{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,2]],"date-time":"2026-09-02T12:08:26Z","timestamp":1788350906320,"version":"build-2803163510"},"reference-count":30,"publisher":"MDPI AG","issue":"5","license":[{"start":{"date-parts":[[2026,3,4]],"date-time":"2026-03-04T00:00:00Z","timestamp":1772582400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Diagnostics"],"abstract":"<jats:p>Background\/Objectives: Early tracheostomy seems favorable in prolonged ventilated patients after surgery. Hence, predicting tracheostomy after cardiac surgery is essential. Recently proposed prediction models aim to support this decision-making process, but their diagnostic validity across other patient populations remains uncertain. Methods: A retrospective single-center study was performed at a university hospital. The patient sample included consecutive patients between 2010 and 2020 who underwent cardiac surgery. Patients who underwent tracheostomy after cardiac surgery were assigned to the intervention group. Control group patients, who had not undergone tracheostomy, were randomly assigned to the group. An existing model was evaluated by receiver operating characteristics curve analysis. Four sets of risk features were chosen depending on results from regression analysis, lasso regularization, random forest or clinical domain knowledge. Newly developed models were created using machine learning methods: random forest, na\u00efve Bayes, nearest neighbor and deep learning. Multiple models were trained with either feature set and then assessed using confusion matrices on an independent test set. Results: A total of 4744 patients were included in this study. One-hundred and eighteen patients were included in the tracheostomy group. Diagnostic accuracy of the existing model showed insufficient discrimination (area under the curve (AUC) = 0.57). Likewise, newly developed models also showed overall poor diagnostic discrimination across all feature sets and algorithms. Conclusions: This study shows the diagnostic limitations of retrospective clinical data for the diagnostic prediction of tracheostomy, thereby informing the design of future prospective diagnostic studies. Training new models should not rely on retrospective data alone. Instead, prospective data collection and integration of physiological or imaging-based diagnostics could likely contribute to the development of a good classifier.<\/jats:p>","DOI":"10.3390\/diagnostics16050771","type":"journal-article","created":{"date-parts":[[2026,3,4]],"date-time":"2026-03-04T16:17:31Z","timestamp":1772641051000},"page":"771","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":1,"title":["Limitations of Retrospective Machine Learning Models for Predicting Tracheostomy After Cardiac Surgery"],"prefix":"10.3390","volume":"16","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-6101-804X","authenticated-orcid":false,"given":"Felix","family":"Wiesmueller","sequence":"first","affiliation":[{"name":"Department of Surgery, University Hospital Greifswald, University of Greifswald, 17475 Greifswald, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Johannes","family":"R\u00f6sch","sequence":"additional","affiliation":[{"name":"Department of Cardiac Surgery, University Hospital Erlangen, Friedrich-Alexander-University of Erlangen-Nuremberg (FAU), 91054 Erlangen, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-2124-3103","authenticated-orcid":false,"given":"Stephan","family":"Kersting","sequence":"additional","affiliation":[{"name":"Department of Surgery, University Hospital Greifswald, University of Greifswald, 17475 Greifswald, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8292-7066","authenticated-orcid":false,"given":"Thomas","family":"Strecker","sequence":"additional","affiliation":[{"name":"Department of Cardiac Surgery, University Hospital Erlangen, Friedrich-Alexander-University of Erlangen-Nuremberg (FAU), 91054 Erlangen, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1968","published-online":{"date-parts":[[2026,3,4]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","first-page":"752","DOI":"10.1093\/ndt\/gfx073","article-title":"Con: Most clinical risk scores are useless","volume":"32","author":"Dekker","year":"2017","journal-title":"Nephrol. Dial. Transplant."},{"key":"ref_2","doi-asserted-by":"crossref","first-page":"2507","DOI":"10.1056\/NEJMp1702071","article-title":"Machine Learning and Prediction in Medicine\u2014Beyond the Peak of Inflated Expectations","volume":"376","author":"Chen","year":"2017","journal-title":"N. Engl. J. Med."