{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,2]],"date-time":"2026-08-02T23:46:50Z","timestamp":1785714410000,"version":"3.56.0"},"reference-count":57,"publisher":"Public Library of Science (PLoS)","issue":"7","license":[{"start":{"date-parts":[[2025,7,8]],"date-time":"2025-07-08T00:00:00Z","timestamp":1751932800000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100001659","name":"Deutsche Forschungsgemeinschaft","doi-asserted-by":"publisher","award":["HE 8077\/2-1, SA 465\/53-1"],"award-info":[{"award-number":["HE 8077\/2-1, SA 465\/53-1"]}],"id":[{"id":"10.13039\/501100001659","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":["www.plosone.org"],"crossmark-restriction":false},"short-container-title":["PLoS One"],"abstract":"<jats:sec id=\"sec015\">\n                    <jats:title>Purpose<\/jats:title>\n                    <jats:p>Machine learning is a powerful tool to develop algorithms for clinical diagnosis. However, standard machine learning algorithms are not perfectly suited for clinical data since the data are interconnected and may contain time series. As shown for recommender systems and molecular property predictions, Graph Neural Networks (GNNs) may represent a powerful alternative to exploit the inherently graph-based properties of clinical data. The main goal of this study is to evaluate when GNNs represent a valuable alternative for analyzing large clinical data from the clinical routine on the example of Complete Blood Count Data.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec id=\"sec016\">\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>In this study, we evaluated the performance and time consumption of several GNNs (e.g., Graph Attention Networks) on similarity graphs compared to simpler, state-of-the-art machine learning algorithms (e.g., XGBoost) on the classification of sepsis from blood count data as well as the importance and slope of each feature for the final classification. Additionally, we connected complete blood count samples of the same patient based on their measured time (patient-centric graphs) to incorporate time series information in the GNNs. As our main evaluation metric, we used the Area Under Receiver Operating Curve (AUROC) to have a threshold independent metric that can handle class imbalance.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec id=\"sec017\">\n                    <jats:title>Results and Conclusion<\/jats:title>\n                    <jats:p>Standard GNNs on evaluated similarity-graphs achieved an Area Under Receiver Operating Curve (AUROC) of up to 0.8747 comparable to the performance of ensemble-based machine learning algorithms and a neural network. However, our integration of time series information using patient-centric graphs with GNNs achieved a superior AUROC of up to 0.9565. Finally, we discovered that feature slope and importance highly differ between trained algorithms (e.g., XGBoost and GNN) on the same data basis.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.1371\/journal.pone.0327636","type":"journal-article","created":{"date-parts":[[2025,7,8]],"date-time":"2025-07-08T13:24:06Z","timestamp":1751981046000},"page":"e0327636","update-policy":"https:\/\/doi.org\/10.1371\/journal.pone.corrections_policy","source":"Crossref","is-referenced-by-count":4,"title":["Edges are all you need: Potential of medical time series analysis on complete blood count data with graph neural networks"],"prefix":"10.1371","volume":"20","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-5094-3566","authenticated-orcid":true,"given":"Daniel","family":"Walke","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2364-598X","authenticated-orcid":true,"given":"Daniel","family":"Steinbach","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7406-4443","authenticated-orcid":true,"given":"Sebastian","family":"Gibb","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Thorsten","family":"Kaiser","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Gunter","family":"Saake","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Paul