Can mutation abundance assess the biological behavior of BRAFV600E-positive papillary thyroid carcinoma?

Yeqin Ni, Ping Song, Xiangfeng Lin, Jingjing Shi, Qian Shi, Yuanhui Li, Siyu Zhu, Tianhan Zhou, Yanping Xun, Shirong Zhang, Xing-chang Ren, Kaining Lu, Fan Wu, Wei Wang, Pan Zhao, Rongjing Zhou, Wenhua Zhang, Dandan Li, Jiaoping Zhang, Chuanghua Chen, Linlin Mao, Li Zhou, Gang Pan, Peng You, Yunxian Yu, Yuying Chen, Ni Rong, Guopei Luo, Yu Shrike Zhang, Jugao Fang, Haitao Zheng, Dingcun LuoJournal of Translational Medicine 2025 · DOI

Abstract ChiCTR, ChiCTR2300071472. Registered 31 July 2016 - Retrospectively registered, https://www.chictr.org.cn/showproj.html?proj=190478. This study investigates whether the quantitative abundance of the $BRAF^{V600E}$ mutation, rather than its mere presence, can predict the aggressiveness of papillary thyroid carcinoma (PTC).

Introduction

Papillary thyroid carcinoma (PTC) is the most prevalent thyroid malignancy, with the $BRAF^{V600E}$ mutation serving as its primary genetic driver. While historically associated with poor prognosis, the increasing sensitivity of genetic testing—such as Droplet Digital PCR (ddPCR)—has revealed this mutation in up to 87% of PTC cases. This high prevalence has led to the "BRAF Paradox," where the mutation's presence alone fails to distinguish between indolent and aggressive tumors. This study explores whether the quantitative mutation abundance can restore predictive power for preoperative risk assessment.

Methods and Study Design

The researchers employed a multi-modal approach integrating molecular biology, digital pathology, and machine learning. The investigation was validated across three distinct patient groups to ensure robustness.

Study workflow and cohort selection for the retrospective, prospective, and external validation datasets.

Analytical Techniques

  • Quantification (ddPCR): ddPCR was used for high-sensitivity detection and quantification of $BRAF^{V600E}$ levels, offering roughly 100 times the sensitivity of traditional sequencing.
  • AI-Assisted Adjustment (DIA): Digital Image Analysis (DIA) with AI was utilized to calculate tumor cell ratios, adjusting raw genetic data to account for non-cancerous cell "dilution" in biopsies.
  • Predictive Modeling (XGBoost): Four models were developed using the Extreme Gradient Boosting algorithm. The primary focus was Model D (The Fusion Model), which combined preoperative clinical data (ultrasound, age, sex) with AI-adjusted mutation abundance.

Results

Abundance as a Key Predictor

The study found that $BRAF^{V600E}$ mutation status (positive vs. negative) did not significantly correlate with aggressive behavior in prospective or external cohorts. In contrast, mutation abundance was strongly associated with tumor diameter, lymph node metastasis, and extrathyroidal extension. This suggests a "threshold effect," where clinical invasiveness only emerges once a sufficient proportion of cells harbor the mutation.

Model Performance

The Fusion Model (Model D) significantly outperformed the clinical-only model (Model C). In the external validation set, Model D achieved an Area Under the Curve (AUC) of 0.82, a 21% improvement over Model C ($P=0.02$), effectively matching the accuracy of postoperative pathological reviews. Multivariate analysis confirmed that adjusted $BRAF^{V600E}$ abundance is an independent risk factor for extrathyroidal extension (OR 1.190) and lymph node metastasis (OR 1.111).

Discussion and Clinical Implications

The findings suggest that a mutation abundance cutoff of 0.167 can effectively stratify patients. Scores below this threshold indicate low-risk PTC, potentially suitable for active surveillance, while higher scores indicate intermediate-to-high risk requiring surgical intervention.

While promising, the study is limited by its focus on a single mutation and its observational nature. Future research should incorporate other genetic markers, such as TERT promoter mutations, and proteomic analysis to further refine risk stratification and surgical planning.

Summary updated

Abstract

ChiCTR, ChiCTR2300071472. Registered 31 July 2016 - Retrospectively registered, https://www.chictr.org.cn/showproj.html?proj=190478 .
BibTeX
@article{ni2025can,
  author = {Yeqin Ni and Ping Song and Xiangfeng Lin and Jingjing Shi and Qian Shi and Yuanhui Li and Siyu Zhu and Tianhan Zhou and Yanping Xun and Shirong Zhang and Xing-chang Ren and Kaining Lu and Fan Wu and Wei Wang and Pan Zhao and Rongjing Zhou and Wenhua Zhang and Dandan Li and Jiaoping Zhang and Chuanghua Chen and Linlin Mao and Li Zhou and Gang Pan and Peng You and Yunxian Yu and Yuying Chen and Ni Rong and Guopei Luo and Yu Shrike Zhang and Jugao Fang and Haitao Zheng and Dingcun Luo},
  title = {{Can mutation abundance assess the biological behavior of BRAFV600E-positive papillary thyroid carcinoma?}},
  journal = {Journal of Translational Medicine},
  year = {2025},
  doi = {10.1186/s12967-025-06493-4},
}
Figure