},{"key":"ref_3","doi-asserted-by":"crossref","first-page":"484","DOI":"10.1038\/nature16961","article-title":"Mastering the game of Go with deep neural networks and tree search","volume":"529","author":"Silver","year":"2016","journal-title":"Nature"},{"key":"ref_4","doi-asserted-by":"crossref","first-page":"1339","DOI":"10.1016\/j.jcin.2019.06.004","article-title":"Machine Learning Is No Magic: A Plea for Critical Appraisal During Periods of Hype","volume":"12","author":"Modine","year":"2019","journal-title":"JACC Cardiovasc. Interv."},{"key":"ref_5","doi-asserted-by":"crossref","first-page":"1394","DOI":"10.1016\/j.athoracsur.2020.05.181","article-title":"Cardiac Surgery in Low- and Middle-Income Countries: A State-of-the-Art Review","volume":"111","author":"Vervoort","year":"2021","journal-title":"Ann. Thorac. Surg."},{"key":"ref_6","doi-asserted-by":"crossref","first-page":"550","DOI":"10.4187\/respcare.04915","article-title":"Prolonged Mechanical Ventilation as a Predictor of Mortality After Cardiac Surgery","volume":"63","year":"2018","journal-title":"Respir. Care"},{"key":"ref_7","doi-asserted-by":"crossref","first-page":"759","DOI":"10.1111\/jocs.14118","article-title":"Predictors of operative mortality among cardiac surgery patients with prolonged ventilation","volume":"34","author":"Fraser","year":"2019","journal-title":"J. Card. Surg."},{"key":"ref_8","doi-asserted-by":"crossref","first-page":"1537","DOI":"10.1016\/j.athoracsur.2020.07.027","article-title":"Impact of Early Tracheostomy on Outcomes After Cardiac Surgery: A National Analysis","volume":"111","author":"Sareh","year":"2021","journal-title":"Ann. Thorac. Surg."},{"key":"ref_9","doi-asserted-by":"crossref","first-page":"679","DOI":"10.1007\/s00405-017-4838-7","article-title":"Timing of tracheostomy in patients with prolonged endotracheal intubation: A systematic review","volume":"275","author":"Adly","year":"2018","journal-title":"Eur. Arch. Otorhinolaryngol."},{"key":"ref_10","doi-asserted-by":"crossref","first-page":"167","DOI":"10.7326\/0003-4819-153-3-201008030-00007","article-title":"One-year trajectories of care and resource utilization for recipients of prolonged mechanical ventilation: A cohort study","volume":"153","author":"Unroe","year":"2010","journal-title":"Ann. Intern. Med."},{"key":"ref_11","doi-asserted-by":"crossref","unstructured":"Andriolo, B.N., Andriolo, R.B., Saconato, H., Atallah, A.N., and Valente, O. (2015). Early versus late tracheostomy for critically ill patients. Cochrane Database Syst. Rev., 1.","DOI":"10.1002\/14651858.CD007271.pub3"},{"key":"ref_12","doi-asserted-by":"crossref","unstructured":"Milman, S., and Ng, T. (2017). Tracheostomy after cardiac surgery: What have we learned?. J. Thorac. Cardiovasc. Surg., 153.","DOI":"10.1016\/j.jtcvs.2017.01.021"},{"key":"ref_13","doi-asserted-by":"crossref","first-page":"1394","DOI":"10.1016\/j.jtcvs.2016.11.040","article-title":"Incidence of sternal wound infection after tracheostomy in patients undergoing cardiac surgery: A systematic review and meta-analysis","volume":"153","author":"Toeg","year":"2017","journal-title":"J. Thorac. Cardiovasc. Surg."},{"key":"ref_14","doi-asserted-by":"crossref","unstructured":"Wang, D., Wang, S., Du, Y., Song, Y., Le, S., Wang, H., Zhang, A., Huang, X., Wu, L., and Du, X. (2021). A Predictive Scoring Model for Postoperative Tracheostomy in Patients Who Underwent Cardiac Surgery. Front. Cardiovasc. Med., 8.","DOI":"10.3389\/fcvm.2021.799605"},{"key":"ref_15","doi-asserted-by":"crossref","unstructured":"Skipper Seabold, J.P. (2010). Statsmodels: Econometric and Statistical Modeling with Python, SciPy.","DOI":"10.25080\/Majora-92bf1922-011"},{"key":"ref_16","unstructured":"Kneusel, R.T. (2025). Practical Deep Learning: A Python-Based Introduction, No Starch Press. [2nd ed.]."},{"key":"ref_17","doi-asserted-by":"crossref","unstructured":"Lee, J.W., Lee, H.J., Jang, H.J., Yun, Y., and Kim, D.Y. (2025). Predicting pneumonia algorithm in stroke patients. Front. Neurol., 16.","DOI":"10.3389\/fneur.2025.1690049"},{"key":"ref_18","doi-asserted-by":"crossref","first-page":"1736","DOI":"10.1002\/ohn.919","article-title":"Predicting Tracheostomy Need on Admission to the Intensive Care Unit-A Multicenter Machine Learning Analysis","volume":"171","author":"Nguyen","year":"2024","journal-title":"Otolaryngol. Head. Neck Surg."