JCO precision oncology · 2026
Integrated Radioproteomic Modeling for Early Recurrence Prediction and Metabolic Characterization in Hepatocellular Carcinoma.
СтатьяНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Zhuang Qiuyu, Xu Tongtong, Xing Xiaohua, Hu En, Zhou Yang, Zheng Xiaoyuan, Bai Lixia, Huang Yao, Liu Xiaolong, Chen Yutao
Аннотация
PURPOSE: Hepatocellular carcinoma (HCC) is a leading cause of cancer-related mortality, with high recurrence rates after surgical resection posing a significant challenge. While deep learning (DL) approaches show promise in predicting HCC recurrence, their clinical translation is limited by poor interpretability and unclear mechanisms. Our study aimed to develop an interpretable DL framework that predicts recurrence while elucidating underlying biology through integration of radiologic imaging and multiomics profiling.
MATERIALS AND METHODS: We developed a DL framework to predict early postoperative HCC recurrence using preoperative multiphase computed tomography imaging and generate an imaging-based early recurrence risk score (ERRS) for risk stratification. To decipher the biological basis of ERRS, we integrated DL features with proteomic data to identify key metabolic alterations, which were validated by metabolomics, immunohistochemistry, and enzymatic assays. Patient-derived organoids (PDOs) were used to assess the therapeutic potential of targeting these alterations.
RESULTS: The Multi-model_NC&ART&PV DL model showed improved performance compared with conventional single-phase models in predicting early postoperative recurrence, with high-ERRS patients exhibiting worse survival and more aggressive features. Mechanistically, radioproteomic analyses linked model predictions to dysregulated pyruvate metabolism, characterized by reduced pyruvate dehydrogenase complex expression/activity and elevated lactate dehydrogenase (LDH) activity. PDOs from patients with HCC were sensitive to LDH inhibitor stiripentol, with high-ERRS tumors showing greater therapeutic vulnerability than low-ERRS tumors.
CONCLUSION: This study bridges artificial intelligence-driven imaging and mechanism-guided therapy by developing a biologically interpretable DL model for HCC recurrence prediction. Radioproteomic integration identified dysregulated pyruvate metabolism as a hallmark of high-risk HCC, enabling the repurposing of stiripentol as a potential therapy. This framework suggests a potential strategy linking noninvasive risk stratification with pathway-guided treatment although further validation and prospective studies are needed to establish its clinical utility.
MATERIALS AND METHODS: We developed a DL framework to predict early postoperative HCC recurrence using preoperative multiphase computed tomography imaging and generate an imaging-based early recurrence risk score (ERRS) for risk stratification. To decipher the biological basis of ERRS, we integrated DL features with proteomic data to identify key metabolic alterations, which were validated by metabolomics, immunohistochemistry, and enzymatic assays. Patient-derived organoids (PDOs) were used to assess the therapeutic potential of targeting these alterations.
RESULTS: The Multi-model_NC&ART&PV DL model showed improved performance compared with conventional single-phase models in predicting early postoperative recurrence, with high-ERRS patients exhibiting worse survival and more aggressive features. Mechanistically, radioproteomic analyses linked model predictions to dysregulated pyruvate metabolism, characterized by reduced pyruvate dehydrogenase complex expression/activity and elevated lactate dehydrogenase (LDH) activity. PDOs from patients with HCC were sensitive to LDH inhibitor stiripentol, with high-ERRS tumors showing greater therapeutic vulnerability than low-ERRS tumors.
CONCLUSION: This study bridges artificial intelligence-driven imaging and mechanism-guided therapy by developing a biologically interpretable DL model for HCC recurrence prediction. Radioproteomic integration identified dysregulated pyruvate metabolism as a hallmark of high-risk HCC, enabling the repurposing of stiripentol as a potential therapy. This framework suggests a potential strategy linking noninvasive risk stratification with pathway-guided treatment although further validation and prospective studies are needed to establish its clinical utility.