❌

Normal view

Received β€” 18 September 2026 ⏭

Artificial Intelligence-Driven Multiomics and Clinical Investigation Identify Macrophage Migration Inhibitory Factor as a Pan-Cancer Biomarker

βœ‡(Multiomics OR Omics) AND (Lung OR gastric OR Hepatocellular)
By: Fengsen Liu Β· Yu Ping Β· Miaomiao Li Β· Kaiyuan Guo Β· Lei Zhang Β· Peng Wu Β· Yijun Tian Β· Jinyan Liu Β· Yi Zhang
17 September 2026 at 18:00

Phenomics. 2026 May 20;6(3):213-229. doi: 10.1007/s43657-026-00322-4. eCollection 2026 Jun.

ABSTRACT

Early cancer detection remains challenging due to the lack of reliable pan-cancer screening methods, particularly blood-based biomarkers. Using a novel three-tiered validation framework combining artificial intelligence (AI)-powered literature mining of 180,000 PubMed articles (1950-2024), multiomics integration across major databases, and extensive clinical validation, we identified macrophage migration inhibitory factor (MIF) as a promising blood-based biomarker for pan-cancer detection. Multiomics analysis revealed consistent MIF upregulation across 21 cancer types at the transcriptional level and across 12 cancer types at the protein level. Clinical validation in independent cohorts (n = 4,269) showed that serum MIF protein levels discriminated effectively between cancer patients and healthy controls (median AUC = 0.994) and between cancer and benign conditions (median AUC = 0.881). Notably, comparative analyses showed that MIF demonstrated superior or comparable performance to established cancer-specific markers, including AFP for hepatocellular carcinoma (MIF AUC = 0.885 vs. AFP AUC: 0.744-0.887) and CA125 for ovarian cancer (MIF AUC = 0.831 vs. CA125 AUC: 0.58-0.71). Meta-analysis of 28 cohorts (n = 5,347) confirmed the diagnostic efficacy of MIF (pooled AUC: 0.782). This cost-effective, blood-based ELISA approach establishes MIF as a valuable tool for broad applications in cancer screening.

SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at https://doi.org/10.1007/s43657-026-00322-4.

PMID:42750739 | PMC:PMC13578188 | DOI:10.1007/s43657-026-00322-4

Received β€” 27 May 2026 ⏭

A Sober Look at Agentic Misalignment in Automated Workflows

βœ‡cs.AI, q-bio.NC updates on arXiv.org
By: Wenqian Ye Β· Bo Yuan Β· Zhichao Xu Β· Yijun Tian Β· Yawei Wang Β· Henry Kautz Β· Aidong Zhang
26 May 2026 at 12:00
arXiv:2605.24197v1 Announce Type: new Abstract: We study a class of emergent misalignment in multi-agent systems (MAS), with a focus on automated workflows, which we refer to agentic misalignment. Although these systems can solve complex tasks, they often fail because agents act according to implicit proxy utilities that do not align with the intended human goals. We formally define these behaviors and analyze them within a Bayesian framework, showing that generic utilities naturally lead to posterior collapse of agents in automated workflows. To address this issue, we propose Agentic Evidence Attribution (AEA), a novel alignment paradigm that improves agent posteriors using context-specific evidence. AEA reasons over agent actions and provides structured evidence to correct misaligned behavior during collaboration. To better understand the role of evidence, we study two instantiations of AEA: self-reflection (internal evidence from the model) and weak-to-strong generalization (external evidence on the agentic trajectory). We show that a small evidence model effectively aligns the MAS by providing orthogonal failure attribution. Our results clarify the sources of agentic misalignment in automated workflows and show that evidence-based alignment can effectively improve agent collaboration and leads to reliable multi-agent systems built on automated workflows.
There are no more articles
❌