},{"key":"ref_19","first-page":"197","article-title":"Predicting risk of tracheostomy after cardiac surgery: A decade of insights from nearly 10,000 patients","volume":"37","author":"Shih","year":"2024","journal-title":"Proc. (Bayl. Univ. Med. Cent.)"},{"key":"ref_20","doi-asserted-by":"crossref","unstructured":"Nagendran, M., Chen, Y., Lovejoy, C.A., Gordon, A.C., Komorowski, M., Harvey, H., Topol, E.J., Ioannidis, J.P.A., Collins, G.S., and Maruthappu, M. (2020). Artificial intelligence versus clinicians: Systematic review of design, reporting standards, and claims of deep learning studies. BMJ, 368.","DOI":"10.1136\/bmj.m689"},{"key":"ref_21","doi-asserted-by":"crossref","unstructured":"Ahmed, F.R., Al-Yateem, N., Nejadghaderi, S.A., Saifan, A.R., Farghaly Abdelaliem, S.M., and AbuRuz, M.E. (2025). Harnessing machine learning for predicting successful weaning from mechanical ventilation: A systematic review. Aust. Crit. Care, 38.","DOI":"10.1016\/j.aucc.2025.101203"},{"key":"ref_22","doi-asserted-by":"crossref","unstructured":"Sterr, F., Reintke, M., Bauernfeind, L., Senyol, V., Rester, C., Metzing, S., and Palm, R. (2024). Predictors of weaning failure in ventilated intensive care patients: A systematic evidence map. Crit. Care, 28.","DOI":"10.1186\/s13054-024-05135-3"},{"key":"ref_23","doi-asserted-by":"crossref","unstructured":"Andaur Navarro, C.L., Damen, J.A.A., Takada, T., Nijman, S.W.J., Dhiman, P., Ma, J., Collins, G.S., Bajpai, R., Riley, R.D., and Moons, K.G.M. (2022). Completeness of reporting of clinical prediction models developed using supervised machine learning: A systematic review. BMC Med. Res. Methodol., 22.","DOI":"10.1186\/s12874-021-01469-6"},{"key":"ref_24","doi-asserted-by":"crossref","unstructured":"Collins, G.S., Moons, K.G.M., Dhiman, P., Riley, R.D., Beam, A.L., Van Calster, B., Ghassemi, M., Liu, X., Reitsma, J.B., and van Smeden, M. (2024). TRIPOD+AI statement: Updated guidance for reporting clinical prediction models that use regression or machine learning methods. BMJ, 385.","DOI":"10.1136\/bmj-2023-078378"},{"key":"ref_25","doi-asserted-by":"crossref","unstructured":"El Emam, K., Leung, T.I., Malin, B., Klement, W., and Eysenbach, G. (2024). Consolidated Reporting Guidelines for Prognostic and Diagnostic Machine Learning Models (CREMLS). J. Med. Internet Res., 26.","DOI":"10.2196\/52508"},{"key":"ref_26","doi-asserted-by":"crossref","unstructured":"Obuchowicz, R., Strzelecki, M., and Piorkowski, A. (2024). Clinical Applications of Artificial Intelligence in Medical Imaging and Image Processing-A Review. Cancers, 16.","DOI":"10.3390\/cancers16101870"},{"key":"ref_27","doi-asserted-by":"crossref","unstructured":"Buriboev, A.S., Muhamediyeva, D., Primova, H., Sultanov, D., Tashev, K., and Jeon, H.S. (2024). Concatenated CNN-Based Pneumonia Detection Using a Fuzzy-Enhanced Dataset. Sensors, 24.","DOI":"10.3390\/s24206750"},{"key":"ref_28","doi-asserted-by":"crossref","unstructured":"Elshennawy, N.M., and Ibrahim, D.M. (2020). Deep-Pneumonia Framework Using Deep Learning Models Based on Chest X-Ray Images. Diagnostics, 10.","DOI":"10.3390\/diagnostics10090649"},{"key":"ref_29","doi-asserted-by":"crossref","unstructured":"Ye, R.Z., Lipatov, K., Diedrich, D., Bhattacharyya, A., Erickson, B.J., Pickering, B.W., and Herasevich, V. (2024). Automatic ARDS surveillance with chest X-ray recognition using convolutional neural networks. J. Crit. Care, 82.","DOI":"10.1016\/j.jcrc.2024.154794"},{"key":"ref_30","doi-asserted-by":"crossref","unstructured":"Obuchowicz, R., Pi\u00f3rkowski, A., Nurzy\u0144ska, K., Obuchowicz, B., Strzelecki, M., and Bielecka, M. (2026). Will AI Replace Physicians in the Near Future? AI Adoption Barriers in Medicine. Diagnostics, 16.","DOI":"10.3390\/diagnostics16030396"}],"container-title":["Diagnostics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/2075-4418\/16\/5\/771\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,3,4]],"date-time":"2026-03-04T16:32:20Z","timestamp":1772641940000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/2075-4418\/16\/5\/771"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,3,4]]},"references-count":30,"journal-issue":{"issue":"5","published-online":{"date-parts":[[2026,3]]}},"alternative-id":["diagnostics16050771"],"URL":"https:\/\/doi.org\/10.3390\/diagnostics16050771","relation":{},"ISSN":["2075-4418"],"issn-type":[{"value":"2075-4418","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,3,4]]}}}