnostic 123 I WBS (Fig. 3, pre-larotrectinib) and SPECT/CT A (Figs. 4A and 5A) in December 2021.Larotrectinib began in June 2022, followed by a 6-month on treatment diagnostic 123 I WBS (Fig. 3, on larotrectinib) with SPECT/CT B (Figs. 4B and 5B) obtained on the same day.The evidence for reinduction of radioiodine uptake supported the repeat 131 I RAI therapy in January 2023.To provide additional support for the decision to proceed with another round of 131 I therapy, 18 F-FDG PET/CT imaging was also conducted.The pre-treatment scan in December 2021 showed low-grade metabolic activity within known pulmonary metastases (FDG PET-CT A, Figs. 6 and 7).Contrastingly, the 18 F-FDG PET/CT scan performed after 6 months of larotrectinib treatment demonstrated a favorable anatomical and metabolic response (FDG PET-CT B, Figs. 6 and 7).This improvement was characterized by a decrease in the size of targeted lesions and the elimination of the previously observed low-grade metabolic activity, in addition to a reduced FDG-PET avidity in the target lesions.
Finally, post-131 I WBS displayed low-grade metabolic activity in the right thyroid bed and multiple RAI-avid    lung metastases with significant uptake, confirming restoration of RAI-avidity (post-131 I treatment WBS, Fig. 3).
In addition, a chest CT conducted in May 2022 (1 month before larotrectinib treatment) visualized target lesion 1 measured at 8 × 6 mm, while the target lesion 2 measured at 8 × 7 mm.By June 2023, a follow-up chest CT scan demonstrated that these target lesions had significantly reduced in size, with the first target lesion decreasing from 8 × 6 mm to 5 × 4 mm and the second target lesion from 8 × 7 mm to 5 × 5 mm, respectively.
At the time of this report, both patients continue to tolerate larotrectinib without toxicity.

RNAseq and determination of thyroid differentiation score
The mean log2 of fold changes for mRNASeq read counts of 16 thyroid differentiation genes, as described in the TCGA study (5), was −0.287 for the TERT positive and −0.060 for the TERT negative tumor, respectively.The apical iodide transporter (SLC5A8) and the sodiumiodide symporter (SLC5A5) gene expression was characterized by a 2.5-fold and a 2.4-fold upregulation of normalized expression values in the TERT-negative compared to TERT-positive tumor, respectively, whereas Pendrin (SLC5A8) showed increased expression (Fig. 8).Log2 fold changes are given in Fig. 9.

Discussion
In summary, mutation and RNA expression analysis of the primary NTRK-rearranged tumors detected a TERT promoter mutation and a low TDS, with downregulation of the apical iodide transporter (SLC5A8) and the sodium-iodide symporter (SLC545) gene expression in the primary tumor of a patient without re-induction of RAI uptake (Fig. 8).In contrast, in the patient showing re-induction of RAI uptake following treatment with larotrectinib, her tumor was negative for TERT promoter mutations and showed high TDS with no downregulation of the apical iodide transporter (SLC5A8) and the sodium-iodide symporter (SLC5A5) gene expression (Fig. 8).Our findings suggest a possible association between additional TERT promoter mutations, alongside low TDS and SLC5A5 mRNA expression in NTRKrearranged metastatic RAI-resistant thyroid cancers and the re-induction failure of RAI uptake following larotrectinib treatment.However, given that these observations stem from a study of only two patients, further research is essential to validate these findings.
The potential mechanism of action for the re-differentiating effect of larotrectinib might be similar to that of MAPK inhibitors (10).In vitro experiments demonstrated larotrectinib-induced restoration of RAI uptake was mediated by sodium/iodide symporter re-expression.This re-differentiating effect was also described for RET rearrangements (9, 10).Corroborating our TDS analysis, in vitro studies conducted by Lee et al. (10) also found that the expression of several genes, notably SLC5A8 and especially SLC5A5 -which are critical for RAI uptake in cells -were expressed at very low levels in patients with two RAI-refractory PTC.This underexpression of SLC5A5 could potentially account for RAI resistance observed in these tumors.Our study findings also revealed a noteworthy 2.5-fold upregulation of SLC5A8 and a 2.4-fold upregulation of SLC5A5 in the TERT-negative tumor as compared to the TERT-positive tumor (Fig. 9).
The lower mRNA expression of the SLC5A5 gene prelarotrectinib therapy in case 1 and the negative post-RAI treatment WBS may be attributed to the presence of the TERT promoter mutation.An analysis of SLC5A5 expression on 378 primary PTCs by Tavares et al. (20) demonstrated a significantly lower SLC5A5 expression in PTCs that harbored TERT, BRAF, or especially dual BRAF and TERT mutations compared to wild-type PTCs (20).In another retrospective study, there was a loss of RAI avidity in 97% of all patients with recurrent disease and the presence of co-existing BRAF and TERT mutation, pointing to a potential synergistic effect (16).The authors also analyzed PTC data in TCGA database for the expression of the thyroid iodide-metabolizing genes, including SLC5A5, TSHR, TPO, TG, and PAX8 in 386 PTC samples that had information available for the analysis and showed lower expression in the TERT-positive group than the TERT-negative group (16).
The TCGA study identified TERT mutations in 9.4% of 384 tumors (5).The impact of TERT mutation was further investigated in patients with distant DTC metastases (21).Of the 66 patients with distant DTC metastases, 15 harbored a TERT mutation, and a rising Tg was observed in 14 patients.Notably, all these patients were classified as RAI-refractory and were associated with an older mean age (~58) at diagnosis, larger tumors, and a greater likelihood of BRAF V600E mutation.
Although the 83-year-old male did not display reinduction of RAI uptake after treatment with larotrectinib, there was a clear clinical and radiological benefit of larotrectinib.This observation is consistent with the respective description of the patient with a lack of re-sensitization post larotrectinib treatment by Groussin et al. (11) and re-demonstrates a difference between the antitumoral and the re-differentiating effect of the NTRK inhibitor.
The Ion AmpliSeq Cancer Hotspot Panel v2 (Life Technologies) was used for the analysis of the previously described NTRK-rearrangement positive metastatic

