Acute respiratory distress syndrome(ARDS)is a severe and life-threatening manifestation of acute lung injury,characterized by widespread pulmonary inflammation and edema,ultimately resulting in acute respiratory failu...Acute respiratory distress syndrome(ARDS)is a severe and life-threatening manifestation of acute lung injury,characterized by widespread pulmonary inflammation and edema,ultimately resulting in acute respiratory failure.Despite advancements in mechanical ventilation and lung-protective strategies,targeted therapies aimed at modulating dysregulated inflammation and promoting tissue repair remain elusive.Extracellular vesicles(EVs),critical mediators of intercellular communication,have emerged as a promising research focus due to their dual regulatory roles in ARDS pathogenesis.Pro-inflammatory EVs,derived from pathogens or injury-stressed cells,exacerbate alveolar macrophage activation and increase endothelial permeability,thereby aggravating pulmonary damage.In contrast,anti-inflammatory EVs originating from mesenchymal stem cells facilitate alveolar barrier restoration and tissue repair by delivering reparative molecular cargo.This review systematically evaluates the dualistic functions of EVs in ARDS from three key perspectives:Molecular mechanisms,clinical translation,and technical challenges.We further discuss the complexities associated with EV heterogeneity,pathogen interactions,and standardization in EV production.Additionally,we propose future directions that integrate engineered EV modifications and multi-omics approaches to address current therapeutic limitations and enhance ARDS management.展开更多
Ulcerative colitis(UC)is a common progressive inflammatory disease whose incidence has increased rapidly in recent years,and can develop into colorectal cancer in severe cases.There are currently no adequate or effect...Ulcerative colitis(UC)is a common progressive inflammatory disease whose incidence has increased rapidly in recent years,and can develop into colorectal cancer in severe cases.There are currently no adequate or effective treatments for UC due to the fact that some patients have found suboptimal results after repeated administration,while others have experienced adverse effects.With the rapid development of nanotechnology,developing innovative colon-targeting platforms is essential to improving efficacy,reducing side effects,and improving patient compliance.In this review,we summarize the pathophysiological characteristics of UC and the most recent status of numerous nanodrug delivery systems based on different targeting mechanisms in treating UC.Oral,intravenous,and rectal drug delivery nanoparticles targeting the colon are discussed,which can provide ideas for the design of colon-targeting nanoparticles for the treatment of colon diseases,especially for the treatment of UC.Last but not least,we provide a glimpse into the future of colon-targeted delivery systems,as well as future advancements in the field.展开更多
BACKGROUND Inflammatory bowel disease(IBD)is a global health burden that affects millions of individuals worldwide,necessitating extensive patient education.Large language models(LLMs)hold promise for addressing patie...BACKGROUND Inflammatory bowel disease(IBD)is a global health burden that affects millions of individuals worldwide,necessitating extensive patient education.Large language models(LLMs)hold promise for addressing patient information needs.However,LLM use to deliver accurate and comprehensible IBD-related medical information has yet to be thoroughly investigated.AIM To assess the utility of three LLMs(ChatGPT-4.0,Claude-3-Opus,and Gemini-1.5-Pro)as a reference point for patients with IBD.METHODS In this comparative study,two gastroenterology experts generated 15 IBD-related questions that reflected common patient concerns.These questions were used to evaluate the performance of the three LLMs.The answers provided by each model were independently assessed by three IBD-related medical experts using a Likert scale focusing on accuracy,comprehensibility,and correlation.Simultaneously,three patients were invited to evaluate the comprehensibility of their answers.Finally,a readability assessment was performed.RESULTS Overall,each of the LLMs achieved satisfactory levels of accuracy,comprehensibility,and completeness when answering IBD-related questions,although their performance varies.All of the investigated models demonstrated strengths in providing basic disease information such as IBD definition as well as its common symptoms and diagnostic methods.Nevertheless,when dealing with more complex medical advice,such as medication side effects,dietary adjustments,and complication risks,the quality of answers was inconsistent between the LLMs.Notably,Claude-3-Opus generated answers with better readability than the other two models.CONCLUSION LLMs have the potential as educational tools for patients with IBD;however,there are discrepancies between the models.Further optimization and the development of specialized models are necessary to ensure the accuracy and safety of the information provided.展开更多
