Case Report: Discordance between RNA-identified ETV6::RET and clinically reportable DNA profiling in salivary gland secretory carcinoma

Abstract

Secretory carcinoma of the salivary gland is a rare, fusion-driven malignancy in which kinase fusions can serve as therapeutic targets. We report a patient with an ETV6::RET-positive tumor who achieved a complete response to pralsetinib and subsequently developed resistance. Research whole-transcriptome sequencing (WTS) of the resistant lesion identified an expressed in-frame ETV6::RET transcript retaining the RET kinase domain, with 300 and 196 split reads at the two breakpoints and three discordant mate pairs; research whole-exome sequencing provided limited concordant DNA-level support. A clinically approved targeted DNA-based comprehensive genomic profiling assay applied to material from the same lesion detected RET intron 11 and ETV6 intron 6 rearrangement signals but did not report actionable ETV6::RET or RET p.L730V. Under the Japanese regulatory framework applicable when selpercatinib was considered, an approved in vitro diagnostic or medical device was required to confirm a RET fusion for non-lung, non-thyroid solid tumors. Research WTS did not constitute an approved companion diagnostic result, and the approved DNA report did not establish an actionable ETV6::RET finding; selpercatinib was therefore not initiated. When DNA profiling yields unresolved rearrangement signals, a clinically validated RNA-based fusion assay and an appropriate regulatory framework for treatment selection may improve access to matched therapy.

Introduction

Salivary gland secretory carcinoma is a rare malignant salivary gland tumor characterized by recurrent ETV6 rearrangements. The canonical driver alteration is ETV6::NTRK3, which has established the value of molecular testing for this disease (1). In an ETV6::RET rearrangement, the RET kinase domain is the druggable oncogenic component, whereas ETV6 contributes the 5-prime fusion segment; this report does not attribute an independent oncogenic role to ETV6. A subset of secretory carcinomas lacks ETV6::NTRK3 and harbors alternative kinase fusions. Identification of the fusion partner is clinically relevant because it may define a therapeutic target.

ETV6::RET has been reported as a recurrent noncanonical fusion in salivary gland secretory carcinoma (2). Alternative fusions, including ETV6::MET and VIM::RET, have been described (3–7). These findings support comprehensive fusion testing in tumors without the canonical ETV6::NTRK3 rearrangement.

Selective RET inhibitors have demonstrated substantial activity in RET fusion-positive cancers. Selpercatinib has shown tumor-agnostic activity across RET fusion-positive solid tumors, including cancers other than lung and thyroid cancer (8). It has also demonstrated systemic and intracranial activity in RET fusion-positive non-small-cell lung cancer (9). Pralsetinib is active in RET fusion-positive non-small cell lung cancer and other RET fusion-positive solid tumors (10, 11).

Acquired resistance to selective RET inhibitors remains an important clinical issue. In the RETgistry Global Consortium, secondary RET kinase mutations were identified in a subset of post-progression specimens, although bypass mechanisms such as MET amplification were more frequent (12). RET p.L730V was reported as an on-target acquired mutation in plasma cfDNA from patients with RET fusion-positive NSCLC who developed resistance to pralsetinib in a conference abstract from the ARROW study (13). L730V/I mutations are located at the roof of the RET solvent-front region and show reduced sensitivity to pralsetinib while retaining relative sensitivity to selpercatinib in preclinical models (14).

We report a patient with ETV6::RET-positive salivary gland secretory carcinoma who achieved a complete response to pralsetinib and subsequently developed RET L730V-positive acquired resistance. The central issue was not a binary RNA-positive/DNA-negative result: targeted DNA profiling of the same resistant lesion detected rearrangement signals involving RET and ETV6, whereas RNA sequencing resolved an expressed actionable ETV6::RET fusion that was not generated as a clinically reportable companion diagnostic finding. RET L730V in resistant tissue and plasma provided additional context for acquired resistance.

Case description

A 49-year-old man presented with a left parotid gland tumor and cervical lymph node metastases. He was diagnosed with salivary gland secretory carcinoma of the left parotid gland, clinically staged as cT2N3bM0. He underwent extended left parotidectomy and left neck dissection of levels I-V. Pathological examination confirmed salivary gland secretory carcinoma. Metastases were present in 50 of 56 ipsilateral lymph nodes: level I, 0/5; level II, 8/8; level III, 4/4; level IV, 12/12; level V, 5/5; and intraparotid lymph nodes, 21/22. Extranodal extension was present in a level II nodal metastasis, which measured 29 x 25 mm. The pathological classification was at least pT3 and pN3b. The surgical margin was focally positive (R1), including the tissue posterior to the mastoid process and a portion of the dissection surface. Re-resection was not performed because it would have required re-entry through the pectoralis major flap reconstruction and was considered excessively invasive. Postoperative radiotherapy (66 Gy in 30 fractions) was selected for local control.