C.","family":"Ahrens","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-9580-740X","authenticated-orcid":true,"given":"David","family":"Broneske","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Robert","family":"Heyer","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"340","published-online":{"date-parts":[[2025,7,8]]},"reference":[{"issue":"7","key":"pone.0327636.ref001","doi-asserted-by":"crossref","first-page":"4514","DOI":"10.1016\/j.jksuci.2020.10.013","article-title":"An optimized XGBoost based diagnostic system for effective prediction of heart disease","volume":"34","author":"K Budholiya","year":"2022","journal-title":"J King Saud Univ comput inf Sci"},{"issue":"6","key":"pone.0327636.ref002","doi-asserted-by":"crossref","first-page":"2131","DOI":"10.1109\/TCBB.2019.2911071","article-title":"XGBoost model for chronic kidney disease diagnosis","volume":"17","author":"A Ogunleye","year":"2020","journal-title":"IEEE\/ACM Trans Comput Biol Bioinform"},{"key":"pone.0327636.ref003","doi-asserted-by":"crossref","unstructured":"Chen T, Guestrin C. XGBoost. In: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York, NY, USA: ACM. 2016. 785\u201394. doi: 10.1145\/2939672.2939785","DOI":"10.1145\/2939672.2939785"},{"key":"pone.0327636.ref004","article-title":"The importance of graph databases and graph learning for clinical applications","author":"D Walke","year":"2023","journal-title":"Database (Oxford)"},{"issue":"1","key":"pone.0327636.ref005","doi-asserted-by":"crossref","first-page":"5","DOI":"10.1023\/A:1010933404324","article-title":"Random forests","volume":"45","author":"L Breiman","year":"2001","journal-title":"Machine Learning"},{"key":"pone.0327636.ref006","volume-title":"Introduction to graph theory","author":"RJ Wilson","year":"1985"},{"key":"pone.0327636.ref007","unstructured":"Hamilton WL, Ying R, Leskovec J. Inductive Representation Learning on Large Graphs.; 07.06.2017 NIPS\u201917: Proceedings of the 31st International Conference on Neural Information Processing Systems. 2017. doi: 10.5555\/3294771.3294869"},{"key":"pone.0327636.ref008","article-title":"Graph Attention Networks.; 30.10.2017","volume-title":"International Conference on Learning Representations (ICLR","author":"P Veli\u010dkovi\u0107","year":"2018"},{"key":"pone.0327636.ref009","unstructured":"Kipf TN, Welling M. Semi-Supervised Classification with Graph Convolutional Networks.; 09.09.2016. In: International Conference on Learning Representations (ICLR). 2017."},{"key":"pone.0327636.ref010","unstructured":"Xu K, Hu W, Leskovec J, Jegelka S. How Powerful are Graph Neural Networks.; 01.10.2018. In: International Conference on Learning Representations (ICLR). 2019."},{"key":"pone.0327636.ref011","doi-asserted-by":"crossref","first-page":"2022","DOI":"10.1145\/3308558.3313562","article-title":"Heterogeneous Graph Attention Network.","volume-title":"The World Wide Web Conference","author":"X Wang","year":"2019"},{"key":"pone.0327636.ref012","doi-asserted-by":"crossref","unstructured":"Hu Z, Dong Y, Wang K, Sun Y. Heterogeneous Graph Transformer. In: Proceedings of The Web Conference 2020. New York, NY, USA: ACM 2020. 2704\u201310. doi: 10.1145\/3366423.3380027","DOI":"10.1145\/3366423.3380027"},{"key":"pone.0327636.ref013","unstructured":"Vaswani A, Shazeer N, Parmar N, Uszkoreit J, Jones L, Gomez AN, et al. Attention Is All You Need.; 12.06.2017. In: Advances in Neural Information Processing Systems 30 (NIPS\u201917). 2017."},{"key":"pone.0327636.ref014","unstructured":"Veli\u010dkovi\u0107 P. Message passing all the way up.; 22.02.2022. In: International Conference on Learning Representations (ICLR). 2022."