The construction and operation of sulfur-containing gas storage are often more difficult than a non-sulfur storage facility due to the need to prevent environmental contamination from H_(2)S leaks,as well as the corro...The construction and operation of sulfur-containing gas storage are often more difficult than a non-sulfur storage facility due to the need to prevent environmental contamination from H_(2)S leaks,as well as the corrosive effects of H_(2)S on production facilities.Rapid elutriation of H_(2)S from the reservoir during the construction of the gas storage is an effective way to avoid these problems.However,the existing H_(2)S elutriation method has low efficiency and high economic cost,which limits the development of reconstructed gas storage of sulfur-containing gas reservoirs.To improve the efficiency of H_(2)S elutriation in sulfur-containing gas reservoirs and enhance the economic benefits,a numerical simulation model of multiphase flow components was established to study the migration law of H_(2)S in the multi-cycle operation of gas storage.Based on the H_(2)S migrate law,the displacement H_(2)S elutriation method was developed,and the elutriation mechanism and elutriation efficiency of the two methods were compared and analyzed.In addition,the main controlling factors affecting the H_(2)S elutriation efficiency were investigated,and the H_(2)S elutriation scheme of H gas storage was optimized.The results indicate that H_(2)S migrates between near-well and far-well regions under pressure differentials.The traditional H_(2)S elutriation method relies on concentration gradient diffusion,whereas the displacement elutriation approach leverages pressure differentials with higher H_(2)S elutriation efficiency.For the displacement elutriation method,higher reservoir permeability enhances the peak-shaving capacity of the gas storage but has a minor impact on H_(2)S elutriation when the formation permeability is between 30 and 100 mD.The elutriation efficiency is significantly higher when wells are drilled in the high structural parts of the reservoir compared to the low structural parts.Longer displacement elutriation time within a cycle improves H_(2)S elutriation efficiency but reduces the working gas volume of the storage.Therefore,the optimal displacement time for H gas storage is 60 days.An optimized H_(2)S elutriation scheme enabled the working gas to meet the national first-class natural gas standard within 10 cycles.This study elucidates H_(2)S migration patterns,H_(2)S elutriation mechanisms,and key influence factors on H_(2)S elutriation efficiency,offering valuable technical insights for sour gas storage operations.展开更多
BACKGROUND Gastrointestinal diseases have complex etiologies and clinical presentations.An accurate diagnosis requires physicians to integrate diverse information,including medical history,laboratory test results,and ...BACKGROUND Gastrointestinal diseases have complex etiologies and clinical presentations.An accurate diagnosis requires physicians to integrate diverse information,including medical history,laboratory test results,and imaging findings.Existing artificial intelligence-assisted diagnostic tools are limited to single-modality information,resulting in recommendations that are often incomplete and may be associated with clinical or legal risks.AIM To develop and evaluate a collaborative multimodal large language model(LLM)framework for clinical decision-making in digestive diseases.METHODS In this observational study,DeepGut,a multimodal LLM collaborative diagnostic framework,was developed to integrate four distinct large models into a four-tiered structure.The framework sequentially accomplishes multimodal infor-mation extraction,logical“chain”construction,diagnostic and treatment suggestion generation,and risk analysis.The model was evaluated using objective metrics,which assess the reliability and comprehensiveness of model-generated results,and subjective expert opinions,which examine the effectiveness of the framework in assisting physicians.RESULTS The diagnostic and treatment recommendations generated by the DeepGut framework achieved exceptional performance,with a diagnostic accuracy of 97.8%,diagnostic completeness of 93.9%,treatment plan accuracy of 95.2%,and treatment plan completeness of 98.0%,significantly surpassing the capabilities of single-modal LLM-based diagnostic tools.Experts evaluating the framework commended the completeness,relevance,and logical coherence of its outputs.However,the collaborative multimodal LLM approach resulted in increased input and output token counts,leading to higher computational costs and extended diagnostic times.CONCLUSION The framework achieves successful integration of multimodal diagnostic data,demonstrating enhanced performance enabled by multimodal LLM collaboration,which opens new horizons for the clinical application of artificial intelligence-assisted technology.展开更多