Three months after surgery, computed tomography revealed recurrent lymph node metastases and multiple pulmonary metastases. The PD-L1 combined positive score (22C3) was 4. Pembrolizumab was selected as first-line therapy at the patient’s preference to avoid cytotoxic chemotherapy-associated toxicities, including alopecia. Pembrolizumab was started on October 19, 2022, and administered for two cycles; progressive disease was confirmed by computed tomography on January 6, 2023.

Diagnostic assessment

After recurrence, archived surgical tumor tissue was analyzed in SCRUM-Japan MONSTAR-SCREEN-2 (UMIN000043899), a nationwide prospective molecular-profiling study for patients with advanced solid tumors that was designed to characterize molecular alterations and support biomarker-guided treatment or trial matching, rather than to evaluate diagnostic accuracy (15). The archived specimen was collected on June 28, 2022, and the MONSTAR-SCREEN-2 test report was issued on November 21, 2022. Caris Life Sciences provided the analysis platform and performed research sequencing, bioinformatics, and variant interpretation at no cost to the study. The study investigators interpreted the findings in the clinical context and made all treatment and trial-eligibility decisions. WES/WTS was performed for research purposes and was not an approved or accredited companion diagnostic in Japan.

Therapeutic intervention

At another institution, the patient enrolled in cohort K of the Tumor-agnostic Precision Immuno-oncology and Somatic Targeting Rational for You (TAPISTRY) trial (NCT04589845), a global, multicenter, open-label, multi-cohort phase II platform trial evaluating the safety and efficacy of biomarker-matched targeted therapies or immunotherapy in patients with advanced solid tumors. Oral pralsetinib was initiated at 400 mg once daily on February 2, 2023. The metastatic lesions regressed markedly, and the patient’s best overall response was an investigator-assessed complete response according to RECIST version 1.1 (Figure 1). Radiographic progressive disease was documented on March 28, 2024, approximately 14 months after pralsetinib initiation.

Figure 1.

Clinical course and longitudinal molecular findings. Pembrolizumab was administered for two cycles, followed by radiographic progression. Pralsetinib was administered for approximately 14 months until radiographic progression. The surgical specimen showed ETV6::RET by research WES/WTS through SCRUM-Japan MONSTAR-SCREEN-2. Pralsetinib induced an investigator-assessed complete response according to RECIST version 1.1. At progression, the resistant lesion retained ETV6::RET and acquired RET p.L730V. Targeted DNA-based comprehensive genomic profiling of material from the same resistant-lesion tissue block detected RET- and ETV6-related rearrangement signals but did not identify actionable ETV6::RET in the clinically reportable result. WES, whole-exome sequencing; WTS, whole-transcriptome sequencing.

Follow-up and outcomes

At radiographic progression, a subcutaneous lesion was biopsied as a resistant metastasis. The resistant tissue was analyzed through the same MONSTAR-SCREEN-2 platform. WTS identified an expressed in-frame ETV6::RET fusion retaining the RET protein kinase domain. The fusion was supported by 300 and 196 split reads at the respective breakpoints and three discordant mate pairs. WES provided limited but concordant DNA-level support for the same rearrangement, with 15 split reads at one breakpoint, no split reads at the reciprocal breakpoint, and no discordant mate pairs (Figure 2; Supplementary Table S1). RET p.L730V was also detected in the resistant lesion. Plasma circulating tumor DNA obtained at progression likewise detected RET p.L730V, confirming its presence in both tissue and plasma (Supplementary Table S2).

Figure 2.

Molecular characterization of the post-pralsetinib resistant lesion. (A) Research WTS demonstrated an expressed in-frame ETV6 exon 6::RET exon 12 fusion retaining the RET kinase domain, supported by 300 and 196 split reads at the respective breakpoints and three discordant mate pairs. (B) Research WES provided limited but concordant DNA-level support for the same rearrangement, with 15 split reads at one breakpoint and no split reads at the reciprocal breakpoint or discordant mate pairs. Breakpoint coordinates in this image are reported on the GRCh38 human reference genome.