},{"key":"pone.0327636.ref015","doi-asserted-by":"crossref","unstructured":"Perozzi B, Al-Rfou R, Skiena S. DeepWalk.; 2014. In: Proceedings of the 20th ACM SIGKDD international conference on Knowledge discovery and data mining. New York, NY, USA: ACM; 2014. p. 701\u201310. doi: 10.1145\/2623330.2623732","DOI":"10.1145\/2623330.2623732"},{"key":"pone.0327636.ref016","doi-asserted-by":"crossref","unstructured":"Grover A, Leskovec J. node2vec: Scalable Feature Learning for Networks.; 03.07.2016. In: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York, NY, USA: ACM; 2016. p. 855\u201364. doi: 10https:\/\/doi.org\/10.1145\/2939672.2939754","DOI":"10.1145\/2939672.2939754"},{"issue":"14","key":"pone.0327636.ref017","first-page":"4693","article-title":"Towards real-time diagnosis for pediatric sepsis using graph neural network and ensemble methods","volume":"25","author":"X Chen","year":"2021","journal-title":"Eur Rev Med Pharmacol Sci"},{"key":"pone.0327636.ref018","doi-asserted-by":"crossref","unstructured":"Fan W, Ma Y, Li Q, He Y, Zhao E, Tang J, et al. Graph Neural Networks for Social Recommendation. In: The World Wide Web Conference. New York, NY, USA: ACM 2019. 417\u201326. doi: 10.1145\/3308558.3313488","DOI":"10.1145\/3308558.3313488"},{"key":"pone.0327636.ref019","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1016\/j.ddtec.2020.11.009","article-title":"A compact review of molecular property prediction with graph neural networks","volume":"37","author":"O Wieder","year":"2020","journal-title":"Drug Discov Today Technol"},{"issue":"5","key":"pone.0327636.ref020","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1145\/3535101","article-title":"Graph neural networks in recommender systems: a survey","volume":"55","author":"S Wu","year":"2022","journal-title":"ACM Comput Surv"},{"issue":"1","key":"pone.0327636.ref021","doi-asserted-by":"crossref","first-page":"45","DOI":"10.1038\/s41698-022-00285-5","article-title":"Cell graph neural networks enable the precise prediction of patient survival in gastric cancer","volume":"6","author":"Y Wang","year":"2022","journal-title":"NPJ Precis Oncol"},{"issue":"3","key":"pone.0327636.ref022","doi-asserted-by":"crossref","first-page":"818","DOI":"10.1109\/JBHI.2020.3004143","article-title":"Disease prediction via graph neural networks","volume":"25","author":"Z Sun","year":"2021","journal-title":"IEEE J Biomed Health Inform"},{"issue":"2","key":"pone.0327636.ref023","doi-asserted-by":"crossref","first-page":"699","DOI":"10.1109\/TCBB.2021.3083566","article-title":"Predicting the survival of cancer patients with multimodal graph neural network","volume":"19","author":"J Gao","year":"2022","journal-title":"IEEE\/ACM Trans Comput Biol Bioinform"},{"issue":"5","key":"pone.0327636.ref024","doi-asserted-by":"crossref","first-page":"379","DOI":"10.1089\/big.2020.0070","article-title":"Graph neural network-based diagnosis prediction","volume":"8","author":"Y Li","year":"2020","journal-title":"Big Data"},{"key":"pone.0327636.ref025","unstructured":"Brody S, Alon U, Yahav E. How Attentive are Graph Attention Networks.; 30.05.2021. In: International Conference on Learning Representations (ICLR). 2022."},{"issue":"3","key":"pone.0327636.ref026","doi-asserted-by":"crossref","first-page":"506","DOI":"10.1093\/clinchem\/hvae001","article-title":"Applying machine learning to blood count data predicts sepsis with ICU admission","volume":"70","author":"D Steinbach","year":"2024","journal-title":"Clin Chem"},{"key":"pone.0327636.ref027","article-title":"sbcdata: laboratory diagnostics from septic and non-septic patients used in the AMPEL project","author":"S Gibb","year":"2023"},{"issue":"1","key":"pone.0327636.ref028","doi-asserted-by":"crossref","first-page":"330","DOI":"10.1111\/imr.12499","article-title":"The immune system\u2019s role in sepsis progression, resolution, and long-term outcome","volume":"274","author":"MJ Delano","year":"2016","journal-title":"Immunol Rev"},{"issue":"3","key":"pone.0327636.ref029","doi-asserted-by":"crossref","first-page":"235","DOI":"10.1007\/s00134-002-1209-6","article-title":"C-reactive protein: a valuable marker of sepsis","volume":"28","author":"P