Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
July 23, 2026Journal of Plastic Reconstructive & Aesthetic SurgeryOpen Access

A preoperative XGBoost model predicted unplanned surgical reintervention within 30 days after free flap reconstruction with an AUC of 0.84 (95% CI 0.74-0.92).

View Full Paper
Ask AI
Bookmark
Share

Why the study?

Flap-related vascular complications requiring surgical reintervention cause morbidity after microsurgical free flap reconstruction, and preoperative risk estimation currently relies on clinical judgement.

Population

650 consecutive adults (649 analysable) undergoing microsurgical free flap reconstruction at two high-volume referral centres

Design

Retrospective cohort study

Follow-up

30 days

Key result

A preoperative XGBoost model predicted unplanned surgical reintervention within 30 days after free flap reconstruction with an AUC of 0.84 (95% CI 0.74-0.92).

Authors

LLLuis E. LagunaAPAlexandra PorrasGMGiovanni Montealegre

Discussion

Loading...

Member takes

Overview

May inform preoperative flap re-exploration risk; leaves open external validation before clinical use.

Study Design

Type

Cohort (n=649)

Multicenter

Yes

Structured PICO

P
Population
649 adults undergoing microsurgical free flap reconstruction evaluated retrospectively to develop a preoperative risk model for 30-day unplanned re-exploration.
E
Exposure
Preoperative artificial intelligence-based risk models (XGBoost, Random Forest, LASSO)
O
Outcome
Unplanned re-exploration within 30 days for arterial, venous, mixed thrombosis, or clinically significant vasospasmcomposite

Main Result

Effect estimate: AUC 0.84 (95% CI 0.74-0.92)

An XGBoost-based preoperative risk model demonstrated good discrimination (AUC 0.84) for predicting unplanned surgical reintervention after microsurgical free flap reconstruction, though external validation is needed.

Limitations

  • The model is not decision-ready
  • Prospective external validation with recalibration is required before clinical adoption

Cite This Study

Laguna et al. (2026) conducted a cohort in microsurgical free flap reconstruction (n=649). Preoperative multivariable risk model (XGBoost) vs. Random forest and LASSO was evaluated on unplanned re-exploration within 30 days for arterial, venous, mixed thrombosis, or clinically significant vasospasm (AUC 0.84, 95% CI 0.74-0.92). A preoperative XGBoost model predicted unplanned surgical reintervention within 30 days after free flap reconstruction with an AUC of 0.84 (95% CI 0.74-0.92).

synapsesocial.com/papers/6a75b4e392e08b1b896a3963https://doi.org/10.1016/j.bjps.2026.07.033
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Predicting reoperation and readmission for head and neck free flap patients using machine learning2024 · 6 citations
  2. 2Risk prediction models for complications after flap repair surgery: a systematic review and meta-analysis2025 · 1 citations
  3. 3Machine Learning‐Based Flap Takeback Prediction Modeling: Theory for a Real‐Time, Patient‐Specific Postoperative Flap Monitoring and Alert System2025
  4. 4Analysis of risk factors for complications after flap reconstruction of head and neck cancer and construction and validation of predictive models2025
  5. 5Effect of Preoperative Medical Status on Microsurgical Free Flap Reconstructions: A Matched Cohort Analysis of 969 Cases2017 · 18 citations