To assess eligibility for selpercatinib, material from the same resistant lesion was evaluated using a clinically approved targeted DNA-based comprehensive genomic profiling assay. The companion-diagnostic section did not report an alteration meeting its eligibility criteria. The tumor-profiling section nevertheless reported two RET intron 11 rearrangements, including RET::CSGALNACT2, and two unpartnered ETV6 intron 6 rearrangements; it did not report ETV6::RET as an actionable fusion or RET p.L730V as a reportable short variant (Table 1; Supplementary Table S3). Research WES/WTS and clinical targeted DNA-CGP used material from the same resistant-lesion tissue block but followed distinct laboratory and analytic workflows. Research profiling reported tumor mutational burden (TMB) 10 mutations/Mb, TP53 p.L257Q, RET p.L730V, CDKN2A p.R22fs, and TERT promoter c.-124C>T, whereas targeted DNA-CGP reported TMB 0 mutations/Mb, TP53 p.L194R, CDKN2A p.R22fs*4, ASXL1 p.K568*, two RET intron 11 rearrangements, and two ETV6 intron 6 rearrangements. Differences in capture regions, local coverage, bioinformatic pipelines, variant-calling and reporting thresholds, and TMB algorithms can account for assay-specific outputs even when material is derived from the same tissue block. The results should not be interpreted as a direct comparison of assay accuracy.

Table 1.

Key targeted DNA-based comprehensive genomic profiling findings.

Category	Finding
Companion-diagnostic interpretation	No alteration was reported as meeting the assay’s companion-diagnostic criteria; tumor-profiling findings are listed below
RET rearrangements	Two RET intron 11 rearrangements; one was annotated as RET::CSGALNACT2
ETV6 rearrangements	Two ETV6 intron 6 rearrangements without assigned partners
Actionable fusion report	ETV6::RET was not reported as an actionable fusion
RET L730V	Not reported as a reportable short variant
Other genomic alterations	ASXL1 K568*; CDKN2A R22fs*4; TP53 L194R
Microsatellite status	Microsatellite stable
Tumor mutational burden	0 mutations/Mb

The companion-diagnostic interpretation and the tumor-profiling section are separate components of the report. The absence of a qualifying companion-diagnostic result does not mean that no genomic alterations or rearrangement signals were detected.

Selpercatinib was considered because RET p.L730V/I has shown greater preclinical resistance to pralsetinib than to selpercatinib (14). Under the Japanese regulatory framework applicable when treatment was considered, the prescribing information for RET fusion-positive advanced or recurrent solid tumors other than non-small cell lung cancer and thyroid cancer required confirmation of a RET fusion with an approved in vitro diagnostic or medical device (16). Research WTS established an expressed in-frame ETV6::RET fusion but was not an approved companion diagnostic result, while the approved targeted DNA-CGP report did not identify ETV6::RET as the companion-diagnostic finding. Accordingly, the diagnostic criterion for routine on-label care at our institution was not met and selpercatinib was not initiated. Cisplatin plus docetaxel was started on April 19, 2024, and discontinued after the fifth cycle on July 22, 2024. The patient subsequently received palliative radiotherapy to the right cervical lesion beginning on July 29, 2024 (45 Gy in 15 fractions), followed by whole-brain radiotherapy beginning on September 2, 2024 (30 Gy in 10 fractions). He died on September 18, 2024.

Patient perspective

The patient preferred an initial treatment strategy that avoided cytotoxic chemotherapy-associated toxicities, particularly alopecia. He subsequently agreed to molecular profiling of tumor tissue and plasma. During his lifetime, he signed the SCRUM-Japan MONSTAR-SCREEN-2 consent form, which included consent for publication of his clinical course and genomic findings. After selpercatinib could not be initiated within routine on-label care, he agreed to proceed with cisplatin plus docetaxel. No direct statement from the patient regarding treatment response, disease progression, or inability to receive selpercatinib was documented; therefore, the Patient Perspective is limited to objective information recorded in the medical record.

Discussion

The principal finding was discordance between research WTS identification of an expressed, kinase-domain-retaining ETV6::RET fusion and clinically reportable targeted DNA-CGP from the same resistant lesion. This was not a binary RNA-positive/DNA-negative result: DNA-CGP detected RET and ETV6 rearrangement signals but did not report actionable ETV6::RET, whereas WTS resolved the expressed in-frame transcript. The assays used different designs and reporting frameworks and were not compared for analytical accuracy.