P\u00f3voa","year":"2002","journal-title":"Intensive Care Med"},{"key":"pone.0327636.ref030","doi-asserted-by":"crossref","first-page":"607952","DOI":"10.3389\/fmed.2021.607952","article-title":"Early prediction of sepsis in the ICU using machine learning: a systematic review","volume":"8","author":"M Moor","year":"2021","journal-title":"Front Med (Lausanne)"},{"issue":"2","key":"pone.0327636.ref031","article-title":"An attention based deep learning model of clinical events in the intensive care unit","volume":"14","author":"DA Kaji","year":"2019","journal-title":"PLoS One"},{"issue":"7","key":"pone.0327636.ref032","doi-asserted-by":"crossref","first-page":"856","DOI":"10.1164\/rccm.201609-1848OC","article-title":"The timing of early antibiotics and hospital mortality in sepsis","volume":"196","author":"VX Liu","year":"2017","journal-title":"Am J Respir Crit Care Med"},{"key":"pone.0327636.ref033","volume-title":"The elements of statistical learning. Data mining, inference, and prediction","author":"T Hastie","year":"2009"},{"key":"pone.0327636.ref034"},{"key":"pone.0327636.ref035","unstructured":"Fey M, Lenssen JE. Fast graph representation learning with PyTorch geometric. arXiv:1903.02428; 06.03.2019."},{"issue":"1","key":"pone.0327636.ref036","doi-asserted-by":"crossref","first-page":"185","DOI":"10.1109\/TSMCA.2009.2029559","article-title":"RUSBoost: a hybrid approach to alleviating class imbalance","volume":"40","author":"C Seiffert","year":"2010","journal-title":"IEEE Trans Syst, Man, Cybern A"},{"key":"pone.0327636.ref037","first-page":"8024","article-title":"PyTorch: An Imperative Style, High-Performance Deep Learning Library.","volume-title":"Advances in Neural Information Processing Systems 32","author":"A Paszke","year":"2019"},{"key":"pone.0327636.ref038","article-title":"Scikit-learn: machine learning in python","author":"F Pedregosa","year":"2011","journal-title":"J Mach Learn Res"},{"key":"pone.0327636.ref039"},{"issue":"8","key":"pone.0327636.ref040","doi-asserted-by":"crossref","first-page":"1735","DOI":"10.1162\/neco.1997.9.8.1735","article-title":"Long short-term memory","volume":"9","author":"S Hochreiter","year":"1997","journal-title":"Neural Comput"},{"key":"pone.0327636.ref041","doi-asserted-by":"crossref","first-page":"799","DOI":"10.1007\/11550907_126","article-title":"Bidirectional LSTM Networks for Improved Phoneme Classification and Recognition.","volume-title":"Artificial Neural Networks: Formal Models and Their Applications \u2010 ICANN 2005","author":"A Graves","year":"2005"},{"key":"pone.0327636.ref042","first-page":"255","article-title":"Convolutional networks for images, speech, and time series","volume-title":"The Handbook of Brain Theory and Neural Networks","author":"Y LeCun","year":"1998"},{"key":"pone.0327636.ref043","unstructured":"Blakely D, Lanchantin J, Qi Y. Time and Space Complexity of Graph Convolutional Networks."},{"key":"pone.0327636.ref044","doi-asserted-by":"crossref","unstructured":"LeCun Y, Kavukcuoglu K, Farabet C. Convolutional networks and applications in vision. ISCAS 2010. 2010 IEEE International Symposium on Circuits and Systems, Nano-Bio Circuit Fabrics and Systems: May 30th-June 2nd 2010, Paris, France. [Piscataway, N.J.]: IEEE; doi: 10.1109\/ISCAS.2010.5537907 2010. pp. 253\u20136.","DOI":"10.1109\/ISCAS.2010.5537907"},{"key":"pone.0327636.ref045","doi-asserted-by":"crossref","DOI":"10.1097\/00003246-200107001-00035","article-title":"Epidemiology of sepsis: an update","volume":"29","author":"DC Angus","year":"2001","journal-title":"Crit Care Med"},{"issue":"4","key":"pone.0327636.ref046","doi-asserted-by":"crossref","first-page":"379","DOI":"10.1007\/s10096-011-1337-4","article-title":"The impact of diabetes on the pathogenesis of sepsis","volume":"31","author":"GCKW Koh","year":"2012","journal-title":"Eur J Clin Microbiol Infect Dis"},{"issue":"1","key":"pone.0327636.ref047","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1038\/s41597-022-01899-x","article-title":"MIMIC-IV, a