Additive manufacturing(AM)methods have garnered considerable attention owing to their flexibility in fabricating complex parts with desirable mechanical properties.However,the poor surface quality of the resulting met...Additive manufacturing(AM)methods have garnered considerable attention owing to their flexibility in fabricating complex parts with desirable mechanical properties.However,the poor surface quality of the resulting metal parts remains a severe challenge for the applications.Here,a novel dual-additive synergy strategy is presented,which simultaneously enhances material removal efficiency and regulates electrode surface reactions during electrochemical polishing(ECP)of AM AlSi10Mg.Theoretical studies and experimental characterizations confirm that NaF promotes selective dissolution at the peaks,while glucose acts as a stabilizer for the surface valleys.This approach effectively facilitates the selective removal of surface protrusions,achieving a smoother and more uniform surface finish,resulting in a surface roughness reduction of approximately 86%,compared to a 63%reduction without additives.This study not only provides a new approach for optimizing surface quality of AM AlSi10Mg but also offers new insights into electrolyte design and the stabilization of metal anodes.展开更多
The measurement field of view of the conventional transmission electron microscopy(TEM)nano-moiréand scanning transmission electron microscopy(STEM)nano-moirémethods is limited to the hundred-nanometer scale...The measurement field of view of the conventional transmission electron microscopy(TEM)nano-moiréand scanning transmission electron microscopy(STEM)nano-moirémethods is limited to the hundred-nanometer scale,unable to meet the deformation field measurement requirements of micrometer-scale materials such as transistors and micro-devices.This paper proposed a novel measurement method based on scanning secondary moire,which can realize cross-scale deformation field measurement from nanometers to micrometers and solve the problem of insufficient measurement accuracy when using only the TEM moire method.This method utilized the electron wave in the TEM passing through the atomic lattice of two layers of different materials to generate TEM moire.On this basis,the TEM was tuned to the STEM mode,and by adjusting parameters such as the amount of defocusing,magnification,scanning angle,etc.,the electron beam was focused on the position near the interface of the two layers of materials,and at the same time,the scanning line was made approximately parallel to the direction of one of the TEM moire fringes.The scanning secondary moire patterns were generated when the scanning spacing was close to the TEM moire spacing.Through this method,the deformation field,mechanical properties,and internal defects of crystals can be detected by a large field of view with high sensitivity and high efficiency.Compared to traditional methods,the advantages of scanning secondary moire method lie in significantly improving the measurement field of TEM moire and STEM moire methods,realizing the cross-scale visualization measurement from nanometers to micrometers,and possessing atomic-level displacement measurement sensitivity.It can also simplify and efficiently identify dislocations,offering a new method for large-area visualization observation of dislocation density in broad application prospects.展开更多
Objective:Deep learning(DL)has become the prevailing method in chest radiograph analysis,yet its performance heavily depends on large quantities of annotated images.To mitigate the cost,cold-start active learning(AL),...Objective:Deep learning(DL)has become the prevailing method in chest radiograph analysis,yet its performance heavily depends on large quantities of annotated images.To mitigate the cost,cold-start active learning(AL),comprising an initialization followed by subsequent learning,selects a small subset of informative data points for labeling.Recent advancements in pretrained models by supervised or self-supervised learning tailored to chest radiograph have shown broad applicability to diverse downstream tasks.However,their potential in cold-start AL remains unexplored.Methods:To validate the efficacy of domain-specific pretraining,we compared two foundation models:supervised TXRV and self-supervised REMEDIS with their general domain counterparts pretrained on ImageNet.Model performance was evaluated at both initialization and subsequent learning stages on two diagnostic tasks:psychiatric pneumonia and COVID-19.For initialization,we assessed their integration with three strategies:diversity,uncertainty,and hybrid sampling.For subsequent learning,we focused on uncertainty sampling powered by different pretrained models.We also conducted statistical tests to compare the foundation models with ImageNet counterparts,investigate the relationship between initialization and subsequent learning,examine the performance of one-shot initialization against the full AL process,and investigate the influence of class balance in initialization samples on initialization and subsequent learning.Results:First,domain-specific foundation models failed to outperform ImageNet counterparts in