RET p.L730V should not be regarded as a common resistance mutation. The RETgistry Global Consortium identified RET G810 mutations most frequently among on-target resistance events and also reported V804M, A764G, and L881V; L730V was not recurrent (12). A preliminary ARROW conference abstract identified patient-derived RET p.L730V in plasma after pralsetinib progression (13), whereas Shen et al. provided peer-reviewed evidence that L730V/I reduced sensitivity to pralsetinib more than to selpercatinib (14). The present tissue and plasma findings support an association with the resistant clone but do not establish causality.

Research WES/WTS and clinical DNA-CGP yielded different TMB estimates and variant lists despite using material from the same resistant-lesion tissue block. Differences in capture regions, local coverage, bioinformatic pipelines, variant-calling and reporting thresholds, and TMB algorithms can produce assay-specific outputs. WES mainly interrogates exonic regions, WTS evaluates the expressed fusion transcript, and targeted DNA-CGP applies its own coverage, annotation, and companion-diagnostic rules. Raw clinical-assay data and locus-level coverage metrics were unavailable for reanalysis; therefore, the precise reasons for the discordant TMB estimates and individual variant calls could not be determined. DNA-CGP detected RET intron 11 and ETV6 intron 6 rearrangements but did not report ETV6::RET as actionable. RET::CSGALNACT2 and unpartnered ETV6 events may represent components of a complex rearrangement, but the available data do not permit reconstruction of its architecture. Long-read sequencing, optical genome mapping, or another structural-variant-resolving method would be required.

RNA sequencing can be useful when a kinase fusion is suspected or a DNA assay yields unresolved rearrangement signals. RNA sequencing has identified targetable kinase fusions not resolved by DNA sequencing in other settings (17), and Guha et al. used research transcriptome analysis to identify ETV6::NTRK3 in a rare pediatric secretory breast carcinoma (18). These reports support transcript-level fusion assessment but do not analytically validate MONSTAR-SCREEN-2, establish superiority of research WTS, or justify routine implementation without a clinically validated and appropriately regulated assay.

The reporting discordance affected treatment access. At the time, Japanese prescribing information required RET fusion confirmation with an approved in vitro diagnostic or medical device for advanced or recurrent solid tumors other than non-small cell lung cancer and thyroid cancer (16). Research WTS did not meet that regulatory role, and the approved DNA-CGP report did not identify ETV6::RET as the companion-diagnostic finding. Thus, the barrier arose from the labeled diagnostic requirement and reportable assay output, not from absence of DNA rearrangement signals. When DNA profiling yields unresolved signals, a clinically validated RNA-based fusion assay and an appropriate regulatory framework may improve access to matched therapy.

Several limitations apply. The platforms and reporting criteria differed, and the outputs were not directly comparable. No orthogonal structural-variant assay or functional validation of the patient-derived ETV6::RET-p.L730V construct was performed. Selpercatinib was not administered, so this case cannot establish clinical benefit or an altered outcome with that drug. The reason RET p.L730V was absent from the clinically reportable DNA-CGP result remains uncertain, and this single case cannot establish assay performance or a general resistance mechanism. The complete response to pralsetinib and detection of RET p.L730V in resistant tissue and plasma support, but do not prove, the proposed biological interpretation.

Conclusion

Research WTS identified an expressed ETV6::RET fusion, whereas clinically approved targeted DNA-CGP from the same resistant lesion detected related RET and ETV6 rearrangement signals but did not report actionable ETV6::RET. The finding concerns complementary assay interpretation and clinical reporting, not analytical superiority of one platform. RET p.L730V in resistant tissue and plasma supports a hypothesis of acquired pralsetinib resistance, but a single case without functional validation cannot establish the mechanism or predict benefit from an alternative RET inhibitor.

CARE guidelines statement

This case report was prepared in accordance with the CARE guidelines.

Funding Statement

The author(s) declared that financial support was not received for this work and/or its publication.

Data availability statement

The data analyzed in this study is subject to the following licenses/restrictions: Deidentified data supporting the findings of this case report are available from the corresponding author upon reasonable request and are subject to applicable ethical and institutional restrictions. Requests to access these datasets should be directed to Shogen Boku, boku.sho@kmu.ac.jp.

Ethics statement

The studies involving humans were approved by the Institutional Review Board of the National Cancer Center Hospital East. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study. Written informed consent was obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.

