BMC Medical Informatics and Decision Making最新文献

{"title":"Deep learning-based automated guide for defining a standard imaging plane for developmental dysplasia of the hip screening using ultrasonography: a retrospective imaging analysis.","authors":"Kyung-Sik Ahn, Ji Hye Choi, Heejou Kwon, Seoyeon Lee, Yongwon Cho, Woo Young Jang","doi":"10.1186/s12911-025-02926-8","DOIUrl":"10.1186/s12911-025-02926-8","url":null,"abstract":"<p><strong>Background: </strong>We aimed to propose a deep-learning neural network model for automatically detecting five landmarks during a two-dimensional (2D) ultrasonography (US) scan to develop a standard plane for developmental dysplasia of the hip (DDH) screening.</p><p><strong>Method: </strong>A model of global and local networks was developed to detect five landmarks for DDH screening during 2D US. Patients (N = 532) who underwent hip US for DDH screening from January 2016 to December 2021 at a tertiary medical center were enrolled. All datasets were randomly split into training, validation, and test sets in a 70:10:20 ratio for the final assessment of landmark detection. The performance of this model for detecting five landmarks for guiding DDH was analyzed using the root mean square error (RMSE) and dice similarity coefficient.</p><p><strong>Results: </strong>The RMSE value for the five landmarks for diagnosing and classifying DDH using global and local networks was 4.023 ± 3.723. The point results using EfficientNetB2 were 1.69 ± 1.26 (first point), 3.34 ± 2.37 (second point), 2.54 ± 1.61 (third point), 5.92 ± 4.25 (fourth point), and 6.61 ± 4.82 (fifth point).</p><p><strong>Conclusions: </strong>Our deep-learning network model is feasible for detecting five landmarks for DDH using ultrasound images. The primary parameters to determine DDH will be significantly detected by applying the deep-learning model in clinical settings.</p>","PeriodicalId":9340,"journal":{"name":"BMC Medical Informatics and Decision Making","volume":"25 1","pages":"91"},"PeriodicalIF":3.3,"publicationDate":"2025-02-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11837398/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143448171","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Uncovering the potential of smartphones for behavior monitoring during migraine follow-up.","authors":"Marija Stojchevska, Jonas Van Der Donckt, Nicolas Vandenbussche, Mathias De Brouwer, Koen Paemeleire, Femke Ongenae, Sofie Van Hoecke","doi":"10.1186/s12911-025-02916-w","DOIUrl":"10.1186/s12911-025-02916-w","url":null,"abstract":"<p><strong>Background: </strong>Migraine is a neurological disorder that affects millions of people worldwide. It is one of the most debilitating disorders which leads to many disability-adjusted life years. Conventional methods for investigating migraines, like patient interviews and diaries, suffer from self-reporting biases and intermittent tracking.</p><p><strong>Methods: </strong>This study aims to leverage smartphone-derived data as an objective tool for examining the relationship between migraines and various human behavior aspects. By utilizing built-in sensors and monitoring phone interactions, we gather data from which we derive metrics such as keyboard usage, application interaction, physical activity levels, ambient light conditions, and sleep patterns. We perform statistical analysis testing to investigate whether there is a difference in user behavioral aspects during headache and non-headache periods.</p><p><strong>Results: </strong>Our analysis of 362 headaches reveals differences in behavioral aspects such as ambient light, use of leisure apps, and number of keystrokes during headache periods and non-headache periods.</p><p><strong>Conclusions: </strong>This exploratory study shows on the one hand that it is possible to monitor various human behavioral aspects using the smartphone sensors and interaction data only. On the other hand it shows that we can observe difference in human behavior between headache and non-headache periods. Our work is a step towards objectively measure the effects that migraine has on people's lives.</p>","PeriodicalId":9340,"journal":{"name":"BMC Medical Informatics and Decision Making","volume":"25 1","pages":"88"},"PeriodicalIF":3.3,"publicationDate":"2025-02-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11837657/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143448176","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Workload in antenatal care before and after implementation of an electronic decision support system: an observed time-motion study of healthcare providers in Nepal.","authors":"Emma Radovich, Seema Das, Sulata Karki, Christian Bottomley, Ona L McCarthy, Abha Shrestha, Loveday Penn-Kekana, Rajani Shakya, Biraj Man Karmacharya, Abha Shrestha, Oona M R Campbell, Giorgia Gon","doi":"10.1186/s12911-025-02868-1","DOIUrl":"10.1186/s12911-025-02868-1","url":null,"abstract":"<p><strong>Background: </strong>Healthcare interventions are shaped by the resources needed to implement them, including staff time. This study, part of a process evaluation, aims to compare time spent on antenatal care (ANC) and related recordkeeping in two rural primary-level health facilities in Nepal, before and after implementation of an electronic decision support system intervention to improve ANC quality that required additional electronic documentation.