❌

Normal view

Received β€” 18 September 2026 ⏭

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

βœ‡(Multiomics OR Omics) AND (Lung OR gastric OR Hepatocellular)
By: Dengyuan Xue Β· Xinyu Gao Β· Qixingmao Zhang Β· Hongxin Li Β· Mengli Chen Β· Xiuzhi Duan Β· Xuchu Wang Β· Pan Yu Β· Zhihua Tao Β· 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 β€” 26 March 2026 ⏭

Not All Tokens Are Created Equal: Query-Efficient Jailbreak Fuzzing for LLMs

βœ‡cs.AI, q-bio.NC updates on arXiv.org
By: Wenyu Chen Β· Xiangtao Meng Β· Chuanchao Zang Β· Li Wang Β· Xinyu Gao Β· Jianing Wang Β· Peng Zhan Β· Zheng Li Β· Shanqing Guo
25 March 2026 at 12:00
arXiv:2603.23269v1 Announce Type: cross Abstract: Large Language Models(LLMs) are widely deployed, yet are vulnerable to jailbreak prompts that elicit policy-violating outputs. Although prior studies have uncovered these risks, they typically treat all tokens as equally important during prompt mutation, overlooking the varying contributions of individual tokens to triggering model refusals. Consequently, these attacks introduce substantial redundant searching under query-constrained scenarios, reducing attack efficiency and hindering comprehensive vulnerability assessment. In this work, we conduct a token-level analysis of refusal behavior and observe that token contributions are highly skewed rather than uniform. Moreover, we find strong cross-model consistency in refusal tendencies, enabling the use of a surrogate model to estimate token-level contributions to the target model's refusals. Motivated by these findings, we propose TriageFuzz, a token-aware jailbreak fuzzing framework that adapts the fuzz testing approach with a series of customized designs. TriageFuzz leverages a surrogate model to estimate the contribution of individual tokens to refusal behaviors, enabling the identification of sensitive regions within the prompt. Furthermore, it incorporates a refusal-guided evolutionary strategy that adaptively weights candidate prompts with a lightweight scorer to steer the evolution toward bypassing safety constraints. Extensive experiments on six open-source LLMs and three commercial APIs demonstrate that TriageFuzz achieves comparable attack success rates (ASR) with significantly reduced query costs. Notably, it attains a 90% ASR with over 70% fewer queries compared to baselines. Even under an extremely restrictive budget of 25 queries, TriageFuzz outperforms existing methods, improving ASR by 20-40%.
There are no more articles
❌