Author contributions

SB: Investigation, Data curation, Writing – review & editing, Writing – original draft, Validation, Conceptualization, Formal analysis. TF: Data curation, Methodology, Conceptualization, Resources, Writing – review & editing, Project administration, Formal analysis, Writing – original draft. TSa: Validation, Data curation, Writing – review & editing, Investigation, Writing – original draft. SK: Writing – review & editing, Writing – original draft. HS: Writing – original draft, Writing – review & editing. HY: Writing – review & editing, Writing – original draft. KO: Writing – review & editing, Writing – original draft. YS: Writing – original draft, Writing – review & editing. TO: Writing – review & editing, Writing – original draft. BK: Writing – review & editing, Writing – original draft. RY: Data curation, Project administration, Investigation, Methodology, Resources, Writing – review & editing, Writing – original draft, Formal analysis. TSh: Resources, Writing – review & editing, Writing – original draft, Formal analysis, Project administration. YN: Writing – original draft, Resources, Writing – review & editing, Formal analysis. HB: Resources, Writing – original draft, Writing – review & editing, Formal analysis. TY: Formal analysis, Writing – review & editing, Supervision, Writing – original draft, Resources. MR: Writing – original draft, Writing – review & editing, Resources, Formal analysis. MY: Writing – original draft, Investigation, Data curation, Writing – review & editing. SO: Writing – review & editing, Writing – original draft.

Conflict of interest

SB reports research funding from Amgen and honoraria from MSD, Ono Pharmaceutical, and Bristol Myers Squibb. TF reports honoraria from Amelieff Co., Ltd., AstraZeneca and MSD. SK reports research funding from Ono Pharmaceutical, Taiho Pharmaceutical, MSD, Janssen Pharmaceutical, Bayer, Eli Lilly, Chugai Pharmaceutical, Daiichi Sankyo, AstraZeneca, and AbbVie; and honoraria from Ono Pharmaceutical, Taiho Pharmaceutical, MSD, Daiichi Sankyo, Merck KGaA, Bristol Myers Squibb, Eli Lilly, Chugai Pharmaceutical, Bayer, Eisai, and Novartis. HS reports research funding from Asahi Kasei, Ono Pharmaceutical, and MSD; and honoraria from Bayer, Bristol Myers Squibb, Chugai Pharmaceutical, Daiichi Sankyo, Eli Lilly Japan, Merck Biopharma, MSD, Ono Pharmaceutical, Sanofi, Taiho Pharmaceutical, Takeda Pharmaceutical, Asahi Kasei, AstraZeneca, Astellas Pharma, Novartis, Tsumura & Co., Guardant Health, Miyarisan Pharmaceutical, Kyorin, Eisai, Terumo, and Yakult Honsha. HY reports research funding from MSD and honoraria from MSD. YS reports research funding from MSD; and honoraria from Bristol Myers Squibb, Eisai, Eli Lilly, Merck Biopharma, MSD, and Ono Pharmaceutical. TO reports research funding from Kyowa Kirin; and honoraria from Bristol Myers Squibb, MSD, Novartis, and Chugai Pharmaceutical. TS reports honoraria from AstraZeneca, Novartis, Exact Sciences, and Nihon Servier. YN reports research funding from Genomedia, Guardant Health Japan, Tempus Labs, Roche Diagnostics K.K., Daiichi Sankyo, PRA Health, and Chugai Pharmaceutical Co., Ltd.; and honoraria from Chugai Pharmaceutical Co., Ltd. HB reports an advisory role for GxD; and honoraria from Ono Pharmaceutical, Taiho Pharmaceutical, and Guardant Health. TY reports research funding from Bristol Myers Squibb, Caris MPI, Chugai Pharmaceutical, Daiichi Sankyo, Eisai, Exact Sciences, FALCO Biosystems, Medical & Biological Laboratories, Merus N.V., Miyarisan Pharmaceutical, MSD K.K., Natera, Nippon Boehringer Ingelheim, Ono Pharmaceutical, Pfizer Japan, Sysmex, Taiho Pharmaceutical, and Takeda Pharmaceutical; honoraria from Chugai Pharmaceutical, Takeda Pharmaceutical, Merck Biopharma, and Ono Pharmaceutical; and consulting fees from Sumitomo Corp. and Indivumed. MR is an employee and shareholder of Caris Life Sciences. SO reports honoraria from Merck Serono, Ono Pharmaceutical, Bristol Myers Squibb, MSD, Bayer Yakuhin, Eisai, Novartis, and Eli Lilly.

The remaining author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declared that generative AI was used in the creation of this manuscript. During the preparation of this work, the authors used ChatGPT (OpenAI) for language editing and manuscript-formatting support. The authors reviewed and edited all output and take full responsibility for the content of the publication.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fonc.2026.1967793/full#supplementary-material