</p><p><strong>Methods: </strong>The study is a before-and-after, observational time-motion assessment. Researchers used the WOMBAT (Work Observation Method By Activity Timing) software to observe and record activities performed by auxiliary nurse midwives providing ANC in two rounds of data collection. We summed the observation time (in minutes) spent on activity categories for each day of observation, in each round of data collection. For each auxiliary nurse midwife, we estimated the proportion of total observation time spent on activities and compared these proportions before and after intervention implementation. We also compared the mean minutes per day spent on ANC and recordkeeping in the two rounds.</p><p><strong>Results: </strong>Six auxiliary nurse midwives were observed over two data collection rounds (41 total observation days). Prior to intervention, providers spent 7% of their workday on ANC and 6% on related recordkeeping, and time spent on these activities did not change after intervention implementation. Only one of the six auxiliary nurse midwives demonstrated a statistically significant increase in time spent on ANC and recordkeeping after implementation. There was considerable day-to-day variation in ANC time, and substantial periods of \"non-work\" time (on break or not engaged in work-related activity). Non-work time reduced from 42% in the first round to 26% in the second round of data collection.</p><p><strong>Conclusions: </strong>Time spent on ANC and related recordkeeping was low and did not change after implementation of the electronic decision support system. ANC and recordkeeping time was sensitive to day-to-day fluctuations in numbers of women attending for ANC at these rural facilities, which may have masked the intervention's effects. However, the large amount of non-work time observed suggests time constraints during the workday were not a major factor inhibiting use of the electronic decision support system.</p>","PeriodicalId":9340,"journal":{"name":"BMC Medical Informatics and Decision Making","volume":"25 1","pages":"87"},"PeriodicalIF":3.3,"publicationDate":"2025-02-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11834578/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143448178","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"How good is your synthetic data? SynthRO, a dashboard to evaluate and benchmark synthetic tabular data.","authors":"Gabriele Santangelo, Giovanna Nicora, Riccardo Bellazzi, Arianna Dagliati","doi":"10.1186/s12911-024-02731-9","DOIUrl":"10.1186/s12911-024-02731-9","url":null,"abstract":"<p><strong>Background: </strong>The exponential growth in patient data collection by healthcare providers, governments, and private industries is yielding large and diverse datasets that offer new insights into critical medical questions. Leveraging extensive computational resources, Machine Learning and Artificial Intelligence are increasingly utilized to address health-related issues, such as predicting outcomes from Electronic Health Records and detecting patterns in multi-omics data. Despite the proliferation of medical devices based on Artificial Intelligence, data accessibility for research is limited due to privacy concerns. Efforts to de-identify data have met challenges in maintaining effectiveness, particularly with large datasets. As an alternative, synthetic data, that replicate main statistical properties of real patient data, are proposed. However, the lack of standardized evaluation metrics complicates the selection of appropriate synthetic data generation methods. Effective evaluation of synthetic data must consider resemblance, utility and privacy, tailored to specific applications. Despite available metrics, benchmarking efforts remain limited, necessitating further research in this area.</p><p><strong>Results: </strong>We present SynthRO (Synthetic data Rank and Order), a user-friendly tool for benchmarking health synthetic tabular data across various contexts. SynthRO offers accessible quality evaluation metrics and automated benchmarking, helping users determine the most suitable synthetic data models for specific use cases by prioritizing metrics and providing consistent quantitative scores. Our dashboard is divided into three main sections: (1) Loading Data section, where users can locally upload real and synthetic datasets; (2) Evaluation section, in which several quality assessments are performed by computing different metrics and measures; (3) Benchmarking section, where users can globally compare synthetic datasets based on quality evaluation.</p><p><strong>Conclusions: </strong>Synthetic data mitigate concerns about privacy and data accessibility, yet lacks standardized evaluation metrics. SynthRO provides an accessible dashboard helping users select suitable synthetic data models, and it also supports various use cases in healthcare, enhancing prognostic scores and enabling federated learning. SynthRO's accessible GUI and modular structure facilitate effective data evaluation, promoting reliability and fairness. Future developments will include temporal data evaluation, further broadening its applicability.