PulseExploreJournal ClubDebatesTrendingResearchersJournals
Instagram
HomeExploreJournal ClubTrending
Synapse
⌘+K
Synapse
March 27, 2026BMC Medical Informatics and Decision Making0 citationsOpen Access

Machine learning prediction of postoperative acute kidney injury in aortic dissection patients using dynamic inflammatory markers and clinical features

YXYansong XuCHChunyan HuangYWYuewu Wang

Key Points

  • The aim is to create and validate a predictive model for acute kidney injury following aortic dissection repair using inflammatory and clinical data.
  • Retrospective cohort study of 720 aortic dissection patients undergoing surgery.
  • Patients randomly divided into training (70%) and validation (30%) cohorts.
  • Variable selection via LASSO regression considering multiple clinical and inflammatory markers.
  • Final model developed using multivariable logistic regression with key predictors.
  • Postoperative AKI occurrence was 17.2%.
  • Key predictors included ΔNLR (OR: 2.23), preoperative PFR (OR: 1.95), open surgery (OR: 5.37), and drinking history (OR: 1.72).
  • Model showed good discrimination (C-statistic of 0.751 in training and 0.732 in validation cohorts).
  • Calibration curves indicated excellent agreement between predicted and actual outcomes.

Abstract

To develop and validate a predictive model for acute kidney injury (AKI) after aortic dissection (AD) repair by integrating the dynamic neutrophil-to-lymphocyte ratio (ΔNLR) with key clinical variables. This retrospective cohort study included 720 patients who underwent AD surgery. Patients were randomly split into training (70%) and validation (30%) cohorts. AKI was defined per RIFLE criteria. Least absolute shrinkage and selection operator (LASSO) regression was used for variable selection from demographics, medical history, imaging, surgical data, and inflammatory ratios (including preoperative, postoperative, and Δ values). Multivariable logistic regression built the final model, evaluated by discrimination (C-statistic), calibration (plots, Hosmer-Lemeshow test), and clinical utility (decision curve analysis). The incidence of postoperative AKI was 17.2%. The final model incorporated four independent predictors: ΔNLR (Odds Ratio OR: 2.23), preoperative platelet-to-fibrinogen ratio (PFR) (OR: 1.95), open surgery (OR: 5.37), and drinking history (OR: 1.72). The model demonstrated good and consistent discrimination, with a C-statistic of 0.751 (95% CI: 0.708–0.794, p < 0.001) in the training cohort and 0.732 (95% CI: 0.673–0.791, p < 0.001) in the validation cohort. Calibration curves showed excellent agreement between predicted and observed probabilities (Hosmer-Lemeshow test p = 0.172). Decision curve analysis confirmed significant clinical net benefit across a clinically relevant range of risk thresholds (approximately 5% to 80%). We developed a robust predictive model for AKI after AD surgery, highlighting the critical value of dynamic inflammation monitoring via ΔNLR. This practical tool facilitates early identification of high-risk patients, potentially enabling timely preventive strategies to improve postoperative outcomes. External validation is warranted to confirm generalizability. Not applicable.

Ask AI
Helpful
Bookmark
Share
View Full Paper

Cite This Study

Xu et al. (2026) studied this question.

synapsesocial.com/papers/69c620d515a0a509bde1967fhttps://doi.org/10.1186/s12911-026-03443-y
Ask AI
Helpful
Bookmark
Share
View Full Paper