• HOME
  • NEWS
  • EXPLORE
    • CAREER
      • Companies
      • Jobs
    • EVENTS
    • iGEM
      • News
      • Team
    • PHOTOS
    • VIDEO
    • WIKI
  • BLOG
  • COMMUNITY
    • FACEBOOK
    • INSTAGRAM
    • TWITTER
Tuesday, November 17, 2026
BIOENGINEER.ORG
No Result
View All Result
  • Login
  • HOME
  • NEWS
  • EXPLORE
    • CAREER
      • Companies
      • Jobs
        • Lecturer
        • PhD Studentship
        • Postdoc
        • Research Assistant
    • EVENTS
    • iGEM
      • News
      • Team
    • PHOTOS
    • VIDEO
    • WIKI
  • BLOG
  • COMMUNITY
    • FACEBOOK
    • INSTAGRAM
    • TWITTER
  • HOME
  • NEWS
  • EXPLORE
    • CAREER
      • Companies
      • Jobs
        • Lecturer
        • PhD Studentship
        • Postdoc
        • Research Assistant
    • EVENTS
    • iGEM
      • News
      • Team
    • PHOTOS
    • VIDEO
    • WIKI
  • BLOG
  • COMMUNITY
    • FACEBOOK
    • INSTAGRAM
    • TWITTER
No Result
View All Result
Bioengineer.org
No Result
View All Result
Home NEWS Science News Health

Machine Learning Model Predicts Postoperative Blood Clots After Lower Leg Fractures

by
November 13, 2026
in Health
Reading Time: 3 mins read
0
Machine Learning Model Predicts Postoperative Blood Clots After Lower Leg Fractures

Machine Learning Model Predicts Postoperative Blood Clots After Lower Leg Fractures

Share on FacebookShare on TwitterShare on LinkedinShare on RedditShare on Telegram

A machine learning algorithm predicted in-hospital postoperative deep vein thrombosis in adults undergoing surgery for tibial shaft fractures with concomitant ipsilateral fibular fractures, but it did not demonstrate a statistically significant performance advantage over conventional logistic regression. The retrospective investigation, conducted at a single medical center, evaluated routine clinical data from surgical procedures performed between January 2015 and December 2025. Across the entire study period, researchers tracked surveillance-detected lower-extremity deep vein thrombosis among adult patients to develop, lock, and temporally evaluate algorithmic risk prediction tools.

Postoperative deep vein thrombosis, which encompasses asymptomatic clot formations identified via diagnostic imaging, represents an important medical complication following orthopedic fixation of combined tibia and fibula injuries. The research team sought to determine whether routinely available preoperative and admission parameters could support reliable prognostic modeling. By leveraging standard clinical measurements gathered during initial patient intake and preparatory surgical planning, investigators aimed to identify individuals at elevated risk prior to or during their inpatient recovery period.

The study analyzed a total cohort of 643 adult surgical patients, among whom 104 individuals, or 16.2 percent, developed in-hospital lower-extremity deep vein thrombosis. To ensure rigorous temporal evaluation rather than a simple random cross-sectional split, the researchers partitioned the records chronologically. Patients admitted between 2015 and 2022 formed the primary development cohort, totaling 473 individuals and 73 recorded thrombotic events. Individuals admitted between 2023 and 2025 formed the separate temporal validation cohort, comprising 170 patients and 31 observed events.

Model construction began with systematic feature selection across baseline parameters, yielding seven final predictive variables: patient age, time elapsed from injury to surgery, hematocrit, white blood cell count, albumin concentration, fibrinogen level, and the albumin-to-globulin ratio. Using cross-validation within the development cohort, investigators compared seven distinct algorithmic frameworks to evaluate prognostic capacity. The chosen support vector machine architecture was subsequently locked in its final configuration prior to temporal assessment, preventing iterative tuning on the future test cases.

In the locked temporal validation cohort, the support vector machine achieved an area under the receiver operating characteristic curve of 0.934, with a 95 percent confidence interval spanning from 0.892 to 0.969, and a Brier score of 0.084. When compared against a conventional logistic regression benchmark, the logistic model demonstrated an area under the curve of 0.918, with a 95 percent confidence interval ranging from 0.871 to 0.957, alongside a Brier score of 0.081. The observed difference in discriminative ability between the two approaches did not reach statistical significance, yielding a p-value of 0.130.

To rigorously scrutinize the stability of the modeling workflow, researchers executed repeated nested cross-validation across the complete development procedure, which produced a mean area under the curve of 0.894 with a standard deviation of 0.004. Beyond discrimination, the initial support vector machine exhibited a tendency to overestimate absolute patient risk. An exploratory cross-validated recalibration procedure successfully reduced the support vector machine’s Brier score to 0.077 and shifted internal calibration metrics closer to ideal reference values.