</p>","PeriodicalId":9340,"journal":{"name":"BMC Medical Informatics and Decision Making","volume":"25 1","pages":"89"},"PeriodicalIF":3.3,"publicationDate":"2025-02-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11837667/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143448175","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A series of natural language processing for predicting tumor response evaluation and survival curve from electronic health records.","authors":"Toshiki Takeuchi, Hidehito Horinouchi, Ken Takasawa, Masami Mukai, Ken Masuda, Yuki Shinno, Yusuke Okuma, Tatsuya Yoshida, Yasushi Goto, Noboru Yamamoto, Yuichiro Ohe, Mototaka Miyake, Hirokazu Watanabe, Masahiko Kusumoto, Takashi Aoki, Kunihiro Nishimura, Ryuji Hamamoto","doi":"10.1186/s12911-025-02928-6","DOIUrl":"10.1186/s12911-025-02928-6","url":null,"abstract":"<p><strong>Background: </strong>The clinical information housed within unstructured electronic health records (EHRs) has the potential to promote cancer research. The National Cancer Center Hospital (NCCH) is widely recognized as a leading institution for the treatment of thoracic malignancies in Japan. Information on medical treatment, particularly the characteristics of malignant tumors that occur in patients, tumor response evaluation, and adverse events, was compiled into the databases of each NCCH department from EHRs. However, there have been few opportunities for integrated analysis of data on both the hospital and research institute.</p><p><strong>Methods: </strong>We developed a method for predicting tumor response evaluation and survival curves of drug therapy from the EHRs of lung cancer patients using natural language processing. First, we developed a rule-based algorithm to predict treatment duration using a dictionary of anticancer drugs and regimens used for lung cancer treatment. Thereafter, we applied supervised learning to radiology reports during each treatment period and constructed a classification model to predict the tumor response evaluation of anticancer drugs and date when the progressive disease (PD) was determined. The predicted response and PD date can be used to draw a survival curve for the progression-free survival.</p><p><strong>Results: </strong>We used the EHRs of 716 lung cancer treatments at the NCCH and structured data of the cases as labels for the training and testing of supervised learning. The structured data were manually curated by physicians and CRCs. We investigated the results and performance of the proposed method. Individual predictions of tumor response evaluation and PD date were not extremely high. However, the final predicted survival curves were nearly similar to the actual survival curves.</p><p><strong>Conclusions: </strong>Although it is difficult to construct a fully automated system using our method, we believe that it achieves sufficient performance for supporting physicians and CRCs constructing the database and providing clinical information to help researchers find out a chance of clinical studies.</p>","PeriodicalId":9340,"journal":{"name":"BMC Medical Informatics and Decision Making","volume":"25 1","pages":"85"},"PeriodicalIF":3.3,"publicationDate":"2025-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11834625/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143440156","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"De-identification of clinical notes with pseudo-labeling using regular expression rules and pre-trained BERT.","authors":"Jiyong An, Jiyun Kim, Leonard Sunwoo, Hyunyoung Baek, Sooyoung Yoo, Seunggeun Lee","doi":"10.1186/s12911-025-02913-z","DOIUrl":"10.1186/s12911-025-02913-z","url":null,"abstract":"<p><strong>Background: </strong>De-identification of clinical notes is essential to utilize the rich information in unstructured text data in medical research. However, only limited work has been done in removing personal information from clinical notes in Korea.</p><p><strong>Methods: </strong>Our study utilized a comprehensive dataset stored in the Note table of the OMOP Common Data Model at Seoul National University Bundang Hospital. This dataset includes 11,181,617 radiology and 9,282,477 notes from various other departments (non-radiology reports). From this, 0.1% of the reports (11,182) were randomly selected for training and validation purposes. We used two de-identification strategies to improve performance with limited and few annotated data. First, a rule-based approach is used to construct regular expressions on the 1,112 notes annotated by domain experts. Second, by using the regular expressions as label-er, we applied a semi-supervised approach to fine-tune a pre-trained Korean BERT model with pseudo-labeled notes.</p><p><strong>Results: </strong>Validation was conducted using 342 radiology and 12 non-radiology notes labeled at the token level. Our rule-based approach achieved 97.2% precision, 93.7% recall, and 96.2% F1 score from the department of radiology notes. For machine learning approach, KoBERT-NER that is fine-tuned with 32,000 automatically pseudo-labeled notes achieved 96.5% precision, 97.6% recall, and 97.1% F1 score.