six out of eight experiments on informative sample selection.Both domain-specific and general pretrained models were unable to generate representations that could substitute for the original images as model inputs in seven of the eight scenarios.However,pretrained model-based initialization surpassed random sampling,the default approach in cold-start AL.Second,initialization performance was positively correlated with subsequent learning performance,highlighting the importance of initialization strategies.Third,one-shot initialization performed comparably to the full AL process,demonstrating the potential of reducing experts'repeated waiting during AL iterations.Last,a U-shaped correlation was observed between the class balance of initialization samples and model performance,suggesting that the class balance is more strongly associated with performance at middle budget levels than at low or high budgets.Conclusions:In this study,we highlighted the limitations of medical pretraining compared to general pretraining in the context of cold-start AL.We also identified promising outcomes related to cold-start AL,including initialization based on pretrained models,the positive influence of initialization on subsequent learning,the potential for one-shot initialization,and the influence of class balance on middle-budget AL.Researchers are encouraged to improve medical pretraining for versatile DL foundations and explore novel AL methods.展开更多
Background:Despite advances in surgical treatment,high recurrence after surgery remains a challenge for patients with hepatocellular carcinoma(HCC).This study aimed to investigate the association between compliance to...Background:Despite advances in surgical treatment,high recurrence after surgery remains a challenge for patients with hepatocellular carcinoma(HCC).This study aimed to investigate the association between compliance to regular follow-up and long-term oncological outcomes among patients undergoing curative resection for HCC.Methods:This multicenter study included patients who underwent curative resection for early-stage HCC between January 2012 and December 2021 at 12 liver surgery centers.Patients were stratified into a regular follow-up group(follow-up every 2–3 months for the first 2 years and every 3–6 months thereafter)and an irregular/no follow-up group.Overall survival(OS),time to recurrence(TTR),and post-recurrence survival(PRS)were compared between the two groups.Results:Among 1544 patients,786(50.9%)underwent regular follow-up during postoperative follow-up.The regular follow-up group had better OS(median:113.4 vs.94.5 months,P=0.010)and PRS(median:37.9 vs.16.3 months,P<0.001)than the irregular/no follow-up group,although TTR was comparable(median:61.4 vs.66.2 months,P=0.161).Furthermore,patients in the regular follow-up group had a lower incidence of tumor beyond the Milan criteria at recurrence(41.6%vs.50.4%,P=0.013)and were more likely to receive curative treatments for recurrence(56.1%vs.49.3%,P=0.061).On multivariate analysis,compliance to regular follow-up was an independent factor associated with better OS[hazard ratio(HR)=0.777,95%confidence interval(CI):0.663–0.910,P=0.002]and PRS(HR=0.523,95%CI:0.428–0.638,P<0.001).Conclusions:Compliance to regular follow-up improved OS and PRS after curative resection for HCC,highlighting the importance of postoperative regular follow-up for early detection of recurrence and timely intervention.展开更多
基金Supported by National Natural Science Foundation of China,No.82374400Heilongjiang Province“Double First-Class”New Round of Construction Disciplines Collaborative Innovation Achievements Construction Project,No.LJGXCG2022-097.
文摘Acute respiratory distress syndrome(ARDS)is a severe and life-threatening manifestation of acute lung injury,characterized by widespread pulmonary inflammation and edema,ultimately resulting in acute respiratory failure.Despite advancements in mechanical ventilation and lung-protective strategies,targeted therapies aimed at modulating dysregulated inflammation and promoting tissue repair remain elusive.Extracellular vesicles(EVs),critical mediators of intercellular communication,have emerged as a promising research focus due to their dual regulatory roles in ARDS pathogenesis.Pro-inflammatory EVs,derived from pathogens or injury-stressed cells,exacerbate alveolar macrophage activation and increase endothelial permeability,thereby aggravating pulmonary damage.In contrast,anti-inflammatory EVs originating from mesenchymal stem cells facilitate alveolar barrier restoration and tissue repair by delivering reparative molecular cargo.This review systematically evaluates the dualistic functions of EVs in ARDS from three key perspectives:Molecular mechanisms,clinical translation,and technical challenges.We further discuss the complexities associated with EV heterogeneity,pathogen interactions,and standardization in EV production.Additionally,we propose future directions that integrate engineered EV modifications and multi-omics approaches to address current therapeutic limitations and enhance ARDS management.