freely accessible electronic health record dataset","volume":"10","author":"AEW Johnson","year":"2023","journal-title":"Sci Data"},{"issue":"6","key":"pone.0327636.ref048","doi-asserted-by":"crossref","DOI":"10.1097\/CCM.0000000000004916","article-title":"Sharing ICU patient data responsibly under the society of critical care medicine\/European society of intensive care medicine joint data science collaboration: the Amsterdam university medical centers database (AmsterdamUMCdb) example","volume":"49","author":"PJ Thoral","year":"2021","journal-title":"Crit Care Med"},{"issue":"3","key":"pone.0327636.ref049","doi-asserted-by":"crossref","first-page":"364","DOI":"10.1038\/s41591-020-0789-4","article-title":"Early prediction of circulatory failure in the intensive care unit using machine learning","volume":"26","author":"SL Hyland","year":"2020","journal-title":"Nat Med"},{"key":"pone.0327636.ref050","doi-asserted-by":"crossref","first-page":"180178","DOI":"10.1038\/sdata.2018.178","article-title":"The eICU collaborative research database, a freely available multi-center database for critical care research","volume":"5","author":"TJ Pollard","year":"2018","journal-title":"Sci Data"},{"issue":"1","key":"pone.0327636.ref051","doi-asserted-by":"crossref","first-page":"4043","DOI":"10.1038\/s41598-023-31122-9","article-title":"Towards a biomarker for acute arterial thrombosis using complete blood count and white blood cell differential parameters in mice","volume":"13","author":"HJ Jang","year":"2023","journal-title":"Sci Rep"},{"issue":"9","key":"pone.0327636.ref052","first-page":"731","article-title":"Leukemia: an overview for primary care","volume":"89","author":"AS Davis","year":"2014","journal-title":"Am Fam Physician"},{"issue":"2","key":"pone.0327636.ref053","doi-asserted-by":"crossref","first-page":"65","DOI":"10.2307\/2985543","article-title":"Matching and prediction on the principle of biological classification","volume":"8","author":"WA Belson","year":"1959","journal-title":"Appl Stat"},{"key":"pone.0327636.ref054","first-page":"217","article-title":"Logistic regression.","volume-title":"Reading and understanding multivariate statistics","author":"RE. Wright","year":"1995"},{"issue":"6088","key":"pone.0327636.ref055","doi-asserted-by":"crossref","first-page":"533","DOI":"10.1038\/323533a0","article-title":"Learning representations by back-propagating errors","volume":"323","author":"DE Rumelhart","year":"1986","journal-title":"Nature"},{"issue":"1","key":"pone.0327636.ref056","doi-asserted-by":"crossref","first-page":"3","DOI":"10.1515\/JLM.2011.004","article-title":"Multizentrische ermittlung von referenzbereichen f\u00fcr parameter des maschinellen blutbildes\/multicentric determination of reference ranges for automated blood counts","volume":"35","author":"T Nebe","year":"2011","journal-title":"LaboratoriumsMedizin"},{"issue":"4","key":"pone.0327636.ref057","doi-asserted-by":"crossref","first-page":"593","DOI":"10.1016\/S0889-8588(18)30148-5","article-title":"The automated CBC. A current perspective","volume":"8","author":"GL Gulati","year":"1994","journal-title":"Hematol Oncol Clin North Am"}],"container-title":["PLOS One"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pone.0327636","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,7,8]],"date-time":"2025-07-08T13:24:23Z","timestamp":1751981063000},"score":1,"resource":{"primary":{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pone.0327636"}},"subtitle":[],"editor":[{"given":"Qiang","family":"He","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"editor"}]}],"short-title":[],"issued":{"date-parts":[[2025,7,8]]},"references-count":57,"journal-issue":{"issue":"7","published-online":{"date-parts":[[2025,7,8]]}},"URL":"https:\/\/doi.org\/10.1371\/journal.pone.0327636","relation":{},"ISSN":["1932-6203"],"issn-type":[{"value":"1932-6203","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,7,8]]}}}