</p><p><strong>Conclusion: </strong>By combining a rule-based approach and machine learning in a semi-supervised way, our results show that the performance of de-identification can be improved.</p>","PeriodicalId":9340,"journal":{"name":"BMC Medical Informatics and Decision Making","volume":"25 1","pages":"82"},"PeriodicalIF":3.3,"publicationDate":"2025-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11831849/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143440142","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Noise is an underrecognized problem in medical decision making and is known by other names: a scoping review.","authors":"Kayla V Dlugos, Mjaye Mazwi, Robert Lao, Osami Honjo","doi":"10.1186/s12911-025-02905-z","DOIUrl":"10.1186/s12911-025-02905-z","url":null,"abstract":"<p><p>Unwanted random variability in day-to-day decision making referred to as 'noise' is associated with unhelpful variation that affects both the reproducibility and quality of decision making. Although this is described in other fields, the prevalence of noise in medical decision making and its effects on patient outcomes and the process and efficiency of care have not been reported and are unknown. This review sought to explore noise as a feature of medical decision making, as well as explore potential sources of noise in this setting. The search generated 2,082 results. Analysis of 14 studies included in the review (11 PubMed, 3 reference mining) suggests noise is a driver of unhelpful practice variation and may have important effects on care efficiency and reproducibility. 7 of the 14 studies demonstrated pattern noise, 3 demonstrated occasion noise, and 5 demonstrated stable pattern noise. The decision making in 8 studies demonstrated level noise, and lastly the decision making in 4 of the studies demonstrated system noise, a combination of both pattern and level noise. Additional study is required to ascertain how to measure and mitigate noise in medical decision making, as well as better understand the sources of noise present. Clinical trial number not applicable.</p>","PeriodicalId":9340,"journal":{"name":"BMC Medical Informatics and Decision Making","volume":"25 1","pages":"86"},"PeriodicalIF":3.3,"publicationDate":"2025-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11834221/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143440239","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"An effective multi-step feature selection framework for clinical outcome prediction using electronic medical records.","authors":"Hongnian Wang, Mingyang Zhang, Liyi Mai, Xin Li, Abdelouahab Bellou, Lijuan Wu","doi":"10.1186/s12911-025-02922-y","DOIUrl":"10.1186/s12911-025-02922-y","url":null,"abstract":"<p><strong>Background: </strong>Identifying key variables is essential for developing clinical outcome prediction models based on high-dimensional electronic medical records (EMR). However, despite the abundance of feature selection (FS) methods available, challenges remain in choosing the most appropriate method, deciding how many top-ranked variables to include, and ensuring these selections are meaningful from a medical perspective.</p><p><strong>Methods: </strong>We developed a practical multi-step feature selection (FS) framework that integrates data-driven statistical inference with a knowledge verification strategy. This framework was validated using two distinct EMR datasets targeting different clinical outcomes. The first cohort, sourced from the Medical Information Mart for Intensive Care III (MIMIC-III), focused on predicting acute kidney injury (AKI) in ICU patients. The second cohort, drawn from the MIMIC-IV Emergency Department (MIMIC-IV-ED), aimed to estimate in-hospital mortality (IHM) for patients transferred from the ED to the ICU. We employed various machine learning (ML) methods and conducted a comparative analysis considering accuracy, stability, similarity, and interpretability. The effectiveness of our FS framework was evaluated using discrimination and calibration metrics, with SHAP applied to enhance the interpretability of model decisions.</p><p><strong>Results: </strong>Cohort 1 comprised 48,780 ICU encounters, of which 8,883 (18.21%) developed AKI. Cohort 2 included 29,197 transfers from the ED to the ICU, with 3,219 (11.03%) resulting in IHM. Among the ten ML methods evaluated, the tree-based ensemble method achieved the highest accuracy. As the number of top-ranking features increased, the models' accuracy began to stabilize, while feature subset stability (considering sample variations) and inter-method feature similarity reached optimal levels, confirming the validity of the FS framework. The integration of interpretative methods and expert knowledge in the final step further improved feature interpretability. The FS framework effectively reduced the number of features (e.g., from 380 to 35 for Cohort 1, and from 273 to 54 for Cohort 2) without significantly affecting prediction performance (Delong test, p > 0.05).