基金financially supported by Beijing Nova Program(Nos.Z211100002121127 and 20220484219)Beijing Natural Science Foundation(No.L212059)+1 种基金Fundamental Research Funds for the Central Universities(No.3332021101)CAMS Innovation Fund for Medical Sciences(CIFMS,Nos.2021-I2M-1-026 and 2021-I2M-1-028).
文摘Ulcerative colitis(UC)is a common progressive inflammatory disease whose incidence has increased rapidly in recent years,and can develop into colorectal cancer in severe cases.There are currently no adequate or effective treatments for UC due to the fact that some patients have found suboptimal results after repeated administration,while others have experienced adverse effects.With the rapid development of nanotechnology,developing innovative colon-targeting platforms is essential to improving efficacy,reducing side effects,and improving patient compliance.In this review,we summarize the pathophysiological characteristics of UC and the most recent status of numerous nanodrug delivery systems based on different targeting mechanisms in treating UC.Oral,intravenous,and rectal drug delivery nanoparticles targeting the colon are discussed,which can provide ideas for the design of colon-targeting nanoparticles for the treatment of colon diseases,especially for the treatment of UC.Last but not least,we provide a glimpse into the future of colon-targeted delivery systems,as well as future advancements in the field.
文摘目的观察超声引导下星状神经节阻滞(SGB)对乳腺癌患者乳腺癌相关淋巴水肿及术后恢复质量的影响。方法选择2022年10月至2023年10月择期全麻下行乳腺癌改良根治术的女性患者80例,年龄18~64岁,BMI 18.5~25.0 kg/m^(2),ASAⅠ或Ⅱ级。采用随机数字表法将患者随机分为两组:SGB组(S组)和对照组(C组),每组40例。S组术前行超声引导下单次右侧SGB;C组术前仅接受右侧星状神经节超声扫描不进行神经阻滞。记录术后3 d内乳腺癌相关淋巴水肿的发生情况。记录术前1 d和术后1、3 d睡眠时间,术后1、3 d 15项恢复质量评分(QoR-15)及Christensen疲劳评分。记录术后2、8、24 h VAS疼痛评分以及术后24 h内舒芬太尼用量、镇痛泵总按压次数及有效按压次数和补救镇痛例数。记录手术时间、拔管时间、PACU停留时间、术后住院时间、术后腹胀和恶心呕吐的发生情况以及术后神经阻滞相关并发症的发生情况。结果与C组比较,S组术后3 d内乳腺癌相关淋巴水肿发生率明显降低,术后1、3 d睡眠时间明显延长,术后1、3 d QoR-15评分明显升高,Christensen疲劳评分明显降低,术后24 h内舒芬太尼用量、镇疼泵总按压次数及有效按压次数明显减少,术后2、8、24 h VAS疼痛评分明显降低,拔管时间、PACU停留时间和术后住院时间明显缩短,术后腹胀和术后恶心呕吐发生率明显降低(P<0.05)。两组术后24 h内补救镇痛率差异无统计学意义。SGB组无一例发生神经阻滞相关并发症。结论在接受乳腺癌改良根治术的女性患者中术前行超声引导下单次SGB,可以降低乳腺癌相关淋巴水肿的发生率,提高术后恢复质量。
基金Supported by the China Health Promotion Foundation Young Doctors'Research Foundation for Inflammatory Bowel Disease,the Taishan Scholars Program of Shandong Province,China,No.tsqn202306343National Natural Science Foundation of China,No.82270578.
