Normal view

Received — 18 September 2026 ⏭

(Multiomics OR Omics) AND (Lung OR gastric OR Hepatocellular)
Non-Invasive Assessment of Microvascular Invasion Risk in Hepatocellular Carcinoma Using Liquid Biopsy: Translational Insights and Clinical Implications Dengyuan Xue · Xinyu Gao · Qixingmao Zhang · Hongxin Li · Mengli Chen · Xiuzhi Duan · Xuchu Wang · Pan Yu · Zhihua Tao · Xiaoxue Cheng
Diagnostics (Basel). 2026 Aug 22;16(17):2686. doi: 10.3390/diagnostics16172686.ABSTRACTMicrovascular invasion (MVI) is a critical prognostic indicator for recurrence and survival in hepatocellular carcinoma (HCC); however, its accurate preoperative assessment remains clinically challenging. Postoperative histopathology is subject to sampling bias and time delays, while traditional imaging techniques lack the molecular specificity required to predict MVI. Liquid biopsy, through the analysis of ci
15 September 2026 at 18:00

Non-Invasive Assessment of Microvascular Invasion Risk in Hepatocellular Carcinoma Using Liquid Biopsy: Translational Insights and Clinical Implications

(Multiomics OR Omics) AND (Lung OR gastric OR Hepatocellular)

By: Dengyuan Xue · Xinyu Gao · Qixingmao Zhang · Hongxin Li · Mengli Chen · Xiuzhi Duan · Xuchu Wang · Pan Yu · Zhihua Tao · Xiaoxue Cheng

15 September 2026 at 18:00

Diagnostics (Basel). 2026 Aug 22;16(17):2686. doi: 10.3390/diagnostics16172686.

ABSTRACT

Microvascular invasion (MVI) is a critical prognostic indicator for recurrence and survival in hepatocellular carcinoma (HCC); however, its accurate preoperative assessment remains clinically challenging. Postoperative histopathology is subject to sampling bias and time delays, while traditional imaging techniques lack the molecular specificity required to predict MVI. Liquid biopsy, through the analysis of circulating tumor DNA (ctDNA), circulating tumor cells (CTCs), circulating tumor RNA (ctRNA), and extracellular vesicles (EVs), provides a minimally invasive approach for capturing tumor-derived molecular and cellular signals associated with vascular invasion. This narrative review comprehensively summarizes the current evidence linking these four liquid biopsy analyte categories to MVI in HCC, evaluates their integration into multi-omics predictive models, including multi-marker, clinicopathological-integrated, and imaging-integrated strategies, and proposes an evidence-level framework that categorizes blood biomarkers according to the strength of their support for MVI prediction, distinguishing direct histopathological validation from indirect associations with aggressive tumor biology. Key challenges are critically examined, including the variable specificity of individual biomarkers for MVI, the lack of head-to-head comparative studies, the absence of standardized pre-analytical and analytical protocols, and the methodological limitations of current prediction models. As a narrative review, this work does not employ systematic review methodology, and the evidence synthesis should be interpreted accordingly. The review provides a framework for understanding how liquid biopsy-based MVI risk stratification may inform surgical and perioperative decision-making following prospective validation.

PMID:42739118 | PMC:PMC13564874 | DOI:10.3390/diagnostics16172686

Received — 26 March 2026 ⏭

cs.AI, q-bio.NC updates on arXiv.org
MedCausalX: Adaptive Causal Reasoning with Self-Reflection for Trustworthy Medical Vision-Language Models Jianxin Lin · Chunzheng Zhu · Peter J. Kneuertz · Yunfei Bai · Yuan Xue
arXiv:2603.23085v1 Announce Type: new Abstract: Vision-Language Models (VLMs) have enabled interpretable medical diagnosis by integrating visual perception with linguistic reasoning. Yet, existing medical chain-of-thought (CoT) models lack explicit mechanisms to represent and enforce causal reasoning, leaving them vulnerable to spurious correlations and limiting their clinical reliability. We pinpoint three core challenges in medical CoT reasoning: how to adaptively trigger causal correction, c
25 March 2026 at 12:00

MedCausalX: Adaptive Causal Reasoning with Self-Reflection for Trustworthy Medical Vision-Language Models

cs.AI, q-bio.NC updates on arXiv.org

By: Jianxin Lin · Chunzheng Zhu · Peter J. Kneuertz · Yunfei Bai · Yuan Xue

25 March 2026 at 12:00

arXiv:2603.23085v1 Announce Type: new Abstract: Vision-Language Models (VLMs) have enabled interpretable medical diagnosis by integrating visual perception with linguistic reasoning. Yet, existing medical chain-of-thought (CoT) models lack explicit mechanisms to represent and enforce causal reasoning, leaving them vulnerable to spurious correlations and limiting their clinical reliability. We pinpoint three core challenges in medical CoT reasoning: how to adaptively trigger causal correction, construct high-quality causal-spurious contrastive samples, and maintain causal consistency across reasoning trajectories. To address these challenges, we propose MedCausalX, an end-to-end framework explicitly models causal reasoning chains in medical VLMs. We first introduce the CRMed dataset providing fine-grained anatomical annotations, structured causal reasoning chains, and counterfactual variants that guide the learning of causal relationships beyond superficial correlations. Building upon CRMed, MedCausalX employs a two-stage adaptive reflection architecture equipped with $\langle$causal$\rangle$ and $\langle$verify$\rangle$ tokens, enabling the model to autonomously determine when and how to perform causal analysis and verification. Finally, a trajectory-level causal correction objective optimized through error-attributed reinforcement learning refines the reasoning chain, allowing the model to distinguish genuine causal dependencies from shortcut associations. Extensive experiments on multiple benchmarks show that MedCausalX consistently outperforms state-of-the-art methods, improving diagnostic consistency by +5.4 points, reducing hallucination by over 10 points, and attaining top spatial grounding IoU, thereby setting a new standard for causally grounded medical reasoning.