</p><p><strong>Conclusion: </strong>The multi-step FS method developed in this study successfully reduces the dimensionality of features in EMR while preserving the accuracy of clinical outcome prediction. Furthermore, it improves the interpretability of risk factors by incorporating expert knowledge validation.</p>","PeriodicalId":9340,"journal":{"name":"BMC Medical Informatics and Decision Making","volume":"25 1","pages":"84"},"PeriodicalIF":3.3,"publicationDate":"2025-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11834488/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143440134","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Prediction of depressive disorder using machine learning approaches: findings from the NHANES.","authors":"Thien Vu, Research Dawadi, Masaki Yamamoto, Jie Ting Tay, Naoki Watanabe, Yuki Kuriya, Ai Oya, Phap Ngoc Hoang Tran, Michihiro Araki","doi":"10.1186/s12911-025-02903-1","DOIUrl":"10.1186/s12911-025-02903-1","url":null,"abstract":"<p><strong>Background: </strong>Depressive disorder, particularly major depressive disorder (MDD), significantly impact individuals and society. Traditional analysis methods often suffer from subjectivity and may not capture complex, non-linear relationships between risk factors. Machine learning (ML) offers a data-driven approach to predict and diagnose depression more accurately by analyzing large and complex datasets.</p><p><strong>Methods: </strong>This study utilized data from the National Health and Nutrition Examination Survey (NHANES) 2013-2014 to predict depression using six supervised ML models: Logistic Regression, Random Forest, Naive Bayes, Support Vector Machine (SVM), Extreme Gradient Boost (XGBoost), and Light Gradient Boosting Machine (LightGBM). Depression was assessed using the Patient Health Questionnaire (PHQ-9), with a score of 10 or higher indicating moderate to severe depression. The dataset was split into training and testing sets (80% and 20%, respectively), and model performance was evaluated using accuracy, sensitivity, specificity, precision, AUC, and F1 score. SHAP (SHapley Additive exPlanations) values were used to identify the critical risk factors and interpret the contributions of each feature to the prediction.</p><p><strong>Results: </strong>XGBoost was identified as the best-performing model, achieving the highest accuracy, sensitivity, specificity, precision, AUC, and F1 score. SHAP analysis highlighted the most significant predictors of depression: the ratio family income to poverty (PIR), sex, hypertension, serum cotinine and hydroxycotine, BMI, education level, glucose levels, age, marital status, and renal function (eGFR).</p><p><strong>Conclusion: </strong>We developed ML models to predict depression and utilized SHAP for interpretation. This approach identifies key factors associated with depression, encompassing socioeconomic, demographic, and health-related aspects.</p>","PeriodicalId":9340,"journal":{"name":"BMC Medical Informatics and Decision Making","volume":"25 1","pages":"83"},"PeriodicalIF":3.3,"publicationDate":"2025-02-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11834192/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143440256","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Machine learning via DARTS-Optimized MobileViT models for pancreatic Cancer diagnosis with graph-based deep learning.","authors":"Yusuf Alaca","doi":"10.1186/s12911-025-02923-x","DOIUrl":"10.1186/s12911-025-02923-x","url":null,"abstract":"<p><p>The diagnosis of pancreatic cancer presents a significant challenge due to the asymptomatic nature of the disease and the fact that it is frequently detected at an advanced stage. This study presents a novel approach combining graph-based data representation with DARTS-optimised MobileViT models, with the objective of enhancing diagnostic accuracy and reliability. The images of the pancreatic CT were transformed into graph structures using the Harris Corner Detection algorithm, which enables the capture of complex spatial relationships. Subsequently, the graph representations were processed using MobileViT models that had been optimised with Differentiable Architecture Search (DARTS), thereby enabling dynamic architectural adaptation. To further enhance classification accuracy, advanced machine learning algorithms, including K-Nearest Neighbours (KNN), Support Vector Machines (SVM), Random Forest (RF), and XGBoost, were applied. The MobileViTv2_150 and MobileViTv2_200 models demonstrated remarkable performance, with an accuracy of 97.33% and an F1 score of 96.25%, surpassing the capabilities of traditional CNN and Vision Transformer models. This innovative integration of graph-based deep learning and machine learning techniques demonstrates the potential of the proposed method to establish a new standard for early pancreatic cancer diagnosis. Furthermore, the study highlights the scalability of this approach for broader applications in medical imaging, which could lead to improved patient outcomes.</p>","PeriodicalId":9340,"journal":{"name":"BMC Medical Informatics and Decision Making","volume":"25 1","pages":"81"},"PeriodicalIF":3.3,"publicationDate":"2025-02-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11830204/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143424873","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}