文摘BACKGROUND Inflammatory bowel disease(IBD)is a global health burden that affects millions of individuals worldwide,necessitating extensive patient education.Large language models(LLMs)hold promise for addressing patient information needs.However,LLM use to deliver accurate and comprehensible IBD-related medical information has yet to be thoroughly investigated.AIM To assess the utility of three LLMs(ChatGPT-4.0,Claude-3-Opus,and Gemini-1.5-Pro)as a reference point for patients with IBD.METHODS In this comparative study,two gastroenterology experts generated 15 IBD-related questions that reflected common patient concerns.These questions were used to evaluate the performance of the three LLMs.The answers provided by each model were independently assessed by three IBD-related medical experts using a Likert scale focusing on accuracy,comprehensibility,and correlation.Simultaneously,three patients were invited to evaluate the comprehensibility of their answers.Finally,a readability assessment was performed.RESULTS Overall,each of the LLMs achieved satisfactory levels of accuracy,comprehensibility,and completeness when answering IBD-related questions,although their performance varies.All of the investigated models demonstrated strengths in providing basic disease information such as IBD definition as well as its common symptoms and diagnostic methods.Nevertheless,when dealing with more complex medical advice,such as medication side effects,dietary adjustments,and complication risks,the quality of answers was inconsistent between the LLMs.Notably,Claude-3-Opus generated answers with better readability than the other two models.CONCLUSION LLMs have the potential as educational tools for patients with IBD;however,there are discrepancies between the models.Further optimization and the development of specialized models are necessary to ensure the accuracy and safety of the information provided.
基金supported by the Science and Technology Research Program of Chongqing Municipal Education Commission(KJQN202401501,KJZD-M202401501).
文摘The construction and operation of sulfur-containing gas storage are often more difficult than a non-sulfur storage facility due to the need to prevent environmental contamination from H_(2)S leaks,as well as the corrosive effects of H_(2)S on production facilities.Rapid elutriation of H_(2)S from the reservoir during the construction of the gas storage is an effective way to avoid these problems.However,the existing H_(2)S elutriation method has low efficiency and high economic cost,which limits the development of reconstructed gas storage of sulfur-containing gas reservoirs.To improve the efficiency of H_(2)S elutriation in sulfur-containing gas reservoirs and enhance the economic benefits,a numerical simulation model of multiphase flow components was established to study the migration law of H_(2)S in the multi-cycle operation of gas storage.Based on the H_(2)S migrate law,the displacement H_(2)S elutriation method was developed,and the elutriation mechanism and elutriation efficiency of the two methods were compared and analyzed.In addition,the main controlling factors affecting the H_(2)S elutriation efficiency were investigated,and the H_(2)S elutriation scheme of H gas storage was optimized.The results indicate that H_(2)S migrates between near-well and far-well regions under pressure differentials.The traditional H_(2)S elutriation method relies on concentration gradient diffusion,whereas the displacement elutriation approach leverages pressure differentials with higher H_(2)S elutriation efficiency.For the displacement elutriation method,higher reservoir permeability enhances the peak-shaving capacity of the gas storage but has a minor impact on H_(2)S elutriation when the formation permeability is between 30 and 100 mD.The elutriation efficiency is significantly higher when wells are drilled in the high structural parts of the reservoir compared to the low structural parts.Longer displacement elutriation time within a cycle improves H_(2)S elutriation efficiency but reduces the working gas volume of the storage.Therefore,the optimal displacement time for H gas storage is 60 days.An optimized H_(2)S elutriation scheme enabled the working gas to meet the national first-class natural gas standard within 10 cycles.This study elucidates H_(2)S migration patterns,H_(2)S elutriation mechanisms,and key influence factors on H_(2)S elutriation efficiency,offering valuable technical insights for sour gas storage operations.
基金Supported by China Health Promotion Foundation Young Doctors’Research Foundation for Inflammatory Bowel DiseaseTaishan Scholars Program of Shandong Province,China,NO.tsqn202306343National Natural Science Foundation of China,No.82270580,No.82070552,No.82270578,and No.82300599.
