A nomogram for predicting free flap necrosis in soft tissue reconstruction of lower limbs: a retrospective cohort...

To identify independent risk factors for free flap necrosis in lower limb soft tissue reconstruction and to develop and internally validate a preliminary nomogram for risk prediction.

A retrospective cohort study was conducted on patients who underwent free flap reconstruction for lower limb soft tissue defects between January 2010 and March 2025. Eligible patients were randomly split into a training cohort (70%) for model development and a validation cohort (30%) for internal validation. Variable selection were performed solely using Least Absolute Shrinkage and Selection Operator (LASSO) regression. A nomogram was constructed based on the identified risk factors, and its performance was rigorously evaluated via bootstrap internal validation (500 repetitions with optimism correction) across three key dimensions: discrimination (optimism-corrected area under the receiver operating characteristic curve, AUC), calibration (calibration slope/intercept and Brier score), and clinical utility (decision curve analysis, DCA).

A total of 220 patients were enrolled in this study. Of these, 154 patients were assigned to the training cohort and 66 patients to the validation cohort. Five independent risk factors were identified: Gustilo-Anderson classification IIIB/IIIC (OR = 3.74, 95% CI: 1.89-7.41), preoperative D-dimer > 0.5 mg/L (OR = 3.16, 95% CI: 1.50-6.64), preoperative albumin < 35 g/L (OR = 2.75, 95% CI: 1.41-5.34), operative time > 6 h (OR = 2.59, 95% CI: 1.35-4.94), and defect size > 50 cm² (OR = 2.05, 95% CI: 1.10-3.83). The nomogram showed promising discriminative ability. The optimism-corrected AUC was 0.87 (95% CI: 0.84-0.90) in the training cohort with excellent calibration (slope = 1.00, intercept = 0.00) and a Brier score of 0.122. In the validation cohort, the AUC was 0.86 (95% CI: 0.81-0.91) with a Brier score of 0.130. The DCA demonstrated that the nomogram had superior net clinical benefit compared to "treat all" or "treat none" strategies.

Our study developed and internally validated a preliminary nomogram incorporating five preoperative factors for predicting free flap necrosis in lower limb soft tissue reconstruction. The model demonstrated robust discrimination, excellent calibration, and meaningful clinical utility in the single-center cohort. However, its generalizability and real-world clinical utility requires confirmation through prospective external validation in multi-center settings.