Figure 1
Figure 1 Tg values with suppressed TSH for case 1. RAI treatments are indicated.Case 1 received his first 150 mCi (5550 MBq) of 131 I in January 2018, followed by another 100 mCi of 131 I in March 2021.Two diagnostic 123 I WBS and SPECT/CTs were performed in September 2021 and June 2023, respectively.Larotrectinib treatment began in June 2021.

Figure 3
Figure 3 Chronological comparison of whole-body scans (WBS) in anterior and posterior viewing panels for TPR::NTRK1 fusion-positive and TERT-negative case 2. Pre-diagnostic 123 I WBS displays the diagnostic 123 I WBS obtained pre-larotrectinib treatment in December 2021.Post-diagnostic 123 I WBS exhibits the diagnostic 123 I WBS obtained at 6 months since the initiation of larotrectinib treatment in December 2022, which showed reinduction of radioiodine uptake in comparison to the December 2021 scan.Post-131 I treatment WBS on the right is the post-150 mCi (5550 MBq) 131 I treatment WBS at 7 months post larotrectinib treatment with re-induction of RAI in pulmonary metastases.
increased 123 I uptake, indicating a positive response to the treatment.

Figure 4 Case 2
Figure 4 Case 2 123 I SPECT/CTs (left panels) and maximum intensity projection (MIP) (right panels) obtained before and after larotrectinib treatment for target lesion 1. 123 I SPECT/CT A represents pre-larotrectinib 123 I SPECT/CT and MIP of the chest in December 2021.In comparison, 123 I SPECT/CT B was obtained post larotrectinib therapy, indicating increased 123 I uptake, in December 2022.

Figure 5 Case 2
Figure 5 Case 2 123 I SPECT/CTs (left panels) and maximum intensity projection (MIP) (right panels) obtained before and after larotrectinib treatment for target lesion 2. Pre-larotrectinib therapy 123 I SPECT/CT and MIP conducted in December 2021 are represented in 123 I SPECT/CT A. In comparison, 123 I SPECT/CT B was obtained post larotrectinib therapy, indicating increased 123 I uptake, in December 2022.

Figure 6 Case 2 FDG
Figure 6 Case 2 FDG PET-CTs (left panels) and maximum intensity projection (MIP) (right panels) obtained before and after larotrectinib treatment for target lesion 1. Pre-larotrectinib therapy FDG PET conducted in December 2021 is represented by FDG PET A. In comparison, FDG PET B was obtained post larotrectinib therapy in November 2022, demonstrating a reduction in size and decreased FDG-PET avidity of the target lesion.

Figure 7 Case 2
Figure 7 Case 2 FDG-PETs and maximum intensity projection (MIP) (right panels) obtained before and after larotrectinib treatment for target lesion 2. Pre-larotrectinib therapy FDG PET CT and MIP of the chest conducted on December 2021 in FDG-PET A. In comparison, FDG-PET B was obtained post larotrectinib treatment, indicating a reduction in size and decreased FDG-PET avidity of the target lesion, in November 2022.

Figure 8
Figure 8 Normalized mRNA expression values for TDS genes for case 1 and case 2. The values in the figure indicate the number of counts for the respective genes.

Figure 9 Log 2
Figure 9Log 2 fold changes (log 2 expression values of case 2 divided by expression values of case 1) for mRNA expression of TDS genes.