文摘BACKGROUND Gastrointestinal diseases have complex etiologies and clinical presentations.An accurate diagnosis requires physicians to integrate diverse information,including medical history,laboratory test results,and imaging findings.Existing artificial intelligence-assisted diagnostic tools are limited to single-modality information,resulting in recommendations that are often incomplete and may be associated with clinical or legal risks.AIM To develop and evaluate a collaborative multimodal large language model(LLM)framework for clinical decision-making in digestive diseases.METHODS In this observational study,DeepGut,a multimodal LLM collaborative diagnostic framework,was developed to integrate four distinct large models into a four-tiered structure.The framework sequentially accomplishes multimodal infor-mation extraction,logical“chain”construction,diagnostic and treatment suggestion generation,and risk analysis.The model was evaluated using objective metrics,which assess the reliability and comprehensiveness of model-generated results,and subjective expert opinions,which examine the effectiveness of the framework in assisting physicians.RESULTS The diagnostic and treatment recommendations generated by the DeepGut framework achieved exceptional performance,with a diagnostic accuracy of 97.8%,diagnostic completeness of 93.9%,treatment plan accuracy of 95.2%,and treatment plan completeness of 98.0%,significantly surpassing the capabilities of single-modal LLM-based diagnostic tools.Experts evaluating the framework commended the completeness,relevance,and logical coherence of its outputs.However,the collaborative multimodal LLM approach resulted in increased input and output token counts,leading to higher computational costs and extended diagnostic times.CONCLUSION The framework achieves successful integration of multimodal diagnostic data,demonstrating enhanced performance enabled by multimodal LLM collaboration,which opens new horizons for the clinical application of artificial intelligence-assisted technology.
基金financially supported by the National Natural Science Foundation of China(Nos.52175444,51905506,21871065 and 22071038)the Sichuan Science and Technology Program(No.2021JDJQ0014).
文摘Additive manufacturing(AM)methods have garnered considerable attention owing to their flexibility in fabricating complex parts with desirable mechanical properties.However,the poor surface quality of the resulting metal parts remains a severe challenge for the applications.Here,a novel dual-additive synergy strategy is presented,which simultaneously enhances material removal efficiency and regulates electrode surface reactions during electrochemical polishing(ECP)of AM AlSi10Mg.Theoretical studies and experimental characterizations confirm that NaF promotes selective dissolution at the peaks,while glucose acts as a stabilizer for the surface valleys.This approach effectively facilitates the selective removal of surface protrusions,achieving a smoother and more uniform surface finish,resulting in a surface roughness reduction of approximately 86%,compared to a 63%reduction without additives.This study not only provides a new approach for optimizing surface quality of AM AlSi10Mg but also offers new insights into electrolyte design and the stabilization of metal anodes.
基金supported by the National Natural Science Foundation of China(Grant Nos.12372178 and 12327801).
文摘The measurement field of view of the conventional transmission electron microscopy(TEM)nano-moiréand scanning transmission electron microscopy(STEM)nano-moirémethods is limited to the hundred-nanometer scale,unable to meet the deformation field measurement requirements of micrometer-scale materials such as transistors and micro-devices.This paper proposed a novel measurement method based on scanning secondary moire,which can realize cross-scale deformation field measurement from nanometers to micrometers and solve the problem of insufficient measurement accuracy when using only the TEM moire method.This method utilized the electron wave in the TEM passing through the atomic lattice of two layers of different materials to generate TEM moire.On this basis,the TEM was tuned to the STEM mode,and by adjusting parameters such as the amount of defocusing,magnification,scanning angle,etc.,the electron beam was focused on the position near the interface of the two layers of materials,and at the same time,the scanning line was made approximately parallel to the direction of one of the TEM moire fringes.The scanning secondary moire patterns were generated when the scanning spacing was close to the TEM moire spacing.Through this method,the deformation field,mechanical properties,and internal defects of crystals can be detected by a large field of view with high sensitivity and high efficiency.Compared to traditional methods,the advantages of scanning secondary moire method lie in significantly improving the measurement field of TEM moire and STEM moire methods,realizing the cross-scale visualization measurement from nanometers to micrometers,and possessing atomic-level displacement measurement sensitivity.It can also simplify and efficiently identify dislocations,offering a new method for large-area visualization observation of dislocation density in broad application prospects.