Decision-curve analyses conducted across clinical threshold ranges revealed broadly comparable net-benefit trajectories for both the support vector machine and the traditional logistic regression model. Additionally, the investigators utilized SHapley Additive exPlanations to quantify and characterize individual predictor contributions across the feature space, establishing how biological metrics such as inflammatory indicators and interval timing shaped the generated predictions.

Despite demonstrating favorable discrimination for surveillance-detected postoperative deep vein thrombosis, the support vector machine failed to prove clear incremental utility beyond standard logistic regression. The authors concluded that the exploratory recalibration results require verification in independent datasets, noting that true clinical utility remains unproven. Consequently, external multi-center validation and prospective impact evaluations remain essential steps before such algorithmic tools can be considered for active clinical deployment.

Subject of Research: Machine learning prediction of postoperative deep vein thrombosis following tibial and fibular fracture surgery

Article Title: Development and temporal validation of a machine learning model for predicting in-hospital postoperative deep vein thrombosis in adult patients with tibial shaft fractures and concomitant ipsilateral fibular fractures

Article References: Duan, W., Li, X., Yang, J., Yang, Y., Hu, H., Zhang, H., Wei, Y., Song, X., Jia, R., & Long, Y. (2026). Development and temporal validation of a machine learning model for predicting in-hospital postoperative deep vein thrombosis in adult patients with tibial shaft fractures and concomitant ipsilateral fibular fractures. European Journal of Medical Research. https://doi.org/10.1186/s40001-026-05464-9

Image Credits: AI Generated

DOI: 10.1186/s40001-026-05464-9

Keywords: deep vein thrombosis, tibial shaft fractures, fibular fractures, machine learning, support vector machine, logistic regression, temporal validation, orthopedic surgery, postoperative complications, risk prediction, clinical calibration, Brier score, SHAP analysis, hematocrit, fibrinogen, albumin

News Source: Teresa Odom. (November 13, 2026). Machine Learning Model Predicts Postoperative Blood Clots After Lower Leg Fractures. Scienmag.

Tags: albuminBrier scoreclinical calibrationdeep vein thrombosisfibrinogenfibular fractureshematocritlogistic regressionMachine Learningorthopedic surgerypostoperative complicationsrisk predictionSHAP analysissupport vector machinetemporal validationtibial shaft fractures
Share12Tweet7Share2ShareShareShare1

Related Posts

Pilot Study Shows High Acceptability and Feasibility for Fellows' Financial Curriculum

Pilot Study Shows High Acceptability and Feasibility for Fellows’ Financial Curriculum

November 17, 2026
German Stroke Deaths Plateau as Analysis Reveals Gaps in Rehabilitation Care

German Stroke Deaths Plateau as Analysis Reveals Gaps in Rehabilitation Care

November 17, 2026

Continuous Wearable Monitoring May Optimize Postoperative Patient Care Pathways

November 17, 2026

FOXR2 and Metabolic Pathway Proteins Correlate With Ovarian Carcinoma Progression

November 17, 2026

POPULAR NEWS

  • Engineered Myeloid Progenitor Platform Overcomes Allogeneic Barriers in CNS Tumors

    Engineered Myeloid Progenitor Platform Overcomes Allogeneic Barriers in CNS Tumors

    29 shares
    Share 12 Tweet 7
  • Physician Specialty Distribution Lags Behind Global Demographic Aging Trends

    29 shares
    Share 12 Tweet 7
  • Rapid In Planta Citrus Transformation System Enables Efficient Genome Editing

    29 shares
    Share 12 Tweet 7
  • Simulated Reinforcement Learning Model Improves 4G LTE Handover Success and Reliability

    29 shares
    Share 12 Tweet 7

About

We bring you the latest biotechnology news from best research centers and universities around the world. Check our website.

Follow us

Recent News

Pilot Study Shows High Acceptability and Feasibility for Fellows’ Financial Curriculum

German Stroke Deaths Plateau as Analysis Reveals Gaps in Rehabilitation Care

Continuous Wearable Monitoring May Optimize Postoperative Patient Care Pathways

Subscribe to Blog via Email

Enter your email address to subscribe to this blog and receive notifications of new posts by email.

Join 88 other subscribers
  • Contact Us

Bioengineer.org © Copyright 2023 All Rights Reserved.

Welcome Back!

Login to your account below

Forgotten Password?

Retrieve your password

Please enter your username or email address to reset your password.

Log In
No Result
View All Result
  • Homepages
    • Home Page 1
    • Home Page 2
  • News
  • National
  • Business
  • Health
  • Lifestyle
  • Science

Bioengineer.org © Copyright 2023 All Rights Reserved.