Yi Liu’s research while affiliated with Beijing Institute of Technology and other places

Alpha-fetoprotein (AFP) is a proven blood biomarker widely used in clinical detection of liver cancer, and its concentration increases immediately after liver injury, with high diagnostic specificity and low sensitivity. However, in patients with AFP-negative hepatocellular carcinoma (AFP-NHCC), also known as small liver cancer, their early stage of AFP expression is usually low or close to the normal range. Compared with AFP-positive HCC, AFP-NHCC is more likely to be subjected to missed diagnosis due to solely dependence on the measurement of blood AFP. Therefore, it is necessary to find a reliable, effective, and economical detection method for the early diagnosis of AFP-NHCC. PIVKA-II is closely related to the occurrence, development, invasion, and metastasis of liver cancer, and is also a new hematological marker widely used in the diagnosis of liver cancer in recent years. PIVKA-II effectively makes up for the limitation of negative AFP. 63.2% – 76.3% of patients with negative AFP showed positive PIVKA-II, and if only PIVKA-II detection was relied on, there would still be missed diagnosis in early patients with AFP-NHCC. In this retrospective study, we selected several commonly used inflammatory indicators to explore the diagnostic efficacy of PIVKA-II, an abnormal form of prothrombin, combined with inflammatory indicators in patients with early-stage AFP-NHCC and analyzed the relationship between some clinical features and hematological indicators in patients with AFP-NHCC. Serum levels of high-sensitivity C-reactive protein (hs-CRP), prealbumin (PA), neutrophil/lymphocyte ratio (NLR), and PIVKA-II were compared among three groups (AFP-NHCC group, benign lesion group, and healthy subjects). The diagnostic efficacy of PIVKA-II alone for AFP-NHCC and the diagnostic efficacy of three inflammatory indicators combined with PIVKA-II for AFP-NHCC were calculated, and their diagnostic specificity and sensitivity were compared. Compared with the other two groups, the AFP-NHCC group showed significant changes in three inflammatory markers and PIVKA-II, which may be related to the inflammatory progression of the tumor. Therefore, we recommend the establishment of a laboratory-based detection approach for diagnosing early-stage AFP-NHCC by combining PIVKA-II with inflammatory indicators hs-CRP, PA, and NLR, to facilitate diagnosis, treatment, and prognosis of AFP-NHCC.

INTRODUCTION Idiopathic normal pressure hydrocephalus (iNPH) and cerebral small vessel disease (CVSD) are age‐related diseases, but their prevalence and clinical relationship are unclear. METHODS This prospective cohort study enrolled 95 patients with probable iNPH in China and evaluated their CSVD burden using magnetic resonance imaging. Linear regression models were used to analyze the association between CSVD scores and clinical outcomes. RESULTS The results showed 78% of the patients had at least one CSVD imaging marker, and higher total CSVD scores were significantly associated with declines in attention, executive function, psychomotor speed, and gait performance after multivariate adjustments. However, the preoperative CSVD score did not affect the post‐shunt improvement in modified Rankin scale or iNPH grading scale scores. DISCUSSION Our findings suggest that CSVD is prevalent in patients with iNPH and is associated with more severe symptoms, but it may not affect shunt outcomes. Future studies are needed to elucidate the underlying mechanisms. Highlights We found that 78% of the patients with idiopathic normal pressure hydrocephalus (iNPH) had at least one type of cerebral small vessel disease (CSVD) imaging marker. The CSVD burden aggravates cognitive and gait impairments in patients with iNPH but may not affect shunt outcomes. The effects of different imaging markers of CSVD on cognition and gait are different and worthy of further investigation.

Background Cognitive fusion MRI-guided targeted biopsy (cTB) has been widely used in the diagnosis of prostate cancer (PCa). However, cTB relies heavily on the operator’s experience and confidence in MRI readings. Our objective was to compare the cancer detection rates of MRI artificial intelligence-guided cTB (AI-cTB) and routine cTB and explore the added value of using AI for the guidance of cTB. Methods This was a prospective, single-institution randomized controlled trial (RCT) comparing clinically significant PCa (csPCa) and PCa detection rates between AI-cTB and cTB. A total of 380 eligible patients were randomized to the AI-cTB group (n = 191) or the cTB group (n = 189). The AI-cTB group underwent AI-cTB plus systematic biopsy (SB) and the cTB group underwent routine cTB plus SB. The primary outcome was the detection rate of csPCa. The reference standard was the pathological results of the combination of TB (AI-cTB/cTB) and SB. Comparisons of detection rates of csPCa and PCa between groups were performed using the chi-square test or Fisher’s exact test. Results The overall csPCa and PCa detection rates of the whole inclusion cohort were 58.8% and 61.3%, respectively. The csPCa detection rates of TB combined with SB in the AI-cTB group were significantly greater than those in the cTB group at both the patient level (58.64% vs. 46.56%, p = 0.018) and per-lesion level (61.47% vs. 47.79%, p = 0.004). Compared with cTB, the AI-cTB could detect a greater proportion of patients with csPCa at both the per-patient level (69.39% vs. 49.71%, p < 0.001) and per-lesion level (68.97% vs. 48.57%, p < 0.001). Multivariate logistic analysis indicated that compared with the cTB, the AI-cTB significantly improved the possibility of detecting csPCa (p < 0.001). Conclusions AI-cTB effectively improved the csPCa detection rate. This study successfully integrated AI with TB in the routine clinical workflow and provided a research paradigm for prospective AI-integrated clinical studies. Trial registration ClinicalTrials.gov, NCT06362291.

Traumatic brain injury (TBI) represents a common and severe medical condition necessitating prompt risk stratification to enhance patient outcomes. Although substantial research has been conducted on the prognostic utility of various biomarkers for TBI, no single biomarker has been definitively recognized as the most precise predictor of disease outcomes. In comparison to other markers, the neutrophil-albumin ratio (NAR) has emerged as a cost-effective and reproducible inflammatory biomarker, demonstrating potential in evaluating the severity of inflammation and prognosticating outcomes in infections and cerebrovascular diseases. This study evaluated the prognostic significance of the NAR in comparison to two other readily accessible and cost-effective composite indices: the Neutrophil-Lymphocyte Ratio (NLR) and the Platelet-Lymphocyte Ratio (PLR) in individuals with TBI. We conducted a retrospective cohort analysis involving 297 hospitalized TBI patients, gathering comprehensive demographic, anthropometric, medical, clinical, laboratory, and imaging data to assess the expression changes of these biomarkers. Our findings suggest that both the NAR and the NLR possess predictive value regarding prognosis following TBI. However, receiver operating characteristic (ROC) curve analysis revealed that NAR outperformed NLR as a prognostic predictor. In conclusion, our examination of blood biochemistry composite indicators indicates that, while both NAR and NLR serve as significant prognostic markers, NAR is a more effective predictor of outcomes in patients with TBI.

Large language models (LLMs) have revolutionized language processing, but face critical challenges with security, privacy, and generating hallucinations — coherent but factually inaccurate outputs. A major issue is fact-conflicting hallucination (FCH), where LLMs produce content contradicting ground truth facts. Addressing FCH is difficult due to two key challenges: 1) Automatically constructing and updating benchmark datasets is hard, as existing methods rely on manually curated static benchmarks that cannot cover the broad, evolving spectrum of FCH cases. 2) Validating the reasoning behind LLM outputs is inherently difficult, especially for complex logical relations. To tackle these challenges, we introduce a novel logic-programming-aided metamorphic testing technique for FCH detection. We develop an extensive and extensible framework that constructs a comprehensive factual knowledge base by crawling sources like Wikipedia, seamlessly integrated into Drowzee. Using logical reasoning rules, we transform and augment this knowledge into a large set of test cases with ground truth answers. We test LLMs on these cases through template-based prompts, requiring them to provide reasoned answers. To validate their reasoning, we propose two semantic-aware oracles that assess the similarity between the semantic structures of the LLM answers and ground truth. Our approach automatically generates useful test cases and identifies hallucinations across six LLMs within nine domains, with hallucination rates ranging from 24.7% to 59.8%. Key findings include LLMs struggling with temporal concepts, out-of-distribution knowledge, and lack of logical reasoning capabilities. The results show that logic-based test cases generated by Drowzee effectively trigger and detect hallucinations. To further mitigate the identified FCHs, we explored model editing techniques, which proved effective on a small scale (with edits to fewer than 1000 knowledge pieces). Our findings emphasize the need for continued community efforts to detect and mitigate model hallucinations.

Oncidium hybridum Lodd. is an important ornamental flower that is used as both a cut flower and a potted plant around the world; additionally, its pseudobulbs serve as essential carriers for floral organs and flower development. The NAM gene family is crucial for managing responses to various stresses as well as regulating growth in plants. However, the mechanisms by which NAM genes regulate the development of pseudobulbs remain unclear. In this study, a total of 144 NAM genes harboring complete structural domains were identified in O. hybridum. The 144 NAM genes were systematically classified into 14 distinct subfamilies via phylogenetic analysis. Delving deeper into the conserved motifs revealed that motifs 1–6 exhibited remarkable conservation, while motifs 7–10 presented in a few NAM genes only. Notably, NAM genes sharing identical specific motifs were classified into the same subfamily, indicating functional relatedness. Furthermore, the examination of occurrences of gene duplication indicated that the NAM genes display 16 pairs of tandem duplications along with five pairs of segmental duplications, suggesting their role in genetic diversity and potential adaptive evolution. By conducting a correlation analysis integrating transcriptomics and metabolomics at four stages of pseudobulb development, we found that OhNAM023, OhNAM030, OhNAM007, OhNAM019, OhNAM083, OhNAM047, OhNAM089, and OhNAM025 exhibited significant relationships with the endogenous plant hormones jasmonates (JAs), hinting at their potential involvement in hormonal signaling. Additionally, OhNAM089, OhNAM025, OhNAM119, OhNAM055, and OhNAM136 showed strong links with abscisic acid (ABA) and abscisic acid glucose ester (ABA-GE), suggesting the possible regulatory function of these NAM genes in plant growth and stress responses. The 144 NAM genes identified in this study provide a basis for subsequent research and contribute to elucidating the intricate molecular mechanisms of NAM genes in Oncidium and potentially in other species.

Plants generate a range of physiological and molecular responses to sustain their growth and development when suffering heat stress. Avocado is a type of tropical fruit tree with high economic value. Most avocado cultivars delete, wither, or even die when exposed to heat stress for a long time, which seriously restricts the introduction and cultivation of avocados. In this study, samples of a heat-intolerant variety (‘Hass’) were treated under heat stress, and the transcriptomics and metabolomics were analyzed, with the expectation of providing information on the variety improvement and domestication of avocados. The differentially expressed genes identified using transcriptome analysis mainly involved metabolic pathways such as plant hormone signal transduction, plant–pathogen interaction, and protein processing in the endoplasmic reticulum. Combined transcriptome and metabolome analysis indicated that the down-regulation of Hass.g03.10206 and Hass.g03.10205 in heat shock-like proteins may result in the reduced Trehalose and Sinapoyl aldehyde content. Metabolomics analysis results indicated that the decrease in Trehalose and Sinapoyl aldehyde content may be an important factor for heat intolerance. These results provide important clues for understanding the physiological mechanisms of adaptation to heat stress in avocados.