文摘Objective:Deep learning(DL)has become the prevailing method in chest radiograph analysis,yet its performance heavily depends on large quantities of annotated images.To mitigate the cost,cold-start active learning(AL),comprising an initialization followed by subsequent learning,selects a small subset of informative data points for labeling.Recent advancements in pretrained models by supervised or self-supervised learning tailored to chest radiograph have shown broad applicability to diverse downstream tasks.However,their potential in cold-start AL remains unexplored.Methods:To validate the efficacy of domain-specific pretraining,we compared two foundation models:supervised TXRV and self-supervised REMEDIS with their general domain counterparts pretrained on ImageNet.Model performance was evaluated at both initialization and subsequent learning stages on two diagnostic tasks:psychiatric pneumonia and COVID-19.For initialization,we assessed their integration with three strategies:diversity,uncertainty,and hybrid sampling.For subsequent learning,we focused on uncertainty sampling powered by different pretrained models.We also conducted statistical tests to compare the foundation models with ImageNet counterparts,investigate the relationship between initialization and subsequent learning,examine the performance of one-shot initialization against the full AL process,and investigate the influence of class balance in initialization samples on initialization and subsequent learning.Results:First,domain-specific foundation models failed to outperform ImageNet counterparts in six out of eight experiments on informative sample selection.Both domain-specific and general pretrained models were unable to generate representations that could substitute for the original images as model inputs in seven of the eight scenarios.However,pretrained model-based initialization surpassed random sampling,the default approach in cold-start AL.Second,initialization performance was positively correlated with subsequent learning performance,highlighting the importance of initialization strategies.Third,one-shot initialization performed comparably to the full AL process,demonstrating the potential of reducing experts'repeated waiting during AL iterations.Last,a U-shaped correlation was observed between the class balance of initialization samples and model performance,suggesting that the class balance is more strongly associated with performance at middle budget levels than at low or high budgets.Conclusions:In this study,we highlighted the limitations of medical pretraining compared to general pretraining in the context of cold-start AL.We also identified promising outcomes related to cold-start AL,including initialization based on pretrained models,the positive influence of initialization on subsequent learning,the potential for one-shot initialization,and the influence of class balance on middle-budget AL.Researchers are encouraged to improve medical pretraining for versatile DL foundations and explore novel AL methods.
基金This study was supported by grants from the National Natural Science Foundation of China(82425049,81972726 and 82273074)Dawn Project Foundation of Shanghai(21SG36)+4 种基金Shanghai Health and Hygiene Discipline Leader Project(2022XD001)Shanghai Out-standing Academic Leader Program(23XD1424900)the Natural Science Foundation of Shanghai(22ZR1477900)Shanghai Science and Technology Committee Rising-Star Program(22QA1411600)the Special Clinical Project of Shanghai Municipal Health Com-mission(20244Y0233)。
文摘Background:Despite advances in surgical treatment,high recurrence after surgery remains a challenge for patients with hepatocellular carcinoma(HCC).This study aimed to investigate the association between compliance to regular follow-up and long-term oncological outcomes among patients undergoing curative resection for HCC.Methods:This multicenter study included patients who underwent curative resection for early-stage HCC between January 2012 and December 2021 at 12 liver surgery centers.Patients were stratified into a regular follow-up group(follow-up every 2–3 months for the first 2 years and every 3–6 months thereafter)and an irregular/no follow-up group.Overall survival(OS),time to recurrence(TTR),and post-recurrence survival(PRS)were compared between the two groups.Results:Among 1544 patients,786(50.9%)underwent regular follow-up during postoperative follow-up.The regular follow-up group had better OS(median:113.4 vs.94.5 months,P=0.010)and PRS(median:37.9 vs.16.3 months,P<0.001)than the irregular/no follow-up group,although TTR was comparable(median:61.4 vs.66.2 months,P=0.161).Furthermore,patients in the regular follow-up group had a lower incidence of tumor beyond the Milan criteria at recurrence(41.6%vs.50.4%,P=0.013)and were more likely to receive curative treatments for recurrence(56.1%vs.49.3%,P=0.061).On multivariate analysis,compliance to regular follow-up was an independent factor associated with better OS[hazard ratio(HR)=0.777,95%confidence interval(CI):0.663–0.910,P=0.002]and PRS(HR=0.523,95%CI:0.428–0.638,P<0.001).Conclusions:Compliance to regular follow-up improved OS and PRS after curative resection for HCC,highlighting the importance of postoperative regular follow-up for early detection of recurrence and timely intervention.
