Prediction of EVT6-NTRK3-Dependent Papillary Thyroid Cancer Using Minor Expression Profile

Area of Science:

• Oncology
• Molecular Biology
• Genetics

Background:

• Rare solid tumors driven by oncogenic neurotrophin receptor (TRK) fusions respond to targeted therapies like entrectinib and larotrectinib.
• Current detection relies on next-generation sequencing to identify chimeric TRK transcripts.
• A hypothesis suggests that chimeric tyrosine kinase protein expression creates a unique tumor cell transcriptomic profile.

Purpose of the Study:

• To investigate if a specific transcriptomic profile can distinguish TRK-dependent tumors.
• To develop a predictive model for identifying NTRK gene rearrangements in papillary thyroid cancer (TC).
• To establish an mRNA profile for classifying TC based on driver NTRK-chimeric TRK genes.

Main Methods:

• Utilized reverse transcription PCR (RT-PCR) to identify EVT6-NTRK3 rearrangements in 215 papillary TC samples.
• Employed machine learning on TCGA data to develop a predictive function based on 10 key genes.
• Validated the predictive function on independent RT-PCR data from TRK-dependent and TRK-independent thyroid tumors.

Main Results:

• Identified 7 papillary TC samples (3.26%) with EVT6-NTRK3 rearrangement.
• Developed a 10-gene recognition function (AUTS2, DTNA, ERBB4, HDAC1, IGF1, KDR, NTRK1, PASK, PPP2R5B, PRSS1) for predicting NTRK rearrangements.
• Achieved high sensitivity (100%) and specificity (70%) on independent validation samples, with TCGA data showing 72.7% sensitivity and 99.6% specificity.

Conclusions:

• A 10-gene mRNA expression profile can effectively classify papillary thyroid cancer concerning the presence of driver NTRK-chimeric TRK genes.
• This molecular signature offers a potentially valuable tool for identifying patients eligible for TRK-targeted therapies.
• The developed method demonstrates acceptable sensitivity and specificity for clinical application in TRK fusion detection.