Quantitative structure-retention relationship (QSRR) is an important tool in chromatography. QSRR examines the correlation between molecular structures and their retention behaviors during chromatographic separation. This approach involves developing models for predicting the retention time (RT) of analytes, thereby accelerating method development and facilitating compound identification. In addition, QSRR can be used to study compound retention mechanisms and support drug screening efforts. This review provides a comprehensive analysis of QSRR workflows and applications, with a special focus on the role of artificial intelligence-an area not thoroughly explored in previous reviews. Moreover, we discuss current limitations in RT prediction and propose promising solutions. Overall, this review offers a fresh perspective on future QSRR research, encouraging the development of innovative strategies that enable the diverse applications of QSRR models in chromatographic analysis.

Quantitative structure-retention relationship(QSRR)is an important tool in chromatography.QSRR examines the correlation between molecular structures and their retention behaviors during chro-matographic separation.This approach involves developing models for predicting the retention time(RT)of analytes,thereby accelerating method development and facilitating compound identification.In addition,QSRR can be used to study compound retention mechanisms and support drug screening ef-forts.This review provides a comprehensive analysis of QSRR workflows and applications,with a special focus on the role of artificial intelligence—an area not thoroughly explored in previous reviews.More-over,we discuss current limitations in RT prediction and propose promising solutions.Overall,this re-view offers a fresh perspective on future QSRR research,encouraging the development of innovative strategies that enable the diverse applications of QSRR models in chromatographic analysis.

Objective To systematically review the application and effectiveness of deep learning(DL)in diagnosis of mild cogni-tive impairment(MCI)among older adults.Methods PubMed,Web of Science,CNKI and Wanfang databases were searched for literatures related to the application of DL in MCI among older adults,from database inception to December,2024.A scoping review was conducted.The literature screening process followed the Scoping Review Report Specification list,and the quality assess-ment was conducted using the cross-sectional study quality evaluation tool developed by the Evidence-based Health Care Center.Results A total of eleven papers were included,from Italy,USA,South Korea,China,India and Switzerland,involving 11 829 elderly participants,publicated mainly between 2014 and 2024,reflecting the rapid development trend of the field in the last decade,which was in line with the timing of the development of DL technology.The quality scores of the included literatures were all six to seven.The types of studies were all cross-sectional studies with significant cross-disciplinary characteristics,mainly originating from the fields of clinical medicine,biology and neuroimaging.The literature data were mainly based on the Alzheimer's disease Neuroimaging Program database and integrated other data resources.In terms of data type,in addition to brain imaging data,one study based on text data was also included.In terms of models used,five of the studies were mainly based on convolutional neu-ral networks,and the rest used different DL modeling frameworks.The task types contained binary and triple classification.In terms of prediction results,the DL models constructed on multimodal data,such as brain imag-es,could be used to construct high-precision prediction models for MCI classification,and the models were all good,with accuracy more than 70%and AUC values more than 0.7.The diagnostic accuracy of some of the mod-els was more than 90%,and the model with the highest prediction accuracy was the one that used the Biceph-Net lightweight framework,with accuracy close to 100%,and the text analysis model based on Transformer made the AUC value of 0.846,which provided new ideas for the diagnosis of non-imaging data.Conclusion DL can not only provide strong support for the accurate identification of MCI in the elderly,but also provide auxiliary prediction tools for clinicians,which can help delay the progression of the disease and improve the prognosis of patients.

Objective To systematically review the application and effectiveness of deep learning(DL)in diagnosis of mild cogni-tive impairment(MCI)among older adults.Methods PubMed,Web of Science,CNKI and Wanfang databases were searched for literatures related to the application of DL in MCI among older adults,from database inception to December,2024.A scoping review was conducted.The literature screening process followed the Scoping Review Report Specification list,and the quality assess-ment was conducted using the cross-sectional study quality evaluation tool developed by the Evidence-based Health Care Center.Results A total of eleven papers were included,from Italy,USA,South Korea,China,India and Switzerland,involving 11 829 elderly participants,publicated mainly between 2014 and 2024,reflecting the rapid development trend of the field in the last decade,which was in line with the timing of the development of DL technology.The quality scores of the included literatures were all six to seven.The types of studies were all cross-sectional studies with significant cross-disciplinary characteristics,mainly originating from the fields of clinical medicine,biology and neuroimaging.The literature data were mainly based on the Alzheimer's disease Neuroimaging Program database and integrated other data resources.In terms of data type,in addition to brain imaging data,one study based on text data was also included.In terms of models used,five of the studies were mainly based on convolutional neu-ral networks,and the rest used different DL modeling frameworks.The task types contained binary and triple classification.In terms of prediction results,the DL models constructed on multimodal data,such as brain imag-es,could be used to construct high-precision prediction models for MCI classification,and the models were all good,with accuracy more than 70%and AUC values more than 0.7.The diagnostic accuracy of some of the mod-els was more than 90%,and the model with the highest prediction accuracy was the one that used the Biceph-Net lightweight framework,with accuracy close to 100%,and the text analysis model based on Transformer made the AUC value of 0.846,which provided new ideas for the diagnosis of non-imaging data.Conclusion DL can not only provide strong support for the accurate identification of MCI in the elderly,but also provide auxiliary prediction tools for clinicians,which can help delay the progression of the disease and improve the prognosis of patients.

Objective:To analyze the risk factors for frailty in older adult patients with chronic obstructive pulmonary disease (COPD) and its correlation with oxidative stress.Methods:A total of 168 patients with COPD aged 60 years and above, who were treated at Pingxiang People's Hospital from August 2022 to August 2023, were selected as the study subjects using the convenient sampling method. The FRAIL scale was utilized to assess frailty status. Patients were divided into two groups based on their FRAIL scale scores: the frail group (≥ 3 points, n = 109), the non-frail/pre-frail group (< 3 points, n = 59). Patients in the non-frail/pre-frail group were sub-divided into the pre-frail group (1-2 points, n = 23), and the non-frail group (0 points, n = 36). Serum levels of 8-hydroxydeoxyguanosine and malondialdehyde, and total antioxidant capacity were measured. One-way analysis of variance was performed to compare differences between groups, and correlation analysis was conducted. Logistic regression was used to identify the risk factors for frailty. Results:The incidence of frailty among 168 older adult patients with COPD was 64.9% (109/168). Multivariable logistic regression analysis revealed that age ( OR = 1.59, 95% CI 1.02-2.25), body mass index ( OR = 4.11, 95% CI 2.02-8.42), comorbidities ( OR = 2.57, 95% CI 1.31-5.02), activities of daily living ( OR = 3.07, 95% CI 1.54-6.06), malnutrition ( OR = 2.97, 95% CI 1.56-5.41), and cognitive impairment ( OR = 2.87, 95% CI 1.42-5.88) were risk factors for frailty in older patients with COPD ( P < 0.05). The frailty scores of older adult patients with COPD were significantly positively correlated with serum levels of 8-hydroxydeoxyguanosine and malondialdehyde ( γ= 0.67, 0.65, P = 0.008, 0.006), and negatively correlated with total antioxidant capacity ( γ= -0.54, P = 0.012). Conclusion:Age, body mass index, comorbidities, activities of daily living, malnutrition, and cognitive impairment are risk factors for frailty in older adult patients with COPD, and the severity of frailty is markedly associated with levels of oxidative stress products.

Objective To develop a deep learning method for volumetric assessment of traumatic intracerebral hemorrhage(TICH)using the Trans-UNet model and to compare its performance with traditional formula-based methods.Methods CT data from 141 TICH patients admitted to Army Medical Center of PLA between May 2018 and May 2023 were collected.A deep learning method based on the Trans-UNet model was established.Manual delineation via picture archiving and communication system(PACS)was served as the gold standard for comparing the accuracy,consistency,and time efficiency of our method against 10 different formula-based methods for measuring the amount of TICH.Results The median volume of TICH,as manual delineation via PACS,was 1.167 mL,with a median measurement time of 135 s per patient.The median percentage error in volume between the deep learning method and manual delineation via PACS was 3.59％.Spearman correlation coefficient was 0.999(P＜0.001),and a median measurement time was only 4.38 s per patient.In contrast,in the formula-based methods,the lowest median percentage error in volume was 16.451％,the highest Spearman correlation coefficient was 0.986(P＜0.001),and the lowest median measurement time was 20 s for a single patient.The statistical differences were observed in percentage error in volume and measurement time between the 2 types of methods(all P＜0.001).Conclusion Our developed deep learning method for volumetric assessment of TICH is superior to the formula-based methods in terms of measurement accuracy and time efficiency.

OBJECTIVES: To evaluate the application value of CT-based radiomics features for the ascending and descending types of nasopharyngeal carcinoma (NPC). METHODS: A total of 217 NPC patients (48 ascending type and 169 descending type), who obtained CT images before radiotherapy in Affiliated Cancer Hospital of Xiangya School of Medicine, Central South University from February 2015 to October 2017, were analyzed retrospectively. All patients were randomly divided into a training set (=153) and a test set (=64). Gross tumor volume in the nasopharynx (GTVnx) was selected as regions of interest (ROI) and was analyzed by radiomics. A total of 1 300 radiomics features were extracted via IBEX. The least absolute shrinkage and selection operator (LASSO) logistic regression was performed to choose the significant features. Support vector machine (SVM) and random forest (RF) classifiers were built and verified. RESULTS: Six features were selected by the LASSO from 1 300 radiomics features. Compared with SVM classifier, RF classifier showed better classification performance. The area under curve (AUC) of the receiver operating characteristic (ROC) curve, accuracy, sensitivity, and specificity were 0.989, 0.941, 1.000, and 0.924, respectively for the training set; 0.994, 0.937, 1.000, and 0.924, respectively for the validation set. CONCLUSIONS CT-based radiomics features possess great potential in differentiating ascending and descending NPC. It provides a certain basis for accurate medical treatment of NPC, and may affect the treatment strategy of NPC in the future.
Humans ; Nasopharyngeal Carcinoma ; Nasopharyngeal Neoplasms ; ROC Curve ; Retrospective Studies ; Sensitivity and Specificity

In the original publication Fig. 1D and supplementary material is incorrect. The correct figure and supplementary material is provided in this correction.

One major strategy to generate genetically modified mouse models is gene targeting in mouse embryonic stem (ES) cells, which is used to produce gene-targeted mice for wide applications in biomedicine. However, a major bottleneck in this approach is that the robustness of germline transmission of gene-targeted ES cells can be significantly reduced by their genetic and epigenetic instability after long-term culturing, which impairs the efficiency and robustness of mouse model generation. Recently, we have established a new type of pluripotent cells termed extended pluripotent stem (EPS) cells, which have superior developmental potency and robust germline competence compared to conventional mouse ES cells. In this study, we demonstrate that mouse EPS cells well maintain developmental potency and genetic stability after long-term passage. Based on gene targeting in mouse EPS cells, we established a new approach to directly and rapidly generate gene-targeted mouse models through tetraploid complementation, which could be accomplished in approximately 2 months. Importantly, using this approach, we successfully constructed mouse models in which the human interleukin 3 (IL3) or interleukin 6 (IL6) gene was knocked into its corresponding locus in the mouse genome. Our study demonstrates the feasibility of using mouse EPS cells to rapidly generate mouse models by gene targeting, which have great application potential in biomedical research.

Recently we have established a new culture condition enabling the derivation of extended pluripotent stem (EPS) cells, which, compared to conventional pluripotent stem cells, possess superior developmental potential and germline competence. However, it remains unclear whether this condition permits derivation of EPS cells from mouse strains that are refractory or non-permissive to pluripotent cell establishment. Here, we show that EPS cells can be robustly generated from non-permissive NOD-scid Il2rg mice through de novo derivation from blastocysts. Furthermore, these cells can also be efficiently generated by chemical reprogramming from embryonic NOD-scid Il2rg fibroblasts. NOD-scid Il2rg EPS cells can be expanded for more than 20 passages with genomic stability and can be genetically modified through gene targeting. Notably, these cells contribute to both embryonic and extraembryonic lineages in vivo. More importantly, they can produce chimeras and integrate into the E13.5 genital ridge. Our study demonstrates the feasibility of generating EPS cells from refractory mouse strains, which could potentially be a general strategy for deriving mouse pluripotent cells. The generation of NOD-scid Il2rg EPS cell lines permits sophisticated genetic modification in NOD-scid Il2rg mice, which may greatly advance the optimization